Research and development of new technology by uajournals

Research and development of new technology

Editors: Eugenia Erica Vera Cervantes Olga Leticia Fuchs Gómez

ISBN BUAP: 978-607-525-651-1 ISBN UA Journals: 978-84-949828-5-9 11.5 MB

Research and development of new technology

Eugenia Erica Vera Cervantes Olga Leticia Fuchs Gómez

Editors

United Academic Journals (UA Journals), Computer Science Faculty

2019

October 2019 ISBN UA Journals: 978-84-949828-5-9 Huelva, España DR © United Academic Journals (UA Journals) Research and development of new technology United Academic Journals (UA Journals) Digital (suministrado electrónicamente) www.uajournals.com

Primera Edición 2019 ISBN BUAP: 978‐607-525‐651‐1

DR © Benemérita Universidad Autónoma de Puebla 4 sur 104, Col. Centro Histórico, Puebla, Pue. CP 72000 Teléfonos: 01 (222) 246 85 59 y 01 (222) 229 55 00 Ext. 5768 y 5764 www.buap.mx

Dirección General de Publicaciones 2 norte 1404, Col. Centro Histórico, Puebla, Pue. CP 72000 Teléfonos: 01 (222) 246 85 59 y 01 (222) 229 55 00 Ext. 5768 y 5764 www.dgp.buap.mx Facultad de Ciencias de la Computación Computer Science Faculty 14 sur esq. Con Av. San Claudio Ciudad Universitaria, Puebla, Pue. Teléfonos: 01 (222) 229 55 00 Ext. 7200 y 7204 www.cs.buap.mx BENEMÉRITA UNIVERSIDAD AUTÓNOMA DE PUEBLA •Rector: José Alfonso Esparza Ortiz •Secretario General: José Jaime Vázquez López •Vicerrector de Extensión y Difusión de la Cultura: José Carlos Bernal Suárez • Director General de Publicaciones: Hugo César Vargas Comsille •Director de la Facultad de Ciencias de la Computación: Marcos González Flores

Hecho en México Made in Mexico

Research and development of new technology

Editors

Eugenia Érica Vera Cervantes Olga Leticia Fuchs Gómez

Research and development of new technology

Reviewers Alexandra Baldaque Faria Marinho Fernandes

U. Portucalense-Portugal

Alexis Vera Sánchez

Universidad Iberoamericana Puebla - México

Ana Isabel Molina Díaz

U. Castilla la Mancha-España

Andres Solano

Universidad Autónoma de Occidente-Cali-Colombia

Carlos Lara Alvarez

CIMAT-México

Eduardo Fernández

Universidad de la República-Uruguay

Francisco Luis Gutiérrez Vela

U. Granada-España

Freddy Paz Espinoza

Pontificia Universidad Católica del Perú

German Lescano

Universidad Nacional Santiago del Estero-Argentina

Germania Del Rocío Rodríguez Morales

Universidad Técnica Particular de Loja-Ecuador

Gisela Esperanza Torres de Clunie

Universidad Tecnológica de Panamá

Ismar Frango Silveira

Universidad Presbiteriana Mackenzie, Brasil

Laura Aballay

Universidad Nacional de San Juan-Argentina

Leandro Florez Aristizabal

Universidad del Cauca-Colombia

Manuel Jesús Ibarra Cabrera

Universidad Nacional Micaela Bastidas de Apurímac- Perú

Manuel Ortega Cantero

U. Castilla la Mancha-España

María Auxilio Medina Nieto

Universidad Politécnica de Puebla - México

Maria Lili Villegas Ramírez

Universidad del Quindío-Colombia

Mario Chacón Rivas

Instituto Tecnológico de Costa Rica

Marta Cecilia Camacho Ojeda

Unicolmayor-Colombia

Mauricio Ramírez

UNAD-Colombia

Natalia Padilla Zea

UNIR-España

Oriana Robles

ApropiACYT-Colombia

Phillipe Palanque

Université Toulouse III-Francia

Raúl Antonio Aguilar Vera

Universidad Autónoma de Yucatán-México

Rosa Vicari

Universidad Federal do Rio Grande do Sul-Brasil

Rosanna Costaguta

Universidad Nacional Santiago del Estero-Argentina

Sergio Albiol Pérez

U. Zaragoza-España

Virginica Rusu

Universidad de Playa Ancha-Chile

INTRODUCTION

Research and development of new technology is a piece of work where you will find novel and current efforts in research to successfully develop interactive systems, with special attention to the educational context and formally described with human computer interaction methodologies. In this particular Crossroad authors are concerned on the unidentified particular needs in complex context such as universities, high schools or elementary schools, always evolving and constantly changing but also by taking into account the technology variable seriously. The goal of this book is to communicate the characteristics and the methodologies needed to address different problems, in different context, applying methodologies that readers could easily reproduce. We do not warranty the effectiveness of the approaches that are presented here but we do encourage you to follow similar paths when it comes to develop interactive systems. The book includes work on generic competences in higher education students considering different aspects such as communication, collaborative work, process management, as a learning objective. The use of games and gamification strategies are also described for reading comprehension, to understand parallelism and digital logic. Virtual reality is used as an immersive environment to learn how to behave and special care the emergency room; this is a good example of how aspects related to technology, learning and social impact are correlated through the different methodologies that are explained in this book. Additionally, you will find the development of electronic system for detection of attention deficit disorder in students of child age. Being a novel device, the design and evaluation can be used as a complementary tool for researchers that can use this kind of technology in their studies to identify attention at first and the disorder ADHD as well. Most of the materials are evaluated with usability or even better new evaluation methods are presented. We are certain that you will enjoy and take advantage of this book as we already did.

Contents Number

Title

Author(s)

Page

SECTION 1 Research and Development of Technological Applications 1

Proposal of a Tool for Usability Inspection in Web Applications

Gabriel Elías Chanchí G., Daniela Pérez Oliveros, Wilmar Yesid Campo Muñoz

11-24

Expectation of ICT development in Mexico within the framework of Latin America

Adrián Sánchez, Agustin Lagunes, Carlos A. Torres, Juan C. Judikis, Francisco López

25-34

Extracting and visualizing a new classification system for Colombia's National Administrative Department of Statistics. A visual analytics framework case study

Pierre Raimbaud, Jaime Camilo Espitia Castillo, John A. Guerra Gómez

35-50

A Comparative Study on Self-Report Methods for User Satisfaction Evaluation

Andrés F. Aguirre, Ángela Villareal Freire, Jaime Díaz Arancibia, Rosa Gil, César A. Collazos

51-67

Methodology for the Analysis of household appliance Consumption

Blanca Nydia Pérez Camacho, Juan Manuel González Calleros

68-79

A Fuzzy-Based Groupware Usability Evaluation Methodology

Luz A. Sánchez Gálvez, Juan Manuel Fernández Luna, Mario Anzures García

80-95

SECTION 2 Computational Linguistics 7

A User-Defined Language for Full-Body GestureBased Interaction

David Céspedes Hernández, Juan Manuel González Calleros

97-111

Discussion on Image Processing for Sign Language Recognition: An overview of the problem complexity

Soraia Silva Prietch, Ivan Olmos Pineda, Polianna dos Santos Paim, Juan Manuel Gonzalez Calleros, Josefina Guerrero García, Roger Resmini

112-126

Nonverbal Communication Mediated through Haptic Technology: A Survey

Hector M. Camarillo Abad, J. Alfredo Sánchez, Oleg Starostenko

127-139

Number

Title

Author(s)

Page

SECTION 3 Educational Research 10

Computer Science Students’ Profile: A comparison between students from Mexico and Colombia

Guillermina Sánchez Román, Josefina Guerrero García, Erika Anabel Martínez Mirón, Mariano Larios Gómez, Adriana Hernández Beristain, María Daniela Trinidad Meneses

141-148

Conócete: Serious game for self-knowledge in the process of Vocational Guidance

Tatiana de Jesús Olvera Pablos, María Teresa Fernández Nistal, Sonia Verónica Mortis Lozoya, David Bonilla, Adriana Peña Pérez Negrón, Emmanuel Guardiola

149-157

Automated tasks for the working memory training

Ismael Esquivel Gámez, Flora Lilia Barrios Martínez, Karina Estela Galvez Buenfil

158-169

Construction-validation of an implement to measure the digital competence of a training teacher in the level of Telesecundaria

Anna Luz Acosta Aguilera, Rubén Edel Navarro, Yadira Navarro Rangel

170-180

“A journey through mathematics”: A serious game used for learning mathematics with ADHD children

Claudia Blanca González Calleros, Josefina Guerrero García, Yadira Navarro Rangel

181-196

Exploring Teenager's Motivations as Digital Natives to Enhance Peace Culture Workshops Adherence Using Gamification

Cristian Jurado, Mónica Llano, Edwin Gamboa, Yuri Bermúdez, Victor Valencia, Maria Trujillo

197-209

ARTUI: An Augmented Reality-Based Application for English Teaching to Children between 10 and 12 years

Liliana Rodríguez Vizzuett, Josefina Guerrero García, Iván Olmos Pineda

210-225

Technological strategy for searching and selecting sources of reliable information on the Internet

Mayra N. Márquez Specia, Josefina Guerrero García, Yadira Navarro Rangel

226-233

Protocol for Creation of a Database: Data Capture for the Automatic Identification of LearningCentered Emotions

Yesenia N. González Meneses, Josefina Guerrero García

234-247

The RASE model as a means of applying communicative language teaching principles online

María de los Milagros Cruz Ramos, Juan Manuel González Calleros, Luz Edith Herrera Díaz

248-263

An Architectural Model for Collaborative Content Production to reduce Digital Divide

Jaime Muñoz Arteaga, José E. Guzmán Mendoza, César Velázquez Amador

264-278

SECTION 1 Research and Development of Technological Applications

The use of top technology is essential in the Knowledge Society to assess and optimize the use of the goods, products and services that a country possesses and thus face current changes and make the best decisions that promote the strengthening of the country. This section contains the research results in Information and Communication Technologies. The studies range from the impact they have on Latin American countries, the tools to study usability and web applications, a comparative study on self-report methods for the evaluation of user satisfaction to the design of methodology to develop demand systems. There are different levels and dimensions in which user satisfaction can be assessed, some reasons depend on the context, others on the type and purpose of the study. In the field of human-computer interaction, it is important to understand not only how satisfied the user is, but also how to meet their needs regarding the use and quality of use of a given system. In this section of the book we will find important works on tools to study usability in web applications. These studies seek to support developers, designers and programmers in the evaluation of websites in order to build more inclusive and usable websites. We will also find the presentation of a proposed methodology to develop demand systems. A demand system is designed to analyze and control the way in which the consumption of household appliances occurs in a context. This type of system has four objectives: reduce consumption, reduce costs, reduce the maximum average ratio and maximize user comfort.

Proposal of a Tool for Usability Inspection in Web Applications Gabriel Elías Chanchí G.1 , Daniela Pérez Oliveros 2 , Wilmar Yesid Campo Muñoz3 1,2

Institución Universitaria Colegio M ayor del Cauca, Popayán, Cauca, Colombia {gchanchi, danielaperez}@unimayor.edu.co 3 Universidad del Quindío, Armenia, Quindío, Colombia wycampo@uniquindio.edu.co

Abstract. To allow website portals to be designed according to the functional needs of users, the concept of usability has gained great importance in recent years. In this sense, according to ISO 9241-11, usability focuses on fulfilling a user’s objectives with effectiveness, efficiency, and satisfaction. At usability level in websites, a set of usability inspection heuristic tests exists, such as Torres-Burriel, which, based on a set of usability criteria and heuristics, allows identifying typical usability problems. In spite of the above, execution of these types of usability tests does not only involve the inspection and evaluation of the criteria, but also the processing and analysis of the data, given that the templates for these tests are usually found in the best case, in spreadsheets. In this paper, we propose as a contribution, a tool to inspect the usability of websites, which allows evaluating the criteria associated with the heuristics and the graphic analysis of the heuristics evaluated. This tool seeks to support developers and experts to carry out usability inspections on websites. Keywords: inspection, usability, usability test, web applications.

Introducción

Website portals are currently the simplest and most widespread way people and companies have to publicize information about a product or service. In relation to the aforementioned, it is necessary for these sites to comply and agree with what users expect to find in them, in such a way that it manages to capture their attention without ignoring the main idea for which the website was created [1]. Due to this, the concept of usability has gained strength with the objective of users being able to fulfill their expectations of use adequately in applications in different application contexts. According to the ISO 9241-11 standard, usability is understood as the degree in which a product can be used by specific users to achieve specific objectives with effectiveness, efficiency, and satisfaction within a given context of use [2]. Various methods of web usability inspection exist among which we can highlight the heuristic test by Sirius [3] and the Torres-Burriel test [4], which, from the definition of heuristics and usability criteria, permit identifying common usability problems in different types of websites. The heuristics defined for websites in both evaluation tests mentioned, start from the general-purpose heuristics by Nielsen and

are specified to the context of the websites [5]. These two types of tests define a different evaluation scale for each of the evaluation criteria considered, permitting to obtain statistics by using documents or spreadsheets, making it necessary for the evaluator to draw graphics manually on the compliance of heuristics in websites. The importance of executing these evaluation methods lies in obtaining recommendations on the website portal evaluated, permitting to provide feedback to the quality of the product software [6]. Although both website usability tests are similar regarding the heuristics and criteria proposed, the website usability test by Torres-Burriel can be considered more complete by bearing in mind a set of accessibility criteria, a currently fundamental element, to guarantee that a greater number of people can access a website independent of the context of use [7]. Thus, in this paper the heuristics of Torres-Burriel are considered, taking into account that they address the main elements of a website, including aspects related to accessibility. This article proposes a tool for usability inspection for web applications, considering the usability test by Torres-Burriel [4]. This tool permits evaluating the compliance of a set of criteria associated with the eleven heuristics of the test (general aspects, identification and information, language and drafting, labeling, navigation structure, page layout, search, multimedia elements, help, accessibility, and control and feedback), obtaining a graphic report on the percentage of compliance and the evaluation average obtained for each heuristic. The tool seeks to become a support to conduct usability inspections on website portals, to identify different elements to improve of usability of a website portal. The rest of the article is organized in the following manner: section 2 presents the different phases of the methodology used for this research; section 3 presents a set of concepts considered to develop this work; section 4 describes the functional structure of the usability inspection tool; section 5 presents a case study conducted through the tool on the website portal of the Puebla Secretary of Education; finally, section 6 presents the conclusions and future work derived from this research.

Methodology

To develop and validate the tool for the usability inspection of websites, the following phases were generated, namely, exploration of the usability test by Torres-Burriel, design of the inspection tool, implementation of the inspection tool, and case study.

P1: Exploration of the usability test by TorresBurriel

P3: Implementation of the inspection tool

P2: Design of the inspection tool

P4: Case study

Figure 1. M ethodology proposal.

Phase 1 - Exploration of the usability test by Torres-Burriel: this phase conducted a detailed study of each of the guidelines and criteria present in the usability test by Torres-Burriel. This was done to identify the different heuristics and criteria, as well as the format of the template used to conduct the test. Phases 2 and 3 - Design and construction of the usability test: starting from the criteria obtained in the Torres-Burriel test and revising its structure in the document in which it is proposed, a tool was designed and constructed using a system of tabs, each of which is associated with a Torres -Burriel heuristic. Phase 4: Case study: once the tool was generated to evaluate the usability of websites, the portal of the Education Secretary of Puebla was chosen, given that by being a government-type portal, it must guarantee usability and accessibility conditions necessary considering the amount of visitors.

Conceptual Framework

This section includes a set of relevant concepts considered to carry out this work. These are: usability, usability test, usability test by Torres-Burriel. 3.1 Usability Usability is understood according to ISO 9241-11 as the measure in which a system, product, or service can be used by specific users to achieve specific objectives with effectiveness, efficiency, and satisfaction within a specific context of use [1]. Likewise, it may be defined as the discipline that studies the way of designing websites for users to interact with them more easily, comfortably, and intuitively [8]. A usability test comprises a series of practices and tests performed on an application or a website to test the comfort, ease or complexity with which it is managed. Expert users or conventional users who explore in detailed manner the functionalities of the interactive system can carry out these tests. The results of the interaction of the users are consigned in a test that accounts for the usability of the software system evaluated [9]. In this work an application for the execution of usability evaluations based on the inspection of Torres-Burriel heuristics is proposed.

3.2 Heuristic Test of Torres-Burriel Torres-Burriel propose a set of heuristics to evaluate websites, which contain the most relevant aspects to keep in mind at usability level, where for each heuristic presented a set of criteria to evaluate is provided within a range from 1 to 5 [4]. These are posed in terms of questions. Said heuristics retake some of the elements proposed in the principles by Nielsen, so that they are distributed in 11 heuristics, as shown in Table 1 [10]. Table 1. Torres-Burriel Heuristics. Heuristic Description Aspects related with the objectives of the website, correct and easy remembrance of the external and internal URLs, H1: General aspects adequate organization and structure of the contents in addition to using clear language, concise and familiar to the user Evaluates aspects related with site identity, as well as the H2: Identification and mechanisms to contact the company , protection of personal information data and authorship of contents. Refers to whether the website speaks the same language as H3: Language and drafting the user. Expresses elements related with the meaning, adequate use H4: Labeling and familiarity of content labels. H5: Structure and Elements referring to adequate organization, disposition, and navigation structure of information, besides the navigation of the site. Aspects related with the distribution and appearance of the H6: Page layout navigation elements and contents in the interface. Evaluates aspects referring to the search system implemented H7: Search in the website related to ease of access, as well as elements related with the effectiveness of searches. Aspects related with the adjustment level of the website H8: M ultimedia elements portal’s multimedia contents. Aspects related with the help available to users during their H9: Help navigation on the site. Aspects related with any user’s ease of use of web pages, which evaluate elements with respect to font size, type, and H10: Accessibility color, weight of page, compatibility with different browsers and elements that permit navigating comfortably. Aspects related with the user’s freedom to undo or redo actions in the navigation, as well as timely and clear H11: Control and feedback information provided to them in their interaction with the website portal.

In keeping with the aforementioned, Table 2 introduces the amount of criteria associated to each of the heuristics presented in Table 1. As mentioned, said criteria are presented to the evaluator in terms of questions.

Table 2. Criteria per heuristic of the test by Torres-Burriel. Heuristic Number of criteria H1: General aspects 9 H2: Identity and information 7 H3: Language and drafting 4 H4: Labeling 5 H5: Structures and navigation 11 H6: Page layout 7 H7: Search 6 H8: M ultimedia elements 4 H9: Help 4 H10: Accessibility 8 H11: Control and feedback 6

Tool proposed

The tool introduced in this work stemmed from a set of usability criteria associated with each of the 11 heuristics proposed by Torres-Burriel (Table 2). These criteria are presented in terms of questions and can be evaluated in a range from 0 to 5, with 0 being the absence of the criterion in the website portal and 5 being maximum compliance of the criterion within the portal. At design level, the tool is based on tabs, each of which comprises a heuristic and its associated criteria. According to the aforementioned, the tool proposed is constituted by 15 tabs organized in the following manner: the first tab consigns the general data of the usability evaluation; tabs 2 to 12 include each of the usability heuristics and their associated criteria; tab 13 presents a table with the average obtained in each of the heuristics evaluated and the general average of the heuristics; tab 14 includes a graphic with the averages of each principle; finally, tab 15 presents a graphic with the percentage of compliance of each of the heuristics by Torres-Burriel. A flow diagram illustrating the functioning of the tool is presented in Figure 2. As observed in Figure 2, the tool allows reviewers to evaluate one-to-one the different criteria associated with the standard’s accessibility principles to, subsequently, generate statistics of the mean value per accessibility principle and of the percentage of compliance of each principle of the application to inspect. In summary, according to Figure 2, the functionality of the tabs is grouped into three: the first instance is related with filling out the test data (first tab); the second instance includes the evaluation of the criteria of the eleven heuristics (tabs 2 to 12); and the third instance contains that related with the analysis of the statistical results and graphics of the evaluation conducted. The results obtained seek to guide evaluators on those heuristics that the website portal evaluated is not complying fully .

Figure 2. Functional diagram of the tool proposed.

To comply with the idea presented in Figure 2, a tool was constructed in Java language through tabs and with different tables per tab, using JTabbedPane (Swing class) and JTable components (Figure 3). Regarding the generation of the analysis graphics, the tool uses the JFreeChart library [11], which in this case permits obtaining bar graphs on the average of the evaluations per principle and their percentage of compliance.

Figure 3. Heuristics evaluation tabs.

As shown in Figure 3, each of the first four tabs has a table associated with three columns: Id, Criterion, and Value. The “Id” column permits distinguishing the different criteria of each heuristic principle, using the first letter of the heuristic and a number. In that regard, the first criterion of the “General” heuristic has the code of G1. The “Criterion” column includes a question that permits evaluators to verify

compliance of the criterion of a given heuristic. Finally, the “Value” column comprises a set of values evaluators can assign to each criterion, depending its level of compliance in the website portal. The scale of values varies between 0 and 5, including the option of N/A, which can be used in case the criterion does not apply for the website portal in question. The 0 value in the scale corresponds to no application of the criterion in the website portal, while 5 corresponds to full application of the criterion in the website portal. Upon filling out basic evaluation data (name of evaluator, URL of site to evaluate, and evaluation description) and ending the inspection by the different heuristics and usability criteria proposed by Torres-Burriel, it is possible to see in the “Results” tab the average obtained in each of the heuristics and general average obtained in the usability inspection (Figure 4). The average obtained within each of the heuristics permits evaluators to identify aspects of the web application that do not comply adequately with the usability criteria.

Figure 4. Heuristics results tab.

Upon generating the heuristic averages and the general average of the evaluation, in the “Graphic – Average” tab it is possible to graphically visualize the average obtained by each of the usability heuristics for websites proposed by Torres-Burriel (Figure 5).

Figure 5. “Graphic – Average” tab.

Likewise, in the “Graphic – Percentage” tab, it is possible to visualize the graphic with the percentage of compliance of each of the usability heuristics for websites proposed by Torres-Burriel (Figure 6).

Figure 6. “Graphic - Percentage” tab.

Case study

This work implements the inspection tool to facilitate application and analysis of the results of the heuristic test by Torres-Burriel and validate the heuristic test by TorresBurriel, which proposes in its different tabs a set of criteria associated with the eleven heuristics website portals and applications must comply according to Torres-Burriel. The inspection tool was applied on the website portal of the Puebla Secretary of Education (http://sep.puebla.gob.mx/) in order to evaluate the usability of it. The Puebla Secretary of Education seeks to articulate public policy to place at their center children, youth, and adults through an educational service with equity, quality, pertinence, and coverage that permits developing programs and actions tending to improve and broaden learning opportunities with holistic, co-responsible, and participative attention strategies [8] (Figure 7). In this sense, considering the diversity of the target public, it is necessary to guarantee inasmuch as possible compliance of certain basic usability guidelines.

Figure 7. Website portal of the Puebla Secretary of Education.

5.1 Results of the inspection The evaluation method used in this article was the inspection method, which took into consideration each of the usability criteria included in the usability evaluation tool, which – in turn – are associated with the different guidelines of the eleven heuristics of the usability test by Torres -Burriel. The evaluation was conducted by a group of three experts in the area of human computer interaction, who inspected the compliance of Torres-Burriel's heuristics on the portal of the Puebla Secretary of Education. Thus, figure 8 presents the average of the evaluations assigned by the evaluators to the criteria of the "Identity and Information" principle.

Figure 8. Evaluation by experts.

Upon finishing the inspection of the Puebla Secretary of Education web page and considering the different criteria by Torres -Burriel, Figure 9 presents one of the graphic results obtained by the usability inspection tool for web pages, which shows the average of the evaluations obtained for each of the eleven heuristics evaluated.

Figure 9. Average graphic of the heuristics.

It may be noted that the portal is poorly usable, considering the inspection results, given that only one of its principles (page layout) reaches an average of 4 in a scale from 1 to 5, remaining < 4 in the rest of the heuristics evaluated, which in mathematical terms means that only one of the heuristics achieves 80% compliance. Likewise, it is possible to analyze that five of the eleven heuristics evaluated have a percentage of compliance > 70%, while four of the heuristics have a percentage of compliance < 60%. The heuristics with the lowest percentage of compliance are “Control and Feedback” and “Search” with less than 55%. Similarly, the percentage of the “Accessibility” heuristic is between 55% and 60%, considered low, bearing in mind the variety of the public accessing the government portal evaluated.

Fig. 10. Compliance of heuristics of web usability .

Finally, the general usability average for the website portal is 3.31, corresponding to 66.3% compliance (Table 3). It is convenient to consider different usability criteria for the portal to have a value > 80%, which can be deemed acceptable. Thus, the inspection of the Puebla Secretary of Education web page helps to generate a set of recommendations keeping in mind the averages per heuristics (Table 3), which seek to be of help to generate possible improvements, considering the number and diversity of users accessing these types of government portals. Table 3. Average per heuristic. Heuristics H1: General aspects H2: Identity e information H3: Language and drafting H4: Labeling H5: Structures and navigation H6: Page layout H7: Search H8: M ultimedia elements H9: Help H10: Accessibility H11: Control and feedback General Average

Average 2.88 3.57 3.75 3.6 3.36 4.0 2.66 3.5 3.75 2.87 2.5 3.31

5.2 Recommendati ons Pursuant to the inspection conducted of the portal of the Puebla Secretary of Education, this section presents a set of recommendations (obtained by the test coordinator) for the eleven heuristics evaluated and considering the criteria obtaining the lowest value through the assessment of the evaluators (see Table 4). Table 4. Recommendations of usability generated.

Heuristic

H1: General aspects

H2: Identity and Information

H5: Structure and Navigation

H7: Search

H8: Multimedia elements H10: Accessibility

H11: Control and Feedback

Recommendati on ● Recommends having a URL that is easy to understand and remember for users ● Recommends using a design aimed at users that permit greater efficiency in user navigation. ● Recommends for the portal to have a periodic update regarding the information presented. ● Recommends for the portal to have a clear and simple slogan. ● Recommends for the portal to have an additional link with more information about it. ● Recommends improving the functionality of user orientation in its different views. ● Images within the portal should have clear titles, indicating contents and to what these refer. ● Recommends making search options more recognizable within the website portal. ● Recommends providing assistance when offering results to the consultations obtained within the portal. ● Recommends diminishing cyclical animations in the portal. ● Recommends improving the adaptation of the portal to different access devices. ● The portal should manage different notifications upon any possible failure, like an error when connecting to the internet. ● Recommends improving response times in the different portal options.

Conclusions and future work

The principal contribution of this research is to propose a tool for the usability inspection of web applications, keeping in mind the usability test by Torres-Burriel. This tool seeks to support developers, designers, and programmers in evaluating websites to construct more inclusive and usable websites. The principal contribution of the tool proposed is the possibility of evaluating each of the criteria associated with the heuristics by Torres-Burriel, as well as

automatically generating the graphic of averages and the graphic of percentage of compliance of the heuristics by Torres-Burriel. The above indicators are intended to be helpful in terms of improving the overall design of the evaluated portal and the increase in its number of potential users. It is worth highlighting that portals, like that of the Puebla Secretary of Education, when aimed at a broad and diverse public, must pay great attention to the criteria of the “Accessibility” heuristic to enable inclusion and access to different users, independent of the context of use. The Java language proved adequate to implement inspection tools, bearing in mind the tab management component (JTabbedPane), the table management component (JTable), and the API to generate JFreeChart statistical graphics. In reflection, it is important to highlight that given that many governmental website portals have been created by using content managers, the websites constructed through these bring along diverse usability problems that do not necessarily depend on the webpage designer, given that these managers were mostly conceived without considering usability criteria. For future work derived from this research, the tool should include the possibility of generating automatic recommendations arising from the statistic results and graphics obtained in the test.

References 1. Jiménez, K., Pérez, D., Rengifo, S.: Propuesta de una evaluación heurística de accesibilidad para sitios web basada NTC 5854. Obras Colectivas en Ciencias de la Computación, p.p. 53-68, Cali-Colombia (2018). 2. Sánchez, W.: La usabilidad en Ingeniería de Software: definición y características. Revista de Ingeniería e Innovación de la Facultad de Ingeniería - Universidad Don Bosco 1(2), 7-21 (2011). 3. Suarez, M .C.: Sistema de evaluación de la usabilidad web orientado al usuario y basado en la determinación de tareas críticas. Editorial Universidad de Oviedo, Oviedo-España (2011). 4. Torres-Burriel Estudio: Test heurístico de Torres-Burriel, http://www.torresburriel.com/weblog/2008/11/28/plantilla-para-hacer-analisisheuristicos-de-usabilidad/ 5. Sánchez, J., Zapata, C., Jiménez, J.: Evaluación heurística de usabilidad de software para facilitar el uso del computador a personas en situación de discapacidad motriz. Revista EIA 14(27), p.p. 63-72, Envigado-Colombia. 6. Covella, G.: M edición y evaluación de calidad en uso de aplicaciones web. Editorial Universidad Nacional de la Plata, La Plata-Argentina (2005). 7. Idrobo, C., Vidal, M ., Chanchí, G.E.: Guía de accesibilidad para el diseño e implementación de sitios web teniendo en cuenta la norma NTC 5854. In Solano, A. & Ordoñez, H. (Eds), Advances in Computing, pp. 572-585, Colombia (2017). 8. Yusef, H.: Introducción a la Usabilidad. Revista No Solo Usabilidad 1, (2002). 9. Gómez, M .: Test de usabilidad en entornos de Realidad Virtual. Revista No Solo Usabilidad 17, (2018).

10. Hurtado S., Pimentel, J., Chanchí, G.E. Estudio comparat ivo de métodos de evaluación de usabilidad para sitios web. Desarrollo e Innovación en Ingeniería, p.p. 129-136, M edellín-Colombia (2018). 11. Librería JFreeChart, http://www.jfree.org/jfreechart/

Expectation of ICT development in Mexico within the framework of Latin America Adrián Sánchez1 , Agustin Lagunes 2 , Carlos A. Torres 3 , Juan C. Judikis 4 , Francisco López5 1

Universidad Veracruzana, Boca del Río 94294, Veracruz, M exico Universidad Veracruzana, Ixtaczoquitlán 94452, Veracruz, M exico 3 Universidad Veracruzana, Veracruz 91780, Veracruz, M exico 4 Universidad de M agallanes, Punta Arenas 01855, Chile 5 Universidad Veracruzana, Boca del Río 94294, Veracruz, M exico adrianvidal2000@hotmail.com, aglagunes@uv.mx, ctorres@uv.mx, juan.judikis@umag.cl, frlopez@uv.mx 2

Abstract. Summary. Since the 90's society has changed with the emergence of the Internet and the development of Information and Communication Technologies, thanks to the constant evolution of electronic devices, which allow easy data transfer and access to information. Unfortunately, they are not impacting in the same way in all countries, due to the economic and social lags that exist. This work aims to estimate the development expectations of ICT in M exico and Latin America, which allow the decision-making to ensure the progress of citizens. For this, through a Logistics Function, the projection of the development of ICTs in Latin American countries was estimated at 30 years, based on the IDI data for the year 2007-2016 provided by the ITU. After having characterized the parameters of the equation for each country that achieve the lowest possible error, with values below 0.03% as a result of the 30year projection, it is estimated that M exico rises only one position in the ranking, while other countries have a higher growth expectation such as Bolivia, Paraguay, and Costa Rica, which are expected to rise 6, 4 and 3 positions, respectively. Knowing these trends allow each country to modify the policies implemented so far, identify the areas of opportunity or replicate what has been done by those with the best growth expectations, so that the trend that has been estimated up to now will be favorably changed, favoring the inclusion of all the citizens in the Knowledge Society. Introduction Keywords: Information and Communication Technologies, Projection, Latin America.

Introduction

Since the 90's, society has changed substantially with the appearance of the internet, as well as with the development of Information and Communication Technologies . This is due to the permanent evolution of electronic devices, with which it is increasingly easy to transfer data and access information (Medina, Lagunes, & Torres, 2018). During an OECD meeting in 2009, the importance of developing the competencies of

students in the classroom using ICT was determined, since it is through them that they achieve socialization today (Navarro, Cuevas, & Martínez, 2017). Unfortunately, the results of the research work of Torres and Valencia (2013) establish that students have poor knowledge of ICT at a basic level, in the public primary schools of Veracruz, Mexico. The above, despite having a better ability in the use of the Internet, since they mainly apply it for issues unrelated to education. This could be a consequence of the evident lack of infrastructure needed to incorporate the use of ICTs present in schools and the Internet in the classroom. As a result of the work of Medina, Lagunes and Torres (2018), favorable opinions were obtained in the perception of high school students regarding the adequacy of ICT in the classroom. They consider that it allows them to achieve a higher level of understanding of the topic seen in class. However, they also believe that to achieve this, it is necessary that the technological equipment is operating in optimal conditions. Supporting previous research, Torres and Kiss (2016) gathered the feelings of university students in Hungary and Mexico regarding the academic use of ICT. They obtained satisfactory results in both cases, even though the countries live different realities. Furthermore, they expressed the perception of ICTs as a means of learning and social communication; as well as a means to obtain information, the latter being the main use employed. It is then apparent that the degree of success of the educational use of ICT is linked to having the appropriate infrastructure in the classroom, which is a common problem for developing countries. An example of this would be Mexico, where the degree of poverty in certain sectors is of considerable levels . However, based on the results of Gabriel y Sánchez's study (2017) regarding the relationship between ICT and poverty in Latin America, they conclude that inequality and poverty in Latin America have been reduced in the last twenty years. In addition, as a consequence, there has been an increase in the use of ICT and the internet. Thanks to the shelter of technological growth and productivity that originates in citizens when using ICT in their daily lives, as a consequence of public policies implemented on purpose (Gabriel & Sánchez, 2017). Therefore, it is of great interest to verify the current state of ICTs in Mexico and America, as well as to estimate their development expectations that give the direction of their course. These will allow adequate decision-making, that assure the progress of the citizens of the knowledge society.

Current State of ICTs

The International Telecommunication Union (ITU) has a tool called the ICT Development Index (IDI or IDT) which is used to measure the level and evolution of ICTs over time. This tool takes into account the situation in developed and developing countries. The IDI is an index made up of 11 indicators that cover access, use and skills in the field of ICTs (ITU, 2011). Its range of possible values ranges from 0 to 10 (ITU, 2016). Based on the results of the IDI published by the ITU, it is observed that Mexico went from place 76 to 92 worldwide from 2007 to 2016, among 159 and 176 countries, respectively. In the American context, this represents a move from position 14

to position 19, as shown by the data in table 1 (ITU, 2010; ITU, 2011; ITU, 2012; ITU, 2013; ITU, 2014; ITU, 2015; ITU, 2016). As can be seen, there are countries in the Americas that are better positioned than Mexico, such as the United States, Canada, Argentina and Brazil. It is important to consider that Mexico has descended 16 places worldwide and 5 in Latin America. Likewise, it is noteworthy that it is below countries with lower eco nomic development, according to data from the World Bank (2016a), despite the fact that some of them have recently been added. On the other hand, Mexico is in the 15th place in the classification of economic development of the countries in Latin America, according to the Gross National Index (GNI). While countries like Dominica, Suriname, and St. Vincent and the Grenadines, are in positions 18, 19 and 20, respectively. However, despite this, they have a better ICT development than Mexico, as shown in table 1. Table 1. Ranking of ICT in America from 2007-2016. Adapted from (ITU, 2016). # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Country United States Canada St. Kitts and Nevis Barbados Uruguay Argentina Chile Costa Rica Brazil Bahamas T rinidad & T obago Dominica Grenada Antigua y Barbuda St. Vincent and the Grenadines Venezuela Colombia Suriname Mexico Panama St. Lucia Ecuador Jamaica Peru Dominican Rep. Paraguay Bolivia

2007 6.33 6.30 3.96 4.13 3.99 3.31 3.49 3.61

4.10 3.33 3.27 3.03 3.39 2.73 3.52 3.03 2.73 2.46 2.39

2008 2010 2011 2012 2013 2015 2016 6.55 6.55 7.35 7.90 8.02 8.06 8.17 6.42 6.42 7.14 7.37 7.62 7.55 7.62 6.23 7.21 5.47 5.47 6.87 7.18 4.21 4.21 5.38 5.92 6.32 6.44 6.79 4.16 4.16 5.06 5.58 5.8 6.21 6.52 4.14 4.14 5.08 5.68 5.92 6.11 6.35 3.45 3.45 4.47 5.64 5.92 6.03 6.30 3.72 3.72 4.59 5.16 5.50 5.72 5.99 5.80 5.98 3.99 3.99 4.54 4.99 5.29 5.48 5.76 5.14 5.71 4.97 5.43 5.41 5.38 4.59 4.59 4.71 5.04 5.17 5.07 5.32 3.73 3.73 4.00 4.68 4.81 5.22 5.27 3.39 3.39 3.89 4.61 4.95 4.98 5.16 4.89 5.09 3.26 3.26 3.78 4.07 4.29 4.45 4.87 3.52 3.52 4.38 4.69 4.75 4.63 4.87 4.68 4.85 2.87 2.87 3.73 4.28 4.56 4.54 4.56 3.06 3.06 3.54 4.01 4.26 4.20 4.52 3.12 3.12 3.58 3.92 4.00 4.23 4.42 2.84 2.84 3.36 3.78 4.06 4.02 4.30 2.66 2.66 3.10 3.56 3.71 3.88 4.08 2.54 2.54 3.08 3.52 3.78 3.49 4.02

The progress obtained is not enough, based on the data provided by the ITU. The perspective of the Knowledge Society in Mexico is not very encouraging, since in recent years it has been experiencing low growth in access, use and skills of ICT. Although this data represents a negative effect on the Information Society, it indirectly impacts the development of the Knowledge Society. Undoubtedly, there are many challenges and actions to take at different levels in order to establish a turning point that allows to change the trend . in such a way that the Knowledge Society can advance and with this, achieve general development and welfare for the country.

Methodology

The classification of the development of ICT is based on the IDI, whose range of values oscillates between 0 and 10. As a consequence, to model its projection a function that is bounded must be used. That is, a function that allows representing a gradual growth, but whose growth rate decreases when approaching its maximum value. Given its nature, the exponential function could model that growth, with the limitation of not being able to adjust to an upper limit. On the other hand, the logarithmic function could be adjusted to an upper limit, but to reserves of not being able to be set to a lower limit of 0 (Larson & Falvo, 2012). However, one of the functions used to model the growth of this kind is the Logistics Function or logistic growth model (Ulloa & Rodríguez, 2010). This being the one selected to perform the projection.

Fig. 1. Projection of a Logistic Equation to a maximum value of 100.

3.1 Logistic Function The Logistic Function is an equation proposed by Verhulst to model the growth of any population with a certain load capacity (Bonilla & Zaragoza, 2011). Figure 1 shows an example of the function that tends to a maximum value of 100. This function is described by the logistic equation: (1) Where:  t = Time where you want to calculate the projection (independent variable).  P0 = Initial value.  P(t) = Value projected at time t (dependent variable).  a, b = Coefficients of the growth rate.  Pmax = Maximum value of the projection determined by the ratio a/b.

With the logistic equation and the ratio of coefficients that limit the growth of the projection to its maximum value, we can obtain the value of b though the equation. Subsequently, it must be adjusted to a known point in the data curve to be projected. (2)

3.2 Procedure It should be noted that only the countries with data for the entire period fro m 2007 to 2016. Well, we look for uniformity in the time interval to be investigated, so 8 of the 27 countries in Latin America were ignored. Therefore, the classification to be analyzed will be as is shown in table 2. Table 2. M odified ranking of ICTs in America from 2007-2016. Adapted from (ITU, 2016). # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Country United States Canada Uruguay Argentina Chile Costa Rica Brazil T rinidad & T obago St. Vincent and the Grenadines Venezuela Colombia Mexico Panama Ecuador Jamaica Peru Dominican Rep. Paraguay Bolivia

2007 6.33 6.30 3.96 4.13 3.99 3.31 3.49 3.61 4.10 3.33 3.27 3.03 3.39 2.73 3.52 3.03 2.73 2.46 2.39

2008 2010 2011 6.55 7.11 7.35 6.42 6.87 7.14 4.21 4.89 5.38 4.16 4.72 5.06 4.14 4.63 5.08 3.45 3.94 4.47 3.72 4.17 4.59 3.99 4.42 4.54 4.59 4.58 4.71 3.73 3.78 4.00 3.39 3.73 3.89 3.26 3.60 3.78 3.52 4.21 4.38 2.87 3.41 3.73 3.06 3.42 3.54 3.12 3.43 3.58 2.84 3.19 3.36 2.66 2.94 3.10 2.54 2.93 3.08

2012 7.90 7.37 5.92 5.58 5.68 5.64 5.16 4.99 5.04 4.68 4.61 4.07 4.69 4.28 4.01 3.92 3.78 3.56 3.52

2013 2015 2016 8.02 8.06 8.17 7.62 7.55 7.62 6.32 6.44 6.79 5.80 6.21 6.52 5.92 6.11 6.35 5.92 6.03 6.30 5.50 5.72 5.99 5.29 5.48 5.76 5.17 5.07 5.32 4.81 5.22 5.27 4.95 4.98 5.16 4.29 4.45 4.87 4.75 4.63 4.87 4.56 4.54 4.56 4.26 4.20 4.52 4,00 4.23 4.42 4.06 4.02 4.30 3.71 3.88 4.08 3.78 3.49 4.02

In order to estimate the projection, first, the value of b for each country was calculated. The foregoing was aimed at adequately adjusting to the data for the period 2007-2016 and showing the lowest possible relative error. This characterization of curves was carried out under the criteria listed below: The initial value P0 for each country was taken as the IDI for the year 2007; the value of t was set at 9, being the 9th year after 2007; P(t) takes the IDI value of the year 2016, and; Pmax is fixed for everyone with a value of 10. Finally, after establishing the accuracy of the logistics function for each country to a degree of certainty, we proceeded to make the projections that will be presented and analyzed below.

Analysis of results

Once the b factor for each country was obtained, the value of a was calculated to be 10 times b. This in order to estimate the IDI with the logistic equation for each year from 2007 to 2016 and compare it with the real data. With this, it is possible to measure the relative error of the calculated value that validates the certainty of the equations. The results obtained are shown in table 3. Table 3. Estimation of the error of the method to be used in the projection. # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Country 2007 United States 0.00% Canada 0.00% Uruguay 0.00% Argentina 0.00% Chile 0.00% Costa Rica 0.00% Brazil 0.00% T rinidad & T obago 0.00% St. Vincent and the G.0.00% Venezuela 0.00% Colombia 0.00% Mexico 0.00% Panama 0.00% Ecuador 0.00% Jamaica 0.00% Peru 0.00% Dominican Rep. 0.00% Paraguay 0.00% Bolivia 0.00%

2008 2010 2011 2012 2013 -0.33% 1.12% 1.41% 5.67% 4.65% -0.65% 1.38% 2.99% 4.03% 5.30% -1.55% -0.64% 2.48% 5.91% 6.83% -5.68% -4.61% -2.96% 1.79% 0.89% -2.64% -3.21% 0.67% 6.46% 5.77% -4.95% -8.51% -3.22% 12.12% 10.49% -0.89% -3.13% 0.18% 5.70% 6.35% 3.81% 2.52% -0.20% 3.97% 4.81% 7.78% 1.70% 1.52% 5.27% 5.00% 5.33% -4.46% -4.09% 6.37% 4.31% -2.21% -3.76% -4.87% 6.90% 8.95% 1.34% -0.17% -0.76% 1.34% 1.50% -0.71% 8.22% 8.05% 10.60% 8.20% -1.42% 3.45% 6.37% 13.61% 14.30% -18.52% -12.40% -11.71%-1.40% 1.91% -1.69% -1.11% -1.13% 3.68% 1.66% -1.64% -0.79% -0.80% 5.75% 7.82% 1.60% -0.37% -0.86% 7.08% 5.81% -0.28% 1.66% 0.72% 7.95% 9.31%

2015 0.67% 0.81% -0.95% -0.96% 0.20% 0.95% 0.08% -0.75% -2.27% 3.29% 0.76% -4.58% -1.50% 4.40% -4.91% -0.61% -2.29% -0.02% -9.42%

2016 0.01% 0.02% -0.02% -0.01% 0.02% 0.03% -0.03% -0.02% -0.03% -0.02% 0.00% 0.00% 0.00% -0.01% -0.02% 0.01% -0.03% 0.00% 0.00%

At the end of the estimation of the IDI for each country, a maximum relative error of 0.03% is obtained. This percentage of error provides a high degree of certainty to the equations obtained to calculate the projections for the next 30 years (table 4). From the results shown in table 4, interesting projections are obtained in some countries. Some examples are marginal estimates, such as the United States, Canada, Costa Rica, among others . The above is presented in figure 2 along with the projection corresponding to Mexico, what allows to establish a clearer comparative. Figure 2 shows that in the next 10 years, it is estimated that Mexico will climb only one position in the classification. As a result, it will surpass Saint Vincent and the Grenadines, going from 12th to 11th place and will manage to maintain this position for the rest of the 30-year period. On the other hand, there are countries that have a higher growth projection in the same period. Some examples are Bolivia, Paraguay and Costa Rica, which are expected to move up 6, 4 and 3 positions, respectively. Bolivia stands out among them, climbing from position 19 to 13, while Paraguay climbs from position 18 to 14. For its part, Costa Rica manages to reach the third position.

31 Table 4. ICT projection in America at 30 years. #

Country

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

United States Canada Uruguay Argentina Chile Costa Rica Brazil T rinidad & T obago St. Vincent and the Grenadines Venezuela Colombia Mexico Panama Ecuador Jamaica Peru Dominican Rep. Paraguay Bolivia

T o 5 T o 10 years T o 15 T o 20 years T o 25 T o 30 years years years years 8.83 9.28 9.56 9.74 9.84 9.91 8.19 8.66 9.01 9.28 9.49 9.63 8.02 8.86 9.37 9.66 9.82 9.91 7.64 8.48 9.06 9.43 9.66 9.80 7.48 8.35 8.96 9.37 9.62 9.77 7.72 8.70 9.30 9.64 9.81 9.91 7.25 8.24 8.92 9.36 9.63 9.79 6.89 7.83 8.54 9.05 9.40 9.62 5.99 6.63 7.21 7.73 8.17 8.55 6.35 7.31 8.10 8.69 9.12 9.42 6.23 7.19 7.98 8.59 9.04 9.36 5.94 6.93 7.77 8.43 8.93 9.28 5.72 6.53 7.26 7.89 8.40 8.81 5.67 6.72 7.62 8.33 8.86 9.24 5.10 5.68 6.24 6.76 7.25 7.69 5.25 6.07 6.83 7.50 8.07 8.54 5.27 6.21 7.07 7.81 8.40 8.85 5.11 6.13 7.06 7.84 8.46 8.93 5.06 6.10 7.05 7.85 8.48 8.95

Furthermore, there are countries with disappointing projections, given that a great setback is expected in their classification. Some examples are San Vincent and the Grenadines, Canada and Jamaica, which will descend 8, 5 and 4 positions, respectively.

Fig. 2. M arginal projections of the development of ICTs in America in a span of 30 years.

It should be recognized that 8 countries added to this clas sification in recent years were omitted in the projection. The positions were modified, and these estimates could change when adding them. As a consequence, these positions are not as relevant as their increasing or decreasing status, given that it reflects the relative growth that ICTs will have in each country.

Conclusions

The methodology based on the Logistics Function made it possible to establish an estimate of the projection development, access, use and aptitudes of ICTs in American countries, in order to establish a comparison criterion that encourages them to improve. As a result of the projection made, it is expected that in the next 30 years, Mexico will have slow progress in the development of ICT. In counterpart, there are other countries with better growth expectations, such as Bolivia, Paraguay, and Costa Rica. In 2005, at the World Summit on the Information Society (WSIS), agreements were established to promote the development of the Information Society focused on the person. Among the agreements was to encourage the use of ICT, through the improvement of Internet and mobile telephony services (ITU, 2015). It is interesting to know how much these commitments have been fulfilled in the different countries and, in turn, to see how this is reflected in their development. The first point of comparison is obtained from data corresponding to the development of mobile telephony (World Bank, 2016b). Where in 2016, the number of mobile phone subscriptions per 100 people in Costa Rica, Paraguay and Boliv ia, were 171, 111 and 92, respectively. On the other hand, in Mexico, the figure of 87 was barely reached, which represents less than one subscription per person, while in Costa Rica it is more than double. Regarding the improvement of internet services, we can rely on the World Bank figures for 2016 regarding the number of broadband subscriptions and internet servers that are secure. The data for every 100 people reveals that in Costa Rica there are 13 of them who subscribe to this service, while in Mexico only 12. There seems to be no such marked difference, however, it is with respect to San Vicente and the Grenadines, where there are about 20, despite having less development in ICT (World Bank, 2016c).On the other hand, in Mexico for every million people there are 115 internet servers that are safe. Meanwhile, in Costa Rica there are more than three times, reaching the figure of 494 (World Bank, 2016d).What reflects a better infrastructure that favors the use of ICT and with it the development of the Info rmation Society. Although, the above shows that the commitments to improve the development of ICT have not been met, in news issues the news is not more encouraging. In 2010, while the government of Mexico invested only 5% of GDP in education, in Bolivia about 8% was spent (World, 2016e). Finally, until now the important role of the government's plans in turn to improve the infrastructure that fosters the development of ICT has been raised. However, educational institutions must also do their part, investing in technological means that help students apply ICT in the academic field, thus developing their digital skills for research. Consequently, it is of utmost importance that both entities, from their sphere of interference, become aware of the importance of their commitment to achieving the objectives set. So that they contribute to the implementation of ICT and internet in different sectors, thus promoting the inclusion of all citizens in the Knowledge Society.

References 1.

10. 11.

12.

13.

Bonilla, A., & Zaragoza, J. (2011). La explotación intensiva y extensiva de la naturaleza. Ideas sobre su posible modelación matemática. Dimensión económica,Instituto de Investigaciones Económicas, UNAM, 2(6), 1-1. Retrieved april 12, 2018, from https://rde.iiec.unam.mx/re Gabriel, J., & Sánchez, A. (2017). ICTs and Poverty in Latin America. Íconos. Revista de Ciencias Sociales(57), 141-160. doi:http://dx.doi.org/10.17141/iconos.57.2017.2095 ITU. (2010). Measuring the Information Society Report 2010. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from https://www.itu.int/en/ITU-D/Statistics/Documents/publications/mis2010/M IS_2010 ITU. (2011). Measuring the Information Society Report 2011. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from https://www.itu.int/ITUD/ict/publications/idi/material/2011/M IS_2011_without_annex_5.pdf ITU. (2012). Measuring the Information Society Report 2012. Luxemburg: International Telecommunication Union. Retrieved september 11, 2017, from https://www.itu.int/en/ITUD/Statistics/Documents/publications/mis2012/M IS2012_without_Annex_4.pdf ITU. (2013). Measuring the Information Society Report 2013. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from https://www.itu.int/en/ITUD/Statistics/Documents/publications/mis2013/M IS2013_without_Annex_4.pdf ITU. (2014). Measuring the Information Society Report 2014. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from https://www.itu.int/en/ITUD/Statistics/Documents/publications/mis2014/M IS2014_without_Annex_4.pdf ITU. (2015). Measuring the Information Society Report 2015. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from https://www.itu.int/en/ITUD/Statistics/Documents/publications/misr2015/M ISR2015-ES-E.pdf ITU. (2016). Measuring the Information Society Report 2016. Luxemburg: International Telecommunication Union. Retrieved september 10, 2017, from http://www.itu.int/en/ITU-D/Statistics/Documents/publications/misr2016/M ISR2016w4.pdf Larson, R., & Falvo, D. C. (2012). Precálculo. M éxico, D. F.: Cengage Learning. M edina, H., Lagunes, A., & Torres, C. A. (2018). Perceptions of Secondary School Students about the use of ICT in their Science Class. Información Tecnológica, 29(4), 259-266. Navarro, L. A., Cuevas, O., & M artínez, J. (2017). A M eta-Analysis of ICT-Based Education in M exico and Latin America. Revista Electrónica de Investigación Educativa, 19(1), 10-19. Torres, C. A., & Kiss, G. (2016). Perceptions of Students towards ICT Competencies at the University. Informatics in Education, 15(2), 319-338.

34 14. 15.

16. 17.

18. 19. 20.

Torres, C. A., & Valencia, L. A. (2013). Uso de las TIC e internet dentro y fuera del aula. Revista de Innovación Educativa, 5(1), 1-1. Ulloa, J. T., & Rodríguez, J. A. (2010). El modelo logístico: Una alternativa para el estudio del crecimiento poblacional de organismos. REDVET. Revista electrónica de Veterinaria, 11(3). Retrieved abril 15, 2019, from http://www.redalyc.org/articulo.oa?id=63613123024 World Bank. (2016a). Latin America & Caribbean. Retrieved september 14, 2018, from https://data.worldbank.org/region/latin-america-and-caribbean?view=chart World Bank. (2016b). Mobile cellular subscriptions (per 100 people). Retrieved M ay 16, 2019, from World Bank: https://data.worldbank.org/indicator/IT.CEL.SETS.P2?view=chart World Bank. (2016c). Fixed broadband subscriptions (per 100 people). Retrieved M ay 16, 2019, from https://data.worldbank.org/indicator/IT.NET.BBND.P2 World Bank. (2016d). Secure Internet servers (per 1 million people). Retrieved M ay 16, 2019, from https://data.worldbank.org/indicator/IT.NET.SECR.P6?view=chart World, B. (2016e). Government expenditure on education, total (% of GDP). Retrieved M ay 16, 2019, from https://data.worldbank.org/indicator/SE.XPD.TOTL.GD.ZS?view=chart

Extracting and visualizing a new classification system for Colombia’s National Administrative Department of Statistics. A visual analytics framework case study Pierre Raimbaud 1,2 , Jaime Camilo Espitia Castillo 1 , John A. Guerra Gomez3 1

Systems and Computing Engineering, Imagine Group, Universidad de los Andes, Bogota, D.C., Colombia p.raimbaud@uniandes.edu.co 2 LiSPEN, Arts et M étiers, Institut Image, Chalon-sur-Saone, France pierre.raimbaud@ensam.eu 3 Northeastern University, San Jose, California Abstract. In a world filled with data, it is expected for a nation to take decisions informed by data. However, countries need to first collect and publish such data in a way meaningful for both citizens and policy makers. A good thematic classification could be instrumental in helping users to navigate and find the right resources on a rich data repository, such as the one collected by the DANE (Departamento Administrativo Nacional de Estadística, i.e. the Colombia’s National Administrative Dep artment of Statistics). The Visual Analytics Framework is a methodology for conducting visual analysis developed by T. M unzner et al.1 that could help with this task. This paper presents a case study applying such framework conducted to help the DANE to better visualize their data repository, and also to understand it better by using another classification extracted from its metadata. It describes the three main analysis tasks identified and the proposed solutions. Usability testing results during the process helped to correct the visualizations and make them adapted to decision-making. Finally, we explained the collection of insights generated from them. Keywords: Visual Analytics, Data Repositories, Open Data.

Introduction

The DANE (Departamento Administrativo Nacional de Estadística, i.e. the Colombia’s National Administrative Department of Statistics) is the Colombian public organization responsible for planning, collecting, analyzing and distributing the country’s national statistics. The total amount of data that this public institution owns is one of the largest in the country (among government institutions), due to the fact that it periodically gathers information about all the major topics, from population statistics, passing by technological literacy, to public access to services, among many others. Because of this, one of DANE’s main goals is that public policies in Colombia become more data-driven [1]. However, this is rarely the case, as public institutions sometimes cannot access the information they need because it is not publicly available, or not well classified and organized. Aware of this, the DANE wants to improve the availability and organization of their data. So, some DANE re presentatives

reached out to our university for help, by applying visual analytics methods, to allow them to deliver better tools to the public policy-making structures stakeholders. Concretely, the DANE owns data coming from both administrative records (called administrative registers (AR) afterwards) and the derived statistical analyses conducted on them i.e. statistical operations (called statistical operations (SO) afterwards) and f. This paper addresses these two main types of data. Note that in this paper we will not consider the final data collected by the DANE when they apply the questions or requests contained in an administrative register or a statistical operation, but we will consider the characteristics of the administrative registers and the s tatistical operations themselves, meaning their attributes, the fields that they address, in other terms all their metadata. So, our original data will be the DANE inventories of statistical operations and administrative registers. Figure 1 illustrates this main distinction.

Fig. 1. DANE’s data distribution and explanation of the problematic: does a new thematic classification coming from the metadata of the administrative registers and statistic operation exist?

Based on these considerations, the main objective of this project is to build a tool to understand and visualize the DANE data, particularly by organizing it through the topics and keywords present among the metadata of different groups of statistical operations and administrative registers, ultimately allowing decision -makers to have right overviews of topics and to find which statistical operations and administrative registers are related to one in particular, thanks to this classification (see Figure 1).

2 2.1

Related work Tamara Munzner’s framework

For this project, we used Munzner’s visualization framework [2] to abstract and understand the data, the users’ tasks and to choose the best idioms that allow the users to complete these tasks. It has three dimensions: the WHAT, the WHY and the HOW. WHAT: It refers to the available information (data) for the visualization. The basic abstractions of the dataset arrangements are tables, networks, fields and geometry. In a dataset, we can find items, or nodes and links, and its attributes, etc. Moreover, data can be static (there are no new data over time) or dat a can be dynamic (typically a data stream) and finally, the items/nodes attributes can be ordered or categorical. WHY: It refers to the tasks abstraction that must feature mainly one action (a verb) and one target (a noun). Here the main objective is to clarify what is the main purpose of a visualization, and its potential secondary purposes. Task abstractions can vary from high to low level (meaning depending on how precise you want to define it), and range from presenting trends (at high level, it would b e to consume data, in opposition to produce data) to identifying outliers (here it is already defined at low level). HOW: It refers to the design decisions made to visualize the data and to perform the required tasks (meaning both visual representations an d interactions if there are some). The two objectives here are to decide which visual channels like size, color, etc. will represent the data, and to choose the right marks, or the visual representations for the data (geometric primitives like lines, points, areas) for the visualization. At this stage, the idea is to choose the visual encoding and the idiom (or representation) that best suits the WHAT and WHY, to finally develop the visualization accordingly. To illustrate this concept, we want to present s ome examples of visualizations that we could use further in this work. First, a bar chart (HOW) allows to summarize distribution (WHY), and to show extremes (WHY) if it also uses order (ascending or descending). Indeed, Elzer et al. [3] showed its efficiency for this kind of tasks, but note that another possible idiom for these tasks is the stacked bar chart, as Indramoto et al. [4] explained it. However, in the stack bar chart, the focus is more on combining both single-attribute and overall-attribute comparisons rather than making only singleattribute comparisons for one or more dataset (this is our case, see section 3). Furthermore, notice that here we derived the original dataset, a table, to a network dataset. In this case, following Munzner’s framework, this kind of dataset is composed by nodes and links (whereas tables are composed by items) - note that it can be relevant to show these links or not, depending on the task. And about our project data, remember that one of our aims is also to discover a new thematic classification. As Ochs et al. [5] showed it, ontologies manipulations and representations are crucial nowadays, but required much work, so they presented a software framework for doing the following tasks: derivation, clustering and visualization as a network. Based on their study, we can note that another visualization for ontologies is the treemap [6]. In our case, we used this last representation, and also the radial force representation (see section 3); note that in both cas es, one of the most critical point is the usage of forces

to separate the nodes, depending on one attribute or relationship. Hilbert et al. explained the usefulness and importance of the forces in a network visualization; indeed, forces allow to separate and form groups, also called clusters [7]. This approach is useful for our work because we want to permit the public policy makers to make decisions based on visualizations that show a new classification, so in th is case it could be shown thanks to the use of clustering (see Figure 2).

Fig. 2. Network visualization with forces for clustering by Hilbert et al.

2.2

Projects with similar issues

In this section, we would like to present some related work that faced the same issues that we are facing in this project, either from the point of view of the policy makers or from the point of view of the designers of visual analytics tools. First, about public policy and data-driven policy making, Petrini et al. [8] were facing an issue similar to the one of the DANE: Brazil had useful data about some activities in their cities, but the authorities were not using them for public-policy making, even more precisely, these data were not used for prioritizing the different public policies over the country. So, they applied an analytic hierarchy process (AHP) that used their data and allowed them to provide some visualizations to the policymakers. As there were evaluating various kind of priorities at the same time (environmental, economic, social), they used a stack bar chart for their visualization as shown in Figure 3.

Fig. 3.Classification of thematic by priorities (multiple values for the priorities) by Petrini et al.

But for these purposes, first clean data and metadata are required: it is a common but complex issue to be dealing with unclean data. As Liu et al. [9] showed it, this is a compulsory phase for creating visualizations. In the same paper, they proposed a framework for cleansing and then creating visualizations, as shown in Figure 4.

Fig. 4. Visual analytics framework for steering data quality by Liu et al.

We can note that their process could be a complementary approach to Munzner’s one, because they gave more importance to the steps of creation and evaluation of the visualization, whereas the Munzner’s framework has more focus on the abstraction of the information related to the visualization (data/task/idiom i.e. what, why, how). Moreover, even when some ontologies have been created especially for policy makers or other final users, there are real needs of availability and accuracy, meaning that, otherwise, these ontologies would be useless, and as a result not be used by the final users. About this issue, Kamdar et al. made a study about the usage/the access of the users to the ontologies in the biomedical field [10], which queries the specialists made, and eventually how they combined the results of various ontologies. So, in this paper, we can see the importance of creating and owning ontologies, the fact they must be user-designed or task/issue-designed, and that they can also be viewed from a “macro” point of view, where ontologies can be combined between themselves. In the Figure 5, we can see how ontologies can be built.

Fig. 5. Analysis of the composition of ontologies and their depth by Kamdar et al.

To sum up, in this section we have seen some common issues with our project: ontologies or classifications are truly required for policy making, sometimes with an additional classification (priority: “meta-classification”, Petrini et al.). Then we have noted that other frameworks for creating visualizations than the Munzner’s one exist, with other focuses than abstraction, for example cleansing (Liu et al.). We have also seen that, to be useful, these ontologies or classifications need to have a good accessibility and accuracy (Kamdar et al.). That leads us to our case study: the classification of the metadata of the DANE data and its usage. Currently, the DANE public policy making tools don’t satisfy the final users of their tools (policymakers), because t he classification used does not fit with policy making, and the visualizations are not appropriate to the tasks that the policymakers want to perform. Particularly, they need to be able to discover (identify) easily which statistical operation or administrative register is the closest and more useful for a specific policy. Figures 6 and 7 show examples of the current classification and visualizations. The classification may be too generic, or specific but not with the right terms, not the ones that really define the information contained in the registers and the operations, and additionally the visualizations don’t show where this information is (which s urveys/census are the most relevant).

Fig. 6. Classification by global/macro topic (in Spanish, source: DANE website).

Fig. 7. Classification by sub-themes, linked to the surveys (in Spanish, source: DANE website).

Applying the visual analytics framework

In this section, we present how we applied the framework for our three main tasks. But, before explaining each task and the work done for each one, we would like to explain shortly how we built a new classification using the available metadata, resulting in a new dataset, derived from the previous ones, so we can build the visualization for task 2(T2) and task 3 (T3) using it. Actually, applying the visual analytics framework, it can be considered as task 0 (T0), called derivation task (from a dataset to new one). For this purpose, we decided to use natural language processing on our original datasets. It was an acceptable solution: the datasets were about 500 lines and 20 columns maximum after cleansing (originally three CSV files). In our approach, we wrote our own natural language processing tool, but it should be also correct using an existing tool. We decided to create our own tool because we wanted to both process the words and build the new dataset in the same program. First, it read all the lines of the file, build a dictionary of keywords that are repeated more than X times (using an exclusion list for the “obvious” words such as determinants or some specific words such as register), and build the nodes and the links of the new dataset, based on the occurrences of the keywords in the metadata of each item of the original dataset. Then, the three main tasks in this case study were: T1: How many statistical operations and administrative registers are there for each topic (in the original classification)? T2: How many statistical operations and administrative registers are there for each new topic, considering a new classification coming from the metadata? T3: Which administrative registers and statistical operations are more related to a specific topic (new classification)? For each of these tasks, we are going to apply the framework by presenting the WHAT, WHY and HOW abstractions, and our prototypes. Here you will find a summary for each task, then in the next subsections we give more details about each one.

Task 1 (T1) What: the original datasets – three tables, two of administrative registers and one of statistical operations (items) with, among others, the following attributes: name (c ategorical attribute) and thematic area (categorical attribute) Why: summarize the distribution (considering the old classification) How: idiom: bar chart; mark: lines; channel: vertical position (and color hue between the two bar charts, one color for the registers, one for the operations) Prototype: see Figures 8 and 9 Principal insight (see more in section 6): large difference between the numbers of administrative registers and statistical operations on the macro -category “Economics” Task 2 (T2) What: a new dataset derived from the previous tables – a network where the nodes (items) are the administrative registers, or the statistical operations, or the keywords of a new classification, and the links represent when an administrative register or a statistical operation matches with a keyword, one or more time; some attributes: name (categorical attribute) and new keyword groups (categorical attribute) Why: summarize the distribution (considering the new classification) How: idiom: treemap; mark: point; channel: color hue, spatial region Prototype: see figure 10 (+11 as an auxiliary visualization for details on demand) Principal insight (see more in section 6): “Labor market” was the penultimate subtheme in the old classification vs. “Companies” is the 4th with our new classification Task 3 (T3) What: a new dataset derived from the original ones (the same as in task 2) Why: identify features/extremes (which nodes are more related to a theme, with the new classification) How: idiom: radial force; mark: points; channel: radial position and color hue Prototype: see Figure 12 Principal insight (more in section 6): with the keyword “Health”, “Individual register of health service delivery– RIPS”, is the most important administrative register. As explained in previous sections, the main objective here is to use and understand better the DANE data to allow decision-making, resulting into two main aims: first, to determine new topics (they must be useful for decision making on public policies) that can emerge from the metadata of the administrative registers and the statistical operations, and to evaluate which different statistical operations and administrative registers are more linked to a topic, and secondly, once found this information, to be useful for the policymakers, to provide some appropriate visualizations, typically through a visual analytics tool. Therefore, we developed for this case study a visual analytics tool, by applying the framework explained above. As there were three main different tasks, it is composed of three main components: first, for the task T1, several context visualizations to analyze the current state of the information held by the DANE (3.1), then for the task T2, a treemap visualization to understand the results of the natural language processing used to understand better the major topics (a new thematic classification useful for decision making) around the DANE’s datasets (3.2), and finally for the task T3, a radial force visualization to navigate between and into

the identified topics and provide a final tool for policy makers by showing which are the most relevant sources of information according to one topic (3.3). 3.1

Task 1: general and contextualization task, on the original dataset

The first set of visualizations aims to represent the current inventory of statistical operations and administrative registers. The main task T1 is to summarize the distributions (WHY) of both datasets, to answer the following questions:  How many statistical operations and administrative registers does the DANE have?  What is the proportion of statistical operations and administrative registers in the three major topics (economics, social and environmental)?  What is the proportion of statistical operations and administrative registers in each of the 30+ specific topics (for example, health, education, etc.)? The datasets that we used were two inventories of adminis trative registers and one inventory of statistical operations (WHAT) (the dataset types were three tables ). Based on the analysis made using the Munzner’s framework, the best visual encoding (HOW) to provide this kind of overview is to use some bar charts where the statistical operations and administrative registers are differentiated by colors. With this chosen encoding, it is easy to provide the user the summary of the DANE’s information inventories. The Figure 8 shows the first three general and context visualizations developed, and the Figure 9 shows two other visualizations developed to present the distribution of the attribute “sub-theme”, allowing to know the global distribution of these sub-themes for all these administrative registers and statistical operations. Note that here we used the same encoding because it is still the best one for summarizing the distributions and to identify extremes (secondary task - as explained before that, here we additionally used the technique “separate order and align” for that purpose). So, about identifying extremes, if considering only the administrative registers (in blue, left), the most present sub-theme is “currency – bank and finance” whereas if considering only the statistical operations (in orange, right), it is “agriculture”.

Fig. 8. Bar chart visualizations based on sub-themes from the original data.

Fig. 9. Bar chart visualizations about original sub-themes, with “separate order and align”.

3.2

Task 2: new dataset, new classification, but which distribution?

Here we want to present our visualization for the task T2: summarize the distribution with the new classification (WHY). Here we used a treemap (HOW). It uses the derived dataset, from T0 task, that contains nodes and links (WHAT), each node being a statistical operation or an administrative register, and each link being a relationship between two nodes, here in particular between a “register/operation” node and a “keyword” node. We used clustering here for grouping them by (new) themes. For separating the clusters, we used an algorithm called force-in-a-box, by J. Guerra (https://github.com/john-guerra/forceInABox). As a result, this visualization allows to have a global vision (summarize distribution) of the new different themes: transport (transporte), research (investigación) etc. It also allows to detect that services (servicio) and credits (crédito) are the themes that most appear (identify extremes).

Fig. 10. Tree map chart visualization on new themes (derived data – DANE data in Spanish).

The following visualization is a table coupled to the previous treemap that give information about one item by clicking on it in the treemap,to get specific information.

Fig. 11. Auxiliary view, a table, of the treemap visualization (DANE data in Spanish)

3.3

Task 3: given a keyword, which are the items more linked to it?

For navigating between the new generated topics and the nodes associated to it, we created a visualization where the user can type a word, and then explore the statistica l operations and administrative registers that feature this keyword in their metadata, in top-down order. As a result, the main task T3 of the visualization is to identify features (WHY).

Fig. 12. Radial force visualization and its auxiliary view on the right (DANE data in Spanish)

In this radial force visualization, after choosing the keyword, the user can see that the statistical operations and administrative registers that contain the keyword are more or less attracted to the center depending on the number of occurrences (the ones that do not contain the keyword at all remain in the border), allowing to identify the extremes. The Figure 12 shows this visualization, where the statistical operations are orange and the administrative registers are blue. In addition, by putting the mouse over any item, the user can see the item name and its type. Moreover, to help identifying the extremes (WHY), right to the visualization, we added an auxiliary view, a table, where the elements are ordered in descendant order,

so the user can find the statistical operation or the administrative register that features the major occurrences of the keyword, considering all its attributes, in its metadata.

4 4.1

Experiment and results Experiment

To validate our work, we organized an experiment where the experts from the DANE were invited to try our tool with all the visualizations created according to the tasks explained before, and according to our application of the visual analytics framework. In total, there were 8 participants, 6 females and 2 males. 5 of them were working in the R&D department (in other words, “our clients”, the people who asked for the tool) and 3 were working in the department respons ible for the planning based on statistics (in other words, the final users of our future tool, apart from the policymakers). All the participants had to follow a story board that we provided to them first. Here we give a short summary of it: “first, try to get new themes about registers and operations (with the treemap now, but actually, during the experiment, with a network visualization, see section 4.2); then get some information about one item (with the coupled table); after that, write a word about one theme of your interest and discover which are the registers and operations more related with this keyword (with the radial visualization); read the name of the most important register/operation in the coupled table”. 4.2

Results

First, note that in this experience, the users were using and judging a previous version of our visual analytics tool that the one in this paper. Actually, thanks to the se results, we have been able to correct our visualizations Nonetheless , these results are interesting because, it shows how our work have evolved, and the fact that to apply the visual analytics framework may often require to be an iterative process with users experiments (and that some tasks may require more iterations). Th e following results come from a questionnaire that they filled after the experiment, where they evaluated the quality of our visualizations (usability and completion of the tasks). Using Likert scale, we asked closed questions, one for each visualization and task. The idea of the questionnaire about this usability testing was to determine if the visualizations were answering to the previously defined issues of the DANE (with the Visual Analytics Framework, visualizations must be built according to some tasks defined thanks to abstractions, coming from the original tasks described by users). Moreover, we asked open questions to get some feedback, for making corrections. Finally, according to the results, that would be discussed in the next section, we created the final visualizations that are presented in the paper in section 3. So notice that in this experience, the tool had only two parts not three (the context visualizations part, for the task T1, was missing), composed of four visualizations: a network visualization (where the forces were not separating the clusters of nodes as good as in the treemap, which is the “evolution” of this visualization after the experience) and its coupled table, for the task T2, and the radial force visualization and its coupled table, for the task T3.

Fig. 13. Results of the experiment with the users (Likert scale graph).

    

Q1Q2Q3Q4Q5-

What is your general impression with this tool? Have you been able to explore a new classification of the items? Have you been able to obtain the detail for one of these items? Have you been able to discover the items in relation with a theme? Have you been able to identify the item more related with a theme?

Discussion

According to the results, both the quality results shown on the Likert scale graph (see Figure 13), and the feedback given by the experts indicated that the task they performed with more difficulty (almost 50% of the participants grade it between 1 and 4 included) was the one asked in Q2: explore the new classification in the network visualization (T2). On the contrary, the easiest task for them (100% of them grade it between 4 and 7 included) was the one asked in Q4: discover the items in relation with one theme in the radial force visualization (T3). As a result, we did more corrections on the visualization used in Q2, so it is the only visualization where we had to completely modify the idiom (visual encoding) used, transforming the visualization from a network visualization at that moment (clusters might not be appearing so clearly, but the relationships/links between elements do appear better) to a treemap now (on the contrary, the focus is clearly on clustering) – to sum up, both idioms use nodes and forces, but the clustering was clearer in a treemap than in a network visualization, at least in our case, so it explains the modification. Additionally, we noticed, both in the comments and the qualitative results (apart from Q2 results, in particular thanks to Q1 results, with 25% of neutral grade, 4), that something might be missing, apart from the current visualizations and their corrections, something that shows better the purpose of our tool and why did we create it. In other terms: to understand where we are going, we should know what were the origi-

nal data? Were there some insights in the original data? The next visualizations are showing different data? What is the difference with the new thematic classification? What are the new insights? All these questions can be considered as a first task for the users (T1). So, that is why we finally added the context visualizations part and organized the tool into three parts and not two: context - T1 (original data), treemap - T2 (derived data) and radial force visualization - T3 (derived data).

Insights The new (corrected) visualizations (presented in the section 3) allow to discover some insights about the statistical data and their classification in new themes. First, thanks to the context visualizations based on the initial data, we discovered than globally more information could be generated because there are more registers than operations (remember that the administrative registers produce and create the statistical operations). More precisely, we noticed this particularly for the macro category “Economics”. Moreover, looking at the visualization by sub -themes, the previous insight is confirmed, and we can observe it with more details: the sub-themes “Currency, banks and finance” and “Accounts and economics” appear as some of the ones where there is much difference between the number of registers and the number of operations, and both are belonging to the macro category “Economics”. Then, thanks to the new derived data and the treemap visualization, we can discover other insights. First, by looking to the terms that appear, on one hand this visualization confirms that the macro category and sub-theme currently used are quite coherent with the metadata. For example, we can find as new keywords “research” (investigación), “credits” (crédito), “market” (mercado), whereas that, in the current subtheme, there were “education and innovation”, “business”, “economy a nd accounts” etc. But, in the other hand, this visualization also suggests that the categories and sub themes used currently may not be so accurate in term of distribution, because in the current classification, “labor market” is the penultimate sub -theme whereas “company” (empresa) is the fourth one with our new thematic classification. Another insight is that this visualization gives us more details about a previous insight from the first visualizations component: in the current category “Economics”, the focus for generating new operations from registers should be done more precisely on the ones that contains in their metadata the words “leasing” or “transactions” because we can observe in this visualization that there are no statistical operations about t hese subjects, only registers. Finally, some other insights have been revealed thanks to the radial force visualization. As our final tool in this study for decision making on public policies, we can notice that thanks to it, people, without being an expert about data manipulation, can identify easily which registers and operations are more related to a theme: for example about “Housing” (the chosen keyword), the most important administrative register is “Housing financing VIS”, or with the keyword “Health”, the most important administrative register is “Individual register of health service delivery – RIPS”. To conclude, this last visualization, by grouping in the center the elements with more relevance but separating them into two categories, for being administrative registers or statistical operations by using two different colors, allows at a glance to notice for one thematic if the relevant elements are mostly of one kind of information (because visually it will be mostly of one color), and as a result, it can confirm that more statis-

tical operations should be generated in this thematic or not (so we could make some “priorities” about statistical operation generations). So, for example, following with the theme “economics”, we can notice for the keyword “credits” that more statistical operations could be generated (there are much more administrative registers).

Conclusion

Finally, applying our approach, we obtained useful visualizations that confirmed that the DANE owns highly relevant information for the country and that they should continue developing more data analysis tools, to provide them to its different stakeholders for maximizing the usage of their data. In particular, the visual analytics tools permit both policymakers and citizens to locate where the relevant information is, and allows its understanding, ultimately enhancing policies and fostering data-driven businesses. So, we might think that our contribution helped the DANE to understand that their data are very valuable, and that with such approaches, these data can be classified and presented in a way that allow the policymakers to use it for public policies making. As written in the previous section, thanks to this study and its different visualizations, the DANE learned, among other insights, that even if they held a large amount of data about economics, they could explore and use it better: more statistical operations could be generated, especially about some topics such as leasing or transactions. This insight might be the most relevant because parts of it appeared in most of the visualizations. About future work, it could be interesting to explore other possibilities about our natural language processing tool to get the keywords that appear in the metadata and how could we grade differently the administrative registers or statistical operations that belong to a theme (currently it is based on the number of occurrences of the keywords in the metadata – the DANE has already confirmed it interest for this avenue). Finally, another possibility of future work could be to study which visualization would be appropriate for showing the relationships between the administrative registers and the statistic operations that are linked (because this register produces this operation) while the visualization shows the clusters of nodes by thematic.

References 1. 2. 3. 4.

OECD, OECD Digital Government Studies Digital Government Review of Colombia Towards a Citizen-Driven Public Sector, OECD Publishing, France (2018) T. M unzner, Visualization Analysis and Design. 1st edn. A K Peters Visualization Series/CRC Press, Natick, M assachusetts, USA (2014). S. Elzer, S. Carberry, I. Zukerman, The automated understanding of simple bar charts, Artificial Intelligence 175(2), 526–555 (2011). Indratmo, L. Howorko, J. Boedianto, B. Daniel, The efficacy of stacked bar charts in supporting single-attribute and overall-attribute comparisons, Visual Informatics 2(3), 155–165 (2018).

50 5.

6. 7. 8.

10.

C. Ochs, J. Geller, Y. Perl, M .A. M usen, A unified software framework for deriving, visualizing, and exploring abstraction networks for ontologies, Journal of Biomedical Informatics 62, 90–105 (2016). B. Shneiderman, Tree visualization with tree-maps: 2-d space-filling approach. ACM Transactions on Graphics, 11(1), 92–99 (1992). M . Hilbert, P. Oh, P. M onge, Evolution of what? A network approach for the detection of evolutionary forces, Social Networks 47, 38–46 (2016). M .A. Petrini, J. Rocha, J.C. Brown, R. Bispo, Using an analytic hierarchy process approach to prioritize public policies addressing family farming in Brazil, Land Use Policy, 51, 85-94 (2016). S. Liu, G. Andrienko, Y. Wu, N. Cao, L. Jiang, C. Shi,Y.S Wang, S. Hong, Steering data quality with visual analytics: The complexity challenge, Visual Informatics, 2(4), 191-197 (2018). M .R. Kamdar, S. Walk, T. Tudorache, M .A. M usen, Analyzing user interactions with biomedical ontologies: A visual perspective, Journal of Web Semantics, 49, 16-30 (2018).

A Comparative Study on Self-Report Methods for User Satisfaction Evaluation Andrés F. Aguirre 1 , Ángela Villareal Freire1 , Jaime Díaz Arancibia 2 , Rosa Gil3 , César A. Collazos 1 1 IDIS Research Group, University of Cauca, Popayán, Colombia {afaguirre, avillarreal, ccollazo}@unicauca.edu.co 2 Department of Computer and Information Sciences, University of La Frontera, Temuco, Chile jaimeignacio.diaz@ufrontera.cl 3 Research Group on Human Computer Interaction and Data Integration, University of Lleida, Lleida, Spain rgil@diei.udl.cat

Abstract. Many studies show that user satisfaction plays a vital role in

the use of any software system. There are different levels and dimensions in which user satisfaction can be evaluated, some reasons depend on the context, others on the type and goal of the study. In the field of Human-Computer Interaction it is important to understand not only how satisfied the user is, but also how to meet their needs regarding the use and quality of use of a given system. Evaluating user satisfaction is not an easy task and there is a great diversity of methods that have been developed for this purpose. Some of the most commonly used methods due to their simplicity and low cost are the self-report methods. These types of methods provide information about the subjective reactions of users, and can become one of the most important inputs to collect and understand the behavior, preferences and perceptions of users. For this reason, choosing the most appropriate method (or combination of methods) to evaluate user satisfaction is a vital fact. In this article we study some of the most common and accepted self-report methods to evaluate user satisfaction in the software industry and academic fields. Through a science mapping study conducted with SciMAT, the differences, approaches, thematic evolution and the relevance of self-report methods such as SUS, SUMI and QUIS are shown. Keywords: User Satisfaction, User Experience, self-reporte methods, subjective methods, SUS, SUM I, QUIS, SciM AT.

Introduction

A few decades ago, digital technology was popular and mostly used by people with advanced technical knowledge [1]. Nowadays, with the appearance of emerging technologies and the evolution of existing technologies, the increase in users is exponential (and growing), since technology covers almost every aspect and context of society [2]. Esta situación ha implicado un cambio radical en la forma en cómo las personas

interactúan con los sistemas e incluso en cómo lo hacen entre ellas mismas a través de distintas herramientas digitales . This fact makes it possible to see the importance of having tools that allow the evaluation of digital technology with respect to the impact and relevance they have on the needs of current users. Although there are user satisfaction evaluation methods widely used in the software industry and in academic environments, it seems that they were not designed with the explicit approaches that support the current technological development. Traditional methods have a pragmatic approach, i.e., they evaluate mainly the efficiency and effectiveness perceived by users in relation to the system with which they are interacting. [3]. However, there are aspects of the user experience (UX), such as fun, entertainment and other subjective characteristics, which affect the general satisfaction of users with equal or greater degree than do it the pragmatic aspects. It is important to consider that user satisfaction is a multimodal construct that includes emotions, evaluations and perceptions of people in the face of the different aspects inherent to their interaction with a system. While it is true that users could differ from what they consider to be more important or that generates greater satisfaction in them, there is a consensus about the multimodal nature inherent to user satisfaction [2]. This aspect highlights the importance of enriching the evaluation of user sat isfaction, especially from an emotional approach, given its importance and the scale of instruments that allow its measurement [4][5]. To achieve this purpose, a deep understanding and attention to every detail about the interaction is required. Although selfreport questionnaires that assess user satisfaction are just one type of instrument among the whole range of options for evaluating UX, they can (and should) be closer to covering a broader and more precise spectrum of the needs that can be evaluated with this type of instruments. This demands a great challenge on the part of the researchers, since their efforts should also converge in emphasizing the intangible (in those aspects of abstract nature such as emotions), of course without neglecting the pragmatic side of the UX as the usability, functionality, efficiency, etc. In this sense, this paper presents a comparative study of different user satisfaction assessment methods that are widely used in the academic and software industry. This in order to contribute to the knowledge of the approaches, differences, limitations and their level of relevance with the current technology, in particular the way in which they evaluate the interactions with said technology. Interaction with technology is not only what people do, but also what people feel [6]; therefore, knowing which methods allow user satisfaction to be measured and analyzed appropriately (or at least approaching that purpose), is prevailing in the current field of UX, since it becomes one of the main inputs in the technology design that is consistent with the different needs of users, their expectations and all aspects related to an interaction.

Research questions

The evaluation of user satisfaction has been a constant challenge for the scientific community and professionals of the area HCI, since the conceptualization of user satisfaction has received different connotations and interpretations over time. For this

reason, it is important to obtain an overview of its evolution, which allows describing how the different self-report instruments have been used to obtain and evaluate those characteristics that affect user satisfaction. Within this context the research questions addressed by this study and that address its goals are: Is it possible identify trends with respect to the topics that have the most relevance in self-report methods? What are the approaches to the self-report methods used to evaluate user satisfaction? Is there a correspondence between the characteristics that evaluate the self-report methods and those found in the study?

3 3.1

Design of the Study Choice of instruments

As mentioned before, there are different user satisfaction evaluation methods. This study takes as a starting point the methods highlighted by Tullis and Albert [7]. From these methods an analysis was made that allowed to show relations among them or to estimate their differences and similarities. The following describes each of the user satisfaction evaluation instruments considered in the present study: System Usability Scale (SUS) is an interesting variation of the traditional questionnaires. It presents a combination of statements written positively and negatively, so that the user really pays attention to each of their answers [8] [9]. SUS consists of a 10-item questionnaire, each with a Likert scale of 5 (or 7) points, which provides an overview of satisfaction with the software [8]. Software Usability Measuring Inventory (SUMI), is a method of evaluating the quality of software that allows measuring satisfaction and assessing user perception [10]. SUMI is a commercially-available questionnaire for assessing usability of software developed, validated and standardized on international databases [10] [11]. This method is referred to in standards ISO 9126 [12] and ISO 9241 [13] as a recognized tool for evaluating user satisfaction via five dimensions of usability. This tool is also available in several languages [10] [11]. Questionnaire for User Interaction Satisfaction (QUIS), is a tool developed by researchers at the University of Maryland Human-Computer Interaction Lab. Designed to assess the subjective satisfaction of users on specific aspects of the human computer interface [14]. The current version, QUIS 7.0, asses ses the user’s overall satisfaction in 6 hierarchically organized facets in each of the nine interface -specific factors defined in this tool. Each facet, in turn, consists of a pair of semantic differentials arranged on a 10-point scale [14] [7]. The questionnaire is designed to be adjusted according to the analysis needs of each interface, in which only sections of interest can be considered. Website Analysis and Measurement Inventory (WAMMI) is an online service that emerged from SUMI. Both were developed at the Human Factors Research Group (HFRG) at University College, Cork. Unlike SUMI, which is designed for the evaluation of desktop software applications, WAMMI focuses on evaluation of web sites [7] [15]. This instrument consists of 20 questions that use 5-point Likert scales as an-

swers [16] [7], and makes it possible to create a questionnaire and link it to WAMMI classification scales [7]. The result of a WAMMI analysis is a measure of ”global satisfaction” [16], that is divided into 5 dimensions . Measuring Usability of Multi-Media System (MUMMS), was developed by the same group that designed SUMI and WAMMI. MUMMS cons ists of a questionnaire that enables assessment of quality of use for multimedia software products [17]. Measurement aspects are the same as those SUMI takes account of and it incorporates a new one related to the user’s emotional perception toward the use of the system. This tries to capture information about the fascination the multimedia application exerts on users [17]. The characteristics that are evaluated in each of the instruments described above are described below in Table 1: Table 1. Characteristics of user satisfaction evaluated by different self-report methods. M ethod Characteristics evaluated SUS Ease of use, Usefulness, Helpfulness, Functionality, and Learnability [7]. SUM I

Efficiency, affection, utility, control, and learning [11].

QUIS

Screen factors, terminology and system feedback, learning factors, system capabilities, technical manuals, online tutorials, multimedia, teleconferencing and software installation [18][19]. WAM M I Attractiveness, control, efficiency, utility and learning, usability [16] [7] M UMMS Attractiveness, Controllability, Efficiency, Helpfulness, and Learnability [7].

Conduct the study

To carry out the comparison of the user satisfaction evaluation methods compiled, a mapping of the science of each of them was done through SciMAT. SciMAT is a powerful open source software tool for analyzing the evolution and relevance of scientific production around a specific scientific area. The steps followed for this study correspond to those that are usually carried out in a science mapping [20] [21]: data retrieval; pre-processing data and applied filters; design and normalize network; science mapping; visualization and analysis. Finally, a visual interpretation of the map of science is presented. 4.1

Data retrieval

The first step is to construct the data set with which the study will be carried out. In this case, the Web of Science (WoS) database is used (one of the most important and significant scientific database that exists). Taking into accoun t the objective of the study, the general criteria that define the search strings, the type of relevant fields for the search and the period of time for the study are established. Consequently, the following criteria are defined:



Search string: given the nature and purpose of this study, the search strings used correspond to the exact names of each instrument, i.e., “System Usability Scale”, “Software Usability Measuring Inventory”, “Questionnaire for User Interaction Satisfaction”, “Website Analysis and Measurement Inventory” y “Measuring Usability of Multi-Media System”. Search field: the search criteria chosen corresponds to "Theme", given that it has a greater coverage than the rest of the available criteria in WoS, and allows the search in key fields for this study such as the title, the abstract, author's keywords and keywords plus. Period: the period for the longitudinal analysis is defined for studies published between 2001 and 2019.



The results obtained in this phase are shown in Table 2 below: Table 2. Record count obtained from WoS per instrument. M ethod Record count obtained from WoS SUS 324 (318 Articles; 6 Proceedings Paper; 4 Reviews, 2 Editorial Material) SUM I

16 (16 Articles; 2 Proceedings Paper).

QUIS

13 (13 Articles; 3 Proceedings Paper)

WAM M I M UM MS

2 (2 Articles). The search did not return any results.

Given the results obtained in this phase, the comparison study will focus primarily on SUS, SUMI and QUIS. It should be noted that the scarcity or absence of records on WAMMI and MUMMS, does not show that they have not been used in the scientific field, simply that they have not been the main object of research or have not had sufficient relevance in a certain study, which motivates their Appearance in the title, abstract or keywords. 4.2

Pre-processing

For the pre-processing of the collected data set, SciMAT is used to detect and eliminate duplicate records. Since the corpus generated with each of the searches performed are relatively small1 , thresholds of citations for the inclusion or exclusion of documents were not established. 4.3

Design network and normalize

In this phase, a network is designed using descriptive terms or words as the unit of analysis; and the type of relationship that is established for the network, is co-

The corpus that are analyzed in scientific mapping studies are usually thousands of registers. That is why it is considered that those used in this study are relatively small.

occurrence 2 . As a measure of similarity to normalize the network, equivalence index is used, which allows the adoption of statistical indices that represent the measure of connection between key words [22]. 4.4

Science mapping

For the generation of the science map, the "simple centers algorithm" is used [23] [24]. This algorithm allows to locate networks of keywords that are strongly related to each other and that correspond to centers of interest that stand out among the scientific community. [25]. 4.5

Visualization and analysis schemes

This study focuses on the behavior of publications, authors and references in different fields that revolve around each of the instruments for evaluating user satisfaction, seeking to detect patterns and trends in related research topics, following the steps below: (1) Identification of research topics, with the normalization of the network, different research topics are identified. This process allows the grouping of keywords in the network that have been strongly linked to the research topics of high interest. (2) Strategic diagram of clusters, this type of diagram shows the groups for each period and orders them in a two-dimensional space determined by centrality (x axis) and density (y axis). The centrality corresponds to the degree of cohesion of a network with other networks (external cohesion of the network). On the other hand, the density represents the level of internal force of the network (internal cohesion of the network). The measures of centrality and density allow to categorize the detected clusters in a certain period [22]. These clusters represent the research topics of a given research field, and are mapped in the two-dimensional space of the strategic diagram. According to the positioning of the cluster in said space, it can be classified in one of the four groups below: (1) upper-right quadrant, it hosts the motor clusters ; (2) upper-left quadrant, it shows highly developed and isolated clusters; (3) lower-left quadrant, it corresponds to emerging topics or declining topics; (4) lower-right quadrant, it contains basic and transversal cluster. In this sense, the strategic diagram of clusters provides essential information to reveal and understand the dynamics of a field of research, and therefore, is one of the main inputs of analysis in this study.

Analysis of results

To carry out the analysis of the information collected, the study period of 19 years (2001-2019) was divided into two subperiods, organized according to the number of publications that mark a milestone in the increase or decrease of publications related to the user satisfaction evaluation methods. Reason for which was sought because the 2

Co-occurrence, when two units of analysis appear together in a set of documents.

years established to delimit the periods, allowed to recognize notable changes in research initiatives, i.e., was designed in such a way that the trends and patterns of use of the instruments, were evident in the change of the period. 5.1

First sub-period

The first subperiod (2001-2010) is characterized by a similarity in the contexts and approaches in which SUS, SUMI and QUIS are used. The behavior of different areas (shown as spheres) for each of the instruments is explained in detail below. Results of the first period. In the strategic diagram of SUS (see Figure 1), two spheres corresponding to usability and functionality are highlighted. In the case of usability, its location and volume allow to see that it has a neutral centrality, but with a high density. This behavior responds to the objective nature of usability, i.e., its relationship with other topics is neutral since it is not an abstract concept like other areas that make up user satisfaction. The high density shown in the diagram on usability, is understood as a strong cohesion of the issues that make up the usability. On the other hand, functionality is characterized by having a strong centrality, but a neutral density. Since the functionality is related to multiple functions that make up the system, it is logical to see that it has a strong centrality. About the density of the functionality shown in the diagram, describes a dynamic behavior of the issues related to functionality, derived from the multiple contexts, applications, functions that are evaluated.

Figure 1. Strategic diagram of SUS of the first period (adaptation of SciM AT).

In the SUMI strategic diagram (see Figure 2), two spheres corresponding to usability and cognitive science are highlighted. As can be seen, the usability retains the location and volume of SUS, i.e., its approaches and interpretation are the same.

With respect to the cluster cognitive science, it shows a strong centrality and a neutral density, and at the same time a small volume compared to usability. Cognitive science is an area that is related to the creation of mental models, design, analysis of experiments, etc. for that reason it can be seen that it has a strong centrality. The density shown by this cluster describes the diversity of areas that make up cognitive science.

Figure 2. Strategic diagram of SUM I of the first period (adaptation of SciM AT).

In the strategic diagram of QUIS (see Figure 3), it shows mainly two spheres that correspond to functionality and interface design. In the case of the cluster of interface design, a neutral centrality with a high density can be observed. The neutral centrality is due to the fact that it is an area that focuses on the style and appearance of the interfaces, which although it is related to other areas, its focus is concrete and limited. The high density that this cluster shows, it is understood how strongly linked are the themes that underlie the interface design field. In the case of the cluster of functionality, it is characterized by having a strong centrality, but a neutral density. Since this behavior is similar to that of SUS in the first period, its interpretation is similar.

Figure 3. Strategic diagram of QUIS of the first period (adaptation of SciM AT).

General analysis of results of the first period. In this period the predominant user satisfaction assessment approaches correspond to the areas of functionality and usability, i.e., these are motors topics and are characterized by their pragmatic approach to evaluation. As can be seen in Table 3, funcionality and usability appear in more than one method, while the areas of cognitive design and interface design are only highlighted in one method (SUMI and QUIS respectively). Table 3. Predominant topics in each method during the first period (2001 - 2010). Functionality Usability Cognitive Science Interface Design SUS X X SUM I QUIS

X X

In short, functionality and usability are the most relevant topics in this period. When analyzing the location of these topics in the strategic diagram, two main aspects can be observed: (1) Spatial location of the salient topics, both the functionality and usability retain their positioning in the diagram, this suggests that the dynamics, treatment and relevance given to these topics, is the same although they have been detected in selfreport methods different. (2) Nature of the salient topics, although functionality and usability are motors topics, they have a different nature. As suggested in previous sections, the high centrality and neutral density that shows the functionality, reveals that it is a transcendental topic and of great importance, but its development continues in effect during this period. On the other hand, the topic of usability is characterized by having a high density with a neutral centrality, which shows that it is a highly developed area, characterized by being too specialized and peripheral.

5.2

Second sub-period

In the second sub-period (2011-2019), the change in the volume of the spheres representing the research topics is clearly seen. However, some disappear and new ones appear. The topics that are preserved change their location according to the patterns and trends of the investigations. The interpretation of the different areas drawn in each strategic diagram for each of the instruments is explained below: Results of the second period. In the SUS strategic diagram, three nodes related to design, acceptance and health are highlighted. The cluster that represents design has a location with relatively weak centrality and density. This is because design is a broad and dynamic concept that can cover different fields, from interaction design, to interface design, which although they can be strongly related, are significantly different. It should also be noted that despite not having a large volume (compared to health), it is considered as one of the motor clusters of this period. The cluster that represents the acceptance, shows a low centrality and density. The location of this cluster in the two-dimensional space, can be interpreted as an emergent area, but little evolved. The cluster that represents health has a centrality and density that are remarkably strong. Due to its volume and location, it is considered as a motor cluster in this period. Its appearance allows to see that it is a context in which the SUS is widely used .

Figure 4. Strategic diagram of SUS of the second period (adaptation of SciM AT).

In the strategic diagram of SUMI (see Figure 5 clusters that belong to functionality and usability are shown. The cluster that represents usability, shows a high centrality and a neutral density. The density neutrality of the cluster describes the discrete approach t hat usability has with respect to other topics with which it can be related. As already mentioned, the strong centrality that characterizes usability in this period shows the internal force with which the areas that make up usability are related.

The functionality cluster is characterized by a strong centrality and a neutral density. Since this behavior is similar to that of SUS in the first period, its interpretation is similar.

Figure 5. Strategic diagram of SUM I of the second period (adaptation of SciM AT).

In the strategic diagram of QUIS, two clusters can be observed that correspond to interface design and cognitive science. In the case of the interface design cluster, it can be observed that it has a neutral centrality and density. It can be understood that the interface design is no longer a cluster engine in the QUIS treatment. In the case of the cognitive science cluster, it is characterized by having a highly strong centrality and density, and it is constituted as a motor cluster in this period. Its appearance allows to see that it is one of the important areas that revolve around QUIS.

Figure 6. Strategic diagram of QUIS of the second period (adaptation of SciM AT).

The diagrams shown above allow observing that the volume of the clusters and their location in the two-dimensional plane, have strong similarities between them. Another aspect evident in the two periods is the conservation of the motor clusters, mainly

those that make up the usability and functionality. Current publications related to SUS, SUMI and QUIS, continue to revolve around these two topics, neglecting other equally or more important aspects that make up user satisfaction in the new technological era. General analysis of results of the second period. Unlike the first period, this period is characterized because none of the topics detected in each of the self-report methods collected in this study is repeated in the other, i.e., that SUS, SUMI and QUIS detect different topics which are not common among the methods (the Table 4 shows the predominant themes in each of the methods). However, not only are features detected but complex areas that cover other areas (better known as umbrella terms), and that broaden the spectrum of topics and features that commonly revolve around user satisfaction. Table 4. Predominant topics in each method during the second period (2011-2019). Functionality Usability Cognitive Interface Acceptance Health Design Science Design SUS X X X SUM I QUIS

X X

The topics detected in this period were diverse and with similar positioning and volume. This makes it difficult to highlight the most relevant topics. However, when analyzing the location of these topics in the strategic diagram, two main aspects can be observed: (1) Spatial location of the salient topics, SUS and QUIS are differentiated from SUMI mainly because in them are detected motors of considerable importance: health (in SUS) and cognitive design (in QUIS), are cases that stand out among others because their positioning in the strategic diagram indicates that these are highly developed and important topics for the construction of the scientific field around the evaluation of user satisfaction through self-report methods. In a similar way, it could be indicated with the design topic but on a smaller scale with respect to the other two, given its size and positioning in the diagram. On the other hand, acceptance is the only cluster that appears in another quadrant different from the others, its location supposes that it is still in development and for that reason it still lacks importance with respect to the rest of the topics that were detected in the study. (2) Nature of the salient topics, although almost all the topics detected in this period are motors (with the exception of acceptance), they have a very diverse nature. The centrality and density measures vary in most cases, i.e., there is no similarity pattern between the results obtained, each one offers diverse information, and therefore its interpretation varies as indicated in the description of each strategic diagram.

5.3

General analysis of the results of the scientific mapping

Is it possible identify trends with respect to the topics that have the most relevance in self-report methods? The change in the volume and location of the spheres that represent the research topics is clearly seen in the transition from one period to another. The volume of a sphere is proportional to the number of documents that belong to it, while its location shows what that topic represents in that period, according to the classification of the quadrants of the strategic diagram. Usability is a subject that has transcended from period to p eriod (detected in SUMI). The change in the volume of the cluster that represents usability, shows that there is a decrease in the works related to the topic. As recent research has indicated, the usability of a product may not be the only, or even the main, determining factor in user satisfaction [26] [27]. Regarding location, it can be observed that centrality increases from one period to another, which shows the strengthening of usability with group structures of other topics, i.e., how it affects and is affected by other aspects inherent in the UX This has also involved a decrease in its density, s ince its understanding has changed over time, and is reflected in its conceptual redefinition, an example of this is the concept established in the standard ISO / IEC 9126-4 [12] and how it has been redefined in the standard ISO/IEC 25010:2011 [28]. Another topic that has remained in the two periods studied, is the interface design (detected in QUIS), its initial volume decreases with respect to the second period, which indicates that there is a smaller number of investigations on the interface design in the published works of QUIS. On the location of this cluster a decrease in its density is observed, which can be related to the evolution of the technology and therefore its interfaces. This constant and accelerated change has involved a restructuring of the related areas, and that is a subject that is in constant development. What are the approaches to the self-report methods used to evaluate user satisfaction? Faced with the issues that were detected in the two periods, in Figure 7 shows which were the topics that transcended from one period to another. In the figure it can be seen that usability and interface design have been themes that have been maintained during the two periods, which suggests that they are key issues in the evaluation of user satisfaction with SUMI and QUIS respectively.

Figure 7. Topics detected in each self-report method during the two periods studied.

On the other hand, SUS presents the most notable changes, since the topics that are characteristic in the first period completely disappear in the second period, in which not only features (such as functionality and usability) are highlighted, but also areas (such as Health) and contexts (such as Health). This is in line with the dynamism and coverage in the conceptualization of user satisfaction. It is worth mentioning that the topics that disappear from a period do not mean that they do not continue to be related to the research fields in which they were initially highlighted, but that they did not have sufficient relevance to be detected and located in some quadrant of the two-dimensional space of the strategic diagram of the later period. Is there a correspondence between the characteristics that evaluate the self-report methods and those found in the study? Another of the main findings in this study is that the topics detected in the scientific mapping carried out on each of the self-report methods compiled (SUS, SUMI and QUIS), do not correspond in their entirety with respect to the ch aracteristics which each of these methods evaluates (see Table 1). This is due in large part to the lack of conceptual clarity that exists about user satisfaction, as a consequence of the different approaches in how user satisfaction is understood, which vary according to the context or purpose of the study. Added to this is the constant technological advance, and the inherent mass use of interactive systems in general. This has brought a constant change and evolution in the interactions between users and technology, and has involved a great challenge to find a consensus on the conceptual structure of user satisfaction. In fact, many of the methods used today for evaluation remain governed by the anachronistic parameters with which user satisfaction was commonly understood. This aspect not only makes it difficult to choose the appropriate method for the evaluation of user satisfaction, but also highlights the urgent need to establish a solid theo-

retical construction that meets current needs and allows objectively and clearly understand user satisfaction.

Conclusions

This study offers a systematic review that allows observing research trends on user satisfaction in the Human-Computer Interaction context, such as the fields of study in which most of the research is focused, the main characteristics of the self-report methods used to evaluate user satisfaction, some of the contexts and areas with which they are most related. This allows to show a general and objective panorama of each one of the methods, which can serve as a point of reference for UX researchers. One of the main findings that were obtained in this study is related to the emotional side of the users. People often link emotionally with the things they use (software systems are not the exception). The presence of a method that covers the emotional spectrum is more than necessary in this new digital era. This aspect is becoming increasingly relevant and invites us to redesign the different mechanisms for evaluating user satisfaction, so as to encompass a holistic view of user satisfaction.

Limitations

It is important to have a significantly larger corpus to perform a science mapping that allows obtaining more revealing and significant findings for a longitudinal and conceptual analysis of a scientific topic. Of course, this depends on the research topic and the publications that have been generated around that topic.

Discussion

WoS and Scopus are considered as the main data sources for bibliometric studies [29]. In fact, the comparison of these two main databases has been an active area of study and discussion, and has produced several scientific publications. For this study, WoS was the main source of data, but given the small size of the corpus obtained, Scopus was also considered. However, the results obtained with Scopus were quite similar to those of WoS. In a report published in [29], it is emphasized that in most cases, the studies that compared WoS and Scopus from a bibliometric perspective obtained similar results between the two databases across a variety of indicators. In this report they also highlight that the data quality obtained with Scopus has a poor performance compared to that of WoS, it should be noted that the quality of the data also has a direct impact on the quality of the bibliometric indicators that can be compiled. This aspect and the similarity of the data obtained with the two bibliographic databases, motivated the authors of this article to stay with WoS.

References 1. B. M oggridge, Designing Interactions. 2007. 2. R. A. Calvo, D. Vella-Brodrick, P. Desmet, and R. M . Ryan, “Positive Computing: A New Partnership Between Psychology, Social Sciences and Technologists,” Psychol. Well. Being., vol. 6, no. 1, p. 10, Dec. 2016. 3. A. F. Aguirre, Á. P. Villareal, R. M . Gil, and C. A. Collazos, “Extending the Concept of User Satisfaction in E-Learning Systems from ISO / IEC 25010,” in Design, User Experience, and Usability: Understanding Users and Contexts, 2017, vol. 1, pp. 167– 179. 4. A. F. Aguirre, Á. P. Villareal, C. A. Collazos, and R. Gil, “Aspectos a considerar en la evaluación de la satisfacción de uso en Entornos Virtuales de Aprendizaje,” Rev. Colomb. Comput., vol. 15, p. 15, 2014. 5. K. Capota, M . Van Hout, and T. Van Der Geest, “M easuring the emotional impact of websites: a study on combining a dimensional and discrete emotion approach in measuring visual appeal of university websites,” in Proceedings of the 2007 conference on Designing pleasurable products and interfaces, 2007, no. August, pp. 22–25. 6. S. D. Shahriar, “A Comparative Study on Evaluation of M ethods in Capturing Emotion,” 2011. 7. T. Tullis and W. Albert, Measuring the User Experience: Collecting, Analyzing, and Presenting Usability Metrics. M organ Kaufmann, 2013. 8. J. Brooke, “SUS - A quick and dirty usability scale,” in Usability evaluation in industry, P. W. Jordan, B. Thomas, B. A. Weerdmeester, and I. L. M cClleland, Eds. London, UK: Taylor & Francis, 1996, pp. 189–194. 9. R. Hartson and P. Pyla, The UX book: Process and guidelines for ensuring a quality user experience. M organ Kaufmann, 2012. 10. J. Kirakowski and M . Corbett, “SUM I: The Software Usability M easurement Inventory,” Br. J. Educ. Trchnology, vol. 24, no. 3, pp. 10–12, 1993. 11. “Software Usability M easurement Inventory (SUM I),” Human Factors Research Group, University College Cork, 1993. [Online]. Available: http://sumi.ucc.ie/index.html. [Accessed: 22-Jan-2015]. 12. “ISO/IEC TR 9126-4: Software engineering -- Product quality -- Part 4: Quality in use metrics.” 2004. 13. “ISO 9241-11: Ergonomic requirements for office work with visual display terminals (VDTs) -- Part 11: Guidance on usability.” 1998. 14. J. P. Chin, V. A. Diehl, and K. L. Norman, “Development of an instrument measuring user satisfaction of the human-computer interface,” … SIGCHI Conf. Hum. …, pp. 213–218, 1988. 15. J. Kirakowski, N. Claridge, and R. Whitehand, “Human centered measures of success in web site design,” in Proceedings of the Fourth Conference on Human Factors & the Web, 1998. 16. G. Lindgaard and C. Dudek, “What is this evasive beast we call user satisfaction?,” Interact. Comput., vol. 15, no. 3, pp. 429–452, Jun. 2003. 17. “M easuring the Usability of M ulti-M edia System (M UM M S),” Human Factors

67 Research Group, University College Cork, 1996. . 18. J. P. Chin, V. A. Diehl, and K. L. Norman, “Questionnaire For User Interaction Satisfaction (QUIS),” Human-Computer Interaction Lab, University of Maryland at College Park, 1988. [Online]. Available: https://isr.umd.edu/news/news_story.php?id=4099. [Accessed: 19-Nov-2018]. 19. T. R. Johnson, J. Zhang, Z. Tang, C. Johnson, and J. P. Turley, “Assessing informatics students’ satisfaction with a web-based courseware system.,” Int. J. Med. Inform., vol. 73, no. 2, pp. 181–7, M ar. 2004. 20. K. Börner, C. Chen, and K. W. Boyack, “Visualizing Knowledge Domains,” Annu. Rev. Inf. Sci. Technol., vol. 37, no. 1, pp. 179–255, 2003. 21. M . J. Cobo, A. G. López-Herrera, E. Herrera-Viedma, and F. Herrera, “Science mapping software tools: Review, analysis, and cooperative study among tools,” J. Am. Soc. Inf. Sci. Technol., vol. 62, no. 7, pp. 1382–1402, Jul. 2011. 22. M . Callon, J. P. Courtial, and F. Laville, “Co-word analysis as a tool for describing the network of interactions between basic and technological research: The case of polymer chemsitry,” Scientometrics, vol. 22, no. 1, pp. 155–205, Sep. 1991. 23. N. Coulter, I. M onarch, and S. Konda, “Software Engineering As Seen Through Its Research Literature: A Study in Co-word Analysis,” J. Am. Soc. Inf. Sci., vol. 49, no. 13, pp. 1206–1223, 1998. 24. M . J. Cobo, A. G. López-Herrera, E. Herrera-Viedma, and F. Herrera, “An approach for detecting, quantifying, and visualizing the evolution of a research field: A practical application to the fuzzy sets theory field,” J. Informetr., vol. 5, no. 1, pp. 146–166, Jan. 2011. 25. M . J. Cobo, M . A. M artínez, M . Gutiérrez-Salcedo, H. Fujita, and E. Herrera-Viedma, “25 years at Knowledge-Based Systems: A bibliometric analysis,” Knowledge-Based Syst., vol. 80, pp. 3–13, 2015. 26. A. De Angeli, A. Sutcliffe, and J. Hartmann, “Interaction, usability and aesthetics: what Influences Users’ Preferences?,” Proc. 6th Conf. Des. Interact. Syst., pp. 271– 280, 2006. 27. A. Sonderegger and J. Sauer, “The influence of design aesthetics in usability testing: Effects on user performance and perceived usability,” Appl. Ergon., vol. 41, no. 3, pp. 403–10, M ay 2010. 28. “ISO/IEC 25010:2011, Systems and software engineering -- Systems and software Quality Requirements and Evaluation (SQuaRE) -- System and software quality models,” Int. Organ. Stand., 2011. 29. C. Grégoire, G. Roberge, and É. Archambault, “Bibliometrics and Patent Indicators for the Science and Engineering Indicators 2016,” SRI Int., no. January, 2016.

Methodology for the Analysis of household appliance Consumption Blanca Nydia Pérez Camacho, Juan Manuel González Calleros Benemérita Universidad Autónoma de Puebla, Av. Sn Claudio y 14 sur Esq. Av. Ciudad Universitaria, C.P. 72570, Puebla, Puebla, M éxico {blancanydia.perezc,jumagoca78} @gmail.com

Abstract. A demand system is designed in order to analyze and control the way in which household appliances consumption occurs in a context. This type of system has four objectives: to reduce consumption, reduce costs, reduce peak average to ratio and maximize the comfort of the users. The design of these systems can be based on the implementation of IoT. When using IoT for the development of demand systems, a layer of information is integrated. The information is showed and obtained through an interface, which let that the user could take decisions. This paper introduces a proposed methodology to develop demand systems. Keywords: M ethodology for Demand System, Internet of things, Analysis of household appliance consumption

Introduction

Until now, a methodology for analyzing electric consumption has not been developed from the perspective of the implementation and use of data in an IoT architecture. Internet of Things (IoT) has been implemented in the way to develo p different application areas, like Smart Warehouse, Smart grid, Smart City, Smart meter, Smart Healthcare and Smart Home [1]. There is a common objective to accomplish between the Smart grid and Smart home that is to develop systems that make an efficient electric consumption. This kind of systems is called Demand Systems (DS). The works focused in the search of the improvement of the electric consumption are about devices or control variables [2-5] buildings consumption [3], houses [4], costs [2,7-10] sensor infrastructure [3] and control algorithms [4,10]. Few works focus on the current challenge of developing solutions supported by IoT. Demand Systems analyses how the electric energy is used, their objectives could be one or more from these: low Peak Average Ratio (PAR), minimize costs, minimize consumption and maximizer user´s comfort [7]. To develop DS requires: define objectives, variables and what kind of factors needs to be cons idered. When DS works with consumption data needs to handle a big quantity of data [8] that could use to make decisions [13, 14] and converts into a multi-objective optimization problem [10]. A metaheuristic is used to find a solution to the multi-objective problem. So, according to the characteristics of the problem, it requires to search for alternative solutions that are appropriate to the context of the problem, which consider

the source of the data, the variables and the implementation of a metaheuristic to search the optimal consumption. Rest of the paper is organized as follows, in Section 2 describes the literature review. A description of the layers that make up the methodology proposed to develop projects that implemented Internet of Things is given in Section 3. In Section 4, a brief discussion is carried out. Both conclusions and work to be done in the future are announced in Section 5.

Literature review

A Demand System (DS) is a system designed to efficiently analyze and control the electricity consumption of a given context. Many researchers around the world worked to make an optimal electric consumption system in different research lines smart meter, smart grid, neural network, metaheuristic, IoT, Genetic algorithm and big data. Nadeem et al. in [11] considered to develop a DS to reduce consumption under a predefined level, costs and waiting time using hybrid metaheuristics schemes based on Teaching-Learning techniques. The metaheuristics used were Optimization Stopping Rule (OSR), Genetic Algorithm (GA) and Firefly Algorithm (FA); in this work is combined OSR-GA, OSR-TLBO, and OSR-FA. This work considered three electric devices (fridge, dishwasher, and dryer), the variables that were considered for every device: costs average per month, the cost reduction, the priority and the delay per hour per day. It is simulated by the implementation of every technique and their hybrids to each of the devices. Yao et al. in [12] said that the energy management problem consists in to solve appliance load scheduling and grid power dispatching under a single optimization framework of a utility grid with dynamic costs, a photovoltaic module and the household appliance with three different types: interruptible, uninterruptible and time varying; in this work was implemented a simulation of a mixed integer linear programming framework. His future consists of two stages, first to implement a genetic algorithm, second to implement a multiobjective optimization framework. Rahim et al. in [11] described the goal of implementing a demand system, which is reduced to the following to reduce the costs, minimization of Peak to average ratio and maximize comfort. In this work is proposed a generic architecture for a demand system, it models the electric consume in a house, here was considered three algorithms genetic algorithm (GA), binary particles swarm optimization (BPSO) and ant colony optimization (ACO). The GA was more efficient than BPSO and ACO in term of consumption reduction, minimizing PAR while is considered user comfort. In this work was concluded that is still an open problem to Minimize PAR, cost, consumption and maximize comfort. Javaid et al. in [13] focused over to control electric consumption and maintenance the comfort taking into consideration the user preferences. Was implemented four algorithms: genetic algorithms (GA), teaching-learning base on optimization (TLBO), enhanced differential evolution (EDE) and enhanced differential teaching-learning algorithm (EDTL). The consumption model took two types of devices flexible and inflexible, device, consumption and schedule of use, this model is implemented in a microgrid context.

On the other hand, Avaid et al. in [3] is presented an electric consumption model that consider houses and devices mount; the kind of devices are interruptible, not interruptible and regular application use. This model is for a demand-side management to reduce peak average to ratio, costs and renewable energy that are considered tariff and time of use. The algorithms that were analyzed are bacterial foraging optimization algorithm (BFOA), genetic algorithm (GA), binary pa rticle swarm optimization (BPSO), wind-driven optimization (WDO). In this work, a genetic binary particle swarm optimization (GBPSO) is proposed. Results showed that GA is better than others in term of PAR reduction and execution time, BPSO is better than the others in term of cost reduction. GBPSO is more efficiently than the others in terms of cost and PAR. Complementary work, is presented by Hao and Wang where they proposed a demand-side management (DSM) in [14], this DSM reduced cost in terms of tariff and costs, devices identified are interruptible and not interruptible. This DSM implemented a game theory which one can reduce costs and help to increase load demand at the off-peak time. The peak time hours were between 8 to 24 hours. Mohsin et al. developed a demand-side management in [9] with the main purpose of reducing costs, PAR and time, maximize comfort in applying a harmony search algorithm (HSA). This work compared his algorithm with binary particle swarm optimization (BPSO), differential evolution (DE), genetic algorithm (GA), ant colony optimization (ACO), Hybrid Differential Evolution – Harmony Search (DE-HS), immune artificial hybrid algorithm, genetic hybrid algorithm, teaching -learning base on optimization (TLBO) and Shuffled Frog Learning (SFL). Devices considered in this works were defined as interruptible, not interruptible, flexible and inflexible. The results showed that HAS is better in terms of costs, PAR and consumption. In addition, Huang et al. in [15] where a set of models of demand response was used, using a gradient (PSO) based on particle swarm optimization, the main objective was to schedule the operation of appliances to save energy and reduce cost considering user convenience. The model is simulated and compared to hybrid PSO algorithm and cooperative PSO algorithm. The proposed alg orithm shows better results in a real-time application. Similarly, Matei et al.in [16] described an IoT architecture, that consist of four layers sensors, physic, digital and meta through that the data flow. In this work were proposed two times of data process, first in the physic layer and second in meta-layer. The first process to select data and reduce computational cost, and in the second process could be implemented in any algorithm or technique, the choose of it depends on objectives. Finally, Mohsin et al. concluded in [9] that a deterministic optimization is inefficient and impractical to handle a big problem. A heuristic optimization is better to implement in big data problems. Also, Silva et al. conclude in [17] that the electric consumption varies according to an hour of the day, the day of the weak and season. Hoon et al. explained in [2], IoT let to monitor, handle and control devices WEB way. The IoT implementation let to get information from data that are generated in real time. There is a challenge for smart meters to develop an intelligent environment (AmI).

A methodology to Analyze Electronic Consumption

A demand system (DS) analyze and control the electric area consumption. The DS objectives [7] are to reduce costs, reduce consumption, reduce PAR and maximizer user comfort. IoT implemented a control and monitoring infrastructure through the web to devices that are in the same red, so it´s possible to develop a DS using this concept. The IoT data flow, its obtained from [16] where describes IoT architecture and [18] describes cyber-physic behavior (see Fig. 1). The IoT data flow, flow through five layers called sensor, there is a sensor architecture; physic, here is implemented an ambient intelligent (AmI), AmI extracts the context and in this stage is made a first process whit sensor and user data; the data extracted pass to information module, where all the context data are gathered; behavioral awareness is processed the data and found a behavior; digital, here is created a logical object that represents the real objects; in the meta-layer all the logical objects data are processed according to particular objectives of the problem to be solved. When is developing a DS and it’s integrated an information layer, which happens when IoT is implemented, the problems became a multi-objective optimization problem and are solved using metaheuristics [19]. Also, is necessary to identify all variables that determine the electric consumption, to know the types of devices, to identify a meta-heuristic algorithm and to propose an electric consumption model.

Fig. 1. DS data flow in IoT, own creation from [18] and [16].

A Home Energy Management System (HEMS) (see Fig.2), it consists of a programming module, the monitor module, a prediction module and a control unit [15]. In the programming module is where the preference data is entered, which is sent to the control unit. The control unit performs communication between the devices, by sending and receiving the signals; and it communicates with the prediction module by sending the data that allows to generate a profile of household

appliances use. The logical control unit sends the consumption of each appliance to the programming module. The development of a HEMS which implements IoT (see Fig.3) consists of a sensor layer, physical layer and a prediction module. The sensor architecture and the monitor module are in the sensor layer. In the physical layer there is the intelligent environment, which is formed by a programming module, logical control unit, and information fusion module. In the prediction module, there are three layers: behavioral awareness, the digital and the goal layer.in the way to develop a HEMS based on the use of IoT, the following variables must be identified: the variables that determine the electricity consumption, the devices, the metaheuristic techniques, and a proposed model of electrical consumption.

Fig. 2. Home Energy M anagement System architecture,[15].

Fig. 3. Home energy management system with IoT, own creation.

The methodology feasibility is going to be illustrated with a case study with a real case scenario of electronic consumption corresponding to every identified device that is frequently used in an office, house, and classroom.

3.1

Consumption and Context of Use

Context of use is a triple composed by the identification of electric consumption device, the environment where the device is used, house, office, classroom, and user’s needs or preferences. For each device, a set of common characteristics have been identified. Accordingly, to the following list:  Minimum power corresponding to the minimum electrical power needed for the device to operate. This value was estimated from public lists of electronic devices publicly available on their websites.  Maximum power corresponding to the minimum electrical power needed for the device to operate. This value was estimated from public lists of electronic devices publicly available on their websites.  time of use. The average daily usage time measured in hours .  kind of device. The devices are classified according to the most common appearance. For example, office supplies, video games, household appliances, electronics, lighting and security devices.  The type of devices corresponding to the a category proposed in related works [6,20,21] that are: interruptible (I), non-interruptible (NI), flexible (F) and notinflexible (NF). An Interruptible system refers to a system that can be turned on and off without any constraint. Interruptible devices are defined like devices that could be used in anytime, and the time of use varies according to the user needs [9,12,13,14,21]. Not-interruptible devices stop their function once they are finished, their consume could be variable or constant [9,11,12,14,21]. Flexible devices, their functions could be stopped and continue at another time, could be on standby too [3,9,12,21] Not-flexible devices, they could not be turned off because is necessary that they have a constant function [9,21].

3.2

Sensor Layer

Sensor Layer stores sensors raw data, also known as primary data, corresponding to the sensors’ consumption readings. In a context the set of devices may vary and the corresponding arrangement as well, we call this setup the Sensor Architecture. So, from this layer the output is a data base with the consumption readings, consumption meters, of each device that is connected to the sensor architecture of the electricity facility under a specific context. Data acquisition nowadays could be the result of using IoT devices, or we could even build a no-invasive potentiometer using Arduino Uno and SCT-013 sensors, or any related technology. Data consumption files are obtained and stored in a DB, a sample set is listed as follows in Table 1.

74 Table 1. Consume list from a laptop. Current Power (Amperes) (Watts) 6.4871

1423.1668

3.9958

878.9938

4.4403

976.8803

4.7000

1059.5591

12.5839

2768.4433

9.4061

2009.3459

8.5427

1894.3894

4.9717

1059.7808

3.8485

846.6782

6.0309

1327.0713

18.2753

4020.5761

6.7826

1492.1634

6.0730

1336.0534

6.6512

1409.5651

3.1939

702.6623

This is an input fusion layer and the classifier. The current work just consider data from a specific context which every household appliance monitored by a non-invasive smart meter that represents a specific consumption that is determined by the occupants’ behavior. In our example, we implemented a database (DB) using MySQL Workbench, DB is in a local server. The DB has 268 registers. Every register is stores: <id_device, device_name, min_ power, max_power, freq_time, shut_down, kind_device>, some registers are showed in Table 2. 3.3

Physical Layer

The physical layer corresponds to the composition of the primary data and a user data that corresponds to a set of constraints. For this purpose, an interactive system is recommended a human-computer interface. The external data is the readings obtained from the room sensors. In a demand system it is identified that the user data have to do with the preferences and times of use of the household appliances that are monitored through the sensor architecture; and the external data are the data obtained from the environmental conditions (temperature, humidity, presence, etc.), that is, all those data that result in a high degree of importance to be considered for the conditioning of the use of electrical devices within the area you want to control. In our case, we have built a website (see Fig. 4) where users select their devices, graphically, and coherent with the proposal of building better Legos [22].

Table 2. M any registers from household appliances DB. ID Device_name device Cannon (three OF1 light) Cannon (three OF2 light) Cannon (one OF3 light) Cannon (one OF4 light) OF5 CPU CPU iM ac OF6 3.06GHz

M in power 6.4

M ax_power Freq_time (use Hrs/day) 400 1.5

Shut_down Kind_device Yes

6.4

1000

1.5

Yes

6.4

220

1.5

Yes

6.4

236

1.5

Yes

14.1

250

Yes

129.6

365

Yes

OF9

Digital decoder

Yes

OF10

Scanner

275

Yes

OF11

Fax

150

Yes

Fig. 4. A web site used to help users to select the physical set of devices available at their facility.

In the website is possible to register new users and unsubscribe them (see Fig. 5). To each one of users can define their context identifying household appliances (see Fig. 6) and make reports and graphs (see Fig. 7).

Fig. 5. High and low users in the web site used.

Fig. 6. High and low household appliances used to define the context.

Fig. 7. M ake reports and graphs of household appliances

In the physical layer, is founded a fusion module, too. This module works as a repository of all the data that must be considered (primary, external and user data) to solve the problem that is being studied; to later be able to group them according to the needs of solution. The output resulting from this module is grouped data.

3.4

Prediction module

The prediction module is formed by three layers, which are the behavioral awareness layer, the digital layer and the meta layer. A behavioral module is in the behavior layer, in which methods or procedures can be implemented to generate the necessary parameters according to the problem to be solved, this module includes both the user's data (which determines the restrictions to consider) as environmental data (which are the conditions for decision making). Demand System in the behavior module, a behavioral model can be imp lemented that identifies the probability and time of use of the devices, giving as output the on and off schedule, and the operating time of each device; and integrate, a module capable of generating the actual consumption data that each device can have. A ll the resulting data pass in vector form to the next layer. The digital layer is where the data from the previous layer arrive and a space is formed with the vectors that represent the possible behavior scenarios of all the objects that are in the real world. The vector space is sent to the next layer. In a demand system, we have a set of vectors formed from the combinations of possible consumptions that each of the devices that are being monitored may have. In the meta layer, procedures are implemented that seek proposals for solutions to the objectives that are to be achieved according to the proposed scenario. A demand system is a multi-objective problem and to implement it, a metaheuristic is implemented.

Discussion

From the methodology described in the previous section it is possible to identify how the flow of the data is carried out when it is integrated into IoT to solve a problem. In the case of implementation of a demand system consists of selection of devices for reading consumption data, the development of an intelligent environment that serves as an user interface and serves to read the consumption data to perform a context extraction, develop a module to merge the data received, develop a behavior model that will be implemented in a behavior module, obtain logical objects that emulate the real object, and implement a metaheuristic whose definition of objective function is based on the objectives (objectives of DS system) to meet the original problem. Has been identified a methodology to identify consumption variable, which consists to identify the household appliances that are the most frequently in a house, office, and classroom; the maximum and minimum power: the frequently time of use according to the area in which demand system is going to be implemented; and the classification to which the device belongs (I, NI, F, NF). With the database, devices are possible to simulate N different scenarios with d devices each one. The data flow in demand system was obtained from two different ideas, one tha t talks about an architecture that would be considered when is implementing an IoT project, and another one when is talking about a context where is necessary to take in a count the information. The selection of metaheuristic to be implemented is based on what has been identified in the literature, in which the constant of selecting a genetic

algorithm is identified against whose results are compared to another metaheuristics implementation.

Conclusion and future work

This methodology was developed after having analyzed articles that have as objective the development of demand systems, where it is appreciated that the development layers in which they focus only cover one or two of the methodology proposed in the present work. It is appreciated that if the methodology proposed in the present work that uses IoT as part of the implementation process is followed so could be to identify and to comply with every one of the requirements to be met. The proposed methodology is used as the basis to develop a demand system, where the appliances were identified as their consumption variables too. As future work, the creation of an intelligent schedule is left to foresee with the probability of use both the schedules and the time of consumption of each device that are contemplated in the scenario. As well as, the implementation of a simulator to propose a household appliance use and the metaheuristic which will made an analysis and will give a user recommendation according to the user-needs and specific objectives.

References 1. A. Parsa, “Implementation of Smart Optimal and Automatic Control of Electrical Home Appliances ( IoT ),” 2017. 2. T. hoon Kim, C. Ramos, and S. M ohammed, “Smart City and IoT,” Futur. Gener. Comput. Syst., vol. 76, no. July 2014, pp. 159–162, 2017. 3. N. Javaid et al., “A new heuristically optimized Home Energy M anagement controller for smart grid,” Sustain. Cities Soc., vol. 34, no. July, pp. 211–227, 2017. 4. S. R and S. K. R, “Data M ining with Big Data,” 2017 11th Int. Conf. Intell. Syst. Control, pp. 246–250, 2017. 5. S. Kakran and S. Chanana, “Energy Scheduling of Smart Appliances at Home under the Effect of Dynamic Pricing Schemes and Small Renewable Energy Source,” Int. J. Emerg. Electr. Power Syst., pp. 1–12, 2018. 6. M . Jahn, M . Jentsch, C. R. Prause, F. Pramudianto, A. Al-Akkad, and R. Reiners, “The energy aware smart home,” 2010 5th Int. Conf. Futur. Inf. Technol. (FutureTech), IEEE., pp. 1–8, 2010. 7. S. Rahim et al., “Exploiting heuristic algorithms to efficiently utilize energy management controllers with renewable energy sources,” Energy Build., vol. 129, pp. 452–470, 2016. 8. D. C. M ocanu, E. M ocanu, P. H. Nguyen, M . Gibescu, and A. Liotta, “Big IoT data mining for real-time energy disaggregation in buildings,” 2016 IEEE Int. Conf. Syst. Man, Cybern. SMC 2016 - Conf. Proc., pp. 3765–3769, 2017. 9. S. M . M ohsin, N. Javaid, S. A. M adani, and S. K. Abbas, “Appliance Scheduling in Smart Homes with Harmony Search Algorithm for different Operation Time Intervals

10.

11.

12. 13. 14. 15. 16.

17.

18. 19. 20.

21.

22.

Appliance Scheduling in Smart Homes with Harmony Search Algorithm for different Operation Time Intervals,” no. February, 2018. C. Fauvel, F. Claveau, P. Chevrel, and P. Fiani, “A flexible design methodology to solve energy management problems,” Int. J. Electr. Power Energy Syst., vol. 97, no. November 2017, pp. 220–232, 2018. Z. Nadeem, N. Javaid, A. W. M alik, and S. Iqbal, “Scheduling appliances with GA, TLBO, FA, OSR and their hybrids using chance constrained optimization for smart homes,” Energies, vol. 11, no. 4, pp. 1–30, 2018. L. Yao, Z. Damiran, and W. H. Lim, “Energy M anagement Optimization Scheme for Smart Home Considering Different Types of Appliances,” 2017. N. Javaid et al., “Demand Side M anagement in Nearly Zero Energy Buildings Using Heuristic Optimizations,” Energies, vol. 10, no. 8, p. 1131, 2017. Y. Hao and W. Wang, “Optimal Home Energy M anagement With PV System in Time of use Tariff Environment,” no. 51641702, pp. 5–9. Y. Huang, H. Tian, and L. Wang, “Demand response for home energy management system,” Int. J. Electr. Power Energy Syst., vol. 73, pp. 448–455, 2015. O. M atei, C. Anton, S. Scholze, and C. Cenedese, “M ulti-layered data mining architecture in the context of Internet of Things,” Proc. - 2017 IEEE 15th Int. Conf. Ind. Informatics, INDIN 2017, pp. 1193–1198, 2017. B. N. Silva, M . Khan, and K. Han, “Load balancing integrated least slack time-based appliance scheduling for smart home energy management,” Sensors (Switzerland), vol. 18, no. 3, 2018. S. Chen et al., “Butler, Not Servant: A Human-Centric Smart Home Energy M anagement System,” IEEE Commun. Mag., vol. 55, no. 2, pp. 27–33, 2017. “Fauvel_flexible design methodology.pdf.” . L. Yao, J. Y. Shen, and W. H. Lim, “Real-Time Energy M anagement Optimization for Smart Household,” Proc. - 2016 IEEE Int. Conf. Internet Things; IEEE Green Comput. Commun. IEEE Cyber, Phys. Soc. Comput. IEEE Smart Data, iThingsGreenCom-CPSCom-Smart Data 2016, pp. 20–26, 2017. S. Kakran and S. Chanana, “Energy Scheduling of Smart Appliances at Home under the Effect of Dynamic Pricing Schemes and Small Renewable Energy Source,” Int. J. Emerg. Electr. Power Syst., pp. 1–12, 2018. S. Jenson, “The Future IoT: Building Better Legos,” Computer (Long. Beach. Calif)., vol. 50, no. 2, pp. 68–71, 2017.

A Fuzzy-Based Groupware Usability Evaluation Methodology Luz A. Sánchez Gálvez

1,2 ,

Juan Manuel Fernández Luna2 and Mario Anzures García1

Benemérita Universidad Autónoma de Puebla, Facultad de Ciencias de la Computación, Puebla, M éxico 2 Universidad de Granada, Departamento de Ciencias de la Computación e Inteligencia Artificial, E.T.S.I. Informática y de Telecomunicación, Granada, Spain sanchez.galvez@correo.buap.mx,jmfluna@decsai.ugr.es, mario.anzures@correo.buap.mx

Abstract. Groupware typically provides a shared workspace, which is displayed through shared user interfaces for group members, with appropriate resources to accomplish a common task. Therefore, the design of this type of applications should always be accomplished considering the users’ interaction and the utilization of the shared resources. Consequently, a commendable design of groupware requires that the shared user interfaces offer a high level of usability. In this paper, a groupware usability evaluation methodology , specifically for the shared user interfaces, is proposed. This methodology provides an evaluation instrument that collects the users' perceptions through four dimensions — effectiveness, efficiency, learnability and satisfaction— In addition, the gap theory of quality service, and a fuzzy linguistic approach using aggregation operators —which operate directly with words— are applied. Consequently, the methodology shows to be a significant, innovative contribution to the research area on groupware usability evaluation. A case study to validate methodology, is presented. Keywords: Usability Evaluation M ethodology, Groupware, Fuzzy Linguistic Approach, Aggregation Operators, SERVQUAL

Introduction

Nowadays, many organizations or groups of persons whose members are geographically distributed, must carry out their work in an effective way. This demands the development of groupware or collaborative application supporting group work. Groupware is a computer-based system that supports groups of people who are engaged in a common task (or goal), and it provides an interface to a shared environment [1]. This definition implicates that the development of groupware must be focused in order to manage and control a user’s group by supplying a s hared workspace, which provides communication, collaboration and coordination among these users. However, many people get frustrated with groupware that do not adequately support them in their collaborative work, in other words, groupware does not match th e users’ tasks and needs. This occurs, since user interface design often relies too much on

trial-and-error techniques. The methods and techniques for user interface design need to be improved in order to systematically develop better groupware systems. Co nsequently, groupware evaluation is necessary in several situations. However, evaluation is frequently done in an ad-hoc manner or not at all. On the other hand, usability is a discipline of applying sound scientific observation, measurement, and design principles to the creation and maintenance of systems in order to bring about the greatest ease of use [2]. The concept of usability is also related to methods to improve the easiness of its use throughout the design process. It is a quality attribute that assesses how simple it is to utilize user interfaces. If an interface is not able to satisfy the needs of their users, they will not be successful in the long term. Therefore, the aim is to design both usable and useful groupware. In such a way that, it is important to base the design on the work that must be done by the users´ interaction. Thus, the users play an important role in acquiring knowledge about their work as well as for usability testing. As a consequence, a groupware usability evaluation methodology, specifically for shared user interface, and a case study of groupware usability evaluation, are presented in this paper, The methodology proposes an evaluation instrument, which collects the users’ perceptions by four dimensions —effectiveness, efficiency, satisfaction and learnability— In addition, the gap theory of service quality and a fuzzy linguistic model using aggregation operators of linguistic information —which operate directly with words— are used. Therefore, the methodology proves to be a significant and innovative contribution to the area of groupware usability evaluation research. The paper is organized as follows: Section 2 briefly introduces to the disciplines —groupware and usability— supporting the groupware usability evaluation methodology. Section 3 shows the background (groupware evaluation, SERVQUAL, linguistic approach). Section 4 details a fuzzy-based groupware usability evaluation methodology. Section 5 presents the application and validation of the methodology. Finally, conclusions and future works are depicted in Section 6.

Disciplines supporting proposed methodology

Two are the disciplines aiding the suggested groupware usability evaluation methodology: groupware, and usability. 2.1 Groupware Groupware provides the shared workspace, where users will perform group work; allowing [3]: Communication, it enables the interaction among persons, which is characterized by the users, the information that is shared, and the means or artefacts used; Collaboration, it denotes a higher degree of users participation for achieving a certain goal, through their express participation in group activities; and Coordination, it manages dependencies among activities carried out into a group to accomplish an objective; providing concurrency mechanis ms to reduce racing conditions and guarantee mutually exclusive resource usage. The session is a term to denote a shared workspace where a group of individuals, geographically distributed, which share a common interest to perform collective tasks.

The session is supported by a manager to control and manage it, this is established by polices that define the group organizational structure to support interaction on such workspace. The users’ interaction is coordinated by the concurrency mechanism to avoid conflicts, due to the cooperative and competitive activities among users, by supplying dynamically generated, temporary permissions to collaborating users. The permissions granted to users depend on the roles that they can play and specify which user can send, receive, or manipulate shared data at a given moment. Furthermore, to cooperate, users must be aware of the presence of other members during the session and of the actions that each one has carried out and is carrying out (group awareness). In such way, a common context be established, and activities coordinated, thereby avoiding surprises and reducing the probability of conflicts in the group. This is achieved by using a mechanism notification, which to ensure the consistency of shared information; indicating to each participant happens it in this shared workspace. As well, in groupware a group memory should be proven, as it is important to count with a historical about use of the shared resources. 2.2 Usability In the literature, there are several definitions of the concept of usability, which provide different and complementary points of view in each research area, which shows the evolution of the term itself from the evolution of knowledge. In Human–Computer Interaction (HCI), the most widely accepted definition of usability is the provided for ISO 9241-11[4]: “the extent to which a product can be used by specified users to achieve specific goals with effectiveness, efficiency and satisfaction in a defined context of use’’. So, usability implies the interaction of users with the software product and can be seen as the product’s capability to meet customer expectations. Nielsen, defines: “usability is a quality attribute that assesses how easy user interfaces are to use” and further breaks down the concept into the following five quality components [5]: Learnability: How easy is it for users to accomplish basic tasks the first time they encounter the design? Efficiency: Once users have learned the design, how quickly can they perform tasks? Memorability: When users return to the design after a period of not using it, how easily can they reestablish proficiency? Errors: How many errors do users make, how severe are these errors, and how easily can they recover from the errors? Satisfaction: How pleasant is it to use the design? These definitions of usability directly affect how it is evaluated, since each method or technique employed in these evaluations may focus on distinct aspects of the term usability (e.g., effectiveness of user task, learnability of user interfaces). Therefore, for establishing the groupware usability evaluation methodology, in this paper, the groupware usability is defined as: “The degree to which a groupware supports the interaction among group users, as well as between the users and the shared workspace” This definition assumes that a groupware system is already usable by examining if the group interaction presents: Effectiveness, Efficiency, Learnability, and Satisfaction (see Table 1).

Table 1. Dimensions of methodology. Dimension Effectiveness Efficiency Learnability

Satisfaction

Sub criteria

Definition

It refers to successful completion of collaborative tasks, where a user group achieve common goals at each stage. It refers to the resources (such as time or effort) required to carry out a successfully collaborative task by the group interaction. T he groupware should be easy to learn and to understand; it should be easy for the group to achieve a collaborative task by using the shared workspace. Easiness of use It refers to the group perception of the use of the shared workspace. Information It is to assess whether the structure, design, and organization of the Organization system reach the group' goals. It refers to the clear labeling of the groupware from the user group Clear point of view, and whether the terminology used is easy to underLabeling stand. Visual It evaluates the groupware design concerning its visual appealing, as Aspect well as the group awareness and group memory provision. Error It has to do with the easiness for recovering from errors made in Recovery each task by the group. It refers to the easiness that group may have to access from one Navigability shared interface to another.

From previous research [3, 6, 7] have been identified the elements assisting to this interaction:  Group organizational structure. It defines the division of labor based on roles that users play, tasks that they performance, resources that use, stages in that roles participate, and user interfaces show.  Group awareness. It allows a user to know the other user is carrying out in the shared workspace —session— through a notification mechanism; which informs happening it in this space. In such a way that a common context is established, allowing the users collaboration in performed activities for avoiding surprises and reducing the probability of conflicts in the session.  Interaction coordination. It presents a concurrence mechanism to manage temporary user’s permissions for using the shared resources, guaranteeing mutually exclusive usage of these and reducing racing conditions.  Groupware security. It refers both the authentication and access control mechanism, which allow to access to groupware (authorized users) and to resources (participants roles in a certain stage). On the other hand, a usability evaluation is a procedure which is composed of a set of well-defined activities for collecting usage data related to end -user interaction with a software product and/or how the specific properties of this software product contribute to achieve a certain degree of usability [8, 9].

Background of the evaluation methodology

The proposed methodology is based on groupware evaluation analysis, theory of gap, and linguistic approach. 3.1 Related work In this section provides a brief background of the groupware usability evaluation methods, presenting some fundamentals ideas about these methods and classifying them in two groups: evaluators’ perception-based evaluation methods and users’ perception-based evaluation methods . 3.1.1 Expert evaluators-based evaluation methods In these methods, the evaluation is based on evaluators’ perception:  Groupware Heuristic Evaluation [10] is an adaptation of the Nielsen’s heuristic evaluation methodology [5, 11, 12, 13] to groupware. A set of eight usability heuristics that evaluators can use to inspect shared workspace and see how users support teamwork. Evaluators identify usability problems recording the violated heuristic, a severity rating and optionally, a solution to the problematic. The problems are then filtered, classified and consolidated into a list, which is used to improve the application.  Groupware walkthrough [14] is a usability inspection technique based on cognitive walkthrough [15], for single-user software that incorporates user descriptions and tasks. It allows designers to evaluate software in the early stages of development. It involves three types of contextual information: a description of the users and the knowledge they possess; descriptions of the tasks the users will perform with the system; and a list of the correct actions that a user must perform to accomplish the tasks with a prototype. It has two components: a group task model for identifying and analyzing real-world collaborative tasks that captures the variability and multiple courses of action in group work; and a walkthrough process for assessing a system’s support for those tasks that guides evaluators as they step through tasks and evaluate the groupware interface. Once scenario specifications and task analyses have been compiled. Evaluators step through the tasks and determine how well the interface supports group members in working toward and achieving the intended outcome.  Human-Performance Models [16] such as the Keystroke-Level Model [17] are based on a cognitive architecture that approximates single-user interaction at a low level of detail (e.g. perceptual, motor, and cognitive processors), which is adapted to groupware: the conventional information flows are considerably changed to reflect collaborative actions, mutual awareness, and interdependence; and the focus and granularity should not remain on the interactions between the user and the physical interface but should significantly change to reflect the interactions among users, mediated by the physical interface. Evaluators define physical interface that may be decomposed into several shared workspaces , and critical scenarios focused on the collaborative actions for shared workspaces. Fi-

nally, they compare group performance in the critical scenarios to p redict execution times.  “Quick-and-dirty” Ethnography [18]. Ethnography refers to the qualitative description of human social phenomena to produce detailed descriptions of the work activities of actors within specific domains. The intention is to see activities as social actions embedded within a socially organized domain and accomplished in and through the day to day activities of participants. Evaluators carry out brief studies of the group work to provide it and suggest to designers the deficiencies of a system, thus allowing existing and future systems to be improved. “Quick and dirty” provides a broad understanding, which is capable of sensitizing designers particularly to issues which have a bearing on the acceptability and usability of an envisaged s ystem rather than on the specifics of design.  Scenario-Based Evaluation [19]. A scenario is “a concrete description of an activity that the user engages in when performing a specific task, a description sufficiently detailed so the design implications can be inferred and reasoned about”. In addition a complete scenario includes an actor, which may be a specified individual but can be represented as a prototypical role; a setting describing the context in which the scenario occurs; task goals describing the motivation behind performing a scenario; and claims, which are statements about the effects or consequences of using the system [20]. Scenarios and claims provide a tangible, traceable basis for both general and concrete analyses, which are based o n actual use of the system. This technique suggests evaluative claims about features of the system, and contextual information about the organization. It is useful for identifying failures to system use that need enhancement.  Knowledge Management Approach [21]. This method is based on knowledge and knowledge flow. It measures whether the tool helps users to detect knowledge flows, to disseminate them, to store previous experience and to reuse it. The knowledge flow process is comprised of six phases (knowledge creation, accumulation, sharing, utilization, internalization), which are also the areas to be evaluated by this approach. To perform evaluation, each area has a list of associated questions, which may be used as a checklist by evaluators. These methods although evaluates the usability, however, they do not make use of information about users and their work contexts. 3.1.2

Users perception-based evaluation methods

In these methods, the evaluation is based on users’ perception:  Groupware Task Analysis [22, 23] combines high-level hierarchical task analysis and field observations for addressing all stages of groupware design. It is based on a conceptual framework including agents, group work, and situation, similar to the work models defined by the Contextual Design approach [24].  Collaboration Usability Analysis [25] is a task analysis technique designed to represent collaboration of a group task in a shared workspace. It is based hierarchical task analysis, which provides flexibility in the ways that tasks are composed and executed. The task hierarchy includes scenarios to describe the high level context of the collaborative situation, tasks to indicate specific goals within

the scenario, and task instantiations to represent individual and collaborative ways of carrying out the task.  Mechanics of Collaboration [26] is a conceptual framework for developing discount usability evaluation techniques that can be applied to shared -workspace groupware. The framework bases on to support mechanics of collaboration: the low level actions and interactions that must be carried out to complete a task in a shared manner. These include communication, coordination, planning, monitoring, assistance, and protection. The framework also includes gross measures of these mechanics: effectiveness, efficiency, and satisfaction. In addition, proposes that various single-user usability testing schemes (such as: questionnaires, walkthrough, heuristic evaluation, task analysis ) can be applied to groupware by having them assess support for the mechanics.  Perceived Value [27]. The approach is centered on a variable –Perceived Value– measuring several external product attributes of meetingware that can be negotiated between developers and users. This evaluation concerns measuring the Perceived Value attributed by the users to the technology. Perceived Value can provide a metric about the alignment of the developers’ and users’ expectations. This metric can then be used to make preliminary decisions about the feasibility of a meetingware project, and intermediate assessments of the development process.  E-MAGINE [28]. The aim is to support groups to efficiently assess the groupware applications that fit them best. The method focuses on knowledge sharing groups and follows a modular structure. It is based on the Contingency Perspective in order to structurally characterize groups in their context, and the new ISOnorm for Information Communications Technology tools, Quality in Use. It comprises two phases. The initial phase leads to a first level profile and provides an indication of possible mismatches between group and application. The formulation of this profile has the benefit of providing a clear guide for further decisions on what instruments should be applied in the final phase of the evaluation process. These proposals evaluate groupware by using the shared workspace context through task analysis, mechanics of collaboration, group setting, and perceived value. However, they do not consider the complete interaction in shared space, quality of service, questionnaires to capture users’ perception, nor the information evaluation in a qualitative way. Therefore, this paper provides a groupware usability evaluation methodology. 3.2 SERVQUAL SERVQUAL [29] evaluation methodology is based on the theory of service quality— assessment applied in the environment of enterprises and organizations. The key element is the customer satisfaction. Service quality is related to diminishing the distance between the customers’ expectations and his final perception. According t o SERVQUAL, customers will evaluate, positively or negatively, the quality of a service where their prior perceptions were either higher or lower than expected. In SERVQUAL to evaluating the service quality, Minimum required level of service; Expected level of service; and the level perceived by the use, are considered. Based

on the users’ feedback, it is possible to define two variables for detecting the strengths and weaknesses i.e. Adequacy of Service (the difference between the perceived value and the minimum value) indicating the areas where the service is below the level expected by the user, and Service Excellence (the difference between the perceived value and the expected value) that identifies areas where it provides a better service than that expected by the user. 3.3 Linguistic approach The information cannot always be evaluated in a quantitative manner, sometimes it is necessary to do it qualitatively. The existence of qualitative variables inherent to human behavior, or external environment elements which are difficult to quantify objectively lead individuals to express their opinions better, by using linguistic terms instead of precise numerical values. A linguistic variable differs from a numerical one in that its values are not numbers but words or sentences in a natural or artificial language [30]. When a linguistic model is used, the existence of a suitable set of terms or labels according to the problem domain is assumed, then, the individuals can express their perceptions. The ordinal fuzzy linguistic model [31] is very useful as it simplifies the computing by eliminating the complexity of having to define a grammar. An ordinal fuzzy linguistic modeling [32] is used in this paper to represent the users’ perceptions with words, based on the linguistic aggregation operators LOWA and LWA [31], in order to evaluate groupware usability. The Linguistic Ordered Weighted Averaging (LOWA) is an operator used to aggregate non-weighted ordinal linguistic information, i.e., linguistic information values with equal importance. The Linguistic Weighted Averaging (LWA) is an operator used to aggregate weighted linguistic information, i.e., linguistic information values have different importance. In order to calculate both operators, this paper follows the definitions established on [30].

A Fuzzy-based groupware usability evaluation methodology

In this work, a groupware usability evaluation methodology, specifically for the shared user interface has been developed. Therefore, this methodology takes an approach that requires the establishment of dimensions to measure usability, based on standards: ISO 9241-11 [4] and Nielsen´s definition [11]. Consequently, it considers four dimensions: effectiveness, efficiency, s atisfaction, and learnability. These in 20 items —which capture the users' perceptions to assess the usability degree of groupware— have been represented. Furthermore, the methodology uses a fuzzy linguistic model by aggregation operators with linguistic information, which handle words directly. They are important to allow sorting and classifying all data from an aggregation process —without any loss of linguistic information—This methodology, also, draws on SERVQUAL that is based on the gap theory of service quality and focused on assessing private sector institutions; by measuring the overall service quality. Deficiencies on groupware usability could be identified through SERVQUAL. Therefore, the proposed methodology of usability evaluation in this paper, could offer

commendations for prioritizing improvements and guaranteeing a proper interface design for groupware, based on users’ preferences.

Application and validation of the methodology

The usability evaluation methodology is based on the development, production, implementation, evaluation and reliability of the questionnaire, as well as the results' analysis, from where a series of recommendations to improve the groupware usability can be provided. For implementing this methodology the following stages are performed: 1. Dimensions of usability. 2. Questionnaire. 3. Evaluation. 4. Results. 5. Commendations.

5.1 Dimensions of usability The usability is a multidimensional concept. In this paper, four dimensions of usability —effectiveness, efficiency, satisfaction and learnability— are proposed to assess groupware usability (see Table 1). 5.2 Questionnaire An important part of the usability research has been the designing of a questionnaire. The items for measuring the groupware usability, were based on the literature on usability evaluation studies. First, it was necessary to establish the dimensions. To be able to generate measurement items; frameworks, guidelines, and usability testing [30, 33, 34] were reviewed. All measurement items chosen, were modified to reflect the unique features of groupware. Thus, twenty items establish the questionnaire (see Table 2) to capture the users' perceptions on the rate of groupware usability, based on the proposed dimensions of usability. Users must answer the questions about their personal experience when interacting with the groupware. 5.3 Evaluati on The evaluation instrument has been applied to the groupware implemented by a semantic methodological approach [6, 7]. This includes two roles: 1) The professor can register himself and enter to this shared workspace, create groups, upload and download files both his and those of the students, publish messages and respond to those made by others; and 2) The students, can register themselves and enter in the shared space and access courses, load and unload files (homework and course ma terials that have been uploaded by the professor), make publications and to reply to them. Therefore, loading files and publishing message are two essentials collaborative interactions, which make use both the shared workspace and the group awareness.

In the case study, the participants were asked to fill out a pre-questionnaire concerning demographic data —level of education, age, and gender, use frequency of the groupware and level of computer skills. A total of 43 users participated in the groupware usability evaluation—including students and professors of undergraduate. The pre-questionnaire was analyzed using frequency charts to determine the type of user, who responded to the survey based on their age, sex, level of education , computer skills, and groupware usage. The usability evaluation questionnaire was examined bearing in mind the following steps: 1. The users expressed their judgment by completing the questionnaires (see Table 1). The scale for this model is S= {VL=Very Low, L=Low, M=Medium, H=High, VH=Very High}. As a result for each one of the users u j (u 1 , u 2,...u n ) and for each questionnaire item ik (i1 , i 2,...im ), m is the total number of questions; there is a tuple (mvjk,pvjk,evjk) of the minimum value —mv—, perceived value —pv— and the expected value — ev—, for each user u j and for each question ik . Table 2. Questionnaire. Dimension

Effectiveness

Efficiency

Learnability

Satisfaction

Global Usability

Item Can you usually complete a shared task by using the groupware? Can you successfully interact through the groupware system? Do the shared resources of the groupware system satisfy your needs of group awareness, and memory? In general, is the groupware system useful to help you achieve the common goal? Can you complete a collaborative task quickly by using the groupware? Do you obtain results quickly by using the groupware system? Does the user interface gives you expeditious access to groupware system shared resources? Is the access to information resources quick and easy to use? Was learning to use the groupware system easy? Are the terms used on groupware easily understandable? Is the groupware system help well organized? Are new users able to utilize the shared resources without considerable effort? What is your main fulfillment when using groupware system? What is the rate of access of groupware system? Is the organization and distribution of information on the groupware system clear? Is the language used by tags on the groupware system appropriate, organized and clearly enough? Is the groupware system visually appealing? Does the groupware system allow an easy error recovery? Are the resources offered by groupware satisfactory? What is the general usability of the groupware system?

2. To compute the global users’ opinion concerning each item ik of the tuple (mv,jk, pvjk, evjk), the following aggregation operators are used: 2.1 LOWA [30] is used if all users are considered to bear the same importance. mvk = Q (mv1k,…,mvnk) pvk = Q (pv1k,…,pvnk) evk = Q (ev1k,…,evnk)

2.2 LWA [30] is used when each user is considered to bear a different level of importance. mvk = Q ((UI(u 1 ,i k),mv1k),…, (UI(u n ,i k),mvnk)) pvk = Q ((UI(u 1 ,i k),pv1k),…, (UI(u n ,i k),pvnk)) evk = Q ((UI(u 1 ,i k),ev1k),…, (UI(u n ,i k),evnk)) Where UI(u j ,i k) S is the level of relative linguistic importance assigned to a user u j for the item ik. 3. The overall review of all questions of the tuple (mv,pv,ev) is calculated similarly to the previous step, by using aggregation operators: LOWA [30] is used when all the items are considered to bear the same importance. mvk = Q (mv1 ,…,mvm ) pvk = Q (pv1 ,…,pvm ) evk = Q (ev1 ,…,evm ) LWA [30] is used when each item is considered to carry a different level of importance. mv = Q ((II(i 1 ),mv1 ),…, (II(i m ),mvm )) pv = Q ((II(i 1 ),pv1 ),…, (UI(i m ),pvm )) ev = Q ((II(i 1 ),ev1 ),…, (UI(i m ),evm )) Where II(i k) S is the level of relative linguistic importance assigned to item ik. 4. The gap theory of service quality is applied to each item. The tolerance zone is located between the minimum and the expected values. The difference be tween the perceived and the minimum values, is called Service Adequacy —SA— and the Service Superiority —SS— is the difference between the expected values and the perceived ones. Therefore, for each item i k , SA k and SSk are computed as follows [30]: SA k = D(pv k , mv k ) SSk = D(ev k , ev k ) In the other hand, it is important to verify the reliability of the evaluation instrument, when questionnaires for evaluating the groupware usability are applied, therefore, it is advisable to use the Cronbach's alpha [35]. Cronbach’s alpha allows to quantify the level of reliability of an evaluation scale, built from k variables observed. Assuming that the variables are related to the qualitative interest data; the k variables should achieve stable, consistent measurements with a high level of correlation among themselves. A questionnaire is considered reliable when Cronbach's alpha is greater than 0.80. The formula for Cronbach's alpha is:

is the item variance i; is the item variance of all observed values; K in the item number of the questionnaire;

5.4

Results

Both a quantitative and a qualitative analysis are accomplished in the groupware usability evaluation. The qualitative analysis focuses on calculating the aggregation operators LOWA and LWA; it is based on proposed linguistic labels on the scale. The LOWA operator requires to obtain the combination of the users’ perception for each item. Thus, Table 3 summarizes the result of the combined aggregation of the users’ perception for the three assessed values: minimum, perceived, and expected values, regarding the groupware as well as their corresponding gap. On the other hand, Figure 1 shows a radar chart that summarizes the user responses to the questionnaire items on the minimum, perceived and expected levels. This type of chart was used, as SERVQUAL encourages its use to display the results obtained with the LOWA operator. As shown there in, the usability of the minimum value is perceived with a medium value by most users, unlike the perceived and the expected values which show a tendency towards a higher level of groupware usability. Table 3. Results of the LOWA operator. Item 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Minimum M L L M M M L H L L H L H VL M M M M M L

P erceived H H H H H H H H H H H H H VL H H VL VL H H

Expected H H H VL H VL H VL H H VL M M VL VL H VL VL H H

Usability Adequacy L L L L L L L L L M M M M M B B M M L L

Usability Excellence VL VL VL VL VL VL VL VL VL VL VL L+ L+ L+ VL VL VL VL VL VL

Fig. 1. Radial chart of the LOWA operator.

The LWA operator allows perceiving the opinion of all users on items with a different level of importance; which is suitable to evaluate the usability on this work,

because it contemplates four dimensions: efficiency, effectiveness, learnability, and satisfaction. So, the level of importance will vary according to the dimension that is being assessed. In case of measuring effectiveness, items 1, 2, 3, and 4 would have a VH (Very High) level of importance, while the remaining items present a VL (Very Low) level of importance as shown in Table 4. Table 4. LWA operator for the effectiveness dimension . LWA: Effectiveness Minimum Perceived M H

Expected H

Fig. 2. Chart of the 20 item.

On the other hand, concerning the quantitative analysis, item 20 has been planted to measure the satisfaction of the user’s overall groupware usability. Figure 2 displays that 19 out of the 43 respondents have a high (H) overall satisfaction when evaluating groupware usability; while 7 of them show a Very High (VH) level; 13 show a Medium (M) value, and the other 4 present a Low (L) value. As mentioned above, the evaluation questionnaire reliability, was calculated using Cronbach's alpha , obtaining a value equal to 0.89 for the minimum value; 0.91 for perceived value; and 0.93 for the expected value, which means such a reliability is fairly acceptable. 5.5 Commendati ons Overall, the groupware usability evaluation results have been satisfactory. However, the adequacy gap indicates that the improvement should focus primarily on two dimensions; learnability and satisfaction. In the former, the adjustments should be directed to improve and simplify the access control, both the shared resources and application itself. So, it is advisable to specify users, tasks, and resources involved in shared workspace in a clear, simply way. While in the latter, the changes should be oriented to improve the navigability in collaborative workspace, group awareness and memory. So, including widgets, menu, messages or icons, that allow to indicate users connected, resources used or changed, state of each user or resource, historical of each resource modified; which clearly inform and orient about the shared workspace situation. As for the interface, three commendations are made: To modify its organization to simplify and improve its navigability; to focus on the resources offered by

the groupware, eliminating or reducing those that are strange to it, and using a clear terminology that would improve the appearance of the groupware. These interface adjustments help to simplify the interaction among users, and these with the system.

Conclusions and future work

A low-cost groupware usability evaluation methodology that can identify shared workspace specific usability problems and can be considerable potential for using it to improve groupware systems usability, has been developed. In the methodology, the basic principles of usability have been combined with models of service quality (the gap theory of service quality has set the pace to suggest a number of commendations for improving systems groupware) and fuzzy logic models as the use of aggregation operators of linguistic information with a measurement scale was raised five linguistic labels specifically applied to groupware usability evaluation. In this research, a measuring instrument that collects users' perceptions based on the four dimensions to evaluate the groupware usability is proposed. Cronbach's alpha for verifying the reliability of the measurement instrument was used. A qualitative study was also carried out using descriptive statistics to compare the results with those obtained by the aggregation operators, which showed that their use for the methodology, is appropriate. Future work will focus on applying diverse linguistic quantifiers to calculate the weights of aggregation operators, which allow us improving this methodology.

References 1. 2. 3.

5. 6.

Ellis CA, Gibs SJ, Rein GL (1991) Groupware: Some Issues and Experiences. Communications of the ACM , vol. 34-1, pp. 39–58. Pearrow M (2000) Web site usability handbook. Rockland, M A: Charles River M edia. Anzures-García M , Sánchez-Gálvez LA, Hornos M J, Paderewski-Rodríguez P (2017) Facilitating the development of collaborative applications with the M VC architectural pattern. Chapter 15, Software Engineering: M ethods, M odeling, and Teaching, (Ed.) Bonaventuriana, vol. 4, pp. 268-290. International Organization for Standardization, Technical Committee of Ergonomics. 1998. Ergonomic requirements for office work with visual display terminals (VDTs): Part 11: Guidance on usability (ISO No. 9241-11). Nielsen J (1993) Usability engineering. Cambridge, M ass. Academic Pr. Anzures-García M , Sánchez-Gálvez LA, Hornos M J, Paderewski P (2018) A workflow ontology to support knowledge management in a group’s organizational structure. An International Journal on Computing Science and Applications, vol. 22-1, pp. 163-178. Anzures-García M , Sánchez-Gálvez LA, Hornos M J, Paderewski-Rodríguez P (2018) Tutorial Function Groupware Based on a Workflow Ontology and a Directed Acyclic Graph. IEEE Latin American Transactions, vol. 16-1, pp. 294-300.

94 8.

10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21.

22.

23. 24. 25.

26.

Fernandez A, Insfran E, Abrahão S, (2011) Usability evaluation methods for the web: A systematic mapping study , Information and Software Technology, vol. 53, pp. 789– 817. Sánchez-Gálvez LA, Fernández-Luna JM (2015) A Usability Evaluation M ethodology of Digital Library. In: The Seventh International Conference on Information, Process, and Knowledge M anagement (eKNOW 2015), Lisbon, Portugal, pp. 22–27. Baker K, Greenberg S, Gutwin C (2002) Empirical development of a heuristic evaluation methodology for shared workspace groupware, In: CSCW ’02, pp. 96–105. Nielsen J, Robert LM (1994) Usability inspection methods. New York, Wiley. Nielsen J, M olich R, (1990) Heuristic evaluation user interfaces, ACM CHI, pp. 249256. Nielsen J (1992) Finding usability problems through heuristic evaluation. In: ACM CHI'92, pp. 372-380. Pinelle D, Gutwin C (2002) Groupware walkthrough: adding context to groupware usability evaluation. In: CHI ’02, ACM Press, pp. 455–462. Polson P, Lewis C, Rieman J, Wharton C. (1992) Cognitive walkthroughs: A method for theory-based evaluation of user interfaces. International Journal of M an-M achine Studies, vol. 36, pp. 741-73. Antunes P, Ferreira A, Pino J (2006) Analyzing shared workspaces design with human performance models. LNCS, vol. 4154, Springer, Heidelberg, pp. 62–77. Card SK, M oran TP, Newell A (1980) The keystroke-level model for user performance time with interactive systems. Communications of the ACM , vol. 23-7. Hughes J, King V, Rodden T, Andersen H (1994) M oving out from the control room: Ethnography in system design. In: CSCW ’94, ACM Press, pp. 429–439. Haynes S, Purao S, Skattebo A (2004) Situating evaluation in scenarios of use. In: CSCW ’04, ACM Press, pp. 92–101. Carroll JM , (1995) Scenario-based design: envisioning work and technology in system development, New York: John Wiley . Vizcaíno A, M artinez M , Aranda G, Piattini M (2005) Evaluating collaborative applications from a knowledge management approach, In: WETICE ’05, IEEE, pp. 221– 225. Van der Veer G, van Welie M (2000) Task based groupware design: Putting theory into practice. In: DIS’00: Proceedings of the conference on Designing interactive systems, New York City, New York, United States, pp. 326–337. Baker K, Greenberg S, Gutwin C (2002) Empirical development of a heuristic evaluation methodology for shared workspace groupware. In: CSCW ’02, pp. 96–105. Beyer H, Holtzblatt K (1998) Contextual design: Defining customer-centered systems. M organ Kaufmann Publishers, San Francisco, CA, USA. Pinelle D, Gutwin C, Greenberg S. (2003) Task analysis for groupware usability evaluation: M odeling shared-workspace tasks with the mechanics of collaboration. ACM Transactions on Computer-Human Interaction, vol. 10-4, pp. 281–311. Gutwin C, Greenberg S (2000) The mechanics of collaboration: Developing low cost usability evaluation methods for shared workspaces. In: WETICE ’00, IEEE, pp. 98– 103.

95 27. 28. 29. 30.

31.

32. 33.

34. 35.

Antunes P, Costa C (2003) Perceived value: A low-cost approach to evaluate meetingware. LNCS, vol. 2806, Springer, Heidelberg, pp. 109–125,. Huis in’t Veld M , Andriessen J, Verburg R (2003) E-magine: The development of an evaluation method to assess groupware applications. In: WETICE ’03, pp. 153-164. Parasuraman A, Zeithaml VA, Berry LL (1988) SERVQUAL: A multiple-item scale for measuring customer perceptions of service quality. J. of Retail, vol. 64, pp. 12-40. Herrera F, Herrera-Viedma E, Verdegay JL (1996) Direct approach processes in group decision making using linguistic OWA operators, Fuzzy Sets and Syst., vol. 79, pp. 175–190. Heradio R, Cabrerizo FJ, Fernández-Amorós D, Herrera M , Herrera-Viedma E (2013) A. fuzzy linguistic model to evaluate the quality of library 2.0 functionalities, Int. Journal of Information M anagement, vol. 33, pp. 642– 654. Zadeh LA (1975) The concept of a linguistic variable and its applications to approximate reasoning. Part III”. Information Sciences, vol.9-1, pp. 43–80. Shackel B (1991) Usability - context, fr0061mework, definition, design and evaluation. In B. Shackel & S. Richardson (Eds.), Human factors for informatics usability, pp. 21-37. Booth P (1989) An introduction to human-computer interaction. Hillsdale, USA: Lawrence Erlbaur Associates. Bojórquez A, López L, Hernández M E, Jiménez E (2013) Utilización del alfa de Cronbach para validar la confiabilidad de un instrumento de medición de satisfacción del estudiante en el uso del software M initab. Innovation in Engineering, Technology and Education for Competitiveness and Prosperity, pp. 14-16.

SECTION 2 Computational Linguistics

Computational linguistics is the study of language and the development of linguistic applications with computational media, as well as the functioning of natural language. Computational linguists focus on developing computer products for the automatic analysis of phonetics, phonology, morphology, syntax and semantics (Sierra, Cuétara, 2015)1. Computational linguistics emerges as an effort to obtain computers capable of automatically translating texts from foreign languages, and applications in the intelligence and armed services. Language engineering is an area of artificial intelligence and applications that aim to close the gap between traditional computational linguistic research and the implementation of potentially real-world applications. This section shows works related to: Language and Knowledge Engineering and its applications, user-defined language for interaction based on full body gestures, a literature study of the three phases of the development process of a system of Automatic Recognition of Sign Language (ASLR) and an analysis of research based on technology to communicate nonverbally through haptic sense. In the first work a full body language is presented to allow gesture -based interaction in different contexts and the development of applications based on fullbody gestures for general purposes. The second work shows a literature study of the three phases of the development process of an Automatic Sign Language Recognition (ASLR) system. In the following work, he analyzes research based on technology to communicate nonverbally through the haptic sense. The analysis shows that using the haptic sense as a means of communication still has many open research areas and possible new applications, and to date it has proven to be an effective mechanism to communicate most of what humans would want: messages, emotions and behavior.

Sierra,G., Cuétara,J.(2015) Lingüística Computacional en M éxico, Investigación y Desarrollo, Instituto de Ingeniería UNAM .

A User-Defined Language for Full-Body Gesture-Based Interaction David Céspedes Hernández, Juan Manuel González Calleros Facultad de Ciencias de la Comp utación, Benemérita Universidad Autónoma de Puebla, M éxico {dcespedesh,jumagoca78}@gmail.com

Abstract. Gesture recognition involves two main stages: gesture identification and association of gestures and meanings. About the semantic meaning of gestures, it is important to consider that due to their cultural and linguistic specificity each of them may be mapped to many concepts and vice versa, making them ambiguous and incompletely defined. From the HCI perspective, there is work in the literature presenting elicitation studies on which researchers find gestures sets to allow user interaction with tailored applications under determined contexts. In this paper, we present a full-body language for enabling gesture-based interaction in different contexts. For this purpose, 70 users were asked to provide the gestures they would use as commands within different existing applications, keeping awareness of the relationship between the tasks they were asked to do on the applications and abstract tasks, resulting on 980 gestures which were compared for generating a reduced set of 68 gestures consisting on their graphic representation, a textual description, an anthropometric characterizations for each of them, and generic labels. Interpretation and insights obtained during the experiment are also reported. Keywords: Gesture-based interaction, Natural user interfaces, user-defined gesture set.

Introduction

Gestures are (static or dynamic) expressive, meaningful body motions involving physical movements of the fingers, hands, arms, legs, head or face. Through gestures, it is possible to communicate in a nonverbal way and enrich verbal communication as well [1]. In the Computer Science field, Gesture Recognition (GR) is currently being applied to several domains such as children-computer interaction; forensic identification and detection; rehabilitation and medical monitoring; navigation and manipulation in virtual environments; distance learning assistance; security; athlete training; and entertainment applications [2-5]. Besides, hardware and software are constantly being developed supporting such interaction modality. GR is a challenging problem for which proposed solutions typically involve at least two stages: gesture identification, and gesture-meaning matching. Gesture identification is commonly achieved through wearable devices, by using vision -based tech-

niques, or through the processing of wireless signals. Interaction through wearables typically requires users to carry uncomfortable devices [6], while vision -based techniques, in turn, content with other problems for the most part related to occlusion [7], and processing of wireless signals has not been thoroughly explored and regularly involves the use of devices like radars and routers [8]. Each data gathering device varies along several dimensions, including accuracy, resolution, latency, the range of motion, user comfort, and cost [1]. Gestures are often specific to a determined language and culture, hence, giving them standard meaning is not possible. Also, since the semantic meaning of the gestures that a person does depends on his/her mental model, there are many -to-one mappings from concepts to gestures and vice versa. Thus, gestures are ambiguous and incompletely specified [9]. Taking this into consideration, the objective of this paper is on reporting the realization of elicitation experiments for proposing a fu ll- body language for enabling gesture-based interaction in different contexts through the use of reified versions of abstract tasks. The rest of the paper is structured into sections. The second section is dedicated to describing the state of the art, including the definition of capability models that were used for identification of the body parts to be considered as well as their movements, and the description of related works on the proposition of user-defined gestures sets. In section three, the followed method and the performed experiments are described, so in the fourth section, the results and its discussion are addressed. Finally, the fifth section is aimed at presenting the conclusions of the project and the future work to be done.

State of the Art

When designing interactive applications, it is important to consider the context in which the to be implemented system is going to be used. This is, where users are going to perform the tasks of interest. The analysis of the context of use has been addressed in several works, most of them from an HCI perspective, agreeing on that it can be separated into three dimensions: user, platform, and environment [10]. For the purposes of this paper, and aligned to the defined problem, this section is focused on describing a method for modeling users and for identifying their capabilities, as well as on presenting related works about the creation of user- defined gestures sets. Originally proposed as part of a framework for modeling virtual users for automatic simulated accessibility and ergonomic testing of applications, in [11] a Generic Virtual User Model (GVUM) and a full-body capability model are described. Considering the GVUM representation, depicted in Fig. 1, it is possible to notice that the human body consis ts of six main components: neck, spinal column, left upper limb, right upper limb, left lower limb, and right lower limb. In a more detailed level, upper limbs include shoulders, elbows, forearms, wrists, and hands containing fingers, while lower limbs include in turn hip, thighs, knees, ankles, and foot toes.

Fig. 1. Generic Virtual User M odel from [11].

Using the GVUM as previously described, it is possible to anatomically model users, but not to represent their interaction with user interfaces. This objective was achieved by defining a capability model, consisting of 108 classes corresponding to systems, body parts, features, and motions. A summary of the body parts that are listed as part of the GVUM related to their pos sible motions according to the capability model is presented in Table 1. Being able to identify all of the possible movements that users can perform in order to interact with user interfaces through full-body gestures, allows to drive experiments focused on the definition of interaction gestures sets and to characterize such gestures. In this order of ideas, in the literature, it is possible to find works from authors addressing the definition of gesture sets for interacting through different means. Some of those works were analyzed prior design and planning of the present project and are described in the following lines. In the context of surface computing, in [12] Wobbrock et al. stated that gestures that are defined by system designers are appropriate only for early prototype testing, while for a production environment it is desirable to have gestures that are reflective of the users’ mental model. That work is only focused on the analysis of gestures of 1 or 2 hands for performing 27 commands, due to the restrictions of the platform of interest. The results were shown as graphic representations along with the level of agreement that was reached for every single command. In that work, it is asserted that the participants preferred 1-hand gestures for 25 of the 27 defined commands, participants employed reversible gestures for dichotomous actions, multiple participants provided gestures using areas out of the screen edges and above it, leading to think

100

that the surface device must have the ability to track gestures using other mechanisms rather than only the touchscreen. Table 1. Body parts from the GVUM related to possible movements from the capability model [11]. Body part

Associated possible motion from the capability model

Neck

Flexion, extension, right lateral flexion, right lateral rotation, left lateral flexion and left

Spinal

Flexion, extension, right lateral flexion, right lateral rotation, left lateral flexion and left

column

lateral rotation.

Shoulders

External rotation, extension, flexion, internal rotation, abduction, and adduction.

Elbows

Extension, Hyperextension, and flexion.

lateral rotation.

Forearms

Supination, pronation

Wrists

Extension, flexion, radial deviation, and ulnar deviation.

Hand

Supination and pronation

Finger

Adduction, extension, abduction, flexion, and hyperextension.

Hip

Adduction, abduction, internal rotation, flexion, external rotation, and extension.

T highs

Flexion and extension.

Knees

Flexion and extension.

Ankles

Dorsiflexion, plantar flexion, eversion, and inversion.

Foot toes

Extension and flexion.

With regards to applying GR in augmented reality applications, in [13] Piumsomboon et al. addressed the definition of a user-defined gesture set for allowing designers to achieve natural interaction in such type of systems. In their experiments, 20 participants were asked to provide gestures for 40 tasks, resulting in 800 gestures that were compared for obtaining levels of agreement. The results of the experiment were reported as graphical representations of the consensus set of gestures along with textual explanations of them and labels for their identification. Also, in this proposal, it is affirmed that the definition of the gesture set was affected by causes such as the mental model of users and their experience on the use of determined platforms. Even though gesture detection was achieved using a depth sensor making it possible to examine full-body gestures, the proposal was only focused on hands and fingers gestures. In the literature, it is also possible to find works addressing the definition of fullbody gesture vocabularies and commonly applied to navigation and control domains. Such is the case of [14] in which users were asked to perform full-body gestures for controlling a humanoid robot. In that study, Obaid et al. proposed navigational commands and analyzed gestures provided by users, reporting agreement scores, time performances, and graphical representations of the consensus set. Along with the obtention of the gestures set, significant insights were discussed like the need of the definition of points of view when trying to recognize gestures for navigation, i.e. if users’ motion is being tracked using cameras in front of them, it is possible to misunderstand the direction of the gestures.

101

From this state of the art and by executing a more exhaustive revision of the literature, it is possible to see that related works follow methodologies consisting on: the definition of the body parts to be analyzed; the selection of participants and description of their characteristics; the implementation of GR software; the obtention of data from users; and the explanation of the gathered results in terms of agreement, time performance, graphical representations of the gestures, gestures labeling, and additional observations. Next section is dedicated to the description of the method that was followed in this specific project and the description of the performed experiment.

Methodology

In accordance with the expressed objective and considering the capability model which was summarized in Table 1, this work is focused on the recognition of fullbody gestures, specifically on movements of the neck tracked through the position of the head, upper limbs observed as arms and hands, and lower limbs to be analyzed considering legs and feet. During preliminary studies, it was noticed that all of the gestures performed by specific users were significantly different from those performed by others for the same tasks. After thoroughly reviewing possible causes for this issue and exploring alternatives, we identified research addressing the influence of creativity on the use and development of technology [15]. Then, we searched on creativity tests and found the Aulive test which is available online and allows to measure users’ creativity delivering an average result and considering eight dimensions: abstraction, connection, perspective, curiosity, boldness, paradox, complexity, and persistence. We then asked users to get their creativity tested and found some of the users getting creativity levels high above or under the average, were the same as those performing gestures that were not consistent with the others. As a total, 70 subjects novice on the use of gesture-based interaction applications were selected for participation and separated into 5 groups with 14 users each. The first group, dedicated to neck gestures, consisted of 9 females and 5 males between 19 and 79 years old. The second group which was created for arm gestures observation consisted of 7 female and 8 males aged between 17 and 67 years. The third group, for hand gestures, was composed of an equal number of female and male users between 20 and 63 years. The fourth group, for legs gestures, involved 7 females and 7 males aged between 13 and 53 years. Finally, the fifth group in which feet gestures were observed, was composed of 7 females and 7 males between 12 and 52 years. Users were selected and equally distributed in such way according to the punctuation they received as average on the Aulive creativity test. Users who received scores wide above, or wide below the average were discarded. Once the participants were selected and the groups were defined, it was time to drive experiments for finding interaction vocabularies for each of the abo vementioned body parts. This kind of elicitation studies is commonly carried out by implementing applications. Nonetheless, the implementation of systems based on gesture interaction represents a challenge itself, and as our objective is on identifying gestures for general purposes, and not on assessing the interaction of a specific appli-

102

cation nor defining commands, the development is out of the scope of this specific work, and it was decided to consider a different approach. The Wizard of Oz technique has been used throughout the history of the development of interactive systems [16], and in particular, in the field of natural interfaces development as it is a way to collect data for mixed reality environments or movement commands for interaction with kids [17] among other application domains. Following this technique, the designed experiment consisted on asking users to use their own body gestures to interact with different applications using only specified body parts and giving them the feeling that thos e commands actually worked on the application, but actually providing the input to the programs through keyboard commands.

Fig. 2. Experiment setup following the Wizard of Oz technique [16] and its disposition in a Gesell chamber.

The setup of the experiment, as shown in Fig. 2, took place in a Gesell room and included two computers (PC1 and PC2), the first one for retrieving users’ movement information from a webcam and allowing to document the experiment; while the second one was connected to a projector for providing the user with feedback and simulating users’ interaction with the system. An observer of the experiment was designated and located next to the user in order to provide support in case any doubt arises. The experiment controller managed the devices to document the experiment and to explain how each user performed the activity. During the experiment, users were asked to provide the gestures they would use to perform 14 commands within different applications and contexts. The 70 tasks that were defined are presented in Table 2. When users carried out gestures for each of the requested commands, data was gathered including response time, a video record of the corresponding gesture, and a score given by the same u ser with respect to how natural they thought the gesture was for the task. As a total, 980 records were acquired and then processed through manual labeling, manual comparison, to finally be assigned with an agreement score. The results of the performed exp eriment and the proposed gestures set are discussed in the next section.

103 Table 2. Tasks to be observed during the experiment. Body part (application/context)

Tasks

Head

Find profiles (N1), Add image (N2), Pick image (N3), T ake photo (N4), Record

(Instagram)

video (N5), View notifications (N6), View profile (N7), Like post (N8), Post story (N9), Add emoji (N10), Add hashtag (N11) Accept follow request (N12), T ag friend (N13), Edit profile (N14)

Arms

Open gate (A1), Close gate (A2), T urn lights on (A3), T urn lights off (A4),

(Smart home)

T urn air conditioner on (A5), T urn air conditioner off (A6), T urn T V on (A7) T urn T V off (A8), T urn T V volume up (A9), T urn T V volume down (A10), Change T V channel (next) (A11), Change T V channel (previous) (A12), T urn water pump on (A13), T urn water pump off (A14)

Hands

View posts (H1), View notifications (H2), View messages (H3), Create a group

(Facebook)

(H4), Like post (H5), Love post (H6), Write a message (H7) Add friend (H8), Upload photo (H9), Tag friend (H10), Block user (H11), Create an event (H12), Start video call (H13), Start call (H14)

Legs

Move down in the main menu (L1), Move up in the main menu (L2), Select option

(Crash Bandicoot

in the main menu (L3), Move forward in 3D (L4), Move left in 3D (L5), Hit a box

game)

in 3D (L6), Jump an obstacle in 3D (L7), Smash a box in 3D (L8), Move forward in 2 ½D (L9), Jump an obstacle in 2 ½D (L10), Hit a box in 2 ½D (L11), Smash a box in 2 ½D (L12), Duck in 3D (L13), Exit minigame in 2 ½D (L14)

Feet

Answer phone call (F1), Hang up a phone call (F2), Go to the home screen (F3),

(Android phone

Start voice search (F4)

control, media play-

Play/Pause (F5), Quit player (F6), Go to previous track (F7), Shuffle play (F8),

er, Web browser)

T urn volume up (F9), T urn volume down (F10) Move cursor to the left (F11), Move cursor to the right (F12), Zoom in (F13), Zoom out (F14)

Results and Discussion

Those gestures with the higher agreement rates for each task were characterized according to the capability model which was presented in the state of the art, described in natural language, depicted in a graphical representation, generically relabeled for their quick identification, and taken as part of a new gestures set which is described in Table 3 and depicted in Fig. 3. In case of having ties on the times in which two or more gestures were performed for a task, the situation would be solved using the grade users gave them and the response time in which they were done.

104 Table 3. Labels, agreement scores, descriptions, and characterization of the user-defined gesture set. Label

Textual description

(Agreement

Characterization according to capability model

score) G001 (5/14)

T ilt the head to the

Right lateral flexion of the neck.

right. G002 (3/14)

T urn the head to the

Right lateral rotation of the neck.

right. G003 (3/14)

T ilt the head to the left.

Left lateral flexion of the neck.

G004 (5/14)

Lift both shoulders.

Flexion of both shoulders.

G005 (2/14)

Lift both shoulders

Flexion of both shoulders twice.

G006 (3/14)

T ilt the head to the

twice. Right lateral flexion of the neck twice.

right twice. G007 (3/14)

Look down.

Flexion of the neck.

G008 (4/14)

Nod twice.

Flexion and extension of the neck twice.

G009 (4/14)

Look up.

Extension of the neck.

G010 (3/14)

T urn the head to the

Left lateral rotation of the neck.

left. G011 (2/14)

Lift the left shoulder.

Flexion of the left shoulder.

G012 (2/14)

Move both shoulders

Adduction of both shoulders.

G013 (2/14)

Lift the left shoulder

forward. Flexion of left shoulder twice.

twice. G014 (2/14)

Lift the left shoulder and tilt the head to the left.

G015 (6/14)

Move the left arm to the left and the right arm to

Flexion of the left shoulder and left lateral flexion of the neck. Extension of both shoulders, extension of both elbows.

the right. G016 (6/14)

Starting with both arms extended, bring them to the

Starting with the extension of both shoulders and elbows, flex both shoulders.

body. G017 (6/14)

Lift the right arm.

Abduction of the right shoulder, extension

G018 (4/14)

Lower the right arm.

Adduction of the right shoulder, extension

G019 (1/14)

Move the right wrist to

Flexion of the right elbow, and internal ro-

G020 (1/14)

Starting with the arm

of the right elbow. of the right elbow. the right. extended, move the right wrist forward.

tation of the right shoulder. Starting with the right elbow extended, flex the right elbow.

105 G021 (4/14)

Move the left wrist to the left.

G022 (4/14)

Starting with the right arm flexed, move right

Flexion of the left elbow, and internal rotation of the left shoulder. Starting with the right elbow flexed, rotate the right shoulder.

wrist forward. G023 (6/14)

Lift the left arm.

Abduction of the left shoulder, extension of

G024 (6/14)

Lower the left arm.

Adduction of the right shoulder, extension

G025 (4/14)

Move the left wrist to

Starting with the left elbow flexed, rotate

G026 (4/14)

Move the left wrist to

G027 (4/14)

Starting with the arm

the left elbow. of the right elbow. the left. the right. flexed and the wrist point-

the left shoulder. Starting with the right elbow flexed, rotate the right shoulder. Starting with the elbow flexed, adduct the shoulder.

ing up, lower the elbow. G028 (4/14)

Starting with the arm flexed and the wrist point-

Starting with the elbow flexed, flex the shoulder.

ing up, move the wrist to point down. G029 (5/14)

Using only the forefin-

Flexion of all fingers but the second one.

ger, point forward. G030 (2/14)

T urn the right wrist from right to left with the

Flexion of all fingers but the second and third ones along with the wrist flexion.

forefinger and the middle finger extended. G031 (1/14)

Pointing up with the forefinger only, flex the

Starting with all fingers flexed but the second one, flex the second finger twice.

fingertip twice. G032 (2/14)

Using both hands,

Extension of the first and the second fingers

reach with the forefinger

of both hands, reaching the second finger of the

each other thumb creating

left hand with the first finger of the right one,

a ‘frame’ form

and the second finger of the right hand with the first finger of the left one.

G033 (13/14)

Use only one thumb to

Flex all fingers and extend the first one.

point up. G034 (8/14)

Join each finger of one

Flex all fingers and reach each one of the

hand to those on the other

fingers of one hand with the same finger of the

creating a heart-like shape. G035 (3/14)

Use both hands for

G036 (4/14)

Using the forefingers of

‘typing’ in the air.

other. Starting with all fingers extended, quickly flex and extend all of them. Extending only the second finger of each

106 both hands, make a ‘plus

hand, use one of them to point up and put the

sign’.

other one perpendicularly in order to make a ‘plus sign’.

G037 (2/14)

With thumb and fore-

Starting with only the first and the second

finger extended, move the

fingers extended, lightly flex the second finger

forefinger down to reach

towards the first one without reaching it.

the thumb without touching it. G038 (2/14)

With the palm forward,

Starting with all fingers extended, hyperflex

touch the thumb with t he

the first and the fifth fingers so they touch each

little finger. G039 (7/14)

Using the forefingers of

other. Extending only the second finger of each

both hands, make an ‘x

hand, use one of them to point up and put the

sign’.

other one perpendicularly in order to make an ‘x sign’.

G040 (1/14)

G041 (2/14)

Using only the forefin-

Extending only the second finger, draw a

ger, make a triangle in the

triangle in the air either rotating the elbow or

air.

the shoulder.

Using only the forefinger, make a circle in the air.

G042 (9/14)

Flex all fingers but the little finger and the thumb.

G043 (3/14)

Slide the left foot

G044 (6/14)

Slide the right foot

Extending only the second finger, draw a circle in the air either rotating the elbow or the shoulder. Flex all fingers and extend only the first and the fifth one. Extend the hip above of the left leg.

backward. Flex the hip above of the right leg.

forward. G045 (5/14)

Stomp.

Flex a knee, then, quickly extend it.

G046 (4/14)

Step forward.

Flex the hip, extend a knee, extend the hip.

G047 (3/14)

Slide the left foot to the

Abduction of the hip above the left leg.

left. G048 (5/14)

Kick.

Flex the hip, flex a knee, then extend it.

G049 (5/14)

Jump using both legs.

Flex the hips, flex the knees, flex the ankles, then extend the ankles and the knees pushing the body up.

G050 (3/14)

Lift a leg and then

G051 (3/14)

Walk forward.

Repeat with each lower limb: Flex the hip,

G052 (7/14)

Rotate the left ankle.

Plantar flexion, adduction, dorsiflexion, ab-

G053 (5/14)

Jump using both legs

Flex the hips, flex the knees, flex the ankles,

stomp.

Flex the hip, flex a knee, then, quickly extend the hip and the knee. extend the knee, extend the hip. duction of an ankle.

107 twice.

then extend the ankles and the knees pushing the body up twice.

G054 (5/14)

Duck.

Flexion of the hips, flexion of the knees,

G055 (3/14)

Slide the left foot to the

Adduction of the hip above the left leg.

plantar flexion of the ankles. left. G056 (4/14)

Slide the left foot to the

Adduction of the hip above the right leg.

right. G057 (6/14)

Get on tiptoes.

Plantar flexion.

G058 (4/14)

Get on tiptoes and

Plantar flexion, extend the hip.

G059 (3/14)

Get on tiptoes / raise

move forward. foot.

Plantar flexion, flex the knee / extend the knee.

G060 (6/14)

Rotate the left foot up.

Eversion of the left foot.

G061 (4/14)

Make an ‘x sign’ with a

Plantar flexion, draw an ‘x sign’ using the

foot.

toe tip.

G062 (4/14)

Rotate both feet in.

Eversion of both feet.

G063 (5/14)

Separate both feet.

Adduction of both ankles.

G064 (7/14)

Rotate the left foot out.

Inversion of the left foot.

G065 (5/14)

Move the left foot for-

Extend the left knee.

ward. G066 (8/14)

Rotate the left foot to

G067 (7/14)

Rotate the left foot to

G068 (5/14)

Split the toes.

the left. the right.

Eversion, dorsiflexion, inversion, plantar flexion. Inversion, dorsiflexion, eversion, plantar flexion. Abduction of all toes.

It was interesting to see that even though users were asked to use only move ments from a determined part of the body and that it would be correct to use same gestures for different commands, they mostly felt more comfortable using different gestures for each command, and particularly in the case of head and feet motion groups, the y commented that the use of the single part of the body was not enough to express what they thought, extending motions to shoulders and ankles respectively. Therefore, the final gestures set contains such type of movements as well. If well the gesture set was created using specific tasks on existing applications, it is possible to see that they may be applicable to other similar contexts as most of them describe abstract activities such as turn on, turn off, open, close, move up, and move down. An interesting insight is on that agreement scores for the consensus gestures from the head (avg. 0.22, stddev. 0.08), legs (avg. 0.32, stddev. 0.09) and feet (avg. 0.36, stddev. 0.11) were not significantly lower when compared to gestures from hands (avg. 0.31, stddev. 0.26) and arms (avg. 0.31, stddev. 0.12) as we expected considering that HCI is commonly being achieved using upper limbs.

108

Fig. 3. Graphic representation of the obtained user-defined gesture set including the identifiers of tasks and their associated gestures.

From the observation during the experiment, this was caused by two reasons, the first one regarding the mental model that users have and the second one concerning that if well it is not common to interact with applications using only head, legs, and feet movements, the degrees of freedom they allow reduce the number of gestures that users feel comfortable doing reducing the options they have and raising the probabilities of having similar movements. Moreover, the standard deviation calculated over the agreement scores showed the highest value for the hands’ gestures. In regard to this, we noticed that gestures which were constantly repeated by users for a specific

109

task, such as G033 for H5, are already being in use in real human-to-human interaction with similar meaning or may have similar representation on widely used platforms and adopted as de-facto standards. This is coherent with the observations in [18] regarding technology and culture influence on gestural in teraction. On the other hand, other tasks that do not have a standard gestural representation such as H2, H3, and H4, got lower agreement scores. We observed that the main cause for this, is on that hands have more degrees of freedom and therefore it is po ssible to have more gestures from them. Expressiveness allowed by upper limbs significantly affected agreement, consider that there were only 4 tasks for which users provided completely different gestures and that such tasks belong to the arms (A5 and A6) and hands groups (H3 and H12). Furthermore, it is also possible to see that gestures G048 and G049 got the highest agreement score (5/14) for two tasks each, hence resulting on the reduction of the vocabulary of gestures from 70 expected gestures to 68. As previously mentioned, response times were also recorded, and users were asked to evaluate the gestures they provided. It was interesting to notice that younger users responded, on average, faster than the older ones did, and that the faster users provided gestures for a specific command, the better they evaluated the movement as it is possible to see in Fig. 4 for head gestures. During the experiment, it was perceived that both of these metrics were affected significantly by the confidence of users towards the interaction modality.

Fig. 4. The relationship between response time and users' age, and between response time and gestures grade for head gestures.

We expected to see relationships existing between the agreement sco re for a gesture and its associated average response time, as well as between the agreement scores and users’ grades for each gesture, however, there does not seem to be any. Final considerations and future work are described in the next section.

Conclusion and Future Work

In this paper, we propose a user-defined language for full-body gesture-based interaction. For this purpose, 70 participants took part in an experiment based on the Wizard of Oz technique, in which they were asked to provide gestures for realizing 14 tasks within specific applications which are aligned to canonical tasks. From the experi-

110

ment, 980 records were obtained, and processed for manual labeling, comparison, and calculation of agreement rate, in order to finally elicit a 68-gesture set containing a graphical representation of each gesture, labels for their identification, textual descriptions, and characterizations of them based on features presented in a capability model. It was noticed that users’ characteristics such as experience, confidence, age, and mental model, as well as additional facets like comfort, expressiveness and the degree of freedom of body parts, affect the gestures, response time, and evaluation of gestures. The impact of additional factors such as creativity and the influence of the use of technology were superficially considered in this work, but it will be necessary to drive additional experiments in order to precisely determine the dimensions of such aspects that affect gestural elicitation. As the tasks from the experiment correspond to abstract tasks, our gesture vocabulary will allow the implementation of full-body gesture-based applications for general purposes, keeping in mind that additional gestures may be necessary for extension and tailoring. As future work, it will be necessary to develop a tool for GR and automatic labeling, as well as means to allow usability and user experience testing on gesture based applications working with user-defined commands.

References 1. 2.

4. 5. 6. 7.

M itra, S.; Acharya, T. Gesture recognition: A survey. IEEE Trans Syst M an Cybern C Appl Rev 2007, 37(3), 311-324. Schlömer, T.; Poppinga, B.; Henze, N.; Boll, S. Gesture recognition with a Wii controller. In Proceedings of the 2nd international conference on Tangible and embedded interaction, Bonn, Germany, 18-20 February; ACM : New York, USA, 2008, pp. 1114. Liang, R. H.; Ouhyoung, M . A real-time continuous gesture recognition system for sign language. Proceedings of the Third IEEE International Conference on Automatic Face and Gesture Recognition, Nara, Japan, 14-16 April; IEEE, 1998, pp. 558-567. Daniloff, J. K.; Noll, J. D.; Fristoe, M .; Lloyd, L. L. Gesture recognition in patients with aphasia. J Speech Hear Disord 1982, 47(1), 43-49. Jia, P.; Hu, H. H.; Lu, T.; Yuan, K. Head gesture recognition for hands-free control of an intelligent wheelchair. Ind Rob 2007, 34(1), 60-68. Norman, D. A. (2010). Natural user interfaces are not natural. interactions, 17(3), 610. M alima, A. K.; Özgür, E.; Çetin, M . A fast algorithm for vision-based hand gesture recognition for robot control. In 14th Signal Processing and Communications Applications, Antalya, Turkey, 17-19 April; IEEE, 2006. Wan, Qian, et al. "Gesture recognition for smart home applications using portable radar sensors." Engineering in M edicine and Biology Society (EM BC), 2014 36th Annual International Conference of the IEEE. IEEE, 2014. Chandler, J.; Schwarz, N. How extending your middle finger affects your perception of others: Learned movements influence concept accessibility. 2009, 45(1), 123-128.

111 10.

11.

12.

13.

14.

15. 16.

17.

18.

Calvary, G.; Coutaz, J.; Thevenin, D.; Limbourg, Q.; Bouillon, L.; Vanderdonckt, J. A unifying reference framework for multi-target user interfaces. Interacting with computers 2003, 15(3), 289-308. Kaklanis, N.; M oschonas, P.; M oustakas, K.; Tzovaras, D. Virtual user models for the elderly and disabled for automatic simulated accessibility and ergonomy evaluation of designs. Univ Access Inf Soc 2013, 12(4), 403-425. Wobbrock, J. O.; M orris, M . R.; Wilson, A. D. User-defined gestures for surface computing. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Boston, USA, 04-09 April; ACM : New York, USA, 2009, pp. 10831092. Piumsomboon, T.; Clark, A.; Billinghurst, M .; Cockburn, A. User-defined gestures for augmented reality. In IFIP Conference on Human-Computer Interaction, Cape Town, South Africa, 2-6 September; Springer: Berlin, Heidelberg, 2013, pp. 282-299. Obaid, M .; Häring, M .; Kistler, F.; Bühling, R.; André, E. User-defined body gestures for navigational control of a humanoid robot. In International Conference on Social Robotics, Chengdu, China, 29-31 October; Springer: Berlin, Heidelberg, 2012, pp. 367-377. M otaghi, Hamed. Creativity and technology in the context of creative industries. Diss. Université du Québec à M ontréal, 2015. Dow, S.; Lee, J.; Oezbek, C.; M acIntyre, B.; Bolter, J. D.; Gandy, M . (2005, April). Wizard of Oz interfaces for mixed reality applications. In CHI'05 Extended Abstracts on Human Factors in Computing Systems, Portland, USA, 2-7 April; ACM : New York, USA, 2005, pp. 1339-1342. Höysniemi, J.; Hämäläinen, P.; Turkki, L. (2004). Wizard of Oz prototyping of computer vision-based action games for children. In Proceedings of the 2004 conference on Interaction design and children: building a community, M aryland, USA, 01-03 June; ACM : New York, USA, 2004, pp. 27-34. Hoff, Lynn, Eva Hornecker, and Sven Bertel. "M odifying Gesture Elicitat ion: Do Kinaesthetic Priming and Increased Production Reduce Legacy Bias?." Proceedings of the TEI'16: Tenth International Conference on Tangible, Embedded, and Embodied Interaction. ACM , 2016.

Discussion on Image Processing for Sign Language Recognition: An overview of the problem complexity Soraia Silva Prietch 1 , Ivan Olmos Pineda 2 , Polianna dos Santos Paim1 , Juan Manuel Gonzalez Calleros 2 , Josefina Guerrero García 2 and Roger Resmini1 1

Universidade Federal de M ato Grosso (UFM T), Rondonópolis, M T, 78.735-901, Brazil 2 Benemérita Universidad Autónoma de Puebla (BUAP), Puebla-PUE, 72592, M éxico soraia@ufmt.br, ivanoprkl@gmail.com, poliannapaim@gmail.com, juan.gonzalez@cs.buap.mx, joseguga01@gmail.com, rogerresmini@gmail.com

Abstract. The goal of this paper is to conduct a literature study of the three phases of the process of developing an Automatic Sign Language Recognition (ASLR) system, in order to discuss the problem complexity, and show result s of conducted tests using Digital Image Processing (DIP) algorithms. The methodological procedures were: brief overview of Brazilian Sign Language (Libras), scenarios description to emphasize the importance of an ASLR system, an exploratory literature review of related works on image processing applied to ASLR problem, and analysis of findings from tests with image processing algorithms using Libras signs. As results, we observed that researches have been conducted in controlled environments with few features or sign language parameters, because real-time recognition requires solutions for many complex aspects of translation of a visual-gestural language. Keywords: Digital Image Processing, Automatic Sign Language Recognition, User Experience.

Introduction

According to WHOSIS [48, p. 07] “More than a billion people are estimated to live with some form of disability, or about 15% of the world’s population”. Considering a specific type of disability, “Deaf people often have trouble accessing sign language interpretation: a survey in 93 countries found out that 31 countries had no interpreting service, while 30 countries had 20 or fewer qualified interpreters” (ibid., p. 10). These data highlight the large number of people who need Assistive Technology (AT) products or services to study, to work, or to enjoy leisure time; in these three situations, to communicate and to interact with other people are required. WHOSIS report [48] also shows that there is a lack of Sign Language (SL) interpreters in many countries, which opens perspectives to investigate alternatives of communication on technology-mediated solution. The trigger to start this work was a Digital Image Processing (DIP) class project, of a doctoral program, in which one of this paper’s authors, motivate d by ones previous work on AT for Deaf people, decided about the application domain of

113

Brazilian Sign Language (Libras) recognition. Thus, the main goal in this paper is to conduct a literature study of the three phases of the process to develop an Automat ic Sign Language Recognition (ASLR) system, to discuss the problem complexity, and to show results of conducted tests using DIP algorithms. In order to achieve the main goal, we present a brief overview of Brazilian Sign Language (Libras) in Section 2. In Section 3, we describe four possible scenarios of ASLR use. An exploratory literature review of related works on image processing applied to ASLR is presented in Section 4. In Section 5, we describe results from tests with digital image processing algorithms using Libras signs, which were carried out with the programming language Python. In Section 6, we present final considerations.

Brief Overview of Brazilian Sign Language (Libras)

Each country has its own sign language [42], in some cases more than one s ign language is used for communication in the same country. In this paper, we focused on Brazilian Sign Language (Libras) to show some examples of DIP signs. In Brazil, most of people who are part of Deaf community use Libras, which became official in the legislation in 2002 (Law n. 10.436). According to Strobel and Fernandes [42], Libras has a manual alphabet, as a loan from oral/written language, with 27 letters, which is used only for fingerspelling proper names and unknown words. This means that it is n ot possible to fully communicate in Libras (or any other sign language) just using manual alphabet. Some examples of Libras manual alphabet that have similar hand shape (e.g., “A” and “S”), or have the same hand shape with different hand orientation (e.g., “G” and “Q”), or have the same hand shape with different hand movement (e.g., “H”, “K” and “P”). Libras dictionaries and glossaries have been created with signs of general (daily life) purpose and with signs of specific domain. As examples, Capovilla et al. [5] published a trilingual dictionary with 13 thousand Libras (general purpose) signs, and Letras-Libras/UFSC (s/d) [22] has jargon specific glossaries in Libras available online. From Marcon et al. [23, p. 24], “[…] grammatical structure of sign languages have their own linguistic levels, phonological, morphological, syntactic and semantic”, also, includes regionalisms, slangs, neologisms, metaphors, classifiers, among other concepts. This means sign languages are not signalized as oral/written language, and phrase constructions are different. Signs in Libras represent words and terms, however, recognizing isolated signs do not mean someone knows how to communicate in Libras. Considering grammatical structure of sign languages [23][41][42], Libras has 05 main parameters: (a) Hand shape: hand configuration to communicate a sign, for which DLibras [10] presents 73 hand configuration types; (b) Hand location (articulation point): place in the head, body or around it, where the sign is communicated, which can be touched, but not necessarily; (c) Hand movement: movement of the hand in space; (d) Hand orientation or orientation: change during

114

communication; and, (e) Non-manual (facial and body) expression: it can define meanings of signs. Rezende, Castro and Almeida [33, p. 01] declare, “One sign is composed of various simultaneous elements”, using all parameters of Libras to communicate a sign or not. For the signs “HOME” (“CASA” in Brazilian Portuguese) and “EXCUSE ME” (“PERDÃO” in Brazilian Portuguese) no movement is necessary to understand signs, although there are start and end positions of arms and hands before and after performing signs (that may be in resting position or be a part of a phrase or conversation), including movements for communication. In the case of the sign “TIRED” (“CANSADO” in Brazilian Portuguese), all five parameters must be used to produce an understanding performance. In this sense, Nel, Ghaziasgar and Connan [25, p. 179] affirm, “The use of a single parameter has meant that only a limited number of signs can be recognized […]”. Therefore, by analyzing only one SL parameter is difficult to determine a specific sign; for instance, in Libras, many signs use the same hand configuration (e.g., “LABOR” and “TELEVISION”, “TRABALHO” and “TELEVISÃ O” in Brazilian Portuguese). Phrase construction in Libras is different from phrase construction in written or spoken Brazilian Portuguese. For Quadros, Pizzio and Rezende [32], syntactic structures of both languages are distinct, once there is variation in word order (subject-verb-object (SVO), subject-object-verb (SOV), verb-subject-object (VSO), object-subject-verb (OSV) as it can happen to any other language. However, Quadros, Pizzio and Rezende [32, p. 34] assure that “the variability observed in Brazilian Sign Language is related to grammatical mechanisms such as the presence of agreement, of topicalization, and of constructions with focus, always associated to the use of specific non-manual marking”. Moreover, according to Strobel and Fernandes [42, p. 16] “[...] articles, prepositions, conjunctions are not used because these connectives are incorporated into sign communication”; and, affirmative, interrogative, exclamatory and negative phrases are emphasized using facial and body expressions.

Scenarios of ASLR use

According to Rojano-Cáceres et al. [34, p. 01-02], in any domain, “we could identify different scenarios where the complexity of the communication varies”. In this sense, authors classified three levels of complexity in a conversation (or interaction) related to the degree of language’ use: low, medium and high levels. High level of interaction refers to the “full use of grammar and syntax in complex interactions”. Below we present four potential scenarios in which a high level of interaction could happen with the use of an ASLR, to generate written or spoken language (as output) from a sign language (as input), making a dialog possible for human -human communication or teaching-learning purposes. ● (Scenario 1) Dictionaries and glossaries , with ASLR system, would be very useful in situations where someone could look up for an unknown sign just by gesturing it to a set of natural data capture. The user can be any person who is learning sign language, or he/she does not know a sign of specific domain, or wants to

115

understand why a sign was used in a specific context of conversation, among others. ● (Scenario 2) The use of ASLR for Real-Time Communication (RTC) would allow several types of interactions, such as, online education (participating as a s tudent or as a teacher/professor), students advisoring, social interaction, research procedures (e.g., interviews, remote usability/ UX/ accessibility tests), among others. ● (Scenario 3) ASLR could be a Mobile App for use in different areas of daily life (education, work, leisure, home), whenever individuals want or need to communicate. This type of technology would allow environment -independent communication; however, it could embarrass the signer being observed through the screen of the cell phone. ● (Scenario 4) In a Context-aware environment, all necessary devices for ASLR would be available. The environment can be a classroom, a government service setting, a medical assistance room, a workplace, among others. This type of technology provides a natural communication between people; however, it is environment-dependent for communication. Also, in a multi-cultural context, as Sign Language may vary depending on the country, even depending on the region the deaf or hard of hearing person live, he/she is viewed to be part of a community or culture [16]. The first scenario (Dictionaries and glossaries) can capture two types of data: (a) manual alphabet or hand configuration to find a sign among a list of signs that starts with a specific hand pose. In this case, a static image work for recognition, because it is not necessary to identify hand movement, hand orientation, hand location or non manual expression; (b) sign search to match with a specific or similar signs stored in database of the dictionary/ glossary. In this case, processing and analysis of SL parameters in a dynamic image are necessary. In Dlibras [10], besides alphabetical word order in Brazilian Portuguese, it is also possible to search for a sign by looking up a list of static images of Libras hand configurations. The other three scenarios allow communication between, at least, two people, in which one of them uses sign language. Usually, when two people who use different languages want to communicate they need an interpreter to intermediate. Unfortunately, not always an interpreter is available at the environment in which one is needed. This can be a great barrier for both sides. Nonetheless, ASLR for these cases would have to support real-time conversations, which accumulates more complexity to this application domain. Discussions about this can be found in next sections. Furthermore, in the case scenarios presented, an ASLR system would recognize and translate a sign language to a written or spoken language, however, it would not reverse the translation. To translate a written or spoken language into a sign language is necessary to generate a 3D avatar to communicate in real-time what is being written or spoken [19][44]. In Brazil, there are four well-known translators of this type: VLibras [45], Rybená [35], HandTalk [14] and ProDeaf [31]. However, this is not the scope of this paper.

116

Another possible scenario is the one presented by Aran et al. [1], the Sign-Tutor system. The system was presented as a sign language teaching -learning environment, offering three modules: training, practice, and feedback. During the practice module, learners can record their sign language performance, which will have three parameters analyzed: hand motion, hand shape, and head motion, through ASLR, in order to verify if the learner’s performance is correct. Furthermore, another possible scenario is the transcription of sign language “videos into a written representation for annotation [HamNoSys] and documentation purposes”, which was the purpose of Borg and Camilleri [3, p. 163]’s work. Although these possible scenarios could be dreams becoming true for many, to allow high level of interaction in a ASLR communication scenario requires dynamic image processing. We present, along Section 4, literature review and, Section 5, examples of static image processing, which show that those scenarios are not easy work to accomplish and academic researches still are not close to reach expected results.

Image Processing applied to Automatic Sign Language Recognition

According to Burger and Burge [4, p. 02], “image processing […] is the conception, design, development, and enhancement of digital imaging programs”, which is different from image editing, commonly used in software applications. Marques Filho and Vieira Neto [24] point out that image processing comprehends three stages: (a) Image Acquisition, considering, e.g., type of the file (image, video, real-time filming), lighting, background noise, among others; (b) Digital Image Processing (DIP), consisting of pre-processing, segmentation, and post-processing; (c) Digital Image Analysis, consisting in attributes extraction from data and patterns recognition and classification. These three stages of image processing, can involve a set of procedures to be performed in order to guarantee goal achievement of automatic sign language recognition. This set of procedures is composed by selection of existent language corpus (database) or creation of a new one, and a sequence of algorithms to prepare static or dynamic images for further proces sing, detection and/or segmentation regions of interest, data extraction and recognition, and classification and analysis of what these data mean. Following subsections, deeply discuss each of the three stages of image processing, taking into account the application domain of ASLR or Sign Language Machine Translation (SLMT). 4.1

Image Acquisition

In terms of image acquisition, to prepare data for further processing and/ or to create a new language corpus of static or dynamic images of sign languages, care mus t be taken with a series of factors, depending on investigation goals. These factors can be: (a) type of data (fingerspelling the manual alphabet, a sign, a phrase or a conversation) [3][17][33], (b) occlusion and self-occlusion [3][9][29], (c) lighting

117

condition [9][17][29]; (d) background and/ or environment conditions [17][29], (e) clothing aspects [17], (f) number of interpreters and signer’s style of communication (e.g., speed of signing, anthropomorphic differences (people sizes), (g) dominant hand) [3][17][33], (h) body posture or body positioning [12], (i) motion blur [3][9], and (j) to determine the exact start and end points of postures among continuous signing [3][17][33]. An example of database is the Libras Corpus of Federal University of Santa Catarina (UFSC) [8]. Computational systems concerned with ASLR are related to Natural User Interface (NUI) and multimodality concepts, since sign languages have as characteristics to be visual and gestural, using multimodal interaction forms (body, head, face, arms, hands and fingers) as natural interfaces for communication [1][6]. Considering types of data capture to be used in ASLR systems, literature review points out two main techniques: sensor-based and vision-based. Concerning sensor-based techniques, there are studies in ASLR domain using gloves with sensors [36][40]; Time-of-Flight (ToF) cameras [7][15][20][43][47][50]; Leap Motion [21]; motion capture [18]; armband (e.g., Myo) [29]. With respect to vision-based techniques, they are represented by different types of cameras [25] to capture static images, videos or real-time filming. This stereo vision technique consists of two simultaneous cameras positioned in a relative spot and angle simulating the human vision [11][46]. In many researches, which c apture images from a dynamic channel, such as videos or real-time filming, it is necessary to convert them into a sequence of static images for processing and analysis [13, p. 175]. There are also some studies, which mention the use of physiological-based techniques, such as, electromyogram (EMG) signals used by Paudyal et al. [29] in a hybrid approach, adding the use of a sensor-based technique of armband (Myo). Furthermore, Perrin, Cassinelli and Ishikawa [30] reported a proposal of a laser-based method for finger gesture recognition. However, authors did not mention sign language as an application; they evaluated results from letters and words written in the air. 4.2 Digital Image Processing (DIP) In this section, we present investigations from literature review of ten related works concerning specifying algorithms used for digital image processing. For Ghanem, Conly and Athitsos [13, p. 173], “The more accurate the detection and extraction is, the better the recognition results become”. The algorithm for skin segmentation was used to identify seven types of features (arms, hands, hand shape, hand palm, hand movement, hand location and body contact) in four different researches [2][3][17][47]. Normalization algorithm was applied to five types of features (hand shape, hand detection, fingers, facial expression and body skeleton) in four investigations [25][29][33][47]. Histograms were cited in two works [9][47] to analyze hand shape and facial expression, respectively. Contour algorithms were used by Barros, Pontes and Almeida [2] and Nel, Ghaziasgar and Connan [25] for arms and hand segmentation and/ or detection, respectively. Nel, Ghaziasgar and Connan [25], Barros, Pontes and Almeida [2] and Deaney et al. [9]

118

applied the Viola-Jones Algorithm to detect faces and facial expressions. Wang et al. [47] and Deaney et al. [9] used binarization for identifying hand configuration and facial expression, respectively. Some algorithms were used only in one paper. Nel, Ghaziasgar and Connan [25] used Gaussian Mixture Model (GMM) to detect arms and face. Barros, Pontes and Almeida [2] to identify fingers, used Convex Hulls (CH). CAMShift Tracking was used for hand detection as reported on Nel, Ghaziasgar and Connan [25]’s work. Borg and Camilleri [3] used three algorithms, not used in other investigations in this section, to recognize hand movement: Kanade-Lucas-Tomasi features, Discrete Cosine Transform, and Multiple hypotheses tracking. Gattupalli, Ghaderi and Athitsos [12] report the use of deep learning based pose estimation methods and a machine learning to improve pose estimation accuracy directly without applying any digital image processing previously. A few features were studied and results were reported by only one research. Wang et al. [47] investigated about the Libras parameter hand shape. Hrúz [17] only studied the Libras parameter hand location. The work of Wiggers, Ré and Porfírio [49] investigated about the Libras parameter hand configuration. Hand palm is not a Libras parameter; however, Barros, Pontes and Almeida [2] used this feature to identify finger location and direction. Body contact was an observed feature only by Hrúz [17], and body skeleton by Wang et al. [47]. The Libras parameters hand orientation and body expression were not investigated in the studies mentioned in this section. Also, researchers use some measurements to achieve better results of recognitions such as “Number of active fingers; Distance between hand palm center and hand; Number of depth points; Presence or not of hollow circumferen ce in the formation of the sign” [2]; identification of body joints [12][47]; physically possible hand angles for hand shapes parameter. Borg and Camilleri [3] defined four classifiers for hand motion (movement), each with its attributes: symmetry, h1 (dominant hand) stationary, h2 (non-dominant hand) stationary and type of motion. 4.3 Digital Image Analysis (DIA) In this section, we present investigations, from the same ten related works described in subsection 4.2, concerning with algorithms used for digital image analysis. The Support Vector Machine (SVM) was used by many authors [2][3][9][25] to recognize different types of features, such as, hand, hand shape, hand palm, hand movement, hand location, fingers and facial expression. Wang et al. [47] proposed Grassmann Covariance Matrix (GCM) associated with SVM (GCM_SVM) to fuse multimodal features (hand shape and body skeleton) of a sign language word and, also compared results from experiments with Hidden Markov Model (HMM), Dynamic Time Warping (DTW), Log-Euclidean Distance with SVM (LED_SVM) and Autoregressive Moving Average with SVM (ARMA_SVM). The k-Nearest Neighbors algorithm (k-NN) was applied for recognition of facial expression and hand movements [3][33], respectively. Hrúz [17] used expectation maximization and Gaussian Mixture Model to recognize hand movements, hand location and body contact features. Borg and Camilleri [3] applied three supervised

119

machine-learning classifiers (SVM, k-NN and eXtreme Gradient Boosting (XGBoost)) to compare results of hand movement. Paudyal et al. [29] proposed a new algorithm called DyFAV (Dynamic Feature Selection and Voting) to recognize fingerspelling in real-time and compared results with four other algorithms: SVM, Naive Bayes, Random Forest and Multi-Layer Perceptron (MLP). Wiggers, Ré and Porfírio [49] applied Kohonen ANN to hand configuration recognition. For human pose, Gattupalli, Ghaderi and Athitsos [12] used Convolutional Neural Networks (CNN) and DeepPose Network to recognize ASL signs. Other relevant information extracted from related works are presented in Table 1, where they are arranged by paper. These information refers to the type of database (created by authors or used existing ones), the database/ corpus size, the image angle, type of data capture (if database was created by authors), and percentage of recognition accuracy, statistic methods and statistic features. Table 1. Other relevant information extracted from related works. Dataset

Images angle

[17]

Author’s ECHO Sign Language A (BU3DFE); B (CK+) ASLID; ChaLearn; FLIC Author’s

[25]

Author’s

[29]

Author’s

[2] [3] [9]

[12]

[33] Author’s [47]

[49]

Hand shape

Corpus size 720

Type of data captured Video-based

Overall accuracy 83.33; 82.35; 87

Upper body

1500

Vision-based

72.21 - 96.54

Facial expression

A=700

Vision-based

66.2

Upper body

Vision-based

Upper body

560

Vision-based

Upper body Armband and EM G Upper body and Facial expression

600

Video-based Sensor and physiologic-based

79 95.36

100

Video-based

80%

ChaLearn; Author’s (A,B,C)

Upper body

A (1850); B (3000); C (7000)

Sensor and videobased

GCM (A=96; B=92.4; C=70.9)

Federal University of Paraná

Hand shape

549

Video-based

85.66

In Table 1, eight out of ten studies used only vision -based techniques to conduct their experiments; six used images from upper body of the human signers; five created their own dataset to perform experiments; five informed the size of their training and testing set (sample size or percentage); and, the corpus (dataset) size used

120

by authors vary from 26 images/ frames to 7000. Also, we can perceive that the higher overall accuracy percentage was obtained in the work of Borg and Camilleri [3] with 96.54%, as a baseline result from the three conducted experiments (SVM, kNN and XGBoost) for h1 (dominant hand) stationary, in which among the three algorithms, XGBoost outperformed with 97.74%. Those results from overall accuracy and other types of measurements are used to understand the performance of the proposed work, depending on what researchers want to observe or to compare. Some statistic methods informed by authors were: Pearsons Correlation Coefficients [17]; Confusion Matrix [3][9][25][29]; Kappa Coefficient, Degree of Agreement, Reliability Test [49]; Chains model [12]; Cross validation, Average Accuracy, Standard Deviation [33]; Cross validation [9][47].

Digital Image Processing of Static Images of Libras Signs

We chose to reduce the scope of the problem, as a first attempt to look for solutions, using gray-scale, two-dimensional and static images. However, even though we tried to simplify input data, there are still issues to be addressed. In this process, eleven static images of a Deaf person performing eleven different signs in Libras were registered, and these pictures were taken with a DSLR (Digital Single-Lens Reflex) camera in a room with eighteen tubular fluorescent lights on the ceiling. Figures 1-5 show images of the picture representing the Libras sign “EXCUSE ME” (“PERDÃO” in Brazilian Portuguese). This sign was chosen because one can understand it by looking a static image. For tests, it was necessary to set a background support environment that consisted on Python 2.7.13 and extra packages, such as OpenCV 4.0, NumPy 1.10.2, Matplotlib 1.5 and Scikit-image; and, the coding environment PyCharm Community 3.4, using existing algorithms.

Fig. 1. Outputs of Thresholding algorithms.

121

Fig. 1 shows the output of Thresholding algorithms that perform binarization on images, being “the simplest way to segment objects from a background” [38]. In Fig. 1(b), Fig. 1(c), Fig. 1(f) and Fig. 1(g), the little finger is not visible, because face and hand assume the same color (white). Fig. 1(b), Fig. 1(c) and Fig. 1(d) show a shadow on the back of the hand and hand on the chin, which can be positive to differentiate the face from the hand, or negative from occlusion perspective. From these images presented, we may observe that binarization is not a good option for face expression extraction.

Fig. 2. Outputs of Corner Detection algorithms.

Fig. 2 shows the output of corner detection algorithms, being Shi-Tomasi algorithm [28] a small modification of Harris [26]. Harris algorithm output, in Fig. 2(a), makes a good mapping on eyes, nose, joint between fingers corner and arm. ShiTomasi algorithm output, in Fig. 2(b), correctly mapped fingertips, however, also mapped eyes and joint between fingers. For this image, Harris algorithm use more points to highlight corners than Shi-Tomasi.

Fig. 3. Outputs of Edge Operators algorithms.

Fig. 3 shows the output of edge operator algorithms for edge detection [39]. Fig. 3(a) and Fig. 3(b) present output from using Roberts and Sobel algorithms pos t-edited in a photo editor with brightness and contrast. They deliver similar results; both delimit hand edges, which facilitates the extraction of hand shape.

122

Fig. 4. Outputs of Canny Edge Detector algorithms.

Fig. 4 shows the output of Canny Edge Detector algorithm, also used for edge detection [37]. The value of sigma (σ) is a parameter applied for blur effect (Gaussian). In Fig. 4(b), the sigma value is 1 and the image gets less Gaussian effect, the output image present many details, hand shape and face expression are well delimited. In Fig. 4(c), the sigma value is 3, with more Gaussian effect, the output image present less details, however, face contour is easier to identify.

Fig. 5. Outputs of Gradient algorithms.

Fig. 5 shows the output of three Gradient algorithms: Laplacian, Sobel X and Sobel Y. Used algorithms can be found in [27]. Fig. 5(b) presents the output of the Laplacian algorithm, which draws the contour of the hand shape and face. Both Sobel X (Fig. 5(b)) and Sobel Y (Fig. 5(c)) present texture in output images, which refers to depth and overlapped features. These tested algorithms are mostly initial steps of digital image processing. The objective of testing them is to help the selection of a set of algorithms that could deliver results that facilitates feature extraction and reduces noise for the next steps. Also, the original picture used in the tests were taken in an inadequate site, with bad lighting, and white background. That influences directly in the results of the algorithms.

123

Final Considerations

The problem complexity can be evidenced by observing different sides of the same domain of application (the user side, the developer side, the dataset side and the device side), which are described below. On the user side, when we talk to them about potential scenarios (Section 3), they usually expect a quick development by the research team and expect an interaction through ideal communication (fast in recognition time, complete and correct translation). However, due to an ASLR system complexity, a project may take many years to result in a slow, incomplete and partially correct translation. In this context , expectation and real results can affect user experience and human -human interaction. Also, considering one of our primary users (deaf or hard of hearing persons that use sign language as their first language), there is a need for low cost devices and accessible environments, since WHOSIS [48, p. 08] affirm, “Disability disproportionately affects vulnerable populations. Results from the World Health Survey indicate a higher disability prevalence in lower income countries than in higher income countries”. On the developer side, to design an ASLR system is still a challenge for researchers as shown in the previous sections, especially when real-time applications is desired. Challenges are present in domain of application (sign language communication), in image acquisition, in digital image processing, and in digital image analysis. One example from our studies is that we did not find an investigation where authors report implementation observing the five sign language parameters together, because these add extra complexity. For example, from related works (Section 4), two parameters of Libras (hand orientation and body expression) were not mentioned to be recognized in any of the ASLR systems proposed. This may delivery wrong translation, resulting in poor communication and frustrating user experience. On the dataset (corpus) side, Rezende, Castro and Almeida [33, p. 01] inform, “There is currently no standardized database containing signs in a format that allows the validation of computational classification s ystems”. Wang et al. [47] also affirm, “publicly available datasets are limited in this area”. Furthermore, we did not find guidelines for creating sign language corpus, and those works who created their own dataset to use during digital image processing and data classification do not mention protocol or guidelines. What we did find were ten problems observed that might be a start to propose best practices. On the device side, many researches have been done and many are ongoing, using vision-based, sensor-based, physiologic-base or laser-based techniques. None of them are perfect, however, there are researchers presenting high accuracy rates in recognition results. Each device has positive and negative factors, for example, glove with sensors or colored gloves or Leap motion facilitate to distinguish hands and fingers for detection and segmentation, and also may eliminate the occlusion problem; however, if we think about natural user interaction, we do not wish to hold or use anything, just gesture (in this case, sign language).

124

As future work we intend to design a protocol for dataset (corpus) creation, develop this dataset following the established protocol, choose a set of digital image processing algorithms and digital analysis methods to test and to compa re results.

References 1.

4. 5. 6.

8. 9.

10. 11.

12.

13.

Aran, O.; Ari, I.; Akarun, L.; Sankur, B.; Benoit, A.; Caplier, A.; Campr, P.; Carrillo, A.; Fanard, F.: SignTutor: An Interactive System for Sign Language Tutoring. IEEE Computer Society 1 (16), pp. 81-93, (2009). https://goo.gl/4TDzCR, last accessed 2018/10/11. Barros, R.; Pontes, A.; Almeida, J.: Reconhecimento de linguagem de sinais: aplicação em LIBRAS. In: V Jornada de Informática do M aranhão on Proceedings, São Luís, M aranhão (2014). https://goo.gl/uWG6y1, last accessed 2018/09/19. Borg, M .; Camilleri, K.: Towards a Transcription System of Sign Language Video Resources via M otion Trajectory Factorisation. In: Proc. of DocEng, Valletta on Proceedings, Valletta, M alta (2017). https://goo.gl/4nWBDk, last accessed 2018/09/19. Burger, W.; Burge, M .: Principles of Digital Image Processing: Fundamental Techniques. 1st edn. Springer, London (2009). Capovilla, F.; Raphael, W.; Temoteo, J.; M artins, A.: Dicionário da Língua de Sinais do Brasil: A Libras em suas M ãos. 1st edn. EdUSP, São Paulo (2017). Céspedes-Hernández D., González-Calleros J.M ., Guerrero-García J., Rodríguez-Vizzuett L. (2019) Gesture-Based Interaction for Virtual Reality Environments Through UserDefined Commands. In: Agredo-Delgado V., Ruiz P. (eds) Human-Computer Interaction. HCI-COLLAB 2018. Communications in Computer and Information Science, vol 847, Springer. Coleca, F.; M artinetz, T.; Barth, E.: Gesture Interfaces with Depth Sensors. In: Grzegorzek, M . et al. (eds.) Time-of-Flight and Depth Imaging 2013, LNCS, vol. 8200, pp. 207-227, (2013). Corpus/Libras UFSC, http://www.corpuslibras.ufsc.br/, last accessed 2018/09/19. Deaney, W.; Venter, I.; Ghaziasgar, M .; Dodds, R.: A Comparison of Facial Feature Representation M ethods for Automatic Facial Expression Recognition. In: SAICSIT ’17 on Proceedings. Thaba Nchu, South Africa (2017). https://goo.gl/sC2pCN, last accessed 2018/09/19. Dicionário da Língua Brasileira de Sinais, www.acessibilidadebrasil.org.br, last accessed 2018/09/24. El-Jaber, M ., Assaleh, K., Shanableh, T.: Enhanced user-dependent recognition of arabic sign language via disparity images. In: M echatronics and its Applications (ISM A) on Proceedings, pp. 1-4, IEEE (2010). Gattupalli, S.; Ghaderi, A.; Athitsos, V.: Evaluation of Deep Learning based Pose Estimation for Sign Language Recognition. PETRA ’16 on Proceedings, pp. 12:1-12:7. ACM , Greece (2016). https://goo.gl/nkjpAj, last accessed 2018/09/24. Ghanem, S.; Conly, C.; Athitsos, V.: A Survey on Sign Language Recognition Using Smartphones. PETRA’17 on Proceedings, pp. 171-176. ACM , Greece. (2017). https://goo.gl/qHzUuF, last accessed 2018/09/24.

125

14. 15. 16. 17. 18. 19.

20.

21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33.

34.

35.

HandTalk, https://www.handtalk.me/, last accessed 2018/12/10. Hansard, M .; Lee, S.; Choi, O.; Horaud, R.: Time of Flight Cameras: Principles, M ethods, and Applications. Springer (2012). Harper, S.; Yesilada, Y. (Eds.). Web accessibility: a foundation for research. Springer Science and Business M edia (2008). Hrúz, M .: Automatic Sign Categorization using Visual Data. In: ASSETS ’11 on Proceedings, pp. 229-230. ACM , Scotland (2010). Huenerfauth, M .; Hanson, V.: Sign Language in the Interface: Access for Deaf Signers. In: Stephanidis, C. (eds.) The Universal Access Handbook, CRC Press (2009). Kipp, M ., Heloir, A., Nguyen, Q.: Sign Language Avatars: Animation and Comprehensibility. In: H. Högni Vilhjálmsson et al. (eds.) IVA 2011, LNAI, vol. 6895, pp. 113-126. Springer, Heidelberg (2011). Kopinski, T.; Geisler, S.; Handmann, U.: Gesture-based Human-M achine Interaction For Assistance Systems. In: 2015 IEEE International Conference on Information and Automation Lijing on Proceedings, pp. 510-517. IEEE, Lijiang (2015). Kumar, P.; Gauba, H.; Roy, P.; Dogra, D.: A multimodal framework for sensor based sign language recognition. Neurocomputing 259, 21-38 (2017). Letras-Libras/UFSC, http://www.glossario.libras.ufsc.br/, last accessed 2018/11/10. M arcon, A., Soares, A., Luna, C., Reveilleau, M ., Souza, T.: Estudos da Língua Brasileira de Sinais. 1st edn. Editora UPF, Passo Fundo (2011). M arques Filho, O.; Vieira Neto, H.: Processamento Digital de Imagens. 1st edn. Brasport, Rio de Janeiro (1999). Nel, W.; Ghaziasgar, M .; Connan, J.: An Integrated Sign Language Recognition System. In: SAICSIT’13, pp. 179-185. ACM , East London (2013). OpenCV, https://goo.gl/dHgEAj, last accessed 2019/02/13. OpenCV, https://goo.gl/sYAa9y, last accessed 2019/02/13. OpenCV Python Tutorials, https://goo.gl/M 9qTdQ, last accessed 2019/02/13. Paudyal, P.; Lee, J.; Banerjee, A.; Gupta, S.: DyFAV: Dynamic Feature Selection and Voting for Real-time Recognition of Fingerspelled Alphabet using Wearables. In: IUI 2017, pp. 457-467. ACM , Limassol (2017). https://goo.gl/vm6f9k, last accessed 2018/09/24. Perrin, S.; Cassinelli, A.; Ishikawa, M .: Gesture Recognition Using Laser-Based Tracking System. In: FGR 2004, pp. 541-546. IEEE, Seoul (2004). ProDeaf, http://prodeaf.net/, last accessed 2018/12/11. Quadros, R.; Pizzio, A.; Rezende, P.: Língua Brasileira de Sinais II. 1st edn. UFSC, Florianópolis (2008). https://goo.gl/wRrKAX, last accessed 2018/10/11. Rezende, T., Castro, C., Almeida, S: An approach for Brazilian Sign Language (BSL) recognition based on facial expression and k-NN classifier. In: 29th SIBGRAPI, Workshop on Face Processing Applications, on Proceedings, pp. 1-2. SBC, São José dos Campos (2016). https://goo.gl/P9dtkg, last accessed 2018/09/19. Rojano-Cáceres, J.; Sánchez-Barrera, H.; M artínez-Gutiérrez, M .; Guillermo, M .; OrtegaCarrillo, J.: Designing an interaction architecture by scenarios for Deaf people. In: XVII International Conference on Human Computer Interaction (Interacción) on Proceedings, pp. 38:1-38:2. ACM , Salamanca (2016). Rybená, https://goo.gl/hsdkxu, last accessed 2018/12/11.

126

36.

37. 38. 39. 40.

41. 42. 43.

44.

45. 46. 47. 48. 49.

50.

Sadek, M . I., M ikhael, M . N., M ansour, H. A.: A new approach for designing a smart glove for arabic sign language recognition system based on the statistical analysis of the sign language. In: NRSC 2017 on Proceedings, pp. 380-388. IEEE, Alexandria (2017). Scikit Image, https://goo.gl/bQFQ49, last accessed 2019/02/13. Scikit Image, https://goo.gl/DGWunx, last accessed 2019/02/13. Scikit Image, https://goo.gl/GqBU6S, last accessed 2019/02/13. Shukor, A. Z., M iskon, M . F., Jamaluddin, M . H., Bin Ali, F., Asyraf, M . F., Bin Bahar, M . B., et al.: A new data glove approach for malaysian sign language detection. Proc. Comput. Sci. 76, 60-67 (2015). Stokoe, W. C.: Sign language structure: An outline of the visual communication systems of the American Deaf. Studies in Linguistics: Occasional papers. T echnical Report 8. Dept. of Anthropology and Linguistics, University of Buffalo (1960). Strobel, K.; Fernandes, S.: Aspectos lingüísticos da LIBRAS. SEED/SUED/DEE, Curitiba (1998). Uebersax, D.; Gall, J.; Bergh, M .; Gool, L.: Real-time Sign Language Letter and Word Recognition from Depth Data. In: IEEE International Conference on Computer Vision Workshops, pp. 383--390. IEEE, Barcelona (2011). https://goo.gl/33kyDX , last accessed 2019/02/14. Vieira, S., Santos, A., Costa, R., M aritan, T., Aschoff, M ., Veríssimo, V.: A Study on the Use of M ultiple Avatars in 3D Sign Language Dictionaries. In: WebM edia’17, pp. 325332. ACM , Gramado (2017). https://bit.ly/2TR8e1r, last accessed 2018/12/11. VLibras, http://www.vlibras.gov.br/, last accessed 2018/12/11. Wang, Q.; Chen, X.; Zhang, L.; Wang, C.; Gao, W.: Viewpoint invariant sign language recognition. Computer Vision and Image Understanding, 108, 87-97 (2007). Wang, H.; Chai, X.; Hong, X.; Zhao, G.; Chen, X.: Isolated Sign Language Recognition with Grassmann Covariance M atrices. ACM Transactions on Accessible Computing 8(4), 14:1-14:21 (2016). https://goo.gl/Vvvkrw, last accessed 2018/09/19. WHOSIS, http://www.who.int/whosis/whostat/2011/en/, last accessed 2018/09/24. Wiggers, K.; Ré, A.; Porfírio, A.: Classificação das Configurações de M ão da Língua Brasileira de Sinais M ediante Rede Neural Artificial Kohonen. RECEN, 16(2), 175-197 (2014). https://goo.gl/Bu48gZ, last accessed 2018/09/18. Zafrulla, Z.; Brashear, H.; Starner, T.; Hamilton, H.; Presti, P.: American Sign Language Recognition with the Kinect. In: ICM I’11 on Proceedings, pp. 279-286. ACM , Alicante (2011). https://bit.ly/2TVAKyX, last accessed 2018/09/10.

Nonverbal Communication Mediated through Haptic Technology: A Survey Hector M. Camarillo Abad 1 , J. Alfredo Sánchez2 , Oleg Starostenko 1 1

Universidad de las Américas Puebla, Ex-Hacienda Sta. Catarina M ártir S/N. San Andrés Cholula, M exico 2 National Laboratory of Advanced Informatics (LANIA), Rébsamen 80, 91090 Xalapa, M exico hector.camarilload@udlap.mx, alfredo.sanchez@lania.edu.mx, oleg.starostenko@udlap.mx

Abstract. Nonverbal communication (NVC) can benefit from any human sense. The sense of touch, or haptic sense, is often used as a channel of NVC. This paper surveys research that relies on technology to communicate nonverbally through the haptic sense. The analysis of reported work and ongoing projects in the area during the last five years has resulted in a taxonomy that is based on three major dimensions, based upon the intent of the messages exchanged: interpretive, affective, and active communication. Analysis shows that using the haptic sense as a communication means still has many open research areas and potential new applications, and has proven to date to be an effective mechanism to communicate most of what humans would want to: messages, emotions, and actions. Keywords: Nonverbal Communication, Haptic Technology, Technology M ediated Communication, Nonverbal Interaction M ediated through Haptic Technology.

Introduction

Even though technology has been changing the ways in which people exchange in formation, the essence of human communication remains the same. Human communication is composed mainly of language and nonverbal communication (NVC). Any signal other than speech or writing is considered NVC, and is important because it can complement, repeat, accent, contradict, regulate, and even substitute language entirely. Arguably, just being human involves the constant use and interpretation of nonverbal signals. NVC was not considered a formal research subject until the 1950’s, when it was recognized as a new field (Schroeder, 2017). NVC is thus a fundamental component of communication as humans, and it must be investigated further from the perspective of Human-Computer Interaction. The role of various technologies in supporting or enabling new forms of NVC deserves special attention, from machine vision or motion sensing to wearable and haptic technologies. We present a survey of existing work in the specific area of NVC mediated by haptic

128

devices, as we consider there is great potential of this technology to support and enhance NVC. Our survey provides perspective of NVC mediated by haptic technologies by following specific criteria (Section 2), organizing knowledge in the field through a taxonomy (Section 3) that considers dimensions derived from the intent of the messages exchanged: affective, active and interpretive communication. We use our taxonomy to contextualize salient projects in each category and to facilitate comprehension of their approaches and interrelationships. We discuss challenges, open issues and research directions for the community interested in advancing the field (Section 4). Finally we present our ongoing work for this research (Section 5).

Survey Criteria

Given the vast amount of work that has been undertaken on NVC, we adopt ed a few criteria in order to narrow down the field of study. Firstly, research relevant to this survey must involve some kind of NVC, even though authors may not mention it explicitly. Secondly, works considered must focus on the haptic sense, either for sending or receiving purposes in the NVC. Haptics includes all touch behavior, and it is considered an area in need of research (Schroeder, 2017). Haptics is the word used for the investigation of human-machine communication using the sense of touch, both as input and output (Hayward et.al. 2004). The scope of the survey also includes multisensory works, with the sole condition that the haptic sense must be present. Thirdly, the communication means employed by the projects in our survey include some form of technology. This technology can either be a robot, wearable technology, or grounded technology. Wearable technology is a small body -worn technology. This implies a new form of human-computer interaction also known as constant user interface (Mann, 1998), and thus it is of significant interest to this survey. Using haptic technology as a nonverbal medium to communicate has, theoretically, applications in every daily life scenario. However, exactly to what extent haptic technology is being used has not been systematically studied. We provide a general overview of all the different applications reported in the literature, organizing them according to their goals and applications. Our proposed taxonomy comprises every application found in our literature review, and the ultimate goal is that any new application can fit into an area, perhaps adding sub-areas, of the taxonomy. Our literature search started by limiting the publications’ years from 2015 to date. The search included all related terms: nonverbal, mediated, movements. Furthermore, results included at least one of the following terms: wearable, wearables, haptic, tactile. Using the variety of applications yielded by the search, we started constructing the taxonomy from the bottom up, adding or removing words in the search to find area-specific works relevant in NVC mediated by technology.

129

Proposed Taxonomy

We have developed a taxonomy for haptic-based NVC mediated by technology (HNVC) that considers mainly the purpose of communication, focusing on reported uses and applications of haptic technology in daily human activities. The taxonomy is intended to classify the wide range of HNVC applications and to inspire researchers to fill any gaps in the depth of the taxonomy. As a starting point to classify research works in the area of HNVC, works are divided into three major categories: interpretive, affective and active (Fig. 1). This classification differs from the work of MacLean et. al. (2005), since they divided communication through touch into three main categories: signaling and monitoring, expression of affect, and sharing control with intelligent systems. Our classification is also inspired on the ultimate goal of the communication , and is intended to further classify current works found in literature, thus making the initial classification more general. If the goal of the research is mainly to help users understand some meaning, it falls under the interpretive area. If it is related to emotion, it falls under the affective area. And if the communication is intended to convey actions or steps to be followed, it falls under the active area. An alternative perspective for this classification considers the most common purposes of verbal communication : sharing a story (interpretive), an exchange of feelings towards something or someone (affective), or instructions to achieve something (active). The next subsections describe the categories in detail, each with their respective subcategories.

Fig. 1. General classification of HNVC.

3.1 Interpretive HNVC Interpretive HNVC refers to work that focuses exclusively on conveying meaning associated to nonverbal cues . Its main intention is that communication can be understood. Examples of interpretive HNVC includes representing words with different haptic sensations instead of words or phrases. Interpretive HNVC can be further subdivided into two subcategories, depending on the granularity of the signals : wordoriented and phrase-oriented, as discussed below. Word-oriented HNVC. Works in the literature that focus on transmitting some meaning using haptic stimuli analogous to a word in verbal communication fall under this category. The work by Enriquez et. al. (2006) investigate how to convey words related to plant and fruits using haptic feedback to users. They discuss that a smarter

130

training method is needed to avoid some interpretation errors made by the users. The work by Chen et. al. (2018) compared guided learning and self-guided learning for acquisition of haptic words. Their findings support the use of guided learning, reporting an accuracy of over 90%. Phrase-oriented HNVC. When research aims to convey haptic complex meanings analogous to phrases or sentences, it falls under this category. The work by Oliveira et. al. (2014) focuses on directional cueing and obstacle detection; hence their meanings are more complex than just one word. The work in Schelle et. al. (2015) involves inclusion of people with cognitive disabilities, and uses a textile pillow that allows for communication between patients with dementia and their families or caregivers. A single pillow is used, and all participants in the communication process must be at the same physical location touching the pillow. The work by Velázquez et. al. (2018) reports research on assisting visually impaired people in their situational awareness . They achieve this by testing the recognition of words coded with haptic stimuli, and then progress to even more complex sentences (involving two, three and four words). Their aim is to understand the learning processes and memory capabilities of the subjects, defining three main tactile concepts: leaning, language, and memory. 3.2 Affective communication mediated through haptic technology The affective communication category involves work with the aim of understanding, emulating, or even creating an emotional reaction of the users. In turn, the affective area can be further subdivided into emotion, mood, and sentiment (Fig. 1). This classification is in line with the theory of Brave et. al. (2003). According to them, emotion is object directed, intentional and short-lived (seconds). Mood is nonintentional and undirected at an object, therefore being diffuse and general. Mood influences appraisal of external events, and are processed over a longer time (hours or days) than emotions. Sentiments last longer (persist indefinitely) than both emotions and moods. Sentiments guide which situations and objects humans seek out or avoid (Brave et. al. 2003). Examples of the affective category include making people feel a certain emotion through vibration, exploring cultural differences in emotion representation, or using robots to elicit certain human emotions. Emotion-oriented HNVC. In this first subcategory, only instantaneous affective perception of humans is considered. The work by Morrison et. al. (2015) enables users to interact through a vest and a wall, both of which are endowed with vibrotactile technology. Their goal was to make people feel a given emotion (relaxed, calm, aware of danger). The research done by Haritaipan et. al. (2016) present a conceptual design of communication using tactile and gestural interactions. Their main objective is to identify the difference between the Japanese and the French cultures. The y conclude that cultural background does have an impact on the representation of emotions. The work by Hirano et. al. (2016) centers on the interaction between humans and robots, and how visual and haptic interaction affects the way the user feels.

131

Mood-oriented HNVC. The work by Ahmed et. al. (2016) shows that users associate moods to the system, and said moods have an influence on them. They researched different forms of touch interaction, vibrotactile and force-feedback, to gain insight of which is referred by users to convey affection. Their setting involves a user interacting with a 3D model of a virtual agent, who interacts through haptic signals sensed by the user´s hand. Their findings also suggest that force-feedback interactions are preferred over vibrotactile for mood and affection purposes. The work by Mazzoni et. al. (2016) explores how a wearable vibrotactile globe can enhance moods watching a film. Their findings are that the intensity and frequency of the vibrotactile stimuli can produce anticipation, tension, and calmness. Sentiment. The work of Goedschalk et. al. (2017) is a good example of a negative sentiment towards a virtual agent. They use haptic feedback to enhance believability to a virtual “bad guy”, trying to inspire negative sentiments on the user. Their haptic feedback was provided by a push sensation on a wearable vest, thus mimicking a common aggression nonverbal cue. Although their results are reported to have non statistical significant effects, they highlight the importance to further research this interaction scenario. The work by Bosse et. al. (2018) also research virtual bad guys, although no real threat was presented to the users, only a threatening n on-functioning device was used. Users were made to believe the device could send a small electrical shock in their finger area. Their findings highlights that only feeling a device that can potentially hurt rises the anxiety sentiments of the users. 3.3 Active communication mediated through haptic technology Active HNVC in our taxonomy refers to emotion-free messages intended not only to be understood, but to be performed as well. In this category users need to understand the message and then perform the required action. The active HNVC category is one we have elaborated with more detail, and hence involves more subcategories. It can be first divided, according to the messages’ main purpose, into either perception-, control- or instruction-oriented HNVC (Fig. 2). Examples of the active category include navigation, control and virtual environments interaction.

Fig. 2. Active HNVC sub-classifications.

132

Perception-oriented HNVC. This category involves all applications focusing on sensing virtual objects, or people sensing their own motions and reactions. Research in this category explores HNVC without regard of emotional or affective implications. The category can be further divided into sensing- and immersion-based HNVC. Sensing-based HNVC. Sensing-based HNVC focuses on either enhancing or complementing any human sense. It is of special interest for inclusion of people with disabilities. Houde et. al. (2016) performed a review of what it is known about deaf people experiencing their own bodies, body-related abilities and body-related processing. Typical users of research in this sub-area are both deaf and blind people. Bălan et. al. (2015) focus on how visually impaired people can benefit from auditory and haptic senses to construct a mental representation of their environment. Their findings include that visually impaired people successfully use other senses, including touch, to acquire and model spatial information. Immersion-based HNVC. Any work that relies on HNVC aiming to remove users from their actual environment, either in a presentation or in a virtual environment, falls under this subcategory. Of particular interest of this sub-section are virtual environments and games. Kruijff et. al. (2016) focus on navigation of particularly large virtual environments or games. They used three senses: haptic, visual, and auditory. The haptic one is achieved through vibrotactile cues under the participant’s feet, using bass-shakers. They compared their approach with other techniques to navigate virtual environments, namely regular seated joystick and standing learning locomotion. They conclude their multisensory method can enhance self-motion perception and involvement/presence. Feng et. al. (2016) worked on enhancing a non -fatiguing walking in a virtual environment. They used tactile cues such as movement wind (the wind resistance felt by walking), directional wind (environmental wind), footstep vibration and footstep sound. They also conclude that tactile cues are of significant improvement for the activity. Bernardet et. al. (2016) developed an open source software framework to construct real-time interactive systems based on movement data. They highlight the importance of recording, sensing, storing, retrieving, analyzing, understanding and displaying movement data through various senses, while all these components communicate with each other. Control-oriented HNVC. Any research focusing on controlling certain technology using the haptic sense, either as a control means or as feedback of the controlled technology is included here. Distinctions in the type of technology lead to two subcategories: logical and physical HNVC. Logical HNVC. Any work that aims to control software is included here. Saunders et. al. (2015) focus on controlling desktop applications through the feet, using discrete taps and kicks. They propose ten design guidelines to be followed, the most general of which are: tapping for frequent actions, consider both feet as dominant ones, prefer toe taps, and sensing techniques should be robust. The work by Costes et. al. (2019) use pseudo-haptics effects to virtually evoke some haptic properties: stiffness, roughness, reliefs, stickiness, and slipperiness . Their results show that haptic properties can be successfully conveyed to a user interacting with a software.

133

Physical HNVC. In contrast to logical HNVC, physical HNVC involves research of controlling robotic mechanisms or hardware in general. Bufalo et. al. (2017) focused on a disturbance rejection task of a controlled element. They considered three different modalities of haptic feedback: Variable Haptic Aid (VHA), Constant Haptic Aid (CHA), and No Haptic Aid (NoHA). Their results show that VHA is the best modality to help users learn the task. D'Intino et. al. (2016) focus on haptic feedback used to learn a compensatory tracking task. Their experiment uses two groups, one without any haptic feedback, and the other training with feedback and evaluating without feedback. Their conclusion is that haptic feedback users learn quicker. Van Oosterhout et. al. (2015) explored the possible effects of inaccuracies in haptic shared control. They found out inaccuracies do have a negative effect in using haptic control tasks, and therefore any system should be accurate to work properly. Xu et. al. (2016) used wearable technology to develop a motion capture of arm movement. They used this data to control a virtual robotic arm. Instruction-oriented HNVC. This category involves all works and research that include HNVC for learning or training. It includes acquisition, improvement, and assistance of both skills and knowledge. Instruction-oriented HNVC is divided into two main types of applications: concept- and skill-oriented. Concept-oriented work focuses on abstract knowledge, while skill-oriented research involves any physical ability. Concept-oriented HNVC. This area does not deal with movement, but with teaching abstract ideas using tangible methods. Robotti (2017) works on helping students with difficulties in algebra, especially the ones with developmental dyscalculia. Help is based on providing meaning to algebraic notations through visual, nonverbal, and kinaesthetic-tactile systems. The work by Magana et. al. (2017) is aimed to improve learning of electricity and magnetism concepts, using a visuohaptic simulation approach. Visuohaptic refers to the use of both visual and haptic sense. Their work compares visuohaptic to visual only and instructional multimedia. Their results show that visuohaptic did better, although not significantly. Their findings include that force feedback is a novelty for the majority of the users, and therefore might contribute to overloading work memory. Skill-oriented NVC. The skill-oriented HNVC category involves human-human interaction mediated through technology, with the purpose of teaching a new skill to the user. According to Yokokohji et. al. (1996) training is a skill mapping from an expert to a learner, also called trainer and trainee, respectively. Their work highlights the importance of skill transfer, because it is time consuming and thus not easy to achieve. Furthermore, motor skills add the challenge of being difficult to describ e with language. Schirmer et. al. (2015) used a tactile interface for eyes -free pedestrian navigation in cities. They used only two actuators and divided their experiments in to compass and navigation modes. Prasad et. al. (2014) used a vibrotactile vest to provide navigation and obstacle detection to the wearer. Their findings include that vibration produces less cognitive load and greater environmental awareness than other navigational approaches. Yang et. al. (2016) worked on a handwriting tutor-tutee

134

scenario mediated by a robot, using haptic feedback for corrections. Their setting involved the computer capturing an expert’s motion skills, and then the robot can handle multiple tutees by itself. Pedemonte et. Al. (2016) also worked on transferring handwriting skills. They researched remote user communication with the goal of improving their handwriting skills. They focus on human-human interaction mediated by a haptic device between novice and expert. Their findings encourage the use of haptic guidance in the handwriting learning process. Choi et. al. (2012) study a simulator for closing a wound with suturing procedures. Their aim is focused not only in medicine, but in nursing as well. They use both hands for the procedure and are aided by haptic feedback to learn the correct procedure.

Research Directions

Based upon our survey of the field, we have generated a taxonomy that provides perspective on existing work. In this section, we discuss directions for research in each of the three major categories of the taxonomy: affective, interpretive and active HNVC. The affective category involves work that started with recognizing emotions, and then evolved to be able to communicate and simulate them. Given the popularity and feasibility of using the haptic sense to convey emotions, recent research directions are shifting towards emotions that linger through time: moods and feelings. It is expected that HNVC will be able to identify and change the affective states of humans even better than we as humans can. The interpretive category has also evolved from users being able to differentiate single words, to understand multiple words that form full sentences. This area has also achieved to convey complex ideas and instructions with iconicity, much like every icon we use to save or open a document. This area is expected to achieve a high level of communication between humans using only the tactile sense. There is still need for research about differentiability of tactile patterns, the cognitive capacity to associate each pattern with a meaning, and if the association lingers in time. The active category has the most variants and subcategories in our proposed taxonomy. This is only natural, as humans perform a variety of activities in daily life, so we cannot turn off the haptic sense for any of them. In the control area the main goal is to achieve full controllability over the technology, being as natural and error-free as possible. Therefore, future research directions need to focus on the mutual understanding of human and technology. Perception strives to enhance our haptic sensory capabilities, possibly to compensate another sense or just to enhance an experience. Future research directions are to what extent our different senses correlate and can compensate each other, and which sensations help users to feel a more realistic virtual immersion in games and simulators, without causing dizziness or distract ions. The instruction area is still very open, because it deals with acquiring new skills using technology. This area needs to benefit from the others, since it needs understanding of interpretive HNVC, and also of some degree of control between human and technology. The concept instruction subarea is still very open, as normal classroom settings do not use currently the haptic sense to help students understand abstract concepts. The health applications have attracted attention because of the high level of training doctors need, and also because of the importance for the well-being of everyone.

135

HNVC has a wide range of application domains. Applications may range from analyzing cultural backgrounds for understanding emotions to using robots for assisting medical doctors in surgery. Understanding this diversity demands a rich taxonomy to encompass all existing research, and also to show the full potential of the human haptic sense to communicate and to perform different tasks. The use of the haptic sense also allows applications to be more inclusive, because visually impaired and deaf people can use their touch sense to compens ate for deficiencies in their other senses. People with some sort of cognitive deterioration (such as dementia) may also benefit from HNVC, because the haptic medium offers them a novel way to express themselves and understand what others are trying to say . Everyone can benefit from advances in HNVC research, therefore there is a current need to continue research in this area. To formalize HNVC protocols and languages is of general importance, because all other subareas will benefit from having a solid basis in the general area.

Ongoing and Future Work

While our taxonomy provides a general overview of HNVC, specific details could potentially have been overlooked. Even though our proposed taxonomy focuses mainly on the purpose and applications of HNVC, we are aware of other transversal attributes that could further characterize the taxonomy, such as human senses or technology used. These attributes could further refine the proposed taxonomy, and potentially give insight of good practices depending on each area and sub-area of the taxonomy. We are currently working on refining our taxonomy by considering those crosscategory dimensions. Since we have focused on finding diverse applications of HNVC, there may be recent works possibly overlooked in the literature review process. Therefore, we are also currently working on an extended literature review, fitting new works that fit under our criteria for HNVC categorization. These new works will expectedly further enrich each sub-area or the taxonomy, and provide more detailed research directions and good practices within each specific area. Overall, the more insight and information about how we as humans use our haptic sense and associated HNVC, the more we will understand ourselves, improve our daily activities, be more inclusive as a society, and develop technology that feels more natural and can understand us better.

References 1. 2.

Schroeder, M ariel Lee. "A research framework for second nonverbal code acquisition." Globe: A Journal of Language, Culture and Communication 5 (2017). Hayward Vincent, Astley Oliver. R., Cruz-Hernandez M anuel, Grant Danny, and Robles-De-La-Torre Gabriel. "Haptic interfaces and devices." Sensor Review 24.1 (2004): 16-29. M ann, Steve. "Humanistic computing:" wearcomp" as a new framework and ap plication for intelligent signal processing." Proceedings of the IEEE 86.11 (1998): 21232151.

136 4.

10.

11.

12.

13.

14.

15.

M orrison, Ann, Hendrik Knoche, and Cristina M anresa-Yee. "Designing a vibrotactile language for a wearable vest." International Conference of Design, User Experience, and Usability. Springer, Cham, 2015. Haritaipan, Lalita, and Céline M ougenot. "Cross-cultural study of tactile interactions in technologically mediated communication." International Conference on CrossCultural Design. Springer, Cham, 2016. Hirano Takahiro, Shiomi M asahiro, Ilo Takasama, Kimoto M itsuhiko, Nagashio Takuya, Tanev Ivan, Shimohara Katsunori and Hagita Norihiro. "Communication cues in a human-robot touch interaction." Proceedings of the Fourth International Conference on Human Agent Interaction. ACM , 2016. Ahmed Imtiaj, Harjunen Ville, Jacicci Giulio and Hoggan Eve. "Reach out and touch me: effects of four distinct haptic technologies on affective touch in virtual reality." Proceedings of the 18th ACM International Conference on Multimodal Interaction. ACM , 2016. M azzoni, Antonella, and Nick Bryan-Kinns. "M ood glove: A haptic wearable prototype system to enhance mood music in film." Entertainment Computing 17 (2016): 917. Goedschalk, Linford, Tibor Bosse, and M arco Otte. "Get your virtual hands off me!– developing threatening IVAs using haptic feedback." Benelux Conference on Artificial Intelligence. Springer, Cham, 2017. Bosse Tibor, Hartmann Tilo, Blankendaal Romy A.M ., Dokter Nienke, Otte M arco and Goedschalk Linford. "Virtually bad: A study on virtual agents that physically threaten human beings." Proceedings of the 17th International Conference on Autonomous Agents and MultiAgent Systems. International Foundation for Autonomous Agents and M ultiagent Systems, 2018. Enriquez, M ario, Karon M acLean, and Christian Chita. "Haptic phonemes: basic building blocks of haptic communication." Proceedings of the 8th international conference on Multimodal interfaces. ACM , 2006. Chen Jennifer, Turcott Robert, Castillo Pablo, Setiawan Wahyudinata, Lau Frances, and Israr Ali. "Learning to feel words: a comparison of learning approaches to acquire haptic words." Proceedings of the 15th ACM Symposium on Applied Perception. ACM , 2018. Oliveira, Victor Adriel de J., and Anderson M aciel. "Introducing the modifier tactile pattern for vibrotactile communication." International Conference on Human Haptic Sensing and Touch Enabled Computer Applications. Springer, Berlin, Heidelberg, 2014. Schelle Kimberly Johanna, Gomez Naranjo Carolina, ten Bhömer M artijn, Tomico Oscar and Wensveen Stephan. "Tactile dialogues: Personalization of vibrotactile behavior to trigger interpersonal communication." Proceedings of the Ninth International Conference on Tangible, Embedded, and Embodied Interaction. ACM , 2015. Velázquez, Ramiro, and Edwige Pissaloux. "Constructing Tactile Languages for Situational Awareness Assistance of Visually Impaired People." Mobility of Visually Impaired People. Springer, Cham, 2018. 597-616.

137 16.

17.

18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

M acLean, Karon E., Jerome Pasquero, and Jocelyn Smith. "Building a Haptic Language: Communication through Touch." Computer Science Department, University of British Columbia TR-2005-29 (2005): 1-16. Brave, Scott, and Clifford Nass. "Emotion in human-computer interaction." The human-computer interaction handbook: fundamentals, evolving technologies and emerging applications (2003): 81-96. Robotti, Elisabetta. "Designing innovative learning activities to face difficulties in algebra of dyscalculic students: Exploiting the functionalities of AlNuSet." Digital Technologies in Designing Mathematics Education Tasks. Springer, Cham, 2017. 193-214. M agana Alejandra J., Sanchez Karla L., Shaikh Uzma A. S., Jones M . Gail, Tan Hong Z., Guayaquil Alejandro and Benes Bedrich. "Exploring multimedia principles for supporting conceptual learning of electricity and magnetism with visuohaptic simulations." Comput Educ J (2017). Kruijff Ernst, M arquardt Alexander, Trepkowski Christina, Lindeman RRobert W., Hinkenjann Andre, M aiero Jens and Riecke, Bernhard E. "On your feet!: Enhancing vection in leaning-based interfaces through multisensory stimuli." Proceedings of the 2016 Symposium on Spatial User Interaction. ACM , 2016. Feng, M i, Arindam Dey, and Robert W. Lindeman. "An initial exploration of a multisensory design space: tactile support for walking in immersive virtual environments." 2016 IEEE Symposium on 3D User Interfaces (3DUI). IEEE, 2016. Bernardet Ulysses, Adhia Dhruv, Jaffe Norman, Wang Johnty, Nixon M ichael, Alemi Omid, Phillips Jordon, DiPaola Steve, Pasquier Philippe and Schiphorst, T. . "m+ m: A novel middleware for distributed, movement based interactive multimedia systems." Proceedings of the 3rd International Symposium on Movement and Computing. ACM , 2016. Saunders, William, and Daniel Vogel. "The performance of indirect foot pointing using discrete taps and kicks while standing." Proceedings of the 41st Graphics Interface Conference. Canadian Information Processing Society, 2015. Costes Antoine, Argelaguet Ferran, Danieau Fabien, Guillotel Philippe, and Lécuyer Anatole. "Touchy: A visual approach for simulating haptic effects on touchscreens." Frontiers in ICT 6 (2019): 1. Houde M . S., Landry S. P., Pagé S., M aheu M ., & Champoux F. "Body perception and action following deafness." Neural Plasticity 2016 (2016). Bălan Oana, M oldoveanu Alin, M oldoveanu Florica and Butean Alexandru. "Auditory and Haptic Spatial Cognitive Representation in the Case of the Visually Impaired People." The 22nd International Congress on Sound and Vibration, Florence, Italy. 2015. Saunders, William, and Daniel Vogel. "The performance of indirect foot pointing using discrete taps and kicks while standing." Proceedings of the 41st Graphics Interface Conference. Canadian Information Processing Society, 2015. Bufalo Francesco, Olivari M ario, Geluardi Stefano, Gerboni Carlo A., Pollini Lorenzo, and Bülthoff, Heinrich H. "Variable force-stiffness haptic feedback for learning a disturbance rejection task." 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC). IEEE, 2017.

138 29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41.

42.

D'Intino Giulia, Olivari M ario, Geluardi Stefano, Venrooij Joost, Innocenti M ario, Bülthoff Heinrich. H., & Pollini Lorenzo. "Evaluation of haptic support system for training purposes in a tracking task." 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC). IEEE, 2016. van Oosterhout Jeroen, Wildenbeest Jeroen G., Boessenkool Henri, Heemskerk Cock J. M ., de Baar M arco R., van der Helm Frans C., and Abbink David A. "Haptic shared control in tele-manipulation: Effects of inaccuracies in guidance on task execution." IEEE transactions on haptics 8.2 (2015): 164-175. Xu Yanbin, Yang Chenguang, Liang Peidong, Zhao Lijun, & Li Zhijun. "Development of a hybrid motion capture method using M YO armband with application to teleoperation." 2016 IEEE International Conference on M echatronics and Automation. IEEE, 2016. Yokokohji Yasuyoshi, Hollis Ralph L., Kanade Takeo, Henmi Kazuyuki, and Yoshikawa Tsuneo. "Toward machine mediated training of motor skills. Skill transfer from human to human via virtual environment." Proceedings 5th IEEE International Workshop on Robot and Human Communication. RO-MAN'96 TSUKUBA. IEEE, 1996. Schirmer M aximilian, Hartmann Johannes, Bertel Sven and Echtler, Florian. "Shoe me the way: a shoe-based tactile interface for eyes-free urban navigation." Proceedings of the 17th International Conference on Human-Computer Interaction with Mobile Devices and Services. ACM , 2015. Prasad M anoj, Taele Paul, Olubeko Ayo, and Hammond Tracy. "HaptiGo: A navigational ‘tap on the shoulder’." 2014 IEEE Haptics Symposium (HAPTICS). IEEE, 2014. Yang Chenguang, Liang Peidong, Ajoudani Arash, Li Zhijun and Bicchi Antonio. "Development of a robotic teaching interface for human to human skill transfer." 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2016. Pedemonte, Nicolò, Thierry Laliberté, and Clément Gosselin. "Bidirectional Haptic Communication: Application to the Teaching and Improvement of Handwriting Capabilities." Machines 4.1 (2016): 6. Choi, Kup-Sze, Sze-Ho Chan, and Wai-M an Pang. "Virtual suturing simulation based on commodity physics engine for medical learning." Journal of medical systems 36.3 (2012): 1781-1793. Brewster, Stephen, and Lorna M . Brown. "Tactons: structured tactile messages for non-visual information display." Proceedings of the fifth conference on Australasian user interface-Volume 28. Australian Computer Society, Inc., 2004. Danjoux, M ichèle. "Choreography and sounding wearables." Scene 2.1-2 (2014): 197-220. Brown, Courtney, and Garth Paine. "Interactive Tango M ilonga: designing internal experience." Proceedings of the 2nd International Workshop on Movement and Computing. ACM , 2015. Giomi, Andrea. "From haptic perception to movement: a phenomenological study on sound and technological mediation in dance pedagogy" Repertório 28 (2017): 342361. Hossny M ., Nahavandi, S., Fielding, M ., M ullins, J., M ohamed, S., Creighton, D., M cCormick J., Vincs K., Vincent J., and Hutchison S. "Haptically-enabled dance vis-

139

43.

44.

45.

46.

ualisation framework for deafblind-folded audience and artists." 2015 IEEE International Conference on Systems, Man, and Cybernetics. IEEE, 2015. Shibasaki, M ina, Youichi Kamiyama, and Kouta M inamizawa. "Designing a haptic feedback system for hearing-impaired to experience tap dance." Proceedings of the 29th Annual Symposium on User Interface Software and Technology. ACM , 2016. Florian Horațiu, M ocanu Adrian, Vlasin Cristian, M achado José, Carvalho Vítor, Soares Filomena, Astilean Adina and Avram Camelia. "Deaf people feeling music rhythm by using a sensing and actuating device." Sensors and Actuators A: Physical 267 (2017): 431-442. Oliveira, Victor Adriel de J., and Anderson M aciel. "Assessment of tactile languages as navigation aid in 3d environments." International Conference on Human Haptic Sensing and Touch Enabled Computer Applications. Springer, Berlin, Heidelberg, 2014. Scheggi, S., A. Talarico, and D. Prattichizzo. "A remote guidance system for blind and visually impaired people via vibrotactile haptic feedback." 22nd Mediterranean Conference on Control and Automation. IEEE, 2014.

SECTION 3 Educational Research

Educational research is an activity that requires continuous reflection and questioning, in which teachers and researchers from different disciplines participate considering behaviors, customs and environment, among many other variables. Its results have led to the knowledge of new learning strategies and the development of more useful and convenient models in this XXI century. The accelerated development of knowledge through the New Information and Communication Technologies (ICT), represents the most important educational impact of recent times. At present, thanks to the presence of these technologies, students have access to other information systems, some of them more updated in the content of the subject being taught, than the teacher himself, a situation that makes the teaching update in this area necessary and cannot be postponed, to be incorporated into the knowledge society. The TICs allow information to a greater number of people and allow the optimization of new communicative and formative possibilities, the incorporation of other procedures of access to the information or the integration of new learning contexts that facilitate the flexibility of the Education, adapting to the characteristics and needs of the students. This book describes research and educational actions supported by ICT as support tools that allow improving the work of the teacher and learning with the student. The works shown in this section are: an investigation that presents a comparison of some of the learning characteristics of Mexican and Colombian students, a serious game that contributes to interpreting a student's personality traits to favor the process of choosing a career professional; a design and development of a cognitive training environment to measure working memory; the construction and validation of an implement to measure the digital competence of a teacher educator at the Telesecundaria level; identification of teenagers' motivations as digital natives using a focus group and a semi-structured interview; design, development and partial evaluation of ARTUI, an application based on Augmented Reality to support English language teaching; a study to know the effectiveness of a sequence of activities designed to develop skills in the search for reliable information on the Internet in university students; the design of a formal protocol for the creation of the database of physiological and behavioral signals of university students while doing a learning activity and the presentation of a collaborative content production model to generate content in order to mitigate the gap digital.

Computer Science Students’ Profile: A comparison between students from Mexico and Colombia Guillermina Sánchez Román, Josefina Guerrero García, Erika Anabel Martínez Mirón 1 , Mariano Larios Gómez, Adriana Hernández Beristain, María Daniel Trinidad Meneses Benemérita Universidad Autónoma de Puebla, 14 sur y Av. San Claudio, Col. San M anuel, M éxico guille.sroman@correo.buap.mx, jguerrero@cs.buap.mx, erika.mtzm@correo.buap.mx, mlg_y@yahoo.com, adrianah.beristain@correo.buap.mx, dtrinidadmeneses@gmail.com

Abstract. For several decades, some deficiencies in programming learning have been noted, especially in the development of algorithmic skills. It is important to contemplate the way in which the student learns on his own to improve the identified weaknesses. This investigation presents a comparison of some of the learning characteristics (students’ previous knowledge, learning style and their reasoning level) between M exican and Colombian students. The complete research aims to create a prototype of Intelligent Tutor System (ITS) to develop algorithmic skills. The methodology used is quantitative, different tests were applied to assess learning styles, as well as scientific reasoning and computational thinking. The sample was integrated by new students entering the degree in computing science in M exico and Colombia. The data obtained from the computational thinking test applied to students was compared to determine the trends of their algorithmic skills. The statistical process is from non - inferential statistics, the results indicate that, in both regions, students present similar tendencies in their learning style, reasoning level and computational thinking. This result contributes with the characterization for the student’s model and the design of activities for the ITS prototype. Keywords: Algorithmic skills, learning style, reasoning, higher education, students’ characteristics.

Introduction

Nowadays, it is important to integrate the Information and Communication Technologies (ICT) that are being generated into education. The science area has been characterized by having little demand and, at the same time, greater difficulty for their learning. In particular, in the area of computer science students have difficulties in the initial courses of programming, which results in high failure rates [1][2]. Studies show that the algorithmic skills of university studen ts who enter careers in Computer Science field are still deficient. In addition, there is an absence of mastery

142

of the cognitive and instrumental aspects related to these skills, both for teaching and evaluation [1][2]. For the integration of teaching-learning strategies it has been considered the gamification to encourage learning and motivation. Gamification is defined as a process of using thought and game mechanics to solve problems and increasing motivation and commitment [3][4]. The algorithmic skills in basic programming will be subsequently integrated into the design of the Intelligent Tutor System (ITS). From the inference of students’ profiles each student will be in a category according to their cognitive style and prior knowledge [1]. The ITS will generate a set of pedagogical and didactic strategies that will customize the learning process according to the particular needs of each student, hoping to improve their algorithmic skills for problem solving, preventing dropouts [2]. In the student module, it is proposed to contemplate the characteristics of the student from three domains: reasoning, learning style and algorithmic abilities. Integrating the features in the learning processes to design recreational activities and according to the skills required for programming. The objective of this research is to characterize Computation area students in Mexico and Colombia from three aspects: a) the algorithmic skills; b) the learning style; c) and the level of reasoning that computer students hav e. A small sample is used to compare the results found in the groups explored in the two regions. Consequently, it is tried to analyze what the skills that must be developed are in the solution of problems and contrasting the data obtained.

Theoretical background (State of the art)

According to the cognitive approach and the instructional theory of Gagné, it is necessary to know how the student obtains previous knowledge, how he acquires representations of the world, how they store and retrieve the data in memory. The model to determine the learning style that gives more information in the science are a according to the literature [5] [6] and has given better results to identify the preferred style, is the one proposed by Felder and Silverman, for what is con sidered This model to evaluate students of computer science to identify how they learn. The instrument of learning styles is applied [7] and evaluates how the student processes, perceives, represents and understands information, and defines four dichotomou s pairs to identify: active-reflexive, sensory-intuitive, visual- verbal, sequential-global that allow to know the predominance of their learning style. Algorithmic thinking is, in a way, a set of skills that are connected to the construction and understanding of algorithms. According to [8], this thinking includes the following capabilities: a) Analyze given problems. b) Specify a problem accurately. c) You will find the basic actions that are adequate to solve the given problem. d) build a correct algorithm to solve a given problem, using the basic actions. e) Think about all possible special and normal cases of a problem.

143

f) Improve the efficiency of an algorithm. This analysis, prior to working with the Scratch programming environment, implies a cognitive activity that involves planning, hypothesis formulation, abstraction, linguistic comprehension; that is, the same skills identified in studies on computational thinking [9], [10]. This implies the ability to clearly define and enunciate a problem; break it down into smaller, more manageable subproblems, and describe a solution to achieve in a well-defined set of steps. The ability to analyze and give a solution to problems, an important skill for every human being, and especially for students [11]. It is fundamental to review the process of the student's cognitive activity, which involves planning, the formulation of hypotheses, abstraction, understanding of the approaches and implementations in a technical language. Therefore, algorithmic abilities are defined based on Bloom's taxonomy in the two dimensions of problem solving and abstraction capacity.

Methodology

A study following the quantitative approach was carried out, the analysis process was developed from non-statistical inference. It is intended to know the characterization of the student that will serve as information to the ITS to create the design of the learning activities to be customized according to the student's profile and develop the skills that need to be reinforced. Sample The study population consisted of new students entering Computer related careers , a global group of 48 students, 28 students from the Faculty of Computer Science (FCC) of the Benemérita Universidad Autónoma de Puebla (BUAP) in Mexico and 20 students from the Faculty of Electronics Engineering and Telecommunications (FIET) from the career Systems Engineering at the University of Cauca (UniCauca) in Colombia. The sample chosen was selected due to both groups were already constituted. Table 1 gathers the information of the students that participated in the study. Their age was between 18 and 20 years old, 81 % of the population were men and 19% were women, approximately. Table 1. Data obtained from M exico-Colombia students (Source: own) Ages M éxico Colombia 18-20 years 25 18 20-23 years 3 2 First course 8 16 Second course 16 4 Women 15% 19% M en 85% 81% Students’ total 28 20

144

Data collection instruments In this research, three instruments were applied: a) a learning style test [7], b) a reasoning test [13] and c) a computational thinking test [12]. The first two tests are already validated and the computational thinking test is in the process of being validated. An adaptation is being made according to the needs of th is research. Below is one of the test questions with the option to choose one of the answers given: When someone gives me the address of a new place, I prefer: a) a map. b) written instructions. The scientific reasoning test consists of 12 questions; each question o wns four possible answers. They are evaluated according to three concrete categories where they must have already been developed empirical-inductive, transition and formal thinking assuming that this thought is already hypothetic-deductive [14]. The type of question is as follows: Six square pieces of wood are placed in a dark cloth bag and mixed. The six pieces are identical in size and shape, but three pieces are red and the rest are yellow. Suppose someone puts their hand inside the bag and without looki ng take out one of the pieces. What is the probability that the extracted piece is red? a) 1/6 (1 chance every 6 events). b) 1/3 (1 chance every 3 events). c) 1/2 (1 possibility every 2 events). d) 1 (1 possibility for each event). e) It cannot be determined with the given information

The proposed Computational Thinking test was adapted in terms of dimensions and algorithmic skills that will be evaluated in agreement with Bloom's taxonomy.

Abstraction Capacity

Table 2. Dimensions to be evaluated of algorithmic skills based on Bloom's taxonomy (Source: Own). Skills

Algorithm development tasks Recognize the syntax from a command. Recognize the semantics from a command

Remember

Interpre t what changes occurred when modifying a command. Translate a given algorithm to pseudocode or flowchart.

Comprehend

Use structures with a clear and concise objective (decisions, cycles). Use routines to develop a specific task. Modify an algorithm.

Abstract

Problem Solving Capacity

145

Predict the possible outcomes when executing an algorithm line by line. Detect possible errors in an algorithm. Evaluate the requirements at a user level, system, organization and develop a coherent algorithm that takes them into account. Compare algorithm solutions that solve the same problem. Verify the set of instructions through desks’ test. Generate a complete algorithm based on the specifications of a problem. Generalize an algorithm.

Analyze

T est

Create

It is proposed to use the Computational Thinking Test [12] to identify the level of algorithmic skills (table 2) about their prior knowledge. The test consists of 28 graphic questions to answer from four options that the student must select. Three categories are posed; rookie, apprentice and expert, for the evaluation according to the number of questions in each category from Bloom’s taxonomy, thus, it is also identified in which skill exactly the student needs to reinforce his learning. The type of questions is as follows (Figure 1): What orders take pac-man to the ghost by the designated path A: Repeat until you reach ...

B: Repeat until you reach ...

C: Repeat until you reach ...

D: Repeat until you reach ...

Fig. 1. Question of the Computational Thinking Test [12]

Design of the study The application of the three instruments was carried out in 2 sessions of 45 minutes each. In one session, the learning style test was applied and the scientific reasoning test was applied in another session. The computational thinking test was applied in the laboratory through an online form.

Results of the comparative study

Below are detailed the results obtained from the learning style test, the reasoning test and the data collected from the computational thinking test. Comparing the data found provides an overview related to the students’ characteristics in college.

146

The Felder and Silverman test is applied online to the two groups according to the four dichotomous pairs: active-reflexive, sensory-intuitive, visual-verbal, sequentialglobal as shown in Figure 2, the group FCC (Mexico) tends to be active, sensorial, visual and global, on the other hand the group FIET (Colombia) has a greater tendency to be active, sensory, visual and sequential. The coincidence of both groups in the way of learning is mostly active, sensory and visual. The information obtained supports the material design integrate the Tutor System that will define the model to deploy the most appropriate materials in concordance to the predominant style in which the student processes, manages and retrieves information.

Fig. 2. Results obtained from the learning style test (Source: own)

The reasoning test that was applied to a group of students’ part of the FCC (Mexico), 73% of the students is in transition, 8% in the concrete level and 9% of them in the form level Table 3). The results gathered from the FIET group (Colombia) are shown in figure 2, where the data possesses similar information such as: 10% in formal level, 20% in concrete level and 70% in transition. Therefore, students need to develop their concrete level or transitional level and move to the formal level. This level is characterized by the fact that students can abstract information deductively from problems raised. Table 3. Comparison Reasoning Level (Source: own).

Reasoning Test

México

Colombia

Concrete Transition Formal

18% 73% 9%

20% 70% 10%

After applying the instrument of Computational Thinking to the FCC students (Mexico), observe Table 4. The first three skills: remembering, understanding and abstracting refer to the level that students possess, this means that a greater percentage correspond to an expert level, a lower percentage is find at a novice level; however, in the last two skills regarding analyzing and testing the expert percentage decreases considerably and the apprentice percentage increase.

147

Table. 4. Computational thinking in FCC students (Source: own).

Computational Thinking Test Remember Understand Abstract Prove

Expert M éxico

40.9 77 0 4.5

Apprentice

Colombia

70 80 55 10

M éxico

45 23 82 50

Colombia

20 15 45 55

Novices M éxico

9 0 18 45

Colombia

10 5 0 35

In the case of Colombian students, they have a greater mastery in the skills of remembering, understanding and abstracting, and they also decrease the percentage in the skills of analyzing and testing at the expert level, however, the apprentice le vel considerably increases. The results obtained from the application of the tests for the characterization indicate that the three instruments that are proposed give us information about the student's disposition in learning and to develop their algorithmic abilities. There are several modeling jobs STI n the student, but none taking into consideration the scientific reasoning and algorithmic skills at the initial level. This work proposes to outline the features that can model the student identifies the level that the new student has to adopt strategies and learning activities that integrate into the tutor system to give a personal and independent monitoring agreement to the progress of the student.

Conclusions

The data presented in graphs pour how students in the two regions of Latin America define the learning style in the area of Computing with greater predominance to be: active, sensory, visual and sequential; In the case of scientific reasoning, both groups showed similar percentages centered mainly on the level of transition and which defines that reaffirming abilities in new students must be reaffirmed. Finally, in the Computational Thinking test students' tendencies for skills: remembering, understanding and abstracting, the students in both groups define an expert level, although in the skills of Higher categories (analyze and test) of Bloom's taxonomy students are still at apprentice level. Therefore, it is considered that the results found in both countries are similar in terms of the type of students they have and their shortcomings in the domain of algorithmic skills necessary for the area of Computer Science. The data pay to define the general characterization of new students to the area of Computing and the need to diagnose them to follow up through a software tool that supports their lea rning autonomously and individually as they are the STI. The research pays the educational area punctually to the area of instruction in the learning of basic programming skills.

148

References 1.

2. 3. 4. 5.

7. 8. 9. 10.

11. 12.

M oroni, N., & Señas, P. Estrategias para la enseñanza de la programación. Primeras Jornadas de Educación en Informática Y TICS en Argentina, pp. 254–258. Bahia Blanca, Argentina, (2005). Rosanigo, Z., & Paur, A., Estrategias para la enseñanza de Algorítmica y Programación. TE&ET’06, pp. 117–124, (2006). Orejuela, H. F., García, A. A., Hurtado, J. A., & Collazos, C., Analizando y Aplicando la Gamificación en el Proceso ChildProgramming, 7–23, (2013). Aparicio, F. A., Gutierérrez, V. L. F., González, J. L., & Isla, M . J. L. (2012). M étodo de análisis y aplicación de la gamificación, (2012) Franzoni-Velázquez, A.L., Cervantes-Pérez, F., Assar, S., A Quantitative Analysis of Student Learning Styles and Teacher, Teachings Strategies in a M exican Higher Education Institution, Journal of Applied Research and Technology, Vol. 10 No.3, pag. 289-308 (2012) Felder M ., Silverman K., Learning and Teaching Styles in Engineering Education, Institute for the Study of Advanced Development, Journal Enginner Education, 78(7), 674–681 (1988). Felder, R. M ., & Silverman, L. K., Learning and Teaching Styles. Engr. Education, 78(7), 674–681, (1988). Futschek, G., Algorithmic Thinking: The Key for Understanding Computer Science. ISSEP (2006), 159 – 168. Wing, J., Computational Thinking. Communications of the ACM 49 (3), 33- 35, (2006). ISTE, International Society for Technology in Education, Operational definition for computational thinking for K-12 education. (2011). Recuperado el 18 de marzo 2015, de http://csta.acm.org/Curriculum/sub/CurrFiles/CompThinkingFlyer.pdf Polya, G., "Como plantear y Resolver problemas", Trillas, 1965. Román-González, M ., Pérez-González, J. C., & Jiménez-Fernández, C., Test de Pensamiento Computacional : diseño y psicometría general. In III Congreso Internacional sobre Aprendizaje, Innovación y Competitividad (CINAIC 2015) . M adrid, España. http://doi.org/10.13140/RG.2.1.3056.5521, (2015).

13.

Lawson, A. E. Science teaching and the development of thinking. Belmont, CA: Wadsworth. pp. 6-19, (1995).

14.

Rodríguez, M . D., M ena, D. A., & Rubio, C. M ., Razonamiento Científico y Conocimientos Conceptuales de M ecánica: Un Diagnóstico de Alumnos de Primer Ingreso a Licenciaturas en Ingeniería. Formación Universitaria, 3(5), 37–46. (2010), http://doi.org/10.4067/S0718-50062010000500006.

Conócete: Serious game for self-knowledge in the process of Vocational Guidance Tatiana de Jesús Olvera Pablos 1 , María Teresa Fernández Nistal1 , Sonia Verónica Mortis Lozoya 1 , David Bonilla 2 , Adriana Peña Pérez Negrón 2 , Emmanuel Guardiola 3 1

Instituto Tecnológico de Sonora, 5 de Febrero #818 Sur, Col. Centro, C.P. 85000, Cd. Obregón, Sonora, M éxico tatiana2105@gmail.com, https://orcid.org/0000-0002-9564-2537 mfernandez@itson.edu.mx smortis@itson.edu.mx 2 Universidad de Guadalajara CUCEI, Blvd. M arcelino García Barragán #1421, C.P. 44430 Guadalajara, Jalisco, M éxico adriana.pena@cucei.udg.mx jose.bcarranza@academicos.udg.mx 3 Cologne Game Lab TH-Köln, Cologne, Germany eg@colognegamelab.de

Abstract. Deciding a career is not always a clear strait forward path for all students; a current tool to help them is the use of questionnaires. Recently, with the study of new technologies, the use of other instrument such as serious games has been introduced to support or substitute such questionnaires. In this paper we present Conócete a serious game that contributes to interpret the personality traits of a student in order to favor the choosing process for a professional career by supporting self-knowledge in the area of vocational guidance. The game design is based on the theory of vocational personalities and work environments of Holland in 1992. The developed is a web application with responsive design that handles 2D and 3D graphics for mobile devices. Keywords. Serious game, Educational games, Holland’s Theory of Vocational Personalities, RIASEC personality and work environment types, Scrum

Introduction

Videogames began to be used in Mexico in the late 70s, causing that nine out of ten young people interacted with them, permeating it use uniformly across all socioeconomic strata [1]. In Mexico in 2012, there were 16 million users in the country who spent half-time 1.8 hours a day to play [2], and in 2015, the number of players increased to 64.8 million, which represents more than half of the Mexican population [3]. In addition, the number of mobile devices (smartphones and tablets), used for digital entertainment by people of all ages, increased to more than two hours a day in 59% of users [3]. Currently, digital games have uses other than for entertainment [4, 5], representing a learning model with multiple advantages compared to actual learning models [6].

150

The implementation of videogames in the educational field is a strategy that has being more evident every day, regardless of the level of education [7]. Research has indicated that self-analysis through the game helps users to become more aware of their own processes of thought, sensory perceptions, values, as well as encourag ing active and critical learning [5, 8]. It also has been found that while people are playing, their hearing, smell, touch and head are excited, creating true learning [9]. The inclusion of Information and Communication Technology (ICT) applications in career and guidance field were developed mid 60s for example: email, chat, newsgroup, websites, software and video-conferencing. These technologies are used for career information, virtual experiences, clients understand themselves and their situations and clients access their networks of support and make moves. The implementation of ICT in career guidance generates new expectations, enriches the quality of school life, contributes to the development of career management skills, and facilitates the integration of the student in the labor and social world [10]. Another technological tool used in career guidance are serious games, they must be designed with clear objectives, so that their results are seriously considered as references for career planning [5]. Serious games in this area, represent an effective solution because they allow users knowing what a career will be like, the work environment, and help them to understand the subjects to master for the necessary vocational skills to carry out some work, which should influence the career options of the students [9]. Likewise, this type of technology improves motivation and interest in the search for professional guidance in the exploration of an adequate professional career [5], allowing students to explore and develop by their own in a virtual situation where their decisions do not directly affect their achievements and happiness. Nevertheless, the technological tools should be used as a complement to the face-to-face attention of a counselor [5, 1114].

Related work

Serious games are based on virtual environments that combines playful and educational objectives [4, 9, 15-16]. Serious games are used in areas such as education, training, health and government [4, 16], allowing players to experience simulated environments that are impossible in the real world due to security, cost, time, etc. [16]. They are used to support a specific area, contribute to the development of skills, strategies, as well as provoke self-awareness, thinking and learning in an effective and innovative way [4, 5, 17]. The theory of vocational personalities and work environments of Holland [18], has the aim of interpreting vocational behavior to help people for vocational choice, changes in employment, vocational achievement, and professional satisfaction. It is based on the principle that vocational alternative is the expression of the personality traits, which gives as a result, that a correct occupation corresponds to the personality of the individual. In addition, there is a relationship between the types of personality and work environments [19]. This approach sustains that for most people there is a relationship to one of six dimensions: Realistic (R), Researcher (I), Artistic (A), Social (S), Entrepreneur (E) and Conventional (C); where, each type of personality is distinguished by preferred: activities, competencies, interests, objectives, values of

151

life, self-beliefs, abilities, and personality traits [18, 20]. One of the major contributions of Holland's theory to the process of vocational guidance is the quantity of instruments elaborated and validated to determine the personality and work environments, as well as the manage of diverse variables (aspirations, expectations, family histories, etc.). There are more than 500 articles, books, chapters and reviews that analyze the functioning of Holland's theory, using experimental tests, its implementation as well as the organization of personal and occupational data [19, 20]. We propose that the use of this basic theory can be implemented in the development of a serious game for vocational guidance [5], representing a more interesting option when compared to the filling of questionnaires, since this modality can be tedious, unattractive for most of the students, and in some cases even with nonsense [5, 19]. Questionnaires have been found to present a number of circumstances that might create bias from the respondent, for example the fact that the res pondent are measuring themselves, the fact that people might try to maintain consistency between their cognition and attitudes, some people need for social approval or acceptance, along with the respondents’ mood state all of which might influence response s [21].

Game Design

The objective of this serious game is to contribute to the self-knowledge and understanding of the type of vocational personality of the user, in order to support the process of choosing a career. The design will follow the methodology proposed by [22] applied on the French game Jeu Serai which aims to capture the psychological profile of the user using the theory of Holland as an example [22]. The game is made of many mini games connected to RIASEC dimensions, each time the player play s one mini game he/she as to rate it with a typical number of stars. And the experiment following the game show that there is a correlation between the best ranked game by the player and their best dimension score on a questionnaire. The development of the digital educational resource is being carried out in collaboration between the Instituto Tecnológico de Sonora (ITSON), the Universidad de Guadalajara (UdeG) and the support of the professor Guardiola that he works for the Cologne Game Lab. The serious game uses a simple and easy-to-understand narrative so that users can perform the tasks presented on it without any especial background. For a successful design, the most of the 11 macrodesign factors of Shi and Shih (2015) were considered: 1) game goals are the foundation of the game; 2) game mechanism are rules and feedback to achieve enjoyable gameplay; 3) interaction are the actions between the player and the game; 4) freedom refers to how many actions and services the players can perform and use; 5) game fantasy considers the environment and background, this element considers the player´s immersion; 6) narrative promote to it telling a story; 7) sensation its refers to the virtual world include the audio and aesthetics; 8) game value is the factor to development motivation and immersion; 9) challenge it’s a factor that motivate people to action; 10) sociality considered the social behavior like communication, cooperation, competition and conflict; 11) mystery this factor involves player curiosity and exploration [23]. The game will start with the opportunity for the user to configure one avatar "your digital self" (see Figure 1 and 2), with the choices of sex, skin tone, costumes,

152

among other features. The story will be based on the development of 18 missions that will lead the user to perform activities, and subsequently indicating the degree of satisfaction while executing them and the level of ability. In such a way that by performing the missions the player can get a better knowledge about the various areas in which he or she can start his/her professional life (see Table 1). The user must perform some missions to reach level three, while accumulating points that can be exchanged for digital material goods (properties, cars, jewelry, books, etc.), allowing to evaluate their tastes and preferences. At the end of the game, the users will be able to see the most suited professions according to their personality traits. The result will be based on the player decisions made during the game. Table 1. Activities of the RIASEC dimensions of serious game. Dimension

Activities “Mechanical repair” It presents a car in poor condition, as well as 5 tools to enable it (gasoline gun, mechanical keys, tire inflator, welding machine and sponge). The user should choose the appropriate tool for each activity and do it. Once the mission is over, the game will allow another similar activity if he/she wants (see Figure 3). “Basic chemistry” The mission consists in carrying out three basic chemistry tests, in each test the name of a substance and four test tubes containing elements that can create the substance are presented. The user must place the correct elements in the Erlenmeyer flask in order to complete the mission (see Figure 4).

Skills

“Assemble the furniture” The user will be shown a dismantled piece of furniture and the mission ends when the user correctly places all the pieces.

“Repair the circuit” The mission consists of the development of two basic electricity practices, in which the user must fix the circuit and allow the passage of electricity through the correct route.

“Designing a t-shirt” The user must design a t-shirt by first se“Photo Studio” lecting the gender of their preference (male The user must take 10 photos or female), then they must choose a backof the objects that appear in ground color (white, black, y ellow, red, the room in which he/she is to blue, pink, gray or green) and finally make later select 5 of them and add strokes with the colors mentioned them to an album. (see Figure 5). “Listening to the patient” A dialogue will be played between the patient and a therapist. Then the user should answer some questions related to the patient's problem.

153

Dimension

Activities

Skills

“Leading a team” Three electric bands are shown that carry eggs and, in each band, there are workers that the leader must direct to collect the largest number of eggs in the shortest time and avoid breaking the eggs. "Sorting the stamps" There will be two sections each with a theme (transport and animals) and in a separate space will appear pictures with objects related to the topics, so that the user will have to classify each stamp correctly to end the mission.

Fig. 3. M echanical repair.

Fig. 1. M an avatar.

Fig. 2. Woman avatar. Fig. 4. Basic chemistry.

Fig. 5. Designing a t-shirt.

For the design of the virtual world, we are looking for design elements with characteristics of the national environment, such as: vegetation, animals, houses, b uildings, tastes, culture, etc., with the purpose of establishing a bond of identity between the users and the serious game. In addition, we look to create a tool that considers gender equity, avoiding violence and racism. For the development, the Scrum method will be applied, this is one of the agile methods preferred for the software development of at present [24, 25]; with the aim,

154

to optimize the process by identifying tasks, administering time more effectively and establishing teams [24]. Regarding particularly on this project development, it follows an interactive incremental design with the stages: analysis, design, development, implementation, validation and maintenance [26] (see Figure 6). In the initial investigation stage, the requirements are obtained. In our case since the design is based on a previous project composed of mini games, we make reverse engineering when t he mini game is part of that project, or interviews when the mission is new or different. A technical brainstorm is applied for the analysis and a list of parallel and sequential activities are determined to fulfill the requirements. Then the workflow and the selection of platforms are settled along with the project architecture in the design stage. In the development stage the software engineers get roles according to their skills and their activities are as follows [27]: The game manager role: is in charge of the project and assigns task to the other roles. The game manager also coordinates the workflow of the project. The game designer role: defines the game mechanics, narratives, sketches, interactive recreational elements that are the basis for the game artist and the game programmer. The game artist role: is in charge of developing the artistic part, visual animation and sound according to the proposal. The game programmer role: is in charge of developing logic by programming the mechanics and incorporating them with the artistic part. In the application, the functional versions go through a testing process. In the testing process the software is verified and validated. In the case that some problems arise at this stage, the project goes to the stage where the problem can be resolved.

Fig. 6. Software development model for video games.

155

Conclusions

The development and implementation of a serious game based on the theory of Holland in 1992, will allow us to determine if it represents an effective, motiv ating, attractive and relaxing pedagogical tool to contribute in the decision -making regarding vocation in students. This project represents an opportunity to generate modern teaching-learning mechanisms taking advantage of the potential of technology. This alternative is sought to attend the tedious, unattractive and sometimes predictive use of questionnaires, generating an option aimed to reduce the biased responses and having the opportunity to obtain a more authentic verdict, based on the possibility of submerging the user in an animated virtual scenario that allows them to forget that they are making an assessment for vocational decisions. We consider that it is necessary to somehow improve the process of making vocational decisions with tools that stimulate the understanding of the students’ personality, aptitudes and values [5], so that they can count on elements to choose with better probabilities of success. There is evidence that shows an association between dissatisfaction with vocational decision and the lack of information about oneself, the workplace and the lack of skills to achieve both school and work success [28]. Once the serious game has been validated, it will be necessary to take the needed steps so that it can be included in the vocation al orientation program of the schools, considering that teachers will have to lead the playful experience towards the application of what has been learned [29]. Serious game can be incorporated into the curricula and make the issue of vocational decision more attractive for students. In this way, this technology can be investigated in the future. Subsequently, it is necessary to work on adjustments to ensure that this serious game complies with a more inclusive education, considering design updates for pe ople with visual disabilities, as well as cognitive, academic, learning styles, for example [4].

Future work To analyze the validity of the serious game, the Game Design Methodology pro posed by Guardiola and Natkin [22] will be used. The questionnaire of Self-Directed Search (SDS) - Form R of John Holland will be applied to 80 high school students. This instrument has been adapted, validated and standardized for the Mexican population [19]. Students must have an age of 16 to 18 years and have taken the subject of vocational guidance. Once the player finishes filling the questionnaire, the same students will be asked to interact with the serious game until the end and obtain the result. Finally, the result of both tools will be related to determine the serious game validity.

156

References 1. González, H.: Veinticinco años de videojuegos en M éxico. Las mercancías tecnoculturales y la globalización económica. Comunicación y Sociedad. 38, 103-126 (2000) page used 104,105,117 2. Barrientos-Gutiérrez, T., Barrientos-Gutiérrez, I., Reynales-Shigematsu, L., Thrasher, J., Lazcano-Ponce, E.: Se busca mercado adolescente: internet y videojuegos, las nuevas estrategias de la industria tabacalera. Salud Pública de M éxico. 54, 3, 303-314 (2012) page used 307 3. Vértiz, C.: El 65% de los mexicanos son usuarios de videojuegos (2016). https://www.proceso.com.mx/427004/el-65-de-los-mexicanos-son-usuarios-devideojuegos. Accessed 06 Jan 2019 4. Cano, S., Arteaga, J., Collazos, C., Gonzalez, C., Zapata, S.: Towards a M ethodology for Serious Games Design for Children with Auditory Impairments. IEEE Latin America Transactions. 14, 5, 2511-2521 (2016) page used 2511 5. Shi, Y., Shih, J.: Game-Based Career Guidance Systems Design Concept. In 2012 IEEE Fourth International Conference On Digital Game And Intelligent Toy Enhanced Learning, Takamatsu, Japan (2012). https://doi.org/10.1109/DIGITEL.2012.53 6. M ontiel-García, D., Cruz-Gómez, D., Santoyo-Rivera, N., Palatto-M erino, N., M anjarrez-Estrada, E.: El potencial de los videojuegos como elementos del aprendizaje para los métodos futuros de enseñanza (online). Revista de Educación en Ingeniería. 13, 26, 42-46 (2018) 7. M arín-Díaz, V., M orales-Díaz, M ., Reche-Urbano, E.: Educational Possibilities of Video Games in the Primary Education Stage According to Teachers in Training. A Case Study. Journal of new Approaches in Educational Research. 8, 1, 42-49 (2019) 8. Altamirano, J.: ¿Esto es un juego? Juegos serios y gamificación, taxonomía y aplicación. Universidad Politécnica de Valencia (2016) 9. He, L., Hu, X., Wei, D.: The case analysis of serious game in Community Vocational Education. In Proceedings of 2011 International Conference on Computer Science and Network Technology, Harbin, China (2011). https://doi.org/10.1109/ICCSNT.2011.6182333 10. Scoda, A., Andrei, A.: Are virtual career guidance centers needed in the Romanian schools?. In The 12th International Conference eLearning and Software for Education, Bucharest, Rumania (2016) 11. Cruz, F., González, C., Cadavid, J.: Diseño preliminar de un juego serio para orientación vocacional a partir de la teoría tipológica de Holland Preliminary design of a serious game for vocational guidance based on Holland’s typology theory. In 2016 XI Latin American Conference on Learning Objects and Technology (LACLO), San Carlos, Costa Rica (2016). https://doi.org/10.1109/LACLO.2016.7751759 12. Gómez, Á.: Una WebQuest para la orientación vocacional y profesional en Bachillerato. Revista Científica de Educomunicación. 16, 32, 215–221 (2009). https://doi.org/10.3916/c32-2009-03-003 13. Héctor, E., M illet, B., Ruiz, L.: Orientación vocacional a través de las TIC: ¿es suficiente?. In XVIII Congreso Internacional EDUTEC “Educación y Tecnología desde

157

una visión Transformadora”, Riobamba, Ecuador (2015) 14. Seng, K., Zeki, A.: Career Guidance and Employment M anagement System. In 2014 3rd International Conference on Advanced Computer Science Applications and Technologies, Amman, Jordan (2014). https://doi.org/10.1109/ACSAT.2014.20 15. M orales, J.: Serious games diseño de videojuegos con una agenda educativa y social. Editorial UOC, España (2015) 16. Susi, T., Johannesson, M ., Backlund, P.: Serious games: An overview (2007) 17. Sandovar, A., Braad, E., Streicher, A., Söbke H.: Ethical Stewardship: Designing Serious Games Seriously. In: Dörner R., Göbel S., Kickmeier-Rust M ., M asuch M ., Zweig K. (eds) Entertainment Computing and Serious Games. Lecture Notes in Computer Science, vol 9970. Springer, Cham (2016) 18. M artínez, J., Valls, F.: Aplicación de la Teoría de Holland a la clasificación de ocupaciones. Adaptación del inventario de Clasificación de Ocupaciones (ICO). Revista M exicana de Psicología. 25, 1, 151-164 (2008) 19. Holland, J., Fritzsche B., Powell, A.: SDS: Búsqueda Autodirigida. M anual técnico. M anual M oderno, M éxico (2005) 20. Nauta, M .: The Development, Evolution, and Status of Holland’s Theory of Vocational Personalities: Reflections and Future Directions for Counseling Psychology. Journal of Counseling Psychology. 57, 1, 11-22 (2010) 21. Podsakoff, P., M acKenzie, S., Lee, J., Podsakoff, N.: Common M ethod Biases in Behavioral Research: A critical Review of the Literature and Recommended Remedies, 88, 5, 879-903 (2003) 22. Guardiola E., Natkin S.: A Game Design M ethodology for Generating a Psychological Profile of Players. In: Loh C., Sheng Y., Ifenthaler D. (eds) Serious Games Analytics. Advances in Game-Based Learning. Springer, Cham (2015) 23. Shi, Y., Shih, J.: Game Factors and Game-Based Learning Design M odel. International Journal of Computer Games Technology, 2015, 1-11 (2015) 24. Dinis-Carvalho J., Ferreira A., Barbosa C., Lopes C., M acedo H., Tereso P.: Effectiveness of SCRUM in Project Based Learning: Students View. In: M achado J., Soares F., Veiga G. (eds) Innovation, Engineering and Entrepreneurship. HELIX 2018. Lecture Notes in Electrical Engineering, vol 505. Springer, Cham (2019) 25. Sobrevilla, G., Hernández, J., Velasco-Elizondo, P., Soriano, S.: Aplicando Scrum y Prácticas de Ingeniería de Software para la M ejora Continua del Desarrollo de un Sistema Ciber-Físico Revista electrónica de Computación, Informática, Biomédica y Electrónica. 6, 1, 1-15 (2017) 26. Osborne-O'Hagan, A., Coleman, G., O’Connor.: Software Development Processes for Games: A Systematic Literature Review, CCIS Vol. 425, Springer-Verlag, June 2014, pp 182-193. (EuroSPI 2014) 27. Rogers, S.: Level Up!, the guide to great video game design. Wiley, United Kingdom, pp 12-15, (2010) 28. Velarde, D., González, D., Camarena, B.: Certeza vocacional en el contexto universitario. ORFILA, M éxico (2013) 29. Valderrama, J.: Los videojuegos: conectar alumnos para aprender. Revista Electrónica Sinéctica. 39, 1-15 (2012)

Automated tasks for the working memory training Ismael Esquivel Gámez, Flora Lilia Barrios Martínez, Karina Estela Galvez Buenfil Universidad Veracruzana, Veracruz, M éxico

Abstract. Cognitive skills are said to be present in every action taken by an individual when he or she executes a series of tasks. Among these tasks, the working memory plays an important role in the early perception through the performance of young learners. Therefore, a cognitive training environment was carefully designed and developed, from automated tasks that were based on previous activities which were used for the measurement of the working memory. The environment consists of six automated tasks hosted into a virtual learning environment course called “M emoWorkout”. A total of 13 third-year high school learners from a private institution, assessed the rank of difficulty of each task in its first level, providing feedback as end users, in order to allow the improvement of the tasks as well as the increase of its possibilities for application in cognitive training programs, oriented to high-school graduates and freshman university learners. The results at least from this sample, allow to verify that the level of difficulty is adequate for the learners of the demographic target.

Keywords: W orking memory, training, automated tasks.

Introduction

This study has been developed taking advantage of the fact that videogames have become one of the central elements of youth entertainment. As a proposal, the creation of a virtual learning environment (VLE) for the reinforcement of the Working Memory (WM) is described, with the intention that both professors and learners from Mexican institutions have access to it. The following section describes the concept of WM, the characteristics of cognitive training, some routines which have been automated as well as their associated experiences. Afterwards, the process followed by the development of the software which is being proposed is documented until the reaching of the assessment from the end user. 1.1 Working Memory Executive functions include several higher cognitive abilities. From this abilities, inhibition of response, working memory and cognitive flexibility are considered to be central functions [1]. Among them, the Working Memory (WM) is responsible for the temporary storage and updating of a small amount of information, keeping such information accessible for subsequent manipulation in real time through the use of

159

strategies [2]. It is of great importance for language understanding, data processing and goals preservation when it comes to solving problems [3-5]. The most referred model for the WM representation is one proposed by Baddeley [6], which is constituted of four components: a central executive, which is subdivided into two subsystems: phonological loop and visuospatial agenda and, the episodic buffer. The central executive is an attentional system which, with certain limitations, chooses and processes data using several control elements, manipulating material in other subsystems and controlling different activities [5] [7]. The function of the phonological loop is to store temporally verbal data. It contains, in turn, two subsystems, the phonological store which keeps the information active, through the sub -vocal repetition, and one of articulatory synthesis which receives words or images and keeps them in the phonological store by sub-vocalization. According to Flores and Ostrosky-Shejet [5], the so-called visuo-spatial agenda is responsible for temporarily maintaining visual and spatial information. It is used to plan movements and it is associated with spatial ability and tasks involving visuospatial memorization [4]. It is constructed with visual components, which proces s patterns and perceive "what" or spatial components responsible for the localization of data in space and its transmission once knowing the "where"; furthermore, a relationship with kinetic components may be possible. The episodic buffer maintains a relationship between the three previous components and the long-term memory. It filters information based on its nature: verbal, visual or perceptual; and, although it is a temporary system, it can access long-term memory, in order to either learn or recover data [8]. The discrepancies within the working memory capacity (WMC) are associated with variations in the performance of different tasks of basic attention. These discrepancies have also been related to differences in: comprehension reading [9], arithmetic skills [10], problem solving [11], simultaneous execution tasks [12], and acquisition of knowledge from hypermedia sources [13]. Additionally, strong associations have been found between the WMC and intelligence [14-15]. 1.2 Cognitive training It is a systematic training that targets the conservation, improvement or development of cognitive abilities. The effects of training can be transferred near, it includes changes in the execution of activities of the same cognitive area (e.g. memory), or far, which constitutes changes in the execution of activities in other associated areas (e.g. arithmetic skills) [16] [17]. The factors which represent the greatest influence on the results of cognitive training are the practice environment, consisting of activities and execution conditions, and the participants profile (age, biological factors, initial status of the skills and personality) [18]. There are three types of training [19]: 1) sessions of psychological therapy, 2) courses lasting one semester or shorter and 3) tasks based on video games designed specifically for cognitive improvement (braingames). Besides, the goal may be: a) intensive training of a single type of cognitive task (e .g. WM), which may include variants in: a) the use of different materials [16] [18], b) The use of software (Cogmed, Cognifit, etc.) to improve a wider cognitive area (e.g., Short-term memory and WM) and, c) The use of multiple tasks for the development of different cognitive skills at the same time (e.g. WM and executive functions).

160

1.3 Cognitive automated training On previous studies, it has been found that the advantages of cognitive skill training, supported by information technology (IT), can be the following: attractive, comfortable, flexible, cheap, and non-invasive. From these studies, Table 1 shows the tasks, the duration of the training, the cognitive skill trained and the size of its effect. To complement the information, in the findings of their study [17]: WM training produces immediate and sustained transferring effects related to both the short-term memory (STM) and WM. Small immediate effects of transference to verbal and nonverbal ability, but not sustainable, were found. The STM and the visuospatial WM showed greater transferring effects than the verbal STM and WM and that the Age (4-71) presents no significant transferring effect. Additional instructions during WM training had no influence on transferring effects. Training both modalities (verbal and visuospatial) of WM could be more beneficial than the training of just one. Better results are found with adaptive exercises of a specific cognitive function and the same effect occurs during the sessions, when there is supervision and monitoring, instead of just the presence of the examiner. Until that time, it did not show the sustained effects of far-transference of WM training to scholar aspects, such as verbal or mathematical skills. So far, the findings could be revealing the implication that not even serious design and variation of training conditions or, more generally, the learning environment has begun.

Software development process

In the beginning, fifteen tasks were conceptually determined, after designing the user interfaces and dialogues , only nine became a prototype. In the end, only six reached the final stage. Once the prototypes were developed, a group of postgraduate specialists in the techno-pedagogy area was asked to execute the tasks and answer a couple of questions with the intention of improving them. The answers were taken into consideration to stablish the appropriate changes. Once the six tasks were completed, a session was organized for high school graduates to play in the initial level and, to their own perception, answer on the degree of difficulty of each task as well as to share the reason for their answers . Below is a description of each of the previous s tages:

161

Table 1. Characteristics of studies in cognitive skills training. Training tasks

Routine’s profile

Cognitive skill – effect size

Dual n-back [20]

8–19 days, for about 25 min daily

 Verbal WM – Adverse effect  Nonverbal ability – Small effect

Letter memory M emory Updating [21]

15–45-min sessions over a period of 5 weeks

 Verbal WM – Strong effect  Verbal ability – Adverse effect  Nonverbal ability – Small effect

Letter memory, Stroop [21]

3 times per week for 5 weeks, each session lasting 45 min

 Verbal WM – Strong effect  Attention - No effect

A mix of 12 tasks – Cogito Study [22]

101 daily sessions

 Verbal WM – Intermediate effect  Visuospatial WM – Small effect  Nonverbal ability – Small effect

Verbal and spatial complex tasks [16] Dual n-back [23]

30–45 min, 5 days per week over 4 weeks

Attention – Small effect Nonverbal ability – No effect

Daily training 5 times per week for a period of 4 weeks

Single n-back (visuospatial) [23]

Daily training 5 times per week for a period of 4 weeks

Verbal WM – Strong effect Nonverbal ability - Strong effect Verbal WM – Strong effect Nonverbal ability - Strong effect

2.1 Conceptual design The tasks were based on the nature of a complex task, which, according to Carruthers [24], have three stages: storing, processing and recalling. In the first stage, a stimulus for memorization is presented, in the second stage, a distractor is handled, such as verifying if a sentence is logical or if an arithmetic operation is correct, and finally in the last one, the participant registers the memorized stimuli. Additionally, it was taken into account that a greater effect is achieved by combining stimuli from both domains in the task [17] as shown in Table 2. The following describes each of the prototypes. Hop on, Hop off stops. An urban bus travels along a route, which involves several stops. In each one, a group of people with different passenger discounts, can get on the bus. The discounts are shown at the beginning of each trip. The number o f passengers is revealed for a moment and then disappears when the bus continues its journey. The goal of this activity is to keep track of the total income and deliver it

162

correctly at the end of the trip, taking into account that parents are not charged for their babies. Soccer. The names of three soccer teams with very similar scores appear on the screen, one of these teams will be relegated to a lower category. Each attempt, distracting images appear first and then the goals that each team scored and received, to calculate the number of points each team will receive, so that the scores must be updated in memory. At the end of several attempts, the teams appear on the screen and the team with the lowest score must be chosen. Fruit seller lady. A woman who sells fruit needs to call a randomly generated number, which appears at the center of the screen and disappears after a while. Then an image of the fruit stand with its prices is shown. A customer arrives and asks for fruit in different quantities and the total must be updated in memory. At the end a bill appears and the correct exchange must be delivered, after that, the number initially shown must be dialed. Crows vs. Pigeons. Initially the alphabet appears on the left. On one side crows will be arranged and on the other side there will be pigeons, which will fly to the opposite side, one and then the other. In each sequence a crow carries a small piece of paper with a letter of the alphabet that cannot be seen until the crow drops it in a box at the end of its flight. At that time, learners must memorize the letter and click on the antipodal letter. If it is correct, the pigeon starts its flight carrying that letter and if not, an error is indicated. The sequence is repeated N times, depending on the level, until the original letters are requested to recover as they appeared and send them as an encrypted message. The score varies according to the letters entered correctly. Win the battle. An image of a military barrack appears at the center. First, soldiers, vehicles and weapons of a certain type enter into the barrack. As time passes, they arrive and leave the barracks . It is necessary to memorize the amount that remains, to register the correct number at the end of the attempt. When they are compared with the correct answers, right and wrong answers will appear, which will affect the score previously obtained. Bingo! A Bingo scorecard is shown with 16 (4x4) squares on it. The squares are randomly numbered (1 to 54). As each of the numbers that are in the scoreboard are heard, it is necessary to remember it along with the previous ones, when the squares are covered in some of the forms allowed (Big frame, Small frame, Diagonal, Column, Row, Cross and Full) click on the corresponding button. The success depend s on whether the buttons have been touched correctly and at the appropriate time or not, which will affect the score. Mexican Lottery. A Mexican lottery scoreboard with 16 colorful images (4x4), randomly assigned, will be displayed on the s creen during the attempt. When a card appears, the player must remember such card and the previous ones, so that when the boards are covered in some of the forms allowed (Big frame, Small frame, Diagonal, Column, Row, Cross and Full) click on the corresponding button. Fin ally, the score depends on how many forms have been indicated correctly with the corresponding buttons. Dancing in the sand. A sand scenario appears, and in the lower left side, a sequence of N steps of a dance figure is shown 3 times. Then a target crab which should be memorized is presented. After a while, crabs of different colors appear and disappear, having to count those of the target crab, to then indicate the quantity that ap-

163

peared. Next, each foot appears with a group of 9 arrows; which indicate th e direction to reproduce the steps shown initially. The total score will depend on the successes in the numbers of crabs and the sequences reproduced. Deactivated! A group of bombs to be deactivated appears on a map. Each one is deactivated by solving an algebraic sum and it is necessary to memorize the total, and an arrow oriented to either the left or right. Once the bombs are deactivated, the mother bomb must be deactivated, in order to do this, the pairs number-arrow and that were memorized must be entered in the correct order. There is a certain time to deactivate each bomb, which will appear in the upper right. If this time ends or incorrect numbers and / or arrows are entered, the corresponding bomb will explode and the attempt will be lost.

Table 2. Domain of the stimuli used by each stage.

Task Hop on, hop off stops Soccer Fruit seller lady Deactivated! Dancing in the sand Crows vs. Pigeons Win the battle Mexican Lottery Bingo!

Storing Verbal Verbal Verbal Both Visuo-spatial Verbal Verbal Both Both

Processing Both Verbal Both Verbal Both Both Both Both Both

Recalling Verbal Verbal Verbal Both Visuo-spatial Verbal Verbal Both Both

2.2 User interface Design Each task manages 3 levels, with 3 attempts to solve in each and with the possibility of automatically adapting them to the level of performance. The layouts size is similar to that of an iPad tablet (1024x768), with the intention of distributing them later on these devices. In the interfaces, constant componen ts that facilitate learning are handled. At the beginning, the instructions appear, followed by a traffic light that prepares the user. Then, during the presentation and recovery, on a black line located at the top, the level, the points obtained and the remaining time appear. At the end of the level, the results layout and a button to continue are displayed. Figure 1 shows the demo screen of the resulting tasks. 2.3 Programming For this stage, following previously designed scripts, the “Software Construct 2” was used, given its integrated image editing tools, the simplicity to create two -dimensional games and exportation to different platforms. The last one was important for this proposal, since the aim was to integrate the automated tasks into a VLE like Moodle in a course called MemoWorkOut.

164

Deactivated!

Dancing on the sand

Crows vs. Pigeons

Win the battle

M exican Lottery

Bingo! Figure 1. End user interfaces.

2.4 Prototype testing At this stage, after having contacted a number of postgraduates from the area in a colloquium, an invitation email was sent, in which their participation was appreciated, the steps to be followed were described and the support was o ffered to clarify any doubts. During the procedure, after sharing the data for accessing Moodle, they were asked to read a brief explanation of the WM, and then on a cyclic basis, for each task they read the instructions, they executed it as many times as they wanted and they responded, in an electronic survey, to the questions: “After executing task X, I think that...” and “to improve it, in addition to eliminating detected errors, it would be good ...”. During this period (month and a half), the ten prototypes were tested with repeated invitations to participate, doubts were clarified and, in the end, responses from ten users were collected, to make the relevant changes, aimed at the development of training tasks of the WM.

165

2.5 Implementing modifications In the next phase of the study, the group of programmers was commissioned a couple of tasks. In order to accomplish these, it was established to finish in a period of five months, departing from the modifications and improve that proceeded, three levels of the tasks, according to the following plan:  Evaluate the actions proposed for improvement and clarify any associated doubt.  Verify that the multimedia resources used were freely distributed and, in case not, adapt them.  Execute the partial and general changes that were required.  Verify that the proposed modifications were actually made and establish them as the base line of the remaining levels .  Ascertain the level of difficulty that corresponded to the product developed .  Plan the characteristic elements of the remaining two levels .  Make and test the modifications, based on weekly feedback.  Design, build and add the software loading layout.  Export the products to HTML5, to be hosted on the Moodle platform. At the end of this stage, only six complete tasks out of nine were obtained, with the distinctive elements of each level shown in table 3. 2.6 End user testing The level of difficulty that each task represented was evaluated by means of an electronic survey with a couple of questions, "After playing task X, I consider that the difficulty (1 to 4) was" and "the reason for this was". Participants. In one session, 13 bachelors studying the sixth semester from a particular institution participated. The average age was 17.62 and 77% of them were men. For one hour they interacted with level 1 on each six tasks. Procedure. Initially, the examiner introduced himself to the students, explaining the reason for the session. The learners also received a general description of the WM and its academic implications. The purpose of the tasks was briefly explained so learners could understand why their point of view was necessary. Then, they were informed of the access data to Moodle; they first answered the survey for “the player profile” and immediately after the "Your opinion is important" survey as they executed the tasks. During the session, they experienced different emotions, which are described later. Once the session was over, it was announced that they would stop working but not everyone did, because they were excited. Finally, students were thanked for their participation.

166 Table 3. Characteristic elements for each level. Task Hop on, Hop off stops

Level 1 4 stops, full fare prices

Soccer Fruit seller lady

3 team names 5 fruits purchased and paid with a $ 50 bill

Deactivated!

4 pairs of arrownumber to remember

Dancing in the sand

A 5-step sequence presented 3 times

Crows vs. Pigeons

3 letters to process and 3 to memorize

Win the battle

4 types of soldiers / vehicles to process and remember

M exican Lottery

10 seconds between each card 10 seconds between each ball

Bingo!

Level 2 5 stops, fare rates in fractions of 50 cents 4 team names 6 fruits purchased and paid with a $ 100 bill 5 pairs of arrow-number to remember A 6-step sequence presented 2 times 4 letters to process and 4 to memorize 5 types of soldiers / vehicles to process and remember 7 seconds between each card 7 seconds between each ball

Level 3 6 stops, fare prices in fractions of 20 cents 3 team shields 7 fruits purchased and paid with a $ 200 bill 6 pairs of arrow-number to remember A 7-step sequence presented 1 time 5 letters to process and 5 to memorize 6 types of soldiers / vehicles to process and remember 5 seconds between each card 5 seconds between each ball

Observations. It can be said that in the case of video games, many young people do not read the instructions or watch the demos. It seems that they prefer to apply the trial and error technique, they only pay attention to the instructions if they cannot find a way to complete the level, even so they would rather ask their classmates to go over the instructions. It is not surprising that some of them chose to continue playing, those mentioned continued on the higher levels. At least two learners felt and stated that the experience provided them with a good challenge. Although they used headphones, they did not focus on their own progress and frequently reviewed the performance of others, they often some time to explain what should be done to move forward. The word antipode was difficult to understand so there was a need to explain it in more detail. Results. Given the similarity of the tasks with the videogames, it was determined to evaluate the profile that players managed, finding that two of them do not usually play and the rest, is auto-perceived with an average ability, playing an average of 24 hours a week, preferably action games. The majority always ends them and 50% usually play more than one game, preferably using a computer or cell phone. As for the degree of difficulty, the questionnaire data were processed, which resulted on the obtaining of the data shown in table 4. The most representative comments about the reason for the scores obtained was also added to the table. As can be seen, the task perceived as the most complicated one is “Bingo”! This mig ht be due to the fact that the handle of sounds is infrequent. On the other hand, the task perceived as

167

the simplest one is “Dancing in the sand” which could be because of the physical reproduction of the steps made with the intention of memorizing them, t his could be explained based on the “enactment effect”, in close association with the embodied cognition. Such cognition suggests that physical attributes in human body, mainly perceptual and motor systems can have an impact on the ability to remember things. This effect is based on two main principles: the actions performed imply representations considerably more elaborated than those from just reviewing visually and, the promulgated actions grasp the motor system while other modes of coding don’t [25]. Overall, for a range from 1 to 4, the average was 2.77. Table 4. Assessment of the level difficulty . Task Deactivated!

Average score 2.85

Dancing in the sand

2.38

Crows vs. Pigeons

2.69

Win the battle

2.85

M exican Lottery

2.62

Bingo!

3.23

It was easy because…

It was difficult because…

The operations are easy for me. The sequence is repeated 3 times, I imitate it with my feet and then we have to count crabs. It is slow and allows you to keep memorizing.

I have not developed the mental calculation ability I do not have agile feet

You just have to pay close attention and count. I have the habit of playing it with my family. It looks a lot like the lottery game.

Its dynamics were very strange because of the antipodes There are many elements that change continuously You need to pay close attention Although similar to the lottery, it is more difficult to handle numbers

The level of perceived difficulty is adequate if, as planned, it represents a challenge that may eventually reinforce the WM. However, when planning the training routines, it is necessary that the coordinators promote and encourage the use of them, giving a continuous follow-up, given the fact that they do not correspond to the type of challenges they commonly face when playing video games.

Conclusions

It is remarkable the behavior of trial and error that several learners had, which could have affected the score they issued on the difficulty, even when seen as an opportunity, the collaboration was strengthened. A small sample of young people has a positive attitude towards tasks that challenge their mind, regardless of whether they lack elements associated with the emotion and satisfaction, which, among other attributes, characterize video games, which have become one of the most pleasurable activities for them. It can be revealing that comments such as " I do not like to memorize", "I do not get the numbers", "I do not like this kind of challenges" , which were issued by

168

some, negatively impact the efforts of the institutions to improve the capacity of this cognitive skill, crucial in their lives . To minimize the impact, it is necessary to reproduce this type of tests in more schools, which could allow, among other things, to improve the software, design the proper training routines, plan a follow-up and close support from the institution, determine the type of personality that has better reception to the challenges and finally, to measure the effectiveness of the software in the training of the WM capacity, in both domains. Additionally, during the training sessions, the registration of the different strategies to memorize the required stimuli, which allow the participants successfully complete the tasks is necessary; in order to, later choose the most effective on es so they can be integrated to the teaching-learning processes, either traditionally or with the incorporation of IT such as in the case of the present study. Furthermore, a more profound study when applying the Enactment effect is needed, since there is evidence of the positive effects of the "remembering doing" strategy, in environments very similar to those found in activities of daily life.

References 1.

4. 5. 6. 7. 8.

10.

M iyake, A.; Friedman, N. P.; Emerson, M . J.; Witzki, A. H.; Howerter, A. y Wager, T. D. (2000). “The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis”, Cognitive psychology, vol. 41, núm. 1, pp. 49-100. Retrieved from: DOI:10.1006/cogp.1999.0734. Jonides, J.; Lacey, S. C. y Nee, D. E. (2005). “Processes of working memory in mind and brain”, en Current Directions in Psychological Science, vol. 14, núm. 1, pp. 2-5. Retrieved from: DOI:10.1111/j.0963-7214.2005.00323.x Baddeley, Alan (2012). “Working memory: theories, models, and cont roversies”, Annual Review of Psychology, núm. 63, pp. 1-29. Retrieved from: DOI: http://doi.org/10.1146/annurev-psych-120710-100422. Etchepareborda, M .C. & Abad-M as, L. (2005). “M emoria de trabajo en los procesos básicos del aprendizaje”, Revista de Neurología, vol. 40, núm. 1, pp. 79-83. Flores, J. & Ostrosky -Shejet, F. (2012). Desarrollo neuropsicológico de lóbulos frontales y funciones ejecutivas, Ciudad de M éxico: M anual M oderno. Baddeley, Alan (2000). “The episodic buffer: a new component of working memory?”, Trends in cognitive sciences, vol. 4, núm. 11, pp. 417-423. Baddeley, A. D. & Eysenck, M .W. (2010). Memoria. M adrid: Alianza Editorial Saeed, T. (2011). A Comparative Study of Working Memory in Children with Neurodevelopmental Disorders (Doctoral dissertation, National University of Ireland M aynooth). Esquivel, I., M artínez, W., Córdoba, R., Reyes, C. M emoria operativa y lectura comprensiva: medición con pruebas de amplitud lectora y tipo cloze en ámbitos pre y universitarios. Apertura. 2016. 8(2): 38-53. Gathercole S, Pickering S, Knight C, Stegmann Z. Working memory skills and educational attainment: evidence from national curriculum assessments at 7 and 14 years of age. Applied Cognitive Psychology . 2003;18(1):1-16.

169 11. 12.

13.

14. 15.

16.

17.

18. 19.

20.

21.

22.

23.

24. 25.

Bühner M , Kröner S, Ziegler M . Working memory, visual–spatial-intelligence and their relationship to problem-solving. Intelligence. 2008;36(6):672-680. Bühner M , König C, Pick M , Krumm S. Working M emory Dimensions as Differential Predictors of the Speed and Error Aspect of M ultitasking Performance. Human Performance. 2006;19(3):253-275. Banas S, Sanchez C. Working M emory Capacity and Learning Underlying Conceptual Relationships Across M ultiple Documents. Applied Cognitive Psychology. 2012;26(4):594-600. Oberauer K, Süβ H, Wilhelm O, Wittmann W. Which working memory functions predict intelligence? Intelligence. 2008;36(6):641-652. Redick T, Broadway J, M eier M , Kuriakose P, Unsworth N, Kane M et al. M easuring Working M emory Capacity With Automated Complex Span Tasks. European Journal of Psychological Assessment. 2012;28(3):164-171. Chein J, M orrison A. Expanding the mind's workspace: Training and transfer effects with a complex working memory span task. Psychonomic Bulletin & Review. 2010;17(2):193-199. Schwaighofer M , Fischer F, Bühner M . Does Working M emory Training Transfer? A M eta-Analysis Including Training Conditions as M oderators. Educational Psychologist. 2015;50(2):138-166. von Bastian C, Oberauer K. Distinct transfer effects of training different facets of working memory capacity. Journal of M emory and Language. 2013;69(1):36-58. Uttal D, M eadow N, Tipton E, Hand L, Alden A, Warren C et al. The malleability of spatial skills: A meta-analysis of training studies. Psychological Bulletin. 2013;139(2):352-402. Jaeggi S, Buschkuehl M , Jonides J, Perrig W. Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences. 2008;105(19):6829-6833. Dahlin E, Nyberg L, Bäckman L, Neely A. Plasticity of executive functioning in young and older adults: Immediate training gains, transfer, and long-term maintenance. Psychology and Aging. 2008;23(4):720-730. Schmiedek F., Lövdén, M . y Lindenberger, U. (2010). Hundred days of cognitive training enhance broad cognitive abilities in adulthood: findings from the COGITO study. Frontiers in Aging Neuroscience. Jaeggi S, Studer-Luethi B, Buschkuehl M , Su Y, Jonides J, Perrig W. The relationship between n-back performance and matrix reasoning — implications for training and transfer. Intelligence. 2010;38(6):625-635. Carruthers, P. “La evolución de la memoria de trabajo”, Ludus Vitalis. 2016; 21(40): 99-124. M adan, C. R., y Singhal, A. (2012). Using actions to enhance memory: effects of enactment, gestures, and exercise on human memory. Frontiers in Psychology, 3, 507.

Construction-validation of an implement to measure the digital competence of a training teacher in the level of Telesecundaria Anna Luz Acosta Aguilera 1 , Rubén Edel Navarro 2 , Yadira Navarro Rangel 3 1 Estudiante del Doctorado

en Sistemas y Ambientes Educativos. Universidad Veracruzana. Av. Jesús Reyes Heroles s/n Zona Universitaria, Fracc. Costa Verde, CP 94294, Boca del Río, Ver. 2 Investigador Tiempo-Completo. Universidad Veracruzana. Av. Jesús Reyes Heróles s/n, Zona Universitaria, Fracc. Costa Verde, CP 94294, Boca del Río, Ver. 3 Investigadora Tiempo-Completo. Benemérita Universidad. Autónoma de Puebla. Calle 4 Sur #104; Col. Centro C.P. 72000; Puebla de Zaragoza, Puebla, M éxico aniluzacosta@gmail.com, redeln@gmail.com, ynavarro44@gmail.com

Abstract. The objective of the present article is to describe the process of designing and validating an implement which allows to measure the digital competence of the students from the bachelor in Secondary teaching in the area of Telesecundaria. First of all, it is set the context and evolution of the concept of digital competence in order to provide the theory to understand the structure of the tool. The design of the questionnaire was made taking as a reference some questionnaires applied by different researchers at a national and international level. 48 items were defined and these were organized into 3 dimensions based on theorical elements from different authors. Three different types of validation were evaluated: 1. content, 2.construct and 3.consistency and reliability. The outcomes indicate high levels of reliability , also the feedback received through the opinion of experts provides valuable and trustworthy elements about the implement. The information obtained in this first procedure bring to light the need to consider technical and logistics aspects in the final application of the questionnaire as well as complement the data with another tool which is not of perception. Key words: Digital competence, initial teacher training, validation of tools.

Introduction

The widespread growth of internet in the mid-nineties, exhibited what it was called by some authors as the technical revolution center in the use of ICT, which reconfigured the material base from society to an accelerated rhythm (5). This social reorganization had a great impact in the teaching field, modifying its practice and therefore the function of teachers, because putting into practice this kind of technology requires a teacher open to innovation and conscious of the methodological implications that involve proposing a teaching environment where virtuality plays an essential rol. This form of being inside the social and school environment, which is called by some authors a digital culture or ciberculture (6) demands different ways to be and

171

behave that also promote different forms of communication, management and school organization, access to information and creation of data and information. In this context of changes and uncertainty arise the concept of digital competence. There is a great deal of definitions, which respond to moments, contexts and specific objectives. However, in order to clarify the characteristics they have, the ones which were considered relevant for the purpose of this investigation were selected. To Peña-López (8) digital competence is the combination of knowledge, skills and abilities, along with values and attitudes to reach objectives effectively and efficiently in context and with digital tools. It is expressed in a strategic control of five great abilities related to the different dimensions of digital competence. He proposes 5 dimensions/aspects: 1)learning dimension, 2) informational dimension, 3) communicative dimension, 4) digital culture dimension and 5) technological dimension. He mentions that having an efficient command in the 5 aspects proposed would mean being digital competent, which would be a goal to achieve for the students and teachers. Some of the dimensions mentioned by Peña-López (8) are closely related to what Jordi Adell exposed (9). He indicates that digital competence is formed by five basic elements which are: 1. Informational competence: set of knowledge, skills and abilities necessary to work with information, search it, have access to it, organize it, manage it, analyze it, criticize it, evaluate it and then create new information and distribute it. 2. Technological competence: It has to do with handling the technological tools. 3. Multiple literacy: Our world is not only the society of the printing texts, since the 50’s with the advent of TV, we live in an audiovisual society, the video language, photography, comics, should be part of the curriculum and not only as spectators but as authors too. 4. Generic cognitive competence, not only having access to information but being able to transform it into knowledge, requires a problem-solving ability. 5. Digital citizenship: it is the preparation to live, into the real and digital world, the preparation for a citizenship to be critic, free, integrated, able to live this society with rules and high standards of behavior in the digital world. According to the European Commission, digital competence is the “safe and critic use of information technology for work, leisure and communication. It is sustained on the basic competence in the subject ICT: the use of computers to obtain, evaluate, store, produce, present and exchange information and communicate and participate in collaborative networks on line”(10, p. 17). This competence was included within the eight key competences established to the accomplishment of permanent learning.

172

In some research analyzed in relation to the work developed by Gisbert, it can be observed a clear tendency to define digital competence according to the contribution its development can have within the teaching-learning process, pointing out that the digital competence is the sum of “skills, knowledge and attitudes in technological, informational, multimedia and communicative aspects which lead to a complex multiple literacy” (11). Therefore, it represents “a set of skills and knowledge that the individual himself must acquire and consolidate as an essential mean to move forward in his studies (at any formal stage and throughout his life)” (12, p. 76). They consider it is “a basic competence to any XXI-century citizen, from which the digital competence of a teacher derives (DCT) specific to any education professional”(12, 57), all of this would require a teacher who had the skills, attitudes and knowledge needed to establish learning environments in context enriched by ICT. Related to the “specific use of knowledge, skills and abilities which has to do with the development of elements and processes that allow to use efficiently, effectively and innovatively the tools and technological resources” (11, p.18). Some other definitions which were found are more general, these are not limited only to the educational field. These defined digital competence as: “the set of knowledge, skills, attitudes, strategies and consciousness the use of ICT and digital resources require to do tasks, solve problems, communicate, handle information, collaborate, create and share contents and also generate knowledge in an effective, efficient, adequate, critic, creative, autonomous, flexible, etic and reflexive way for work, leisure, participation, learning, socialization, consumption and empowerment.” (1, p.6) All of the above matches what it was stated by Castillejos, Torres and Lagunes (13) who say that digital competence is a key and transverse aspect which every digital citizen should develop, since they face every day the need to take advantage of technology. “Promoting them involves the critical, creative and safe us e of ICT, either for work, school or every day activities” (p. 57). Cabello, Cuervo, Puerta y Serrano complement this definition arguing that so as to be digitally competent it is necessary to have the abilities to get, process, transmit information and transform it in knowledge. This means having the control of basic specific languages, technological and expressive resources to transmit the information, strategies to identify and solve software and hardware problems as well as logical thinking skills. All that has been described so far allow us to identify the complexity and diversity of the concepts analyzed, which represent a foundation to the establishment of criteria according to the interest, purpose and context in which they are examined.

173

Methodology

Prior to the description of the methodological characteristics and the procedure for preparing the survey, the variables that support it and the hypothesis established for this research work are described, it must be mentioned that the information presented is part of a more comprehensive research related to the pedagogical usability of ICT in initial teacher training. 2.1 Hypothesis The general hypothesis of this research is the following: H1 The pedagogicall usability of the ICT improves the digital competence of the future teachers-to-be in Telesecundaria. 2.2 Variable: digital competence The digital competence of the student was selected as a variable of the study after a thorough search in relation to the theoretical construct developed up to date, as well as the objectives set. These pretend to identify the perception the students have on their own digital competence and at the same time analyze the type and level of competence they actually have, under de assumption that this knowledge is essential to design strategies which allow teacher trainers to offer better educational experiences in relation to the use of ICT to their students. 2.3 Construction of the implement The questionnaire at issue was elaborated after making an important search and identification of some tools that have been applied by other researchers both at national and international level, even when the exact questions were not transcribed they were considered as an important basis for the final version that it is presented. The works from Zúñiga (2), Pech (3) and Esteve (4)were analyzed in more detail since they have developed their investigations in higher education contexts related to the educational field, that is how the three dimensions and fourteen variables of “digital competence” intended to be measured with this implement were chosen. Another critical moment in the making of the implement came out when the multiple choice options in scale Likert were determined. Ho wever, after analyzing what De Villis (15) says to this matter, only four-option answers were chosen obtaining with this decision a higher reliability in the questionnaire. Finally, a pilot implement was designed, with 48 items, which answer-options were presented in a Likert scale. These include the options “Incapable, A bit capable, Quite capable and Very capable”. 2.4 Subjects/population The participants were chosen randomly considering the following inclusion, exclusion and elimination criteria:

174

1. They have to be students of the Bachelor degree in Secondary Teaching from the area of Telesecundaria. 2. They have to be registered in the subjects of: Secondary level Teaching Basic Issues II and Educational Attention to teenagers in threatening situations. 3. They have to show interest and be willing to participate.

2.5 Pilot test The tool was applied at the Benemérita Escuela Normal Veracruzana (BENV) “Enrique C. Rébsamen” to a sample of 60 students from two different groups from the Bachelor in Secondary Teaching in the area of Telesecundaria in order to test the functioning of the tool and therefore to have elements to determine if the questions were clear and easy to understand by the participants. The implementation of the tool was self-administered online through the use of LimeSurvey which is free software to elaborate and apply surveys and is simple an d easy to use. The data is stored on the page www.limesurvey.org where they can be handled without difficulty. In spite of being a self-administered questionnaire, the conducting was guided and during school time, which required to ask permission from the teachers in advance to finish their lessons 30 minutes before scheduled. With the support of a teacher from the research area of the BENV, it was installed a computer and a modem in which a kind of intranet was configured so that the students were able to answer in their phones or laptops without using up their download volumes. It is important to mention that the participants were explained the purpose of the investigation and they were ensured its reliability. The implementation took place in two days at different times. There were no difficulties during the application.

Data Analysis

The analysis consisted in estimating descriptive statistic for each item (graphs, percentages, among others). After that it was carried out the factor analysis, the main component analysis and the reliability/internal consistence of the items analysis throughout the Cronbach’s Alfa. To do so it was used the Statistical Program for Quantitative Analysis SPSS version 24.

3.1 3.1.1

Validity Validity of content

175

Validity of content refers to the need of proving that the items that form the tool represent the construct intended to be evaluated, one of the most common procedures to carry out this validation is to turn to external criteria for example, the opinion of experts (16). This is why it was requested the assessment of three experts in digital competence, who were given a check list through which they could evaluate item by item the validity of the tool.

3.1.2

Validity of construct

Validity of construct refers to the consistence itself has the carrying of the data in the evaluated construct (16) that means, it allows to know if it is really being measured what it is supposed to be measured, to carry out this validation it was used the KaiserMeyer-Olkin test (KMO) and the Bartlett’s sphericity test. 3.2

Reliability analysis

The factor analysis enables to explore the suitability and validity of the proposals made by the researcher about the acceptability of the theoretical variables which intends to build. Its objective is to reduce the number of variables for a better scientific explanation and simpler to reality. In the same way, it clusters the data produced by many empirical variables into few theoretical variable or latent variables, or factors independent among them.

Results

4.1. Validity of content Initially it was determined who would be the expert judges to whom support in the validation of the tool would be requested. It was sent to three experts whose career and knowledge in the area of digital competence and teacher training is outstanding. They were contacted via email to ask their support. After that the tool and the corresponding check list were sent. Even when there were some comments related to the structure of the indicators in general, the judges considered appropriate and relevant the items designed for each dimension. Verbal feedback was possible only with one of them, this was carried out through a video chat system (WebEx). He was pointing out the observations he had made, supporting each one with a reason and suggesting options to a greater understanding.

4.2 Validity of construct According to Castro and Galindo (17), the Kaiser-Meyer-Olkin test (KMO) and the Barlett’s sphericity test help prove if the sample is acceptable for the factoring of the variables.

176

KMO and Bartlett test Sampling adequacy measurement Kaiser-Meyer-Olkin .765 Bartlett’s sphericity test Aprox. Chi-square 2560.611 gl 990 Sig. .000 From the previous results it can be concluded that the sample is acceptable for the factoring of the variables by reaching a KMO index of 0.765 (quite close to 0.8 which is the one normally demanded). At the same time the correlation matrix turned out spherical (p<0.01), and as a consequence, analyzable from the factorial point of view. 4.3 Factorial analysis In order to do the factorial analysis, first of all the correlation matrix is obtained with the aim of identifying if the variables share common factors which explain its variance. In the correlation matrix most of correlations resulting were highly significant (p<0.01), this indicates the existence a strong correlation in most of them, this would assume the presence of common factors which explain the variance of such variables and bring together the information provided in the factors hypothetically operationalized. 4.4 Reliability analysis The Cronbach’s Alfa coefficient was used to do the reliability analysis, obtaining a value of 0.976 as a result the tool is considered reliable to measure what is intended to be measured. Table 1. Cronbach’s Alfa tool. Reliability statistics Cronbach’s Alfa Cronbach’s Alfa based on standardized elements .976 .976

N of elements

In the analysis of correlations between elements that is presented in the following table, it is observed that each item appears with a positive correlation, which denotes consistency and reliability in the survey, likewise it can be observed that it is not necessary to eliminate any element, since none has a low correlation value, which shows the internal uniformity of the instrument. Table 2. Total statistics of the element.

Average scale if the element has been removed

Variance of scale if the element has been removed

Total correlation of elements revised

Cronbach’s Alfa if the element has been removed

177

D101 D102 D103 D104 D105 D106 D107 D108 D109 D110 D111 D112 D113 D201 D202 D203 D204 D205 D206 D207 D208 D209 D210 D211 D212 D213 D214 D215 D216 D217 D218 D301 D302 D303 D304 D305 D306 D307 D308 D309 D310 D311 D312

127.82 128.03 127.87 127.82 127.85 127.72 127.65 127.85 127.73 127.95 127.85 128.28 128.33 127.92 128.38 127.47 127.53 127.73 127.73 127.98 127.98 127.92 127.83 127.87 128.23 127.48 127.57 127.50 127.68 128.17 127.95 127.90 127.90 127.63 127.88 127.83 128.07 127.78 127.63 127.65 127.57 127.80 128.15

545.745 544.507 552.660 545.949 547.316 548.478 546.672 545.587 544.877 543.981 541.248 541.562 543.277 548.247 542.274 545.202 544.795 542.402 548.301 542.254 544.118 542.145 544.548 544.694 549.301 545.000 547.436 546.254 544.830 548.040 542.964 540.227 541.956 545.999 542.105 543.362 538.911 547.596 547.355 548.096 548.419 553.485 548.096

.728 .622 .512 .676 .661 .642 .680 .737 .765 .712 .792 .661 .666 .644 .637 .739 .779 .722 .658 .709 .697 .779 .754 .782 .568 .645 .603 .672 .756 .648 .764 .774 .769 .769 .797 .790 .817 .646 .595 .660 .593 .466 .530

.975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .975 .976 .975

178

D313 D314

127.70 127.82

554.654 548.356

.432 .571

.976 .975

If we observe in detail the previous chart, we can identify there is consistence in the value assigned to each item, only 2 of them show an increase of .001 which is not considerable enough to think of removing it from the tool. In that manner no questions were eliminated.

Discussion and future works

Designing a questionnaire that allows to know and analyze the digital competence that the future teacher of Telesecundaria has is the first step for the design of a proposal of integration of the ICT that foster the pedagogical use of the ICT through the establishment of methodologies, strategies and activities that meet the specific needs of the future teacher of Telesecundaria. This statistical work was a first approach to the analysis of validity and reliability of the implement. These first outcomes allow to distinguish some changes or adjustments that will be necessary to improve the measurement scale, such as revising the structure of dimensions to reorganize the questions in each of them. During the factorial analysis it can be observed that in the results, the variables are not distributed in the factors or dimensions identified, inst ead of that they are concentrated in two out of nine. One of the causes of this result could be redundancy in the items or lack of differentiation within the questions, as well as the location of each set of questions in the dimensions built a priori (Learning environment, Information and communication environment and Digital culture environment). We have to be very careful at the moment of integrating each dimension and the name that is given to them. It is also important to check that they are not include d one another and that they are well defined. Additionally, after implementing the tool errors in the structure of the document could be identify. These were pointed out by the participants and included typing and spelling mistakes along with a repeated item in dimension 1 and 2. Apart from the conclusions expressed before, it is important to state that the results obtained during the pilot stage permitted not only knowing relevant aspects of the context of implementation but establishing a precedent to:  Learning how to measure the digital competence of a training teacher of Telesecundaria.  Identifying the levels of digital competence from the participants and the possible relation between this and the pedagogical usage of ICT in the initial teacher training.  Identifying areas of opportunities which allow to strengthen the digital competence of teacher-training students at a curricular level. In the future it would be appropriate to develop a quantitative tool which enables to reinforce this perception questionnaire and at the same time evaluates the performance of the participants, by contrasting or complementing their opinion to what they really know how to do.

179

References 1. Punie, Y, Brecko, B (2014) DIGCOM : M arco Europeo de competencias digitales. European Commission. http://jakintza.eus/wp content/uploads/DIGCOM P_Donostia_ES-Rev.pdf 2. Zúñiga, J (2016) Las competencias digitales en el perfil universitario: el caso de la Facultad de Pedagogía de la Universidad Veracruzana. Veracruz, M éxico. Universidad Veracruzana; Facultad de Pedagogía. https://cdigital.uv.mx/bitstream/123456789/41455/1/Zuniga.pdf 3. Pech, S. (2016) Competencia digital docente. Cuestionario en línea. Ciudad Real: España. https://docs.google.com/forms/d/e/1FAIpQLSeoPt3bZlU_mUIhaUCq4GjGt0AYAI5 XpmcM 7niSogPg3uxI1Q/viewform 4. Esteve, F (2015) La competencia digital docente Análisis de la autopercepción y evaluación del desempeño de los estudiantes universitarios de educación por medio de un entorno 3D. Departamento de Pedagogía, Universitat Rovira i Virgili. Tarragona. España. http://francescesteve.es/tesis/ 5. Castells, M (2000) La era de la información: economía, sociedad y cultura. Volumen I. La Sociedad Red. 2ª ed. M adrid: Alianza Editorial, S.A. pp. 656. 6. Levy, P (2011). Cultura escrita y tecnocultura contemporánea: mediaciones cognitivas en la formación universitaria, en: Revista Nexus, No. 10, Escuela de Comunicación Social de la Universidad del Valle. 7. Ramírez-M artinell, A, Casillas, M A (Coords.) (2017). Saberes digitales de los docentes de educación básica. Una propuesta para la discusión desde Veracruz. Veracruz: Secretaría de Educación de Veracruz. https://www.uv.mx/blogs/brechadigital/files/2017/04/Saberes-Digitales-SEV-librofinal.pdf 8. Peña-López, I (2009). La competencia digital: una propuesta. [Comentarios en una wiki en línea]. https://competenciadigital.wikispaces.com 9. Adell, A (2010) ¿Qué es la competencia digital? http://carraud.blogspot.mx/2010/10/jordi-adell-que-es-la-competencia.html 10. Lion, C (2012). Desarrollo de competencias digitales para portales de la región. Banco Interamericano de Desarrollo y Red Latinoamericana de Portales Educativos. https://documentop.com/queue/desarrollo-de-competencias-digitales-para-portalesde-la-region-relpe_5a0d2f6a1723dd99282eedfb.html 11. Gisbert, M , Esteve, F (2011) Digital leaners: la competencia digital de los estudiantes universitarios. La cuestión universitaria. Boletín electrónico de la Cátedra UNESCO de Gestión y Política Universitaria. No. 7. http://polired.upm.es/index.php/lacuestionuniversitaria/article/view/3359/3423 12. Silva. J, M iranda, P, Gisbert, M , M orales, J, Onetto, A (2016) Indicadores para evaluar la competencia digital docente en la formación inicial en el contexto chilenoUruguayo. Revista Latinoamericana de Tecnología Educativa RELATEC. Vol. 15, No. 3 ISSN: 1695-288X. https://relatec.unex.es/article/view/2807/1982 13. Castillejos, B, Torres, C, Lagunes, A (2016) La seguridad en las competencias digitales de los M illennials. Apertura. Revista de Innovación Educativa. Universidad

180

14.

15. 16. 17.

de Guadalajara. Vol. 8, No. 2, ISSN 2007-1094. http://www.udgvirtual.udg.mx/apertura/index.php/apertura/article/view/914/586 Cabello, J. L., Cuerva, J., Puerta A. & Serrano J. (2013). Tratamiento de la información y competencia digital en la educación. [Comentarios en una wiki en línea]. En red: https://sites.google.com/site/lascompetenciasbasicas/ticd De Villis, R (2017) Scale Development. Theory and Applications. Fourth Edition. Applied Social Research M ethods Series Volume 26. USA: SAGE. M uñiz, J, Fidalgo, A, García-Cueto, E, M artínez, R, M oreno, R (2005) Análisis de los ítems. Editorial La M uralla S.A. pág. 184 Castro, J., Galindo, M . (2000) Estadística M ultivariable: análisis de correlaciones, Amarú: Ediciones, Salamanca: España. ISBN 848196137.

“A journey through mathematics”: A serious game used for learning mathematics with ADHD children Claudia Blanca González Calleros 1 , Josefina Guerrero García2 , Yadira Navarro Rangel1 1

Facultad de Ciencias de la Electrónica, Benemérita Universidad Autónoma de Puebla, Av. San Claudio y 18 sur. Puebla, M éxico 2 Facultad de Ciencias de la Computación, Benemérita Universidad Autónoma de Puebla, Av. San Claudio y 14 Sur, Puebla, M éxico claudiablan.gonzalez@alumno.buap.mx, [joseguga01, ynavarro44]@gmail.com

Abstract. A journey through mathematics is a serious game designed for children with Attention Deficit Hyperactivity Disorder (ADHD), to achieve better learning in the acquisition of new concepts for solving mathematical problems, attending to their special needs. We focus on a two-dimensional design and the implementation of activities for an interactive learning, this through a narrative of how mathematics has been important, in the timeline of different civilizations, to solve problems of everyday life. The activities focus on improving current conditions in the learning of children with ADHD. We present preliminary results related to the efficiency of game design through the analysis of usability tests. Keywords: Serious games, gamification, ADHD, teaching-learning of mathematics, model driven development

Introduction

Currently, and as part of globalization, a social and cultural reorganization is necessary to face exclusion, rejection and any type of discrimination that many children may suffer at the basic level of education. As an important part of the process, pedagogical strategies that adapt to social circumstances and to the evolution of a world centered on new technologies and the era of knowledge and information must be designed and developed. To give attention to children with ADHD there are different treatments ranging from psychological therapy, medical-based treatment, as well as educational therapy; the most effective way to achieve a high pedagogical quality for these children is through methods that guarantee the development of skills in the teacher, as well as a quality educational program, dynamic and adjusted to the needs of the children. children with ADHD [1]. Currently, ICT as educational tools are part of the pedagogical strategies necessary to facilitate learning in school-age children. The use of serious games has proven to be highly efficient so that children with ADHD achieve better learning and develop

182

their academic potential. Through didactic strategies, teachers must facilitate training and learning with the use of technology and didactic methods that allow building knowledge in a creative and dynamic way. It is essential to know the strategies that are currently used in the classrooms at the basic level, to know how we can develop strategies to meet the educational demand and thus achieve the inclusion of children with ADHD in the classroom. The use of technology in the classroom, as a didactic tool, has proven to be highly effective and is supported by various publications, studies and pilot classrooms [2]. Similarly, there are works on the inclusion of gamification and serious games to promote the learning of specific topics, positively affecting the level of concentration of children who use them, favoring their learning, and providing motivation equal to or higher than other methods regarding on the acquisition of knowledge [3], [4] and [5]. Considering the efforts made with the incursion of technology in education, comes the motivation to develop tools supported on gamification techniques and serious games, for the Mexican context and focused on favoring the teaching of mathematics to children with ADHD. The document is organized as follows: Sections 2 and 3 provide basic in formation about ADHD, a review of the literature on serious games for ADHD, and the reasons for using virtual agents to promote the learning of mathematics. Section 4 Model Driven Development for the development of Serious Games. Theoretical foundation. In Sections 5 Serious game development and 6 Conclusions and future work.

ADHD

ADHD is the most common neurodevelopmental disorder in children, it has a worldwide prevalence of 5.29% in school-age children [6], and specifically in Mexico, where is estimated that there are approximately 33 million children and adolescents, of which 1.5 million could be diagnosed with ADHD [7]. This disorder presents a state characterized by an inappropriate level of inattention, hyperactivity and impulsivity with repercussions on behavior and emotional, cognitive, academic and social functions [8]. ADHD has a high psycho-social impact, which is reflected in the deterioration of the child's functioning in family, school and social life. Although it has been shown that it is not a cause of learning problems, it is related to poor organization of tasks, problems in waiting their turn, issues of socialization, suspensions and expulsions, anxiety disorders, behavior problems and depression. Some of the cognitive difficulties that occur in these children are low school performance, as well as problems to stand still and concentrate in class [8]. Children with ADHD have difficulties to respect the rules and keep an adequate behavior in every moment and such disabilities promote a slow learning [9]. According to the Diagnostic and Statistical Manual of Mental Problems (DSM -V), the disorders that influence the academic performance of children with ADHD are: difficulty in reading, written expression and mathematical difficulty [10]. In the classroom, children are liable to academic failure, low self-esteem, rejection by their classmates and learning alterations; however, when subjected to stimuli and special education, they can achieve significant learning [11].

183

Some of the cognitive disabilities that occur in children with ADHD are low school performance, although most of them have a high intellectual level, and problems to stand still and concentrate in class [8]. On the other hand, the children with ADHD have difficulties to respect the rules and to maintain an adequate behavior in each moment and such disabilities propitiate a slow learning [12]; they tend to be assessed negatively due to constant conflicts and reprimands ; doubt their abilities and underestimate their possibilities. The teacher must know which tasks can be more successful and use them to motivate them and teach them to value themselves. If the teacher trusts the possibilities of the student, he will learn to believe in himself [12]. Educational inclusion [13] refers to the use of strategies with which "it is possible to include all students and ensure that all people have the same opportunities to progress in the educational field", because these strategies "favor relationships social, cooperative learning, the development of new skills and new ways of building knowledge and promoting the capacities of creativity, communication and reasoning" [14]. Therefore, comes the need to create innovative teaching proposals that promote the training processes of teachers to improve the quality of education received by this vulnerable sector of the population.

Systematic review of the literature

The impact on the use of serious games on learning in school-age children has increased considerably in recent years; it has been identified that , through these, it is possible to improve the cognitive abilities of the students, reflecting in an effective learning. The use of technology in the classroom, as a didactic tool, has proven to be highly effective and is supported by various publications, studies and pilot classrooms [15]. Similarly, there are works on the inclusion of gamification and serious games to promote the learning of specific topics, positively affecting the level of concentration of children who use them, favoring their learning, and providing motivation equal to or higher than that of other methods with respect to the acquisition of knowledge [30]; The need to offer better learning and, considering the efforts made about the incursion of technology in education, arises the motivation to develop tools based on gamification techniques and serious games, for the Mexican context and focused in favoring the teaching of mathematics to children with ADHD. The objective of this section is to systematize the previously performed works and experiments related to the use of serious games as a pedagogical strategy for the meaningful learning of mathematics for children with ADHD in order to identify gaps in knowledge. 3.1 Serious games The term serious games was defined since 1970 by Clark Abt, he explained that his interest in this type of games is that they are not intended to be played solely for fun but have an explicit educational purpose and carefully planned. For the design of this type of educational resource is necessary the theoretical understanding of learning

184

such as behaviorism, cognitivism and constructivism because they will be immersed, and they will be the column of game mechanics [16]. It can be observed that serious games have great advantages in the educational field, among which the improvement in vision, increase in self-esteem, interactive learning is favored, at the same time as it promotes learning through the challenge , they allow to improve social abilities, of the language, reading of rules and menus, and basic mathematics, as well as the articulation of an abstract thought [17]. A series of characteristics that must be taken into account for the design of a serious game are: the character, a narrative description, challenges, restrictions, rewards, punishments, interactive elements and feedback [17]. Serious games have a mechanic for recording the progress of the player and for managing punctuations, as well as simulation elements such as challenges, rules, scenarios and interactive objects [18]. 3.2 Serious games to care for children with ADHD Some studies show the effectiveness of serious games as a strategy for teaching children with ADHD. [19] The elements that contain serious games can help this population to direct and monitor their activities and regulate their emotions. Serious games are positive, not only for children's learning, but also for personal development, inclusion in society and, at the same time, didactic entertainment that brings children closer to the use of modern technologies [20]. Serious games can be effective in reducing symptoms related to the disorder, the foregoing through a systematic review and a controlled meta-analysis of the literature [21]. The use of serious games for children with ADHD can be beneficial, because through these the educational practice can be more flexible, adjusting to the rhythm of work of children. In addition, serious games provide stimuli to achieve optimal performance and to attract their attention. An improvement in self-esteem is also achieved, because children can see their performance in the game and verify that they perform the activity correctly. Thus, through serious games, it is possible to favor co operation and self-control, increase motivation and improve self-concept [22]. Now, we present some success stories using these games in the treatment of children with ADHD reported in the literature. In their research, in [23], they found that serious games are fun and functional for children with ADHD; in [24], through his research examines the benefits of adding game elements to the standard computerized working memory; examined whether the elements of the game would improve the motivation and performance of children with ADHD. Therefore, in [25], they make use of serious games as part of behavioral therapy in children with ADHD. With this research it was observed that there was a more effective response in children who used serious games than in those who followed traditional therapies. In [26], they present a game whose objective is to improve in children aspects such as: waiting capacity, planning capacity, ability to follow instructions and ability to achieve objectives. The game is presented as a tool for the training of sustained attention in children with ADHD through the neuromodulation of the Beta and Theta waves through an electrode located in the central part of the frontal lobe of the brain.

185

The processing of an electroencephalographic signal occurs automatically within the video game, which allows generating a report of the evolution of the theta / beta relationship, a biological marker. [19] designed a game with which to promote behavioral learning and the use of strategies for self-control in daily life, such as time management, planning, organization and social skills that are known, are problematic for children with ADHD. Serious games have worked as a tool for the multi-modal treatment of children with ADHD, who often have problems related to motivation and react differently to other children in the face of rewards. Through serious games, a balance can be achieved between the elements of motivation, learning and behavioral and cognitive challenges, thus achieving greater and better participation in class. In the 1990s [27] he compared the effects of using two software packages for math and two readings to maintain the attention of children with ADHD. The study was conducted with children of third and fourth degree diagnosed. The objective of the study was to determine if the software could increase the attention of these children. The results indicate that attention was increased using a game format when the animation was not excessive. However, in this study an evaluation of the learning achieved was not carried out. In [28], they present a video game supported with augmented reality for the learning of logical-mathematical skills, demonstrating that the performance in the game is the same to children with special needs and achieving their integration in the learning process. On the other hand [29] He found in his experiment that computer games had a significant effect on increasing speed and attention in mathematical calculations; however, the effect of computer games on mathematical learning and learning stability was not significant, suggesting that computer games can be used as training aid along with teacher training. As a result of this systematic review, a gap was detected regarding research on the use and design of serious games focused on the achievement of a meaningful learning of mathematics for children with ADHD, which is why interest arises. for the design of these games attending the special needs of this sector of the population in order to achieve better learning and be at the level of their peers.

Model Driven Development for the development of Serious Games. Theoretical foundation

Model Driven Development (MDD) is a software development methodology, in which the software is developed not by writing code directly in implementation languages, but by constructing high-level logic or logic models to represent aspects of the software, through which it is possible to transform models into more refined software artifacts. [30] [31] [32]. Through modeling, it is possible to understand a complex problem and its possible solutions through abstraction. Through MDD the automatization of many tasks is po ssible, reuse at the domain level, increase the quality of the models, make improvements, reduce costs using an automated process [33] [34].

186

"Model-driven development is simply the notion that we can construct a model of a system that we can then transform into the real thing" [33]. The MDD concept is related to the based-on models’ architecture that consists of a paradigm that incorporates a standard for the establishment of the components, which integrates a methodology for the development of interactive systems. These components are: models, language, software and approach [35]. The MDD for serious games can be beneficial for the dozens as it allows them to focus on the creation of models that represent their didactic skills. Serious games have a positive effect, because when played a high level of motivation is achieved in the player to continue and, this is beneficial in education and a high interest in learning is developed. Teachers use serious games to reach students, managing to stimulate their senses to achieve learning through the motivation to continue playing [36].

Serious game development

The present research is focused on the development of a serious game for meaningful learning in solving mathematical problems for children with ADHD. For the de sign of the game A journey through mathematics we have been working with 13 children between 8 and 12 years old, diagnosed with ADHD and 6 teachers who are responsible for the groups to which these children belong. The children are enrolled in a public sector school on the south side of the city of Puebla, Mexico, the work that has been done with them consists of semi-structured interviews with parents, group teachers, specialists in the diagnosis and attention of this disorder, as well as observation of the classes in order to understand the way in which they learn. This is how the idea of playing a game to reinforce learning in the solving of mathematical problems arises, through a narrative of how mathematics has been important in the timeline of different civilizations and how they were used. to solve situations of everyday life. Considering the elements that should be considered a serious game, as well as the elements of gamification. There is a main character that is Pythagoras who is telling the timeline in different civilizations. The design of the problems is aligned to the thematic contents of the Ministry of Public Education, which are adapted to the timeline. Some of the elements that must be considered during the design of a serious game are: Narrative description, player, interactive objects, player actions simulation environment feedback [37]. From the information described above, the elements considered for the design of the game are described below Table 1. Table 1. Serious game design elements. Element

Narrative description

Description A journey through mathematics is a journey through 9 civilizations presented by a main character Pythagoras (who makes a brief description of the importance of mathematics in each of the civilizations). The play er will discover the importance of the use of mathematics throughout history, through the

187

Player Interactive objects Player actions

Simulation environment

Challenges

M echanics

resolution of problems that are related to basic arithmetic operations. Children with ADHD who can choose a character based on gender. Character of Pythagoras doing the different narrations. Elements allusive to different civilizations Solve mathematical problems according to the indications given in each of the civilizations The game is centered on a two-dimensional design, the background and the elements are changing according to civilization. For example, a cave, a river and an Ishango bone appear in prehistory. The challenges in the game are to solve different problems whose complexity is increasing, with the purpose that the student acquires a significant learning in the acquisition of new mathematical notions For each problem that the character is solving, he will win 3 coins, otherwise he will not receive any. Once the resolution of the problem contemplated in each of the civilizations is concluded, it is credited with a suit alluding to it.

Once this information is available, we proceeded to design the task model under the Cameleon reference framework, whose structure is composed of four critical stages for the definition of user interfaces: tasks and concepts, abstract user interface, concrete user interface and final user interface. Figure 1 shows the model for the "Start of Game" task in which the user chooses between the subtasks of selecting civilization on the left and selecting options on the right. The subtask of selecting the option in turn is divided into nine interactive subtasks for the selection of one of the Prehistoric, Sumerian, Greek, Etruscan, Rome, Egyptian, Chinese, Indian and Mayan civilizations. On the other hand, the subtask selection option is subdivided into the tasks of requesting help, returning and leaving. Figure 2 shows the model for the task "student registration", here the user will be registered in the system, at the start you have the option to choose between the interactive tasks of selecting options, selecting to play or choosing the abstract task of entering data, where you must enter your name, surname, username, school name, age, select gender and enter the group.

188

Fig. 1. Game start task model.

Fig. 2. Student registration task model.

Figure 3 presents the model for the "Solve problem" task, the user will be able to visualize the problem, as well as the gamification elements such as time and rewards; At the same time the user will solve the problem.

Fig. 3. Problem solving task model.

From each of the elaborated task models, a concretization process is carried out to process the abstract user interface, considering that it is, like the task model, inde-

189

pendent of any modality or platform and being a graphic representation to better understand the behavior of the task model [35]. Figures 4, 5 and 6 show the abstract user interfaces to represent each of the options that the user can access from the main menu to carry out the activities outlined above.

Fig. 4. Abstract user interface task “start game”.

Fig. 5. Abstract user interface task “student registration”.

190

Subsequently, the user's concrete interface is made, which resembles a model of the final view. Figures 7, 8 and 9 show the concrete user interfaces developed following the concretization process on abstract-type interfaces [35].

Fig. 6. Abstract user interface task “solve problem.

Fig. 7. Concrete user interface “start game”.

191

Fig. 8. Concrete user interface “user registry”.

Fig. 9. Concrete user interface “solve problem”.

Finally, we proceeded to design the final interfaces of th e serious game for the acquisition of new notions in solving math problems for children with ADHD, the game is developed with UNITY which is an independent version of the internet, in figure 10 we show the initial window of the serious one, which links the different civilizations.

192

Fig. 10. Final user interface “start game”.

Figure 11 shows the student registration interface, which aims to register the student by collecting personal data, it is also possible that the user selects his gender and with it a character with whom he will play. Figure 12 shows the problem interface, in which a problem is presented, and the user can go on counting the elements writing their response and receive feedback.

Fig. 11. Final User interface user record.

193

Fig. 12. Final user interface end user record.

Conclusions and future viewings of the work

In this work “A journey through mathematics” was presented, a serious game for the acquisition of new mathematical notions for children with ADHD, through Model Driven Development and from the elements that should be considered for the design of a serious game. The activities presented on the game were designed based on the thematic contents of the Secretaria de Educación Pública, adjusted to a timeline of how mathematics has been important throughout history for solving problems in everyday life. At the same time, this game has been designed considering the special characteristics and requirements that are necessary to improve attention and achieve significant learning in this sector of the population. As future work, the evaluation of this game will take place using the Product Emotion Measurement Instrument (PrEmo) [37] through this instrument it is possible to measure more than one emotion experienced simultaneously. This ev aluation will be applied to the 13 children with whom they are working to measure their level of satisfaction when playing the seriously designed game.

References 1.

Bakker, L., Rubiales, J.: Autoconcepto en niños con trastorno por déficit de atención con hiperactividad. PSIENCIA. Revista Latinoamericana de Ciencia Psicológica, 4(1) (2012). Gutiérrez M aldonado, J., Letosa-Porta, À., Rus-Calafell, M ., & Peñaloza Salazar, C.: The assessment of Attention Deficit Hyperactivity Disorder in children using contin-

194

4. 5.

7. 8. 9.

10. 11. 12.

13. 14. 15.

16. 17.

18.

uous performance tasks in virtual environments. Anuario de Psicología, 2009, vol. 40, num. 2, p. 211-222 (2009). Hamari, J., Koivisto, J., Sarsa, H.: Does Gamification Work A Literature Review of Empirical Studies on Gamification. In HICSS Vol. 14, No. 2014, pp. 3025-3034 (2014), Breuer, J., Bente, G.: Why so serious? On the relation of serious games and learning. Journal for Computer Game Culture, 4, 7-24 (2010). Wouters, P., Van Nimwegen, C., Van Oostendorp, H., Van Der Spek, E. D.: A metaanalysis of the cognitive and motivational effects of serious games. Journal of educational psychology, 105(2), 249 (2013). Polanczyk, G., De Lima, M . S., Horta, B. L., Biederman, J., Rohde, L. A.: The worldwide prevalence of ADHD: a systematic review and metaregression analysis. American journal of psychiatry, 164(6), 942-948 (2007) Artigas-Pallarès, J., Narbona García, J.: Trastornos del neurodesarrollo (No. 616.8053.2). Viguera, (2011). Joselevich, E., Quirós, B. D.: AD/HD: qué es, qué hacer: recomendaciones para padres y docentes (2003) M enéndez, S. A., & González, A. P.: Trastorno por déficit de atención con hiperactividad en mi aula de infantil//Attention deficit hyperactivity disorder in my preschool classroom. REOP-Revista Española de Orientación y Psicopedagogía, 26(3), 141-152 (2015). Santos Ruiz, J. L., Sanz Rodríguez, L.: DSM -5 Trastornos del Neurodesarrollo. DSM -5: Novedades y Criterios Diagnósticos (2013) Leung, A. K., Lemay, J. F.: Attention deficit hyperactivity disorder: an update. Advances in therapy, 20(6), 305-318 (2003). M enéndez, S. A., González, A. P.: Trastorno por déficit de atención con hiperactividad en mi aula de infantil//Attention deficit hyperactivity disorder in my preschool classroom. REOP-Revista Española de Orientación y Psicopedagogía, 26(3), 141-152 (2015) UNESCO: Guía para Asegurar la Inclusión y la Equidad en la Educación. Paris: UNESCO, 2017, Francia. (2017) Abarzúa, A., Cerda, C.: Integración curricular de TIC en educación parvularia. Revista de Pedagogía, 32(90) (2011). Gutiérrez M aldonado, J., Letosa-Porta, À., Rus-Calafell, M ., & Peñaloza Salazar, C.: The assessment of Attention Deficit Hyperactivity Disorder in children using continous performance tasks in virtual environments. Anuario de Psicología, 2009, vol. 40, num. 2, p. 211-222 (2009). Abt, C. C.: Serious games. University press of America (1987). Prieto, R; M edina, N; Diaz, D; Jahiel, J. y García, C. Videojuegos serios en educación infantil y primaria) X Congreso Español sobre M etaheurísticas, Algoritmos Evolutivos y Bioinspirados (M AEB 2015) (2015). Céspedes-Hernández, D., Rodríguez-Vizzuett, L., González-Calleros, J. M ., M uñozArteaga, J.: Extension of a User M odel for Promoting the Development of Applications to Support Auditory Rehabilitation. In HCI for Children with Disabilities (pp. 53-70). Springer, Cham (2017).

195 19.

20. 21.

22. 23. 24.

25.

26.

27.

28.

29.

30.

31. 32.

33. 34.

Bul, K. C., Franken, I. H., Van der Oord, S., Kato, P. M ., Danckaerts, M ., Vreeke, L. J., M aras, A.: Development and user satisfaction of “Plan-It Commander,” a serious game for children with ADHD. Games for health journal, 4(6), 502-512 (2015). Echeverry Chaves, N.: Diseño de un videojuego didáctico de educación cívica para niños autistas, TDAH y discapacidad cognitiva (2014). Lau, H. M ., Smit, J. H., Fleming, T. M ., & Riper, H.: Serious games for mental health: are they accessible, feasible, and effective? A systematic review and metaanalysis. Frontiers in psychiatry, 7, 209 (2017). González, T.: TIC y TDAH. III Congreso Internacional Virtual sobre la Educación en el Siglo XXI (M arzo de 2018) (M arzo): 631-38. (2018) Kim, J. Y., & Bae, J. H.: A Study on Serious Game Technology Based on BCI for ADHD Treatment (2014). Prins, P. J., Dovis, S., Ponsioen, A., Ten Brink, E., Van Der Oord, S.: Does computerized working memory training with game elements enhance motivation and training efficacy in children with ADHD?. Cyberpsychology, behavior, and social networking, 14(3), 115-122 (2017). M artínez, F., Barraza, C., González, N., González, J.: KAPEAN: Understanding Affective States of Children with ADHD. Journal of Educational Technology & Society, 19(2) (2016). M uñoz, J. E., Lopez, D. S., Lopez, J. F., Lopez, A.: Design and creation of a BCI videogame to train sustained attention in children with ADHD. In 2015 10th Computing Colombian Conference (10CCC) (pp. 194-199). IEEE (2015). Ford, M . J., Poe, V., Cox, J.: Attending behaviors of ADHD children in math and reading using various types of software. Journal of Computing in Childhood Education (1993). M uñoz, H. T., Navarro, S. B., Fabregat, R.: Gremlings in M y M irror: An Inclusive AR-Enriched Videogame for Logical M ath Skills Learning. In 2014 IEEE 14th International Conference on Advanced Learning Technologies (pp. 576-578). IEEE (2014). M ahmoudi, H., Koushafar, M ., Saribagloo, J. A., Pashavi, G.: The effect of computer games on speed, attention and consistency of learning mathematics among students. Procedia-Social and Behavioral Sciences, 176, 419-424 (2015). Pham, H., Ferworn, A., M ahmoud, Q. H., Sadeghian, A.: Applying model-driven development techniques to the development of search and rescue systems. In 2007 IEEE International Conference on System of Systems Engineering (pp. 1-6). IEEE (2007). Selic, B.: The pragmatics of model-driven development. IEEE software, 20(5), 19-25 (2003). Tang, S., Hanneghan, M ., Carter, C.: A Platform Independent Game Technology M odel for M odel Driven Serious Games Development. Electronic Journal of eLearning, 11(1), 61-79 (2013). M ellor, S. J., Clark, T., Futagami, T.: M odel-driven development: guest editors' introduction. IEEE software, 20(5), 14-18 (2003). Atkinson, C., & Kuhne, T. M odel-driven development: a metamodeling foundation. IEEE software, 20(5), 36-4 (2003).

196 35.

36. 37.

Céspedes-Hernández, D., Pérez-M edina, J. L., González-Calleros, J. M ., Rodríguez, F. J. Á., & M uñoz-Arteaga, J.: SEGA-ARM : a metamodel for the design of serious games to support auditory rehabilitation. In Proceedings of the XVI International Conference on Human Computer Interaction (p. 10). ACM (2015). Thillainathan, N.: A model driven development framework for serious games. Available at SSRN 2475410 (2013). Desmet, P.: M easuring emotion: Development and application of an instrument to measure emotional responses to products. In Funology 2 (pp. 391-404). Springer, Cham (2018).

Exploring Teenager's Motivations as Digital Natives to Enhance Peace Culture Workshops Adherence Using Gamification Cristian Jurado, Mónica Llano, Edwin Gamboa, Yuri Bermúdez, Victor Valencia, Maria Trujillo Universidad del Valle, Cali, Colombia {cristian.camilo.jurado, monica.llano, edwin.gamboa, yuri.bermudez, valencia.victor, maria.trujillo}@correounivalle.edu.co

Abstract. Colombian teenagers are exposed to multiple violent situations. Thus, the Colombian foundation M ultipropaz tries to prevent such exposure by conducting formative workshops to promote a culture of peace and peaceful coexistence among teenagers. M ultipropaz uses an active methodology in which teenagers' continuous participation is key to achieve the expected results. However, M ultipropaz has detected a lack of adherence from teenagers to the workshops, which is evidenced by their intermittent participation. In this paper, teenager's motivations as digital natives are identified using a focus group and a semi-structured interview. The collected data is analysed using the thematic analysis method. The results of the analysis indicated that teenagers may feel motivated by interactivity, digital content, playfulness, competition, customisation and music. The results were validated with a group of teenagers and teachers from a school using a prototype, in which four game mechanics were included to address five of the identified motivations. The validation indicated that the participants considered that the prototype fulfil those motivations effectively. Therefore, the employed approach to identify teenager's motivations and to address those in an application might be an effective way to use gamification in social contexts. The developed gamified mobile application is planned to be extended to generate a more interactive environment for teenagers during formative workshops. The purpose of using this application is to meet teenagers' needs as digital natives to improve their motivation and adherence to workshops. Keywords: Teenagers, M otivation, Adherence, Peace promotion workshops, Gamification, Technology enriched workshops.

Introduction

Multipropaz is a non-profit organization born 16 years ago from the social project “Jóvenes Constructores y Constructoras de Paz Multipropaz” and led, at the time, by a group of former teenagers and teachers at the Institución Educativa Multipropósito in Cali, Colombia. Multipropaz leads social processes that foster teenager's participation and peaceful coexistence. The foundation focuses on working at the Comuna 20

198

of Cali Colombia, which is a zone composed of 12 neighbourhoods that belong to the working class; i.e., 1 (83%), 2 (14%), 3 (3%) Colombian social strata. Only 2.2% of that population reaches higher education and the majority (42.2%) only accomplish basic primary education level [18]. Consequently, academic and professional opportunities for them are minimal. In addition, teenagers from this zone are exposed to social conflicts such as drug addiction, thefts, gangs and violence. Multipropaz intervention is mainly based on preventing those problems by conducting formative workshops that foster the acquisition of conflict resolution skills through activities with an active and dynamic methodology to improve the leading, communication and to foster a peace culture among the teenagers. These workshops are conducted by members of the foundation (workshop facilitators) who are responsible for planning and leading the activities, in which the teenager’s continuous participation is key to achiev e the expected results. However, Multipropaz has detected a lack of adherence by teenagers to the workshops, which is evidenced by their intermittent participation. According to Robert Dubin [5], motivation is the complex set of forces starting and keeping a person at work in an organization. Motivation is something that moves a person to action and continues him/her in the already initiated course of action. This paper presents a study conducted to identify the teenager's motivations towards attending Multipropaz workshops. Those motivations were studied to design a user-centred gamified application as a complementary tool for the formative workshops conducted by Multipropaz. The application may be an interactive environment that satisfies the teenager's motivations as digital natives [19] and improves their adherence to the formative process.

Related Work

The following is a summary of relevant gamified platforms used in educative and social environments: 2.1 Cartas por la paz Created by Gestionet and Corporación Universitaria Minuto de Dios [12], Cartas por la paz is a trans-media project that teaches interactive concepts related to peace culture and conflict resolution, using two elements: physical (a set of cards) and virtual (a mobile application and a web portal). Besides peace concepts, the application includes augmented reality and audio-visual testimonies of diverse types of violence. The goal of the project is to generate a context that facilitates the search of a collaborative solution. Cartas por la paz consists of three components: 1. A Package of Physical Cards: This component is already implemented and consists of a card game where participants act as conflict managers. 2. Mobile App: This component expects to offer an experience of augmented reality, presenting examples about conflicts for the participants to propose possible solutions. 3. Web Portal: This component expects to be a meeting place for users to share their experiences.

199

2.2 Duolingo Duolingo is a free language learning platform created by Luis Von Ahn and Severin Hacker. According to its website, it has more than 30 million registered users and is the most downloaded education application in the world [6]. Duolingo uses some gamification elements (e.g., badges, leaderboards, points and rewards) to motivate and engage users [17]. 2.3

Tinycards

Tinycards by Duolingo is a gamified educational app designed to facilitate learning of words or concepts [9]. In Tinycards, users see the front and back of a set of cards; later, they answer a set of questions. During a Tinycards lesson travel, users get a package of cards, each one having a hidden side revealed when users tap on the card. During a lesson, the app will distribute different kinds of questions, to make sure that users memorise the expected information. 2.4

ClassDojo

ClassDojo is a communication application that connects teachers, parents and students during a school day, allowing to give feedback of student's interactions in realtime through a web or mobile device [20]. ClassDojo offer badges, gold stars or points, as extrinsic motivators. ClassDojo provides three accounts types, interconnected as follows: 1. Teacher Account: teachers register to access a free account with a virtual classroom, where they can manage the score of each student, the class story and parent's interaction. The Teachers can change the actions for which the students obtain feedback to adapt to the needs of the class or of the school. 2. Students Account: students start using the application using an access code. After creating an account, students can customise avatars, see classes and see individual progress. 3. Parent account: parents can also access ClassDojo when they are invited by teachers, each parent can supervise one or more students, check student's information and send messages to the teacher.

3 3.1

Methods Aim of the study

To identify teenager's motivations as digital natives to design a gamified prototype to assist the formative workshops led by Multipropaz.

200

3.2

Instruments

The study was conducted along two stages, data collection and analysis. Data collection instruments included a semi-structured interview [13] and a focus group [7]. The data analysis was conducted using the thematic analysis method [16]. 3.3

Data Collection

Participants. Ten teenagers who are regular participants of Multipropaz workshops and two workshop facilitators participated voluntarily during the data collection (seven males (58.3%) and five females (41.7%)). The ages of the participants ranged between 12 and 17 years (Mean: 14.5 years, Standard Deviation: 1.3 years). Eleven (91.7%) of them had a smartphone. The number of participants depended on the number of teenagers that attended to the activities conducted by the foundation th e day on which the study took place. Focus Group. A focus group was conducted to identify teenager's motivations regarding their participation in Multipropaz workshops. According to Denscombe [4], a focus group consists of a small group of people, who meet to explore attitudes, perceptions, feelings and ideas about a topic. One of the advantages of focus groups is obtaining information without using an instrument, such as a questionnaire or a survey, which may cause discomfort to interviewees and make them feel analysed. As a result, an environment of interaction and reliability with the group of teenagers was achieved. The focus group was conducted following the approach presented in [7], which is summarised as follows: 1. Opening and presentation: the focus group is carried out in a quiet place within the school where Multipropaz conducts its work. First, we introduced ourselves and explained the study purpose. The workshop facilitators did not participate in this session to foster teenagers’ freedom of speech. Also, we explained to the participants that their answers will be used anonymously and for academic purposes only. The session lasted approximately 30 minutes. 2. The beginning of the conversation: to generate an appropriate environment in which teenagers feel comfortable and confident, we started the co nversation using a question (i.e., What do you like to do in your free time?) and continued according to the participants’ answers and the structure presented in Table 1. We tried to encourage participation by showing a tolerant and accessible attitude to any response. 3. Script structure: after formulating the first question, we continued formulating questions based on the script showed in Table 1. 4. End of interview or closure: we finished the interview by thanking teenagers for their time and attitude.

201 Table 1. Focus group script structure.

Conversation topics Favourite free-time activities (listen to music, study, etc) Attraction to mobile applications Liked and disliked mobile apps Social networks use Expectations towards a gamified application

Specification Why? Which and Why? Which and Why? How often? Which and Why?

Interview. Two workshop facilitators at Multipropaz were interviewed individually to recognise their experience and knowledge regarding teenager's motivations towards workshops. Both participants were informed about the purpose of the study and its use in the development of a gamified mobile application. The interview was guided using the following questions: 1. In your opinion, what are teenager's motivations to participate in a workshop? 2. What activities are the most attractive to teenagers during a workshop? 3. In your opinion, what is more motivating for teenagers, a group o r an individual activity? 4. Would you use a digital application as an assisting tool in the workshops you facilitate? 5. Do you think that a digital application may motivate teenagers to participate in a workshop? 6. What elements should have a digital application to assist workshops? 3.4

Analysis

Participants. Four 5th -year systems engineering students, a 2nd -year computer science master student and a 1st -year computer science doctoral student participated in the thematic analysis of the collected data. Procedure. The collected data was transcribed in a document (as a list of separated sentences), which was analysed using the thematic content analysis method [16]. First, each sentence was encoded using a brief description of what was said by the participants. Then, these sentences were grouped iteratively using an affinity diagram [2]. Affinity diagrams allow developing a system to organise a large amount of information [3]. To achieve the goal of this study, the analysis was guided by the following research question: “What teenager's motivations may be fostered by an application that assists Multipropaz workshops?”

Findings

After coding the collected data and grouping them using an affinity diagram, we obtained six possible motivations, which are described below and illustrated in Fig 1.

202

Fig. 1. Motivations identified through an affinity diagram. 4.1

Interactivity

During the focus group, the participants expressed that they use the internet frequently (i.e., more than six hours per day), mostly in social networks, such as Facebook, Instagram, WhatsApp and Twitter. They like sharing images and videos, reacting to content (i.e., using the reactions provided by each social network, such as Like or Love), creating and consuming memes, and communicating with other people. On the other hand, the two facilitators claimed that teenagers also use social networks constantly during the workshops, leading to distraction. However, they state that using social networks in a gamified application may result in fostering teenagers’ motivation during the workshops lead by the foundation. Furthermore, the facilitators said that a mobile application may allow creating a more interactive environment during the workshops. They would like an assisting application to allow them to assign, review and track activities. For instance, they would like teenagers to be able to answer questions and submit evidence of their work etc. 4.2

Digital Content

The two facilitators claimed that teenagers constantly use their smartphones during Mutipropaz workshops. Similarly, the teenagers explained that they use their smartphones during these workshops to review and update their social network profile, watch videos, share images, read and communicate with friends and family. In addition, they stated that they are used to using smartphones and computers every day for completing academic tasks (e.g., making animated presentations and writing essays), learning about the world, finding out national news and getting entertained.

203

4.3

Playfulness

The interviewed teenagers showed an interest in playing sports (e.g., soccer and basketball) and play on their smartphones. Some of them use Duolingo [6] to learn English since they consider learning through games more enjoyable than in a traditional context. Also, they prefer to play on a smartphone since it offers portability and access to a wide variety of games for free. The games they play the most are Free fire [1]; Minecraft [15] and Piano tiles [21]. On the other hand, workshop facilitators believe that digital games could improve teenager's motivations and could create a playful environment in which teenagers learn while having fun during Mutipropaz workshops. 4.4

Music

During the focus group, the teenagers remarked that they listen to music frequently and use Youtube [10] to watch music videos and create playlists of their favourite songs (mainly on their smartphones). According to them, music allows them to express themselves and relax during their daily activities. Also, they expressed that they would like to hear their favourite songs on mobile educational applications. Similarly, the workshop facilitators confirmed that teenagers use headphones frequently during Mutipropaz workshops to listen to music and that they use music to express their thoughts during these workshops . 4.5

Customisation

All the teenagers who participated in the focus group have at least one social network profile, which they personalise to express themselves and their interests (e.g., profile pictures, selfies, videos, photos, posts, etc.). They like to invite their friends to view their profile and comment on their posts.

Validation

To validate the findings of the previously presented study, we conducted a second focus group involving 2 facilitators and 16 teenagers at the same school. Two of them participated during the initial interviews. We provided the participants with the final identified themes after the thematic analysis along with sample sentences extracted from the interview and focus group sessions. Furthermore, a gamified mobile prototype was developed to assess whether five of the motivations identified previously (i.e., digital content, interactivity, playfulness, competition and customisation) may be fulfilled using gamification. The motivation that was not addressed in the prototype (i.e., musical) will be included in a future phase of the application. Finally, a user experience (UX) evaluation was conducted to validate whether the motivations were fulfilled or not by the prototype. The details of the prototype are presented in Section 5.1 and the evaluation is presented in Section 5.2.

204

5.1

Prototype Description

A prototype of a gamified mobile application was developed for the Android operating system. It includes an avatar and a core mechanic called RaceQr, which includes two sub-mechanics (i.e., points and a leader-board). The relation between those mechanics and the addressed motivations are illustrated in Fig 2 and described below: 1. Avatar: this mechanic is a graphical representation of a user. In the prototype, this mechanic will allow teenagers to represent themselves using customisation features. It is accessed through the main menu of the application. 2. RaceQr: it is the core mechanic of the prototype. The RaceQr was based on the need of workshop facilitators to evaluate knowledge acquisition of teenagers regarding the concepts delivered during workshops. It consists of a time -trial race of questions, in which facilitators can create a set of questions and the associated QR codes. Meanwhile, when teenagers access a RaceQr, they should find and scan the codes to answer the questions. This mechanic is intended to meet the competition, interactivity and playfulness motivations of teenagers. Also, the RaceQr includes the following sub-mechanics: a. Points: this sub-mechanic is a numerical representation of the progress of teenagers using the prototype. During the race, teenagers score p oints when they answer a question correctly. These points will be used in future versions of the application to personalise the avatar included in the prototype. b. Leaderboards: this sub-mechanic consists of a board presenting the positions of a group of participants in a competition. In this prototype, the board presents the order of arrival of participants during a race. It means that all participants can reach the same score if they answer all the questions correctly.

Fig. 2. Used mechanics to support motivations (Icons made by Freepik, Flat Icons and DinosoftLabs from www.flaticon.com is licensed by CC 3.0 BY).

205

5.2

UX Evaluation

Participants. Sixteen teenagers of grades 6th and 7th (6 Females and 10 Males), participated in the evaluation voluntary. These teenagers are different from the group of teenagers who participated during the design phase; however, they are from the same school. The participants are between 12 and 14 years old (Mean: 12.5 years, Standard Deviation: 0.87 years). Evaluation settings. The evaluation was conducted at Multipropósito school during a workshop led by Mutipropaz and included the following stages (See Table 2): 1. Before interaction: first, we introduced ourselves and explained the participants that the goal of the evaluation was to test the prototype and not their skills or knowledge. Then, an “icebreaker” activity was carried out to foster a comfortable environment. 2. During interaction: later, the main activity called RaceQr (See Section 5.1) was carried out. To execute this activity, we asked the teenagers to work in groups and select a leader as requested by the facilitators. Each leader was given a smartphone with the prototype installed. The activity lasted approximately 20 minutes. 3. After interaction: this stage was divided into two sections. First, a focus group was conducted involving the teenagers to know their opinions, feelings and suggestions about the prototype and the performed activity. The focus group lasted approximately 15 minutes and was recorded for further analysis. After that, the participants filled out a demographic questionnaire and a 5-point Likert scale (See Table 3) to assess to what extent each identified motivation was fulfilled by the prototype. Table 2. Evaluation settings of the interaction.

Evaluation Moment Before Interaction During Interaction

After Interaction

Employed Instruments

Evaluati on Goal

Ice breaking activity Video recording Field observation [11] Question asking protocol [13] Focus group [7] Likert scale questionnaire [14] Demographic questionnaire

Evaluation preparation Take evidence of UX Gather feedback regarding interaction Gather feedback regarding motivations Collect demographic data

Table 3. M otivation approached with 5-point likert scale questionnaire.

Id Q1 Q2 Q3 Q4 Q5

Question Including a smartphone in the activity was fun It was satisfying to perform a Mutipropaz activity using a smartphone I would like to carry out other activities using a smartphone It was fun to answer the questions through the mobile application The application allowed me to interact with other classmates

Motivati on Digital Content

Interactivity

206

Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14

5.3

The mobile application was easy to use The application generated a playful and learning environment The Mutipropaz workshop was more enjoying using the mobile application I liked getting rewards when I finished the activity The application allowed me to compete against my classmates sanely Competing against others was fun I would like to compete against others again liked that the application allowed me to have an avatar Personalising an avatar allows me to express myself

Playfulness

Competition

Customisable

Findings

The results of the 5-point likert were analysed as presented in Table 4. Those are described in the following sections. Table 4. Interpretation and assessment of motivations fulfilment based on participants answers

Assessment Properly fulfilled Not properly fulfilled

Interpretation Very well fulfilled Well fulfilled Poorly fulfilled Very poorly fulfilled

Provided Answer Strongly agree Agree Neutral Disagree Strongly disagree

Motivations fulfilment. As shown in Fig 3, the findings suggest that most teenagers (87,5%) considered that their motivation towards digital content was properly fulfilled by the developed prototype. Only 12,5% of them considered that is was not properly fulfilled. We obtained similar results for the interactivity motivation, 93,75% of the teenagers assess it as properly fulfilled, while the remainder (6,25%) did not. Similarly, most of the participants provided a positive feedback regarding playfulness (93,75%) and competition (87,5%). However, some of them considered that those motivations were not properly fulfilled (6,25% for playfulness and 12,5% for competition). In comparison with the above motivations, the feedback regarding customisation is quite more moderate. 68,75% of the teenagers considered that it was properly fulfilled, whereas 31,25% of them did not.

207

Fig. 3. Participants answers by motivation fulfilled

Feedback obtained by participants. During the focus group, the participants stated that using a mobile application with game elements as an evaluation tool is much more motivating than using a written exam or paper-based exercises. They also found it positive, that moving around the school was a core mechanic of the RaceQr. They consider it as a fun way to work in groups, answer questions and get distracted from the common academic environment while using a mobile application. On the other hand, the avatar included in the prototype was not perceived by most teenagers. Some relevant excerpts from their feedback are given below: 1. “I liked being evaluated through a question race, because I had fun and didn't get nervous”. 2. “I liked doing a race on my smartphone, it was so much fun”. 3. “The avatar should show up at the beginning of the application, I didn't find it”. In general, they consider that the application should be used during the Mutipropaz workshops since they would like to carry out other activities in a similar way.

Discussion

Our findings showed that teenagers may be motivated by digital content, interactivity, playfulness, competition and customisation and music. A gamified prototype was developed to address those motivations except for the last one. The prototype was validated with a group of teenagers. In general, they considered that the addressed motivations were fulfilled by the prototype in a fun way. The most relevant reason for this result may be the use of a user-centred design approach. The prototype was developed in collaboration with Mutipropaz’s members and considering the feedback provided by a group of teenagers who participated in the study to identify their motivations. The result of this was a mobile application that addresses facilitators’ and participants’ needs. Also, the application integrates properly into current facilitators’ methodologies, rather than requiring a complete change. Moreover, we combined well known gamification mechanics (e.g., points and leader boards) with an augmented

208

realty based QR race to generate and interactive environment, which may be novel for teenagers since they are used to conventional mobile apps (mainly games and s ocial networks). Finally, they expressed that the prototype was intuitive and easy to use, which may contribute to the positive findings. On the other hand, the results regarding customisation were quite moderate. Some of the participants did not even notice that the prototype included an avatar, which may be expected since the central activity was focused on using the RaceQr mechanic. Also, the avatar customisation may be a mechanic to use individually and maybe not during the workshops. Moreover, the participants could interact with the application during 20 minutes and the application was mainly controlled by the group leader. Finally, teenagers and facilitators expressed their eagerness to continue using the prototype in future sessions and request to include more mechanics to enrich the experience. 6.1

Limitations and Future Work

The most relevant limitation of our work is that the data was collected from one evaluation session of a gamified prototype. Consequently, most of the findings may be short-term responses that may change after a repeat use of the prototype. To assess teenagers' long-term motivation and adherence to Mutipropaz’s workshops a study monitoring the use of the application for a significant period (e.g., one academic year) may be required. Also, more mechanics may be required to avoid overusing the RaceQr, which in turn requires a deeper knowledge of Mutipropaz’s context to facilitate capturing particularities to develop proper and relevant content. Regarding the obtained positive results, we argue that some of the feedback might be biased. First, only one smartphone was available per group and the whole interaction lasted 20 minutes. Thus, individual interaction times might be short. Additionally, the group of teenagers who tested the prototype did not have previous experiences with a similar technology; therefore, the novelty of the application may have biased their point of view. Although all participants belong to the same school and social context, there is an age gap between the teenagers who participated in the study to identify motivations (12 -17 years old) and those who tested the prototype (12-14 years old). Thus, the view of older teenagers is to be validated. Finally, the number of teenagers and facilitators who participated in t he identification and validation studies is still limited. Similarly, the design protype was intended to meet the needs of the workshop facilitators. Therefore, our results cannot be generalised.

Conclusion

This paper presented a qualitative study to identify teenagers’ motivations toward attending Mutipropaz workshops. We confirmed that design a user-centred strategy could be useful to fulfil those motivations inside an application. The evaluation participants considered that a user-friendly and intuitive interface and the mechanics included in the application allowed meeting Mutipropaz expectations. Finally, this

209

study provides a first approach to how digital tools could improve teenagers’ interest to peace promotion processes aimed at specific social contexts.

Acknowledgements The authors would like to acknowledge the facilitators at the Mutipropaz foundation and the teenagers at the Multipropósito school for their engagement and feedback provided during this study.

References 1. 111dots Studio: Garena Free Fire. Best survival Battle Royale on mobile! 2. Cambridge Online Dictionary: Significado de GIT en el Diccionario Cambridge inglés. 3. Dam, R., Siang, T.: Affinity Diagrams – Learn How to Cluster and Bundle Ideas and Facts | Interaction Design Foundation. 4. Denscombe, M .: The Good Research Guide For Small Scale Research Projects. (2010). https://doi.org/10.1371/journal.pone.0017540. 5. Dubin, R.: Human Relations in Administration. PrenticeHall of India. (1974). 6. Duolingo Inc: Duolingo: Aprende inglés, español y otros idiomas de manera gratuita. 7. Elejabarrieta, F. et al.: Tractaments penitenciaris per fases : la visió dels afectats. Cent. d’Estudis Jurídics i Form. Espec. la General. Catalunya. (1991). 8. Foundation, I.D.: How to Conduct Focus Groups | Interaction Design Foundation. 9. Frederic Lardinois: Duolingo wants to reinvent flashcards with Tinycards | TechCrunch. 10. Google et al.: YouTube. 11. Hancock, B.: Trent Focus for Research and Development in Primary Health Care: An Introduction to Qualitative Research. Res. Dev. Gr. NHS Exec. Trent. June, 1–27 (2011). https://doi.org/10.1136/ebnurs.2011.100352. 12. Johnny López: Presentamos “Cartas por la paz”: un proyecto realizado para la universidad colombiana M inuto de Dios. 13. Leech, B.L.: Asking questions: Techniques for semistructured interviews. In: PS Political Science and Politics. (2002). https://doi.org/10.1017/S1049096502001129. 14. M cLeod, S.: Likert Scale | Simply Psychology. 15. M ojang: Sitio oficial | M inecraft. 16. M ortensen, D.: How to Do a Thematic Analysis of User Interviews | Interaction Design Foundation. 17. M unday, P.: THE CASE FOR USING DUOLINGO AS PART OF THE LANGUAGE CLASSROOM EXPERIENCE. RIED. Rev. Iberoam. Educ. a Distancia. (2015). https://doi.org/10.5944/ried.19.1.14581. 18. M unicipal, C.D.A. de P. et al.: Una mirada descriptiva a las comunas de Cali. (2007). 19. Prensky, M .: Teaching Digital Natives: Partnering for Real Learning. (2010). 20. Rip Empson: ClassDojo Lands $1.6M From Paul Graham, Ron Conway To Help Teachers Control Their Classrooms | TechCrunch. 21. Zeng, H.W.: Piano Tiles. Cheetah M ob. (2014).

ARTUI: An Augmented Reality-Based Application for English Teaching to Children between 10 and 12 years Liliana Rodríguez Vizzuett 1 , Josefina Guerrero García2 , Iván Olmos Pineda2 1

Facultad de Ciencias de la Electrónica, Facultad de Ciencias de la Computación Benemérita Universidad Autónoma de Puebla, M éxico 2

liliana.rodriguezvizz@alumno.buap.mx,{joseguga01, ivanoprkl}@gmail.com

Abstract. One of the main challenges associated to the foreign language teaching-learning process is lack of motivation from students. In the literature it is reported that it is possible to promote motivation towards such process through the realization of educational activities on familiar environments created by using animations and digital sounds. In this paper, the process that was followed for the implementation of an Augmented Reality -based application to support the development of the listening comprehension skill in children between 10and 12-years learning English as foreign language is presented, considering educational content provided by certification institutions and technological elements. A preliminary assessment of interface prototypes by driving an experiment using the Wizard of Oz technique was also addressed. Keywords: Augmented Reality, Tangible User interfaces, English Language Learning.

Introduction

According to information from the Mexican organization Instituto Mexicano para la Competitividad [1], only 5% of Mexican population is able to understand English language. Moreover, Mexicanos Primero (http://www.mexicanosprimero.org/) reports that 14.7% of the 50,000 English language teachers in public institutions do not understand such language, and that 23.8% of them reach the A1 level from th e Common European Framework of Reference for Languages (CEFR) [2] which is the expected level for students enrolled in the fourth degree of the elementary school. English language teaching was not considered a mandatory subject in the Mexican educational model for 2016 [3], however, it is in the 2017 version implemented since 2018, leading to think that new strategies and tools are required as means to support teaching tasks. The Mexican national program for English language teaching in elementary education has determined that the abilities to be developed during the fifth and the sixth grades of elementary school are equivalent to the level A2 of the CEFR [2]. For the purposes of this research work, only abilities related to the listening comprehension skill presented in Table 1 are considered.

211 Table 1. Listening comprehension ability for the A2 CEFR level. Listening comprehension ability Can understand phrases and expressions related to areas of most immediate priority (e.g. very basic personal and family information, shopping, local geography, employment) provided speech is clearly and slowly articulated. Can understand enough to be able to meet needs of a concrete type provided speech is clearly and slowly articulated.

Considering the above-mentioned problematic, and with the purpose of provid ing a tool to aid to solve it, in the literature it is possible to find works reporting the advantages of using Augmented Reality (AR) to promote motivation and as a mean to reinforce knowledge acquisition [4] [5]. Furthermore, in [6] it is reported that t he use of AR in the education domain, allows the creation of contents otherwise unviable, consents the realization of activities in in-door environments, and promotes collaboration and interactivity.AR is an interaction paradigm, which is result of the combination of the processing capabilities of computers, and elements from the user’s real environment [7]. In recent years, a number of prototypes and applications relying on AR had been created in areas such as computer graphics, Human Computer Interaction (HCI), computer vision, and collaborative work including the following sectors: architecture, engineering, planning, games, and education [8-10]. Aligned to the exposed problematic, considering the characteristics of the Mexican context, and inspired by the positive insights regarding motivation and aiding on the diminution of negative effects like anxiety and frustration, the objective of this work is to present the design of an AR-based application for English language teaching to children between 9 and 12 years specifically to develop listening comprehension skills. The viability of the solution is supported by the fact of that in Mexico, the introduction of commercial smartphones made AR-based applications available to several population strata. Information provided in 2015 by the Mexican national statistics and geography institute (INEGI) shows that 55.2% of Mexicans aged 6 and above, already had access to smartphones making them potential users of such type of applications. The remainder of this paper is structured into sections. Section 2 presents the state of the art including important concepts for the realization of this work. In section 3, we present the design process of ARTUI describing user stories and the description of the functionalities of each of the modules to be developed. Section 4 is dedicated to the design of an assessment for the developed application and to describe the gathered results. And finally, in section 5 conclusions and future work are provided.

State of the Art

In this section, a literature review on relevant concepts for understanding the proposal is presented including technology applied to the English teaching domain, AR applied to the education field, tangible user interfaces, gamification, and related work.

212

2.1 Technology Applied to the English Language Teaching Domain During the decades of 1970 and 1980, computer assisted language learning was developed with the support of teaching methodologies which encourage this process through predefined exercises [11]. According to [12], this type of learning may be addressed from three approaches: the behaviorist approach, based on observation, practice, reinforcements, and habits formation; the communicative approach, supported by cognitive theories which emphasized that learning is a process of discovery, expression and development; and the integrative approach, using multimedia as support to the teaching activities, and relying on the Web, emphasizing the use of the language in a real-life context [13]. Among the technological developments implemented, it is possible to identify electronic books, management tools, webquests, and forums. Later, between 1990 and 2000, the terms e-learning presenting tools for remote education [14], and b-learning as a merge of in-classroom and virtual education were introduced [15]. During this decade, platforms for the development of abilities and social networks for sharing information were created. Finally, after 2010 m-learning was presented as a teaching modality which allows to create remote learning environments through the use of mobile devices [16]. The development of mobile-based applications for language teaching enables student interaction and motivates them through the use of gamification components. The massification of mobile devices has encouraged the implementation of ARbased applications for educational purposes which are addressed in the following subsection. 2.2 Augmented Reality Applied to the Educati on Domain With the purpose of understanding the impact of AR when applied to the educational context, we performed a systematic literature review. The results of such review in terms of the objectives and results of the analyzed works are reported in Table 2. 2.3 Tangible User Interfaces The term Tangible User Interfaces (TUI), coined by Fitzmaurice and defined as a physical identifier for accomplishing virtual tasks were first described as comprehensive User Interfaces [31]. TUIs augment the real physical environment and add digital information to common physical objects [32]. The objective pursued with the implementation of TUIs is then, to manipulate digital information through the use of everyday objects which are present in users’ contexts. TUIs allow to realize physical representations through technology, making virtual environments more real and intuitive for users. Multiple authors refer that physical action is important during learning processes and that tangible objects may provide opportunities to think over the application of the obtained knowledge and the real world.

213 Table 2. Results of the literature review on AR and the educational domain. Works [1719]

[2128]

[29] [30]

O bjectives T o create AR-based tools to support the learning of different concepts and to perform studies for understanding the advantages of the use of AR for teaching different subjects. T o create augmented books supported by 3D resources, to develop tools to support teaching processes, and to implement applications overlapping virtual resources on real objects. T o create tools for teaching sciences. T o compare the traditional teaching processes with AR assisted processes.

Results Improvement of the language learning process. AR-based applications enable teachers to perform better presentations. By enriching educational processes with the use of AR, students report to have fun while learning, making it easier for them. AR is useful as support for t eaching as it allows the development of tools which use motivat es students. It was observed that attention levels and general satisfaction were increased in educational environments supported by AR applications. Proposition of a development process for allowing AR use in educational domains using free Software. Children making use of AR-based applications are more engaged as they are encouraged to create and control instead of only analyzing and describing.

TUIs are employed for reaching a better performance in learning tasks and may also be an alternative to Graphic User Interfaces as they allow direct manipulation, enabling users to control applications and to navigate through data by selecting and posing physical objects and not only representations of them [33]. The use of TUIs makes it possible to obtain information through: augmented physical surfaces like walls, desks, ceilings, and windows; graspable objects e.g. building blocks, models and instruments; and environmental features such as light, sound, air flow, and water flow. In the domain of education, TUIs provide benefits like reduction in the time to learn the use of the interface; diminution of the cognitive effort when learning, giving the opportunity of focusing on the content rather than on the use of the application; and a continuous representation of objects of interest by applying quick, reversible and incremental actions which are immediately visible [33]. Finally, TUIs enable multiple users to interact with an application by sharing objects and hence, promoting collaborative work and natural interaction [33]. A specific type of TUI requires mobile computing and wireless technology like Bluetooth, Radio Frequency Identification cards, Near Field Communication (NFC) tags, infrared beacons, WiFi, etc. In this specific proposal NFC tags are being used. NFC was developed by Philips® and Sony® in 2002 and consists on the integration of mobile phones and radio frequency technology and provides simple and safe communication between devices. NFC allows to automatically identify products, components, animals, people and everyday objects using tags [34]. 2.4 Related Work The introduction and massive availability of mobile devices make AR-based applications accessible to different population strata. Considering this fact along with the already described problematic regarding the impact of AR on the motivation towards the learning of a foreign language [5], and the intention of proposing an strategy based on the use of such applications as a means to reinforce the learning process, in Table 3 it is possible to find a summarized comparison of related works on the use of

214

AR and its application to the English language teaching domain in terms of their objectives and the obtained results. Table 3. Related works of AR applied to the English language teaching domain. Works [5, 35-39] [4042]

[4345]

O bjetives T o develop AR-based tools to incentivize reading and for teaching basic vocabulary and grammar to nonnative English speakers. T o implement AR-based systems in order to analyze teachers’ perception towards such type of technology and for enhancing students’ English level, and to create a virtual classroom with the use of AR based on learning theory. T o create teaching material based on the content of text books combining AR with ordinary resources.

Results Motivation towards language learning was increased and educational material was created. It was concluded that the supervised use of this type of tools reports benefits. Students reported having fun when interacting with such type of technology, they feel less pressured and seem to be more interested in the class. ARbased games showed a positive impact in the educational process, specifically for progressive recognition of words and concepts. Learning was improved and enriched as motivation increased. Besides, researchers found that the use of AR is recommended as support to learning processes independently of the educational level.

2.5 Gamification The concept gamification is derived from the word ‘game’ and was coined during the decade of 2000, being exploded mostly during the second half of the decade of 2010. Gamification is the use of game elements in no-game contexts, and it is aimed at making tasks that may result boring and unattractive into interesting and entertaining ones without modifying their main objectives. Games make use of mechanics and dynamics which, when well applied, are responsible of making them attractive to players. Mechanics on one hand, are strategies and tools used for satisfying players’ needs, this is, rules, techniques and actions used for making games attractive to players. On the other hand, dynamics refer to special needs and motivations that users may have for using the created game [46]. Gamification relies on that those dynamics and mechanics may be implemented in other areas for changing user attitudes in different contexts. In the educational domain, gamification consists of enhancing learning processes through the addition of game-like resources. In the specific case of computer supported learning, it is intended to apply resources typical of videogames for developing effective teaching processes the integration and motivation towards the content of the applications while promoting creativity [47]. In the next section, we address the design of ARTUI, an AR-based application relying on TUI and which content was created according to material provided by language certification entities and incorporating the use of gamification resources .

Application Design

As part of a spiral Software development process model, in order to gather requirements, interviews with final users (teachers and students of English language) within three educational institutions were scheduled. The interviews, following a qualitative

215

approach were created specifically to allow understanding the main challenges that actors face during the English language educational process. In this sense, teachers were also asked about the didactic material that they currently use when teaching and about their perception towards the use of instructive Software to support their classes. From the interviews, the following requirements and clarifications were identified: 1. The application will be used by children between 9 and 12 years as they are completing elementary school and according to the official educational program, they must reach an equivalent to the A2 level of the CEFR. 2. The application to be implemented will be focused on the development of listening comprehension skills. 3. The content to be used should be based in material provided by entities for preparing language certifications. 4. The typography to be used must be understandable by children. 5. The used vocabulary should be adequate to the level A2 from the CEFR. 6. Colors to be used in the application should attract children’s attention. 7. The application should not rely on the use of an Internet connection. 8. It is necessary to provide a teacher role in the application in order to allow analysis of the progress of each student individually and as part of a team. Table 4. Sign up and Login User Stories. Id

User story

Description

User sign up (teacher)

User sign up (student)

As a teacher, I want to sign up to the application entering my personal data, selecting the school to which I’m enrolled and the group to which I teach. Besides, I want to choose an avatar, so my profile can be easily identified. As a student, I want to sign up to the application entering my personal information, selecting the group and school to which I belong to, and selecting an avatar to distinguish my profile. As a registered student, I want to login to the application with my user name. As a registered teacher, I want to login to the application with my user name. As a logged student, I want to see the available activities within the application and be able to select any of them. As a logged student, I want to be able to choose my team mates from a list, name my team, and select an avatar so we can be distinguished from other teams.

4 5 9

Gamification resource Avatar

Avatar

NA NA NA Avatar

From the above-mentioned requirements along with the characterization of students and teachers and after observation of English language teaching sessions, it was decided to develop an application entitled ARTUI. ARTUI was conceptualized as a mobile application enclosing two main activities. The first one is a word to 3-D model dictionary in which, supported by AR, students may be able to find words associated to the A2 level of the CEFR taking Cambridge’s “Flyers” book as reference, and that are useful for interacting with the other module within the app. The second activity within of ARTUI consists on telling the student a story and evaluating if he/she was able to comprehend it by asking related questions. Students should answer using NFC tags stuck to each of the faces of a cube which works as TUI.

216

After gathering and describing functional and non -functional requirements, selfexplanatory user stories were created in order to have a more organized and precise representation of them. Such stories, presented in Tables 4 to 8, resulted on the implementation of 4 modules in the first iteration: user sign up and login, “Dictionary” activity, “Story” activity, and evaluation; as well as on additional modules to be implemented in a second iteration. Table 5. Dictionary Activity User Stories. Id

User story

Description

“Dictionary” activity selection “Dictionary” activity presentation

As a logged student, I want to select the “Dictionary” activity from a list. As a logged student, I want to see a list of words in English in the screen, and that when I select one of them, either a related 3D model, video, image, or audio is displayed.

Gamification resource NA NA

Table 6. Story Activity User Stories. Id

User story

Description

“Story” activity selection

“Story” activity presentation

As a logged student, I want to select the “Story” activity from a list. As a logged student, I want to be able to select the story that I want to listen to. As a logged student, I want to listen to the story which I previously selected.

Gamification resource NA NA NA

Table 7. Evaluation of the Story Activity User Stories. Id

User story

Description

Select evaluation of the “Story” activity Show evaluation of the “Story” activity Answer evaluation of the “Story” activity

As a logged student, I want to advance to the evaluation section of the listened story.

13 14

Gamification resource NA

As a logged student, I want to listen to questions about the listened story.

As a logged student, I want the system to show answer options to the asked question. I want to be able to select any of those options and receive either feedback or rewards depending on the correctness of my answer.

Feedback, use of rewards.

Table 8. User Stories to be implemented in the second iteration. Id

User story

Description

Get reports

Select level of difficulty

As a logged teacher, I want to be able to see reports about activities by student and by group. As a logged student, I want to be able to select a level of difficulty and to see it reflected on the available stories and their related questions.

Gamification resource NA NA

Next subsections are dedicated to describing the implemented modules according to the presented user stories.

217

3.1

This module is divided into two parts. The first one allows new users to sign up to the application. To this purpose, users s hould specify whether they are students or teachers. As it can be seen in Fig. 1, in case a student is signing in, the application as ks him/her for his/her first name, last name, a user name, age, gender, the group t o which they belong to, and an avatar with which the student will be represented. As shown in Fig. 2, if the user is a teacher, he/she will be asked to provide first name, last name, user name, the school in which he/she is enrolled, and the groups to which he/she teaches.

Fig. 1. Student’s sign up interface.

The second part of this module consists on allowing users to log into the application using the username they provided when creating their accounts. The use of this functionality is simple, when starting the application users should input their user names into a text box in the screen and tap on the “Login” button. The application will validate the existence of the user in the database. The implementation of this module required the design and development of a database using MySQL, the connection to the database is managed using PHP and integrated to the Unity game engine in which the project runs.

Fig. 2. Teacher's sign up interface.

218

3.2

Dictionary Module

In the dictionary module, which user interface is presented in Fig. 3, a word-object dictionary supported by AR is provided. Once users are logged in and t hey selected the “Dictionary” activity from the menu, they may see a list of available words, as well as the camera view. For using this feature, a TUI is given to users in the form of a cube. Two of the cube faces show ARTUI’s application logo. After selecting a word and pointing the camera to any of the above referred cube faces, users are able to listen to the correct pronunciation of the word, and to see a 3D model associated to it.

Fig. 3. Dictionary module interface with cube as TUI.

The development of this module was achieved using the Vuforia platform and its integration to the Unity game engine, as well as with the addition of free 3D models available in online libraries. The audio resources were recorded by native speakers of the language. The technical challenge which was faced, was on associating one single marker with multiple resources depending on users’ choice. 3.3 Story Presentation Module The story presentation module, illustrated in Fig. 4, presents multiple stories to users and is intended to aid on the improvement of the listening comprehension ability. To this purpose, images and audio resources are used. The stories included in the ARTUI application are extracted from material provided by certification an d educational institutions such as Cambridge’s ‘Fun for Flyers’, McMillan education’s ‘I’m Ready’, University of Dayton’s ‘Do it!’, and Pearson’s ‘Sunshine’.

219

Fig. 4. Story presentation module interface.

The implementation of this module was completely achieved using the Unity game engine. All of the audio resources within the module are recorded by native speakers of the English language taking care of the speed and diction considering the users skills. 3.4 Evaluati on Module In the evaluation module, which user interface is presented in Fig. 5, user is provided with a series of questions related to the previously listened story along with the elapsed time for completing the activity. Such questions are presented in text and audio making it easier to students to understand them. In the same way, answer options are presented. This module considers the provision of a cube with four NFC tags attached to the user. The faces of the cube containing NFC tags are identified using colors and images attached to them. Such colors are consistent with the presented options and allow users to associate each of the cube faces to the options displayed. For choosing an option, users must first take the cube and locate the NFC sensor in the mobile device, then, take the desired face of the cube close to the sensor and wait for the application to detect the NFC tag. The application will then display a message letting the user know whether the option selected was correct or wrong. Depending on the attempts that users take to provide a correct answer, they are rewarded with gold, silver, or bronze trophies. As a first assessment of the developed application, we drove an experiment with English language teachers from different institutions. The design of such evaluation along with the obtained results are presented in the following section.

220

Fig. 5. Evaluation module interface and cube with NFC tags as TUI.

Design of Assessment and Results

As a parallel process to the development of the application and taking advantage of the model-based approach applied, we decided to evaluate the created user interfaces and workflow in an early stage in order to be able to correct any potential failures in posterior testing. For this experiment, we recruited 6 teachers from three schools. As up to this moment we only counted on user interfaces, the experiment was performed in compliance to the Wizard of Oz technique [48]. To record data from the experiment, an instrument including the list of tasks to be performed and cells for documenting the execution time and mistakes made as the one described in Table 9 was created. Table 9. Tasks to be performed during the experiment along with the average results. Task ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Task Start application Sign up as student Sign up as teacher Login to the application Select “Dictionary” activity Select word from the dictionary Go back to main menu Select “Story” activity Play story Repeat story Go to evaluation Select right answer Select wrong answer Exit application Go to credits section

Execution time 3s 7.8s 5.4s 2.4s 3s 1.8s 2.6s 5.6s 2.2s 1.6s 5s 2.6s 4.2s 2s 3.2

Mistakes 0 0.4 0 0 0 0 0.2 0 0 0.2 0 0 0 0 0

221

Once the experiment was performed and the gathered information was analyzed, it was possible to see that in average the tasks that took more time to be performed are “Sign up as student” and “Select story activity”. Also, the task in which more mistakes were made (2) was “Sign up as student”, leading to think that users needed some time to get used to the user interface and that this interface should be simplified for optimizing its use. After the interaction with the user interfaces, participants were asked to answer the CSUQ questionnaire [49] for measuring usability in terms of four dimensions: System usability, information quality, interface quality, and general satisfaction. The processed data reported that our application received for system usability an average score of 6.77 out of 7 (stddev = 0.24), for information quality an average score of 6.5 out of 7 (stddev = 0.27), for interface quality an average score of 6.47 out of seven (stddev = 0.6), and finally for general satis faction an average score of 6.6 out of seven (stddev = 0.54). Based on the presented results, in the next section we describe the identified conclusions and future work.

Conclusions and Future Work

In this paper, the design, development and partial assessment of ARTUI, an AR-based application for support to the English language teaching, is presented. Following a traditional Software development process, the requirement gathering stage was achieved by interviewing English teachers and through direct observa tion of inclassroom activities. The collected requirements were expressed as user stories making them easier to be developed and tracked. After generating the user stories, a model-based approach was followed for the implementation of user interfaces and their associated functionalities. In order to take advantage of the followed approach and considering that the application of the Wizard of Oz technique would allow to drive an experiment for understanding issues related to navigation and interaction, prio r termination of the development phase, a partial assessment was done with teachers from three institutions. During the evaluation, considering that high scores were obtained in the CSUQ questionnaire, it was possible to notice that teachers appreciate the proposal and that its use results easy and intuitive, and to see that the use of ARTUI in classroom sessions is viable and may positively impact English learning. The development of ARTUI involved technical challenges such as the management of the database and its integration to the Unity engine, the use of the Vuforia platform in Unity and its configuration to work as expected, and the detection and identification of NFC tags from the Unity engine. As future work, it was identified that following a spiral Software process model, a second iteration is required in order to implement functionalities like the addition of difficulty levels, the implementation of reports and the creation of additional content adequate to other skills. Also, if well the partial assessment was encouraging, content and interaction evaluation must be done with more teachers and students in order to assure quality and pertinence of the solution.

222

References 1.

3. 4.

6. 7.

10.

11. 12. 13. 14.

15.

16.

Piñón, N. de los A. G.; Sánchez, S. del C. H.; Nares, Y. del C. P.: Las TIC en la enseñanza del inglés en educación básica. Revista Electrónica Sobre Tecnología, Educación Y Sociedad, 4(7). (2017). De Europa, C.: M arco común europeo de referencia para las lenguas. Strasburgo: Consejo de Europa, M inisterio de Educación, Cultura y Dep orte/Instituto Cervantes. (2002) Secretaría de Educación Pública. El Modelo Educativo 2016. (2016). Lee, L. K.; Chau, C. H.; Chau, C. H.; Ng, C. T.: Using Augmented Reality to Teach Kindergarten Students English Vocabulary. In Educational Technology (ISET), 2017 International Symposium on (pp. 53-57). IEEE. (2017). Hsieh, M .-C.: Teachers’ and Students’ Perceptions towards Augmented Reality M aterials. 2016 5th IIAI International Congress on Advanced Applied Informatics (IIAIAAI), (1), pp.1180–1181. (2016). Prendes Espinosa, C. Realidad aumentada y educación: análisis de experiencias prácticas. Píxel-Bit. Revista de Medios y Educación, 46, 187-203. (2015). Azuma, R.; Baillot, Y.; Behringer, R.; Feiner, S.; Julier, S.; M acIntyre, B. Recent advances in augmented reality. IEEE computer graphics and applications, 21(6), 34-47. (2001). Dubois, E: Chirurgie Augmentée: un Cas de Réalité Augmentée; Conception et Réalisation Centrées sur l'Utilisateur(Doctoral dissertation, Université Joseph-FourierGrenoble I). (2001). Kaufmann, H.; Schmalstieg, D.: M athematics and geometry education with collaborative augmented reality. In ACM SIGGRAPH 2002 conference abstracts and applications(pp. 37-41). ACM . (2002). Klopfer, E.; Perry, J.; Squire, K.; Jan, M . F.: Collaborative learning through augmented reality role playing. In Proceedings of th 2005 conference on Computer support for collaborative learning: learning 2005: the next 10 years! (pp. 311-315). International Society of the Learning Sciences. (2005). Gündüz, N.: Computer Assisted Language Learning. Journal of Language and Linguistic Studies, 1(2), pp. 193–214 (2005). Warschauer, M .: Computers and language learning: an overview. Language Learning, 31, pp.57–71. (1998). Beatty, K.: Teaching & researching: Computer-assisted language learning. Routledge. (2013). Cesteros, A.: Las Plataformas E-Learning Para La Enseñanza Y El Aprendizaje Universitario en Internet. Las Plataformas De Aprendizaje. Del Mito a La Realidad, pp. 45–73. (2010). Chérrez, E.: El B-learning como estrategia metodológica para mejorar el proceso de enseñanza-aprendizaje de los estudiantes de inglés de la modalidad semipresencial del departamento especializado de idiomas de la Universidad Técnica de Ambato, p.503 (2014). M ontoya, M . S.: Dispositivos de mobile learning para ambientes virtuales: implicaciones en el diseño y la enseñanza. Apertura, pp.82–96 (2008).

223 17.

18.

19.

20.

21. 22. 23.

24.

25.

26.

27.

28.

29.

30.

31.

M oralejo, L.; Sanz, C.; Pesado, P. M .; Baldassarri, S.: Avances en el diseño de una herramienta de autor para la creación de actividades educativas basadas en realidad aumentada. TE & ET. (2014). Gupta, N.; Rohil, M . K.: Exploring possible applications of augmented reality in education. In Signal Processing and Integrated Networks (SPIN), 2017 4th International Conference. IEEE, pp. 437-441 (2017). Kiat, L. B.; Ali, M . B.; Halim, N. D. A.; Ibrahim, H. B. Augmented Reality, Virtual Learning Environment and M obile Learning in education: A comparison. In eLearning, e-Management and e-Services (IC3e), 2016 IEEE Conference. IEEE, pp. 23-28 (2016). Lee, H. S.; Lee, J. W.: M athematical education game based on augmented reality. In International Conference on Technologies for E-Learning and Digital Entertainment (pp. 442-450). Springer, Berlin, Heidelberg. (2008). Gazcón, N. F.; Larregui, J. I.; Castro, S. M .: La Realidad Aumentada como complemento motivacional. TE & ET. (2016). Eras M ontaño, E. J.: Realidad aumentada como propuesta metodológica para la enseñanza en un entorno de aprendizaje escolar (Tesis de Licenciatura). (2016). Freitas, R.; Campos, P.: SM ART: A System of Augmented Reality for Teaching 2nd grade students. In Proceedings of the 22nd British HCI Group Annual Conference on People and Computers: Culture, Creativity, Interaction-Volume 2, pp. 27-30. BCS Learning & Development Ltd. (2008). Kirner, T. G.; Reis, F. M . V.; Kirner, C.: Development of an interactive book with augmented reality for teaching and learning geometric shapes. In Information Systems and Technologies (CISTI), 2012 7th Iberian Conference on (pp. 1-6). IEEE. (2012). Di Serio, Á.; Ibáñez, M . B.; Kloos, C. D.: Impact of an augmented reality system on students' motivation for a visual art course. Computers & Education, 68, pp.586-596 (2013). Gutiérrez, R. C.; M artínez, M . D. V. D. M .; Bravo, J. A. H.; Bravo, J. R. H.: Tecnologías emergentes para la enseñanza de las Ciencias Sociales. Una experiencia con el uso de Realidad Aumentada en la formación inicial de maestros. Digital Education Review, (27), pp. 138-153 (2015). Peula, J. M .; Zumaquero, J. A.; Urdiales, C.; Barbancho, A. M .; Sandoval, F.: Realidad Aumentada aplicada a herramientas didácticas musicales. Málaga: Grupo ISIS. (2007). Delgado, R. G.; Parra, N. S.; Trujillo, P. M . N.: AR-Learning: libro interactivo basado en realidad aumentada con aplicación a la enseñanza. Tejuelo: Didáctica de la Lengua y la Literatura. Educación, (8), pp. 74-88 (2013). Sánchez, S. Á.; M artín, L. D.; González, M . Á. G.; García, T. M .; M enéndez, F. A.; M éndez, C. R.: El Arenero Educativo: La Realidad Aumentada un nuevo recurso para la enseñanza. EDMETIC, 6(1), pp.105-123 (2016). Kerawalla, L.; Luckin, R.; Seljeflot, S.; Woolard, A.: “M aking it real”: exploring the potential of augmented reality for teaching primary school science. Virtual reality, 10(3-4), pp.163-174 (2006). Fitzmaurice, G. W.; Buxton, W.: Graspable user interfaces. University of Toronto. (199/)

224 32.

33. 34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Ishii, H.; Ullmer, B.: Tangible bits: towards seamless interfaces between people, bits and atoms. In Proceedings of the ACM SIGCHI Conference on Human factors in computing systems (pp. 234-241). ACM . (1997). Xu, D. Tangible user interface for children-an overview. In Proc. of the UCLAN Department of Computing Conference. (2005). Tapia, D. I.; Cueli, J. R.; García, Ó.; Corchado, J. M .; Bajo, J.; Saavedra, A.: Identificacion por radiofrecuencia: fundamentos y aplicaciones. Proceedings de las primeras Jornadas Científicas sobre RFID. Ciudad Real, Spain, 1-5. (2007). Lee, L. K.; Chau, C. H.; Chau, C. H.; Ng, C. T.: Using Augmented Reality to Teach Kindergarten Students English Vocabulary. In Educational Technology (ISET), 2017 International Symposium on (pp. 53-57). IEEE. (2017). Dalim, C. S. C.; Dey, A.; Piumsomboon, T.; Billinghurst, M .; Sunar, S.: TeachAR: an interactive augmented reality tool for teaching basic English to non-native children. IEEE International Symposium on Mixed and Augmented Reality Adjunct Proceedings, pp.344–345 (2016). Amaia, A. M .; Iñigo, A. L.; Jorge, R. L. B.; Enara, A. G.: Leihoa: A window to augmented reality in early childhood education. 2016 International Symposium on Computers in Education, SIIE 2016: Learning Analytics Technologies. (2016). Barreira, J.; Bessa, M .; Pereira, L. C.; Adão, T.; Peres, E.; M agalhães, L.: M OW: Augmented Reality game to learn words in different languages: Case study: Learning English names of animals in elementary school. In Information Systems and Technologies (CISTI), 2012 7th Iberian Conference on (pp. 1-6). IEEE. (2012). M artínez, A. A.; Benito, J. R. L.; González, E. A.; Ajuria, E. B.: An experience of the application of Augmented Reality to learn English in Infant Education. In Computers in Education (SIIE), 2017 International Symposium on (pp. 1-6). IEEE. (2017). Liu, T. Y.; Tan, T. H.; Chu, Y. L.: 2D barcode and augmented reality supported english learning system. Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science, ICIS 2007; 1st IEEE/ACIS International Workshop on E-Activity, IWEA 2007, (Icis), 5–10. (2007). Zhou, L.; Zhang, S.:Design Research and Practice of Augmented Reality Textbook. 2014 International Conference of Educational Innovation through Technology, 16– 20. (2014). Yang, M . T.; Liao, W. C.; Shih, Y. C.: VECAR: Virtual english classroom with markerless augmented reality and intuitive gesture interaction. Proceedings - 2013 IEEE 13th International Conference on Advanced Learning Technologies, ICALT 2013, pp. 439–440. (2013) Hsieh, M .-C.; Koong Lin, H.-C.: A Conceptual Study for Augmented Reality Elearning System based on Usability Evaluation. Communications in Information Science and Management Engineering, 1(8), pp.5–7 (2011). Vate-U-Lan, P.: The seed shooting game: An augmented reality 3D pop -up book. 2013 2nd International Conference on E-Learning and E-Technologies in Education, ICEEE 2013, pp. 171–175. (2013). He, J.; Ren, J.; Zhu, G.; Cai, S.; Chen, G.: M obile-based AR application helps to promote EFL children's vocabulary study. In Advanced Learning Technologies (ICALT), 2014 IEEE 14th International Conference on (pp. 431-433). IEEE. (2014).

225 46. 47. 48.

49.

Belló, C: Diseño y desarrollo de aplicaciones software para la creación de actividades docentes con elementos de Gamificación (Bachelor's thesis). (2014). M arín-Díaz, V.: La Gamificación educativa. Una alternativa para la enseñanza creativa. Digital Education Review, (27). (2015). Llisterri, J.: Las tecnologías del habla: Entre la ingeniería y la lingüística. In Actas del Congreso Internacional “La Ciencia ante el Público. Cultura humanística y desarrollo científico y tecnológico”, Universidad de Salamanca, Instituto Universitario de Estudios de la Ciencia y Tecnología (pp. 44-67). (2003). Lewis, J. R.: IBM computer usability satisfaction questionnaires: psychometric evaluation and instructions for use. International Journal of Human‐ Computer Interaction, 7(1), 57-78. (1995).

Technological strategy for searching and selecting sources of reliable information on the Internet Mayra N. Márquez Specia, Josefina Guerrero García, Yadira Navarro Rangel Doctorado en Sistemas y Ambientes Educativos Facultad de Ciencias de la Electrónica, Benemérita Universidad Autónoma de Puebla, Ciudad Universitaria, Av. San Claudio y 14 sur Puebla, Pue. M éxico mayra.specia@correo.buap.mx, joseguga01@gmail.com, yadira.navarro@correo.buap.mx

Abstract. This is an experimental study with pre-test and post-test design whose main objective is to know the effectiveness of a sequence of activities designed to develop the skills in the search of reliable information on the Internet, this, in university students of the first semester in a degree in the Chemical Biological area. Throughout this study, ICT are used as a tool for the search of specialized information, through the use of databases as well as for the coevaluations of information sources carried out by the students. To obtain results in this research, a comparison is made of the data obtained in the coevaluations, so it is divided into two stages, one before and one after the application of the sequence of activities, to verify the effectiveness of the sequence the comparison of means is made by the t-Student statistic. Keywords: Information Search, Higher Education, Co-evaluation of sources, C.A.R.S. Checklist

Introduction

In the educational field there is a great diversity of competences, in the case of this study, it refers to one of the primary stages in the teaching of the research process and the development of research competencies, that is, the search and eva luation of reliable information, particularly in students of the Chemical Biological Sciences area. To enter context, we understand by investigative competence the set of skills and attitudes required for the development of a research project [1], on the other hand, the search for reliable scientific information is undoubtedly one of the most important elements in the teaching and development of skills for research. So for the development of these skills and abilities, students go through a process where they acquire epistemological, methodological and instrumental foundations, in order to build scientific knowledge [2]. In such a way that the teaching of the research process should be adapted in the way of producing and using scientific knowledge, through the application of didactic research strategies, presenting significant moments in the process of teaching scie ntific research [ 3].

227

The curriculum of university students should evolve along with current professional needs, where the context is the knowledge society, since professionals must not only master their area of knowledge but must know the use and potential of Info rmation and Communication Technologies (ICT). According to Fernández and Villavicencio [4] the support to research nurseries is of importance for students to develop the necessary skills for research among which the skills in information search and bibliometric analysis are considered, which is a requirement to achieve a training based on research. From this point of view, higher education institutions require the definition and application of ICT integration strategies and innovation in didactic - pedagogical mediation, to help students develop a profile of scientific competencies that guide the search for pertinent information and construction of significant knowledge [5]. A task of great importance when doing a documentary search is, without a doubt, to evaluate the quality of the information sources that are consulted, since in the case of the Internet, it is a powerful resource of free access, so that the user is obliged t o recognize the sources of reliable and quality information [6]. The initial phase of any research process is the search and review of information, so obtaining reliable sources with scientific support becomes a primary task of any researcher, therefore developing the necessary skills for students to perform this task in an effective way should be a priority in educational programs. For the development of this research, an experimental study is contemplated, with a pre-test post-test design, taking as sample students of the Pharmacobiologist Chemistry Degree from the Benemérita Universidad Autónoma de Puebla (BUAP), with whom a sequence of activities was applied, focused on developing in students the necessary skills for the search and evaluation of information sources with the support of ICT. In such a way, that the main objective of this research is based on knowing the e ffectiveness of the sequence of activities through the comparison of the co -evaluations of information sources obtained from the internet in the pre and post test. 1.1 C.A.R.S. of Harris Checklist for the co-evaluation of information sources Some of the tasks of university students in all disciplines is that they write essays and research reports, being an important requirement of these tasks the use of evidence to support the assertions that appear in the work, so teachers consider the eva luation of the sources with their students with greater relevance, however, students have problems to understand what these concepts mean [7]. As the critical attitude in the search for information has already been raised, the use and knowledge of pertinent search tools and the use of reliable scientific information are important requirements for the development of research competences in university students of any area. In order to develop these competencies and skills in information management, students must have access to various search tools, whether search engines or specialized databases, however, the evaluation of the information obtained is still of gre at importance, since that not only reliability is important, but also its accuracy. Regarding the evaluation of information sources, there are several models or checklists that allow the evaluation of information sources in an effective way, being

228

the C.A.R.S. of Harris Checklist one of the most used, which has been previously implemented in the BUAP. The C.A.R.S. (Credibility, Accuracy, Reasonableness, Support) Checklist is d esigned to help researchers evaluate the sources of information, probably few sources will comply with each of the criteria in this list, but if you learn to use and apply these criteria, it will be possible to separate the quality information from the bad quality information [8]. To clearly understand the C.A.R.S. Checklist, a summary of the four elements that comprise it is presented (see Table 1). Table 1. Summary of the elements of the CARS Checklist for the evaluation of information sources [9].

Element Credibility

Accuracy

Reasonableness

Support

Description Trustworthy source, author’s credentials, evidence of quality control, known or respected authority, organizational support. Goal: an authoritative source, a source that supplies some good evidence that allows you to trust it. Up to date, factual, detailed, exact, comprehensive, audience and purpose reflect intentions of completeness and accuracy. Goal: a source that is correct today (not yesterday). Fair, balanced, objective, reasoned, no conflict of interest, absence of fallacies or slanted tone. Goal: a source that engages the subject thoughtfully and reasonably, concerned with the truth. Listed sources, contact information, available corroboration, claims supported, documentation supplied. Goal: a source that provides convincing evidence for the claims made.

Methodology

As part of the design of a didactic strategy, a series of activities focused on develo ping specific research skills in university students who are studying the first semester of an educational program in the Chemical Biological area are being implemented. The application of the sequence of activities has been carried out with two groups of new students to the degree of Chemist Pharmacobiologist, the 2 groups are made up of 20 and 23 students respectively, the tests that are currently perfo rmed correspond to a general piloting of the methodology proposed for a doctoral study in order to detect areas of opportunity and necessary adjustments to it. Among the main characteristics of the participants in this study, we found that 66% are female and 34% male, where 82% come from public higher secondary education, while 18% from private education, within the sample of the 43 students, the predominant ages are 18 and 19 years. It is worth mentioning that the selection of the sample was made for convenience since the two groups available for the spring 2019 period were used.

229

Considering what has been established in the syllabus proposed for the subject and approved by the Faculty of Chemical Sciences, topics from 1.1 to 1.4 corresponding to Unit 1 have been selected (see Table 2). Table 2. Subjects corresponding to Unit 1 of the syllabus of the subject Computational Tools.

Unit

Topics 1.1 How to define the need for information 1.2 Reliability of information: selection of sources 1.3 Tools for searching information on the internet

1. Search for reliable information

1.3.1 Google Scholar 1.3.2 Electronic resources provided by the general management of BUAP libraries 1.4 Databases: SciELO, SpringerLink, SCOPUS, etc. 1.5 Information search strategies

The subject of Computational Tools is taught in one of the computer labs of the Faculty of Chemical Sciences of the BUAP, this laboratory has a total of 36 computers with Internet access, and since the total number of students is divided into two groups, each student has access to a computer in each session, as well as an own account in the BlackBoard institutional platform, which is being used to store the virtual portfolios of the students divided into teams of 3 to 4 members. To test these activities a total of 5 sessions of two hours each per group were used, the indications and links necessary for the development of each activity are published by group in the announcement section of BlackBoard. The work and the topics were divided into two stages, the first where the students, select a scientific topic by team, use their own strategy and the tools they know to search for information and obtain 2 sources of information per person, later they are divided into pairs, selecting a person from a team different from their own, to perform the coevaluation applying the C.A.R.S. Checklist for the evaluation of information sources on the internet, which was adjusted for its application in Google forms, which raises a scale from 1 to 5, being 1 = Very poor, 2 = Poor, 3 = Fair, 4 = Good and 5 = Excellent. For stage 2, it began with the activities proposed for each topic, making readings and visiting university pages to collect the most important points to define the need for information, reliability of the sources and types of information. Subsequently, the students entered Google Scholar and databases available in the electronic resources of the web page of Libraries of the BUAP, as well as databases of free access. Having covered the topics from 1.1 to 1.4 of Unit 1, the students again seek sources of information regarding the previously selected scientific topic and make a comparison between the first sources obtained and those obtained from the databases. Finally, they carry out a second evaluation of the so urces obtained in stage 1 through the C.A.R.S. Checklist, for the post-test, a different form was used than the pre-test, the same elements of the C.A.R.S. Checklist however, questions such as: Are there changes in your way of evaluating the web reference? This question is applied

230

to the two sources obtained, the next item asks for the explanation of the previous answer, the last question: Which of the points of the C.A.R.S. evaluation checklist seems more important to determine the quality of the information we find on the internet? Seeks to know which element is most important for students.

Results

For the results obtained, a comparison of the co-evaluation of sources is presented through the CARS Checklist in the first and second stages (see Figure 1), where the students' approach to the elements proposed in the model of evaluation of information sources obtained from the Internet changed, it is notorious, this is reaffirmed when observing the results obtained in the questions referring to the students ch anged their way of evaluating sources 1 and 2 with respect to the co -evaluation carried out in stage 1 , to which 100% of the students answered affirmatively, that is, all the st udents modified their evaluation in the second stage of the study. With respect to the comments made by the students about the explanation of this change, they agree that in the first stage they did not understand the concept of reliability, they did not consider determining factors such as the date of publication, a uthor's data, source support and type of s ource among others, generally refer to that they did not know the importance of the reliability criteria and already knowing and understanding their importance, the vision they have on the reliability and type of sources obtained has changed.

Fig 1. Comparison for results obtained in the co-evaluation using the C.A.R.S. Checklist for stages 1 and 2.

231

Regarding the importance that students give to the elements of the C.A.R.S. Checklist 78.6% consider that all elements have the same importance to determine the quality of information sources, 11.9% believe that support is the most important element, fo llowed by accuracy and credibility. Subsequently, to verify if there is a significant difference between the results obtained from the co-evaluation performed in the first stage and those obtained in the second stage, a t-Student test was performed for paired samples, for which the IBM SPSS Statistics statistical software [10] was used, considering the means obtained for each of the four elements of the information source evaluation model, prior to the t test, a normality test was performed where it is verified that due to the P-value obtained , the data of the two stages of the study come from a normal distribut ion (see Table 3), in contrast to the value of α = 0.05. Table 3. Shapiro-Wilk test for normality distribution in data.

Kolmogorov-Smirnov a Stage 1 Stage 2

Statistical .232 .266

Shapiro-Wilk gl 4 4

Sig. . .

Statistical .915 .952

gl 4 4

Sig. .511 .727

Regarding the results obtained in the t-Student test for paired samples (see table 4), in the comparison of the means obtained in the co -evaluations of the pre-test and the post-test, it was found that there is a significant difference in the results before and after the sequence of activities carried out with the students, so it is concluded that the sequence does have significant effects on the students' understanding regarding the reliability of the sources of information, this in contrast with the P-value obtained and the value of α = 0.05. Table 4. t test of paired samples, for comparison of means. Paired differences 95% confidence interval of the differStandard Standard M an Dev. error ence Inferior Superior Stage 1– .08129 .04064 1.16835 1.42705 Stage 1.29770 2

Sig. (bilateral)

31.928

.000

Conclusions

In accordance with the main objective of this research which focuses on knowing the effectiveness of the sequence of activities through the comparison of the co evaluations of information sources obtained from the internet in the pre and post test, it is concluded that the the activities proposed for the sequence have had sufficient

232

effect to generate a change in the understanding and application of the students on the 4 elements addressed in the Harris´ C.A.R.S. Checklist. Since in addition to obtaining a significant difference according to the statistics made, the total of the students participating in the study claim to have modified their criteria to perform the co-evaluation, since within the activities it includes the verification of the central points to determine the reliability of the sources, visit specialized databases, search and obtain information about these resources and compare the type of information obtained in stages 1 and 2 of the study, which allows the students themselves to determine the differences in the information obtained and reflect on the type of information required for their university education. It is clear the change that the students had when co -evaluating the sources obtained through the search tools that they use on a regular basis, being predominant the use of Google, since by understanding the importance of the origin of the sources of information reflect on the use of tools for the search of specialized information, and the use of available resources including search time, so that from the activities carried out they prefer to spend the greatest amount of time in determining the descriptors for the search. Finally, it should be mentioned that for each of the elements of the model: Credibility, Accuracy, Reasonableness and Support, there was a difference between the means obtained in the pre and post test, which indicates that the decrease in the points achieved for each element is because the s tudents were more rigorous and careful in the application of each criterion, adhering to the quality standards of the specialized information that is requested for their professional training, the results obtained in this study are part of the preliminary analysis of data for the doctoral research of which it is a part.

References 1.

Jaik Dipp, A.; Competencias investigativas: una mirada a la educación superior. 1st ed. M éxico: red Durango de investigadores educativos a. C. Redie, pp.3-16, 50-60.: http://redie.mx/librosyrevistas/libros/competenciasinvestigativas.pdf (2013). Last accessed 2018/10/23. Rojas Soriano, R.; Formación de investigadores educativos Una propuesta de investigación (12th ed.). M éxico, D.F.: Plaza y Vadés. http://raulrojassoriano.com/cuallitlanezi/wpcontent/themes/raulrojassoriano/assets/libros/formacion-investigadores-educativosrojas-soriano.pdf (1992). Last accessed 2019/02/03 Rojas-Betancur, H.; M éndez-Villamizar, R.; Cómo enseñar a investigar. Un reto para la pedagogía universitaria. Educ.Educ, 16(1), 95-108. http://dx.doi.org/10.5294/edu.2013.16.1.6 (2013). Last accessed 2019/02/05 Fernández-Espinosa, C. E.; Villavicencio-Aguilar, C. E.; Habilidades investigativas para trabajos de graduación. ACADEM O Revista De Investigación En Ciencias Sociales Y Humanidades, 4(1). http://revistacientifica.uamericana.edu.py/index.php/academo/article/view/99 (2017). Last accessed 2019/03/20

233 5.

8. 9. 10.

M arín, F. V.; Inciarte, A. de J.; Hernández, H. G.; Pitre, R. C.; Estrategias de las Instituciones de Educación Superior para la Integración de las Tecnología de la Información y la Comunicación y de la Innovación en los Procesos de Enseñanza. Un Estudio en el Distrito de Barranquilla, Colombia. Formación universitaria, 10(6), 29-38. https://dx.doi.org/10.4067/S0718-50062017000600004 (2017). Last accessed 2019/03/23 Landgrave-Ibáñez, S.; Rosas, E. P.; Esquivel, L. B.; Coria, A. I.; Jiménez Galván, I.; Sámano Sámano, A. Uso de la Web e internet como herramientas para la búsqueda de información médica científica. Archivos en M edicina Familiar, 18(4), 95-106. https://www.medigraphic.com/cgi-bin/new/resumen.cgi?IDARTICULO=72102 (2017). Last accessed 2019/03/19 Auberry, K. Increasing students’ ability to identify fake news through information literacy education and content management systems. The Reference Librarian, 59(4), 179-187. https://doi.org/10.1080/02763877.2018.1489935 (2018). Last accessed 2019/03/28 North Hennepin Community College. CARS Checklist. https://nhcc.edu/studentresources/library/doinglibraryresearch/cars-checklist. Last accessed 2019/03/23 Harris, Robert. “Evaluating Internet Research Sources.” VirtualSalt. https://www.virtualsalt.com/evalu8it.htm (2015). Last accessed 2019/03/23 IBM SPSS Statistics for Windows. IBM Corp, Armonk, NY: IBM Corp (2013)

Protocol for Creation of a Database: Data Capture for the Automatic Identification of Learning-Centered Emotions Yesenia N. González Meneses , Josefina Guerrero García BUAP, Facultad de Computación, Av. San Claudio y 14 Sur, Col. San M anuel Edificio CCO3, Ciudad Universitaria Puebla, Pue., M éxico. {yeseniaglez0, joseguga0}@gmail.com

Abstract. In this article we define the elements that should be considered in the design of a formal protocol for the creation of the database of physiological and behavioral signals of college students while doing a learning activity. The goal of designing a formal protocol for data capture is to provide an adequate database for the study of learning-centered emotions. This is a fundamental task for the recognition of emotions in specific contexts and is part of the data treatment stage in the methodology of the research project in which it is intended to automatically identify emotions in educational environments. For data capture we will use acquisition technologies of physiological and behavioral signals and machine learning algorithms to identify relationships between emotions and learning. It is considered to collect data from four of the main learning-centered emotions: interest, boredom, confusion and frustration, according to [1]. For the execution of the protocol it is proposed to merge data from two technologies for the acquisition of physiological signals and two behavioral signals with the idea of integrating a vast set of data. So, the design of the experiment’s protocol is described (elements, considerations and formalization). Keywords: Database, Physiological Signals, Behavioral Signals, LearningCentered Emotions, Learning Activity.

Introduction

In the area of affective computation research, the automatic recognition of emotions is a relevant problem, since it is the beginning of the study and development of humancomputer interaction systems sensitive to the emotions of human beings. The human computer interaction can be implemented from the use of physiological signal sensors (parameters that can be measured and that allow to regulate the biological functioning of a human being) for the acquisition of data. Another approach has to do with the observation of human behavior through the analysis of voice, images of the face, eyes, head or body movements of people, as mentioned by [2]. The data obtained from different devices must be processed and classified based on a specific objective. In this investigation they will be captured in real educational environments when students are performing a learning activity and subsequently used

235

to recognize learning-centered emotions. This activity can be done using any electronic device such as computer, tablet or cell phone, navigating in an intelligent tutorial system or MOOC (Massive Open Online Courses), interacting with an educational video game, or using traditional devices; whether watching a video, studying, researching or reading. In this work we do not consider face-to-face classes with the intervention of a lecturer. The objective is to build a database of physiological and behavioral signals, starting with the capture of data and ending with the selection of useful features for the identification of learning-centered emotions. The signals should be captured in an educational environment, in order to collect data specifically from learning activities. These data will be the basis for the implementation of a model of emotion identification with an acceptable level of accuracy, more in line with the reality within an educational context. Data tagged with specific emotions and/or the process of automatic identification of emotions can be integrated into the module of student modeling in educational software sensitive to students' emotions, such as intelligent tutorials, MOOCs or educational games. This with the idea of complementing the student model and consequently reinforcing the decision making of teaching strategies and assignment of activities that contribute to better levels of learning.

State of the Art

A database of emotional expressions is a collection of images or video clips , speech and physiological signals related to a wide range of emotions. Its content corresponds to emotional expressions related to the context in which they were captured and then tagged, this is essential for the training, testing, and validation of algorithms for the development of expression recognition systems. The tagging of emotions can be done with a discrete or continuous scale. Most of the databases are usually based on the emotions theory of [3], which assumes the existence of six basic emotions on a discrete scale (anger, fear, disgust, surprise, joy, sadness) and an approximate variety of 22 secondary emotions. However, in some databases the emotions are tagged in continuous arousal-valence scale [4]. Other databases include the AU (Action Units) based on FACS (Facial Action Coding System)[5]. The databases of expressions of emotions are mostly formed only by facial expressions and these are classified into posed and spontaneous. In posed expression databases, the participants are asked to display different emotional expressions, while in spontaneous expression database, the expressions are natural. Spontaneous expressions differ from posed ones remarkably in terms of intensity, configuration, and duration. In most cases, the posed expressions are exaggerated, while the spontaneous ones are subtle and differ in appearance. Apart from this, synthesis of some AUs are barely achievable without undergoing the associated emotional state so it is not possible to capture physiological data since these cannot be controlled by people and therefore do not correspond to the emotion that is acted out. Next, we show a compilation of databases of facial expressions, although our interest lies in hybrid databases that contain facial images and physiological data, so far, we have only found one database of this type publicly available. The DEAP database [6] contains physiological recordings (of EEG) and facial video of an experiment where 32 volunteers watched a subset of 40 music videos. In this experiment partici-

236

pants are also asked to rate each video according to the emotion it caused in them. Therefore, the analysis of the databases shown in Table 1 only includes databases of facial expressions available to the public. These databases correspond to facial expressions of basic emotions, which have been captured while people perform various activities. Most of the databases are comprised of viewers watching media content (i.e. ads, movie trailers, television shows, animated gifs and online viral campaigns) and of posed facial expressions. Out of these, the most robust database for the amount of data stored is Affectiva [25]. Table 2 shows details of databases of spontaneous facial expressions. These correspond to learning-centered emotions that have been captured while students perform some learning activity in a wild setting. Some of them contain physiological data but unfortunately, they do not make them available to the public or provide information about the characteristics of these data, let alone details of the processing they do with them. Out of these, four databases include physiological data. The most complete is the one mentioned in [41] in which 67 students participate and use three physiological data sensors and a video camera. In general, they have been created with data from very few participants. From this analysis, we have identified as an area of opportunity the creation of a physiological and behavioral database supported by a formal pro tocol for data collection. The definition of the protocol allows us to run a controlled experiment in a natural environment and that can be replicated the number of times necessary to create a robust database.

Problem Definition

The problem involved in the construction of a database of physiological and behavioral signals for the automatic recognition of emotions has been an area of investigation highly active in the last years. Regardless of this, a clear solution, which is within reach of most people, is still far away. Several drawbacks have influenced the construction of an appropriate solution from a computational point of view. On one hand, a factor that affects the performance of emotion recognizers in real contexts is the difficulty to generate databases with spontaneous emotions. Generally, works are made with acted databases which provide portraits of emotions representing prototypical and intense emotions that facilitate the search of correlations and the subsequent automatic classification. This kind of databases are usually captured in a controlled environment, which decreases problems in the processing of information (noise, for example). In addition, it can guarantee a balanced number of samples per class. As consequence, there have not been good results when translating the knowledge extracted from these databases to real contexts [7].

237 Table 1. Databases facial expressions of basic emotions. Database

Number Subjects

Facial expression

Number of images/videos

Type

32-channel 512Hz EEG peripheral physiological signals and 22 face videos

Spontaneous

DEAP [6]

Analysis of continuous emotion model.

Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) [9]

Speech: Calm, happy, sad, angry, fearful, surprised, disgusted, and neutral. Song: Calm, happy, sad, angry, fearful, and neutral.

7356 video and audio files

Posed

F-M FACS 3.0 (EDU, PRO & XYZ versions) [10]

neutral, sad, surprised, happy, fearful, angry, contemptuous and disgusted

4877 videos and images sequences

Posed and Spontaneous

Extended Cohn-Kanade Dataset (CK+) [11]

neutral, sad, surprised, happy, fearful, angry, contemptuous and disgusted

123

593 image sequences (327 sequences having discrete emotion labels)

Posed; spontaneous smiles

Japanese Female Facial Expressions (JAFFE) [12]

neutral, sad, surprised, happy, fearful, angry, and disgusted

213 static images

Posed

MMI Database [13]

Disgusted, happy and surprised

1280 videos and over 250 images

Posed and Spontaneous

Set 1 (disgusted, fearful, amused, frustrated, surprised)

114

570 video clips

Set 2 (disgusted, fearful, amused, frustrated, surprised, angry, sad)

650 video clips

Set 3 (disgusted, amused)

180 video clips

Belfast Database [14]

DISFA [15]

fearful,

Natural Emotion

4,845 video frames

Spontaneous

Multimedia Understanding Group (MUG) [16]

neutral, sad, surprised, happy, fearful, angry, and disgusted

1462 sequences

Posed

Indian Expression (ISED) [17]

sad, surprised, happy, disgusted

428 videos

Spontaneous

Spontaneous Database

and

Radboud Faces Database (RaFD) [18]

neutral, sad, contemptuous, surprised, happy, fearful, angry, and disgusted

Three different gaze directions and five camera angles (8*67*3*5=8040 images)

Posed

Oulu-CASIA Database [19]

surprised, happy, sad, angry, fearful and disgusted

Three different illumination conditions: normal, weak and dark (total 2880 video sequences)

Posed

FERG (Facial Expression Research Group Database)DB for stylized characters [20]

angry, disgusted, fearful, joyful, neutral, sad, surprised

55767

Frontal pose

Affectnet [21]

neutral, happy, sad, surprised, fearful, disgusted, angry, contemptuous

IMPA-FACE3D [22]

neutral frontal, joyful, sadness, surprised, angry, disgusted, fearful, opened, closed, kiss, left side, right side, neutral sagittal left, neutral sagittal right, nape and forehead (acquired sometimes)

534 static images

FEI Face Database [23]

neutral, smiley

200

2800 static images

200

~1,250,000 annotated

NIR-VIS

Aff-Wild [24]

Affectiva [25]

7 emotions (angry, joyed, sad, surprised, fearful, disgusted, contemptuous), 20 expressions and 13 emojis

~450,000 annotated ~ 500,000 annotated

manually automatically

7,860,463 Faces analyzed

Wild setting

Posed

Posed manually

In-the-Wild setting

Spontaneous

238

Table 2. Databases of facial expressions that correspond to learning-centered emotions. Database

Facial expression

Number Subjects

Number of images/videos

Type / Software Tool

Own database of [26]

Frustrated, confused, sad, joyful, anguished and fearful

39 students

30 to 60 minutes for student

Spontaneous / Intelligent Tutoring System

Own database of [27]

Interested, bored, frustrated, confused, surprised, pleasured, curios, happy and neutral

22 college students

Images

Spontaneous / Learning administrator system

Own database of [28], [29] y [30]

Interested, bored, frustrated and excited

8 students

Video and EEG (discrete emotion labels)

Spontaneous / Intelligent tutor

Video and self-report

Spontaneous / Computational learning environment to program in python

67 students

Video and self-report

Spontaneous / Intelligent multi-agent tutorial system for teaching the human circulatory system

28 - 30 students

Video of eyes tracking, dialogue notes with the tutor, facial expressions video, body postures (body pressure measurement system), mouse pressure sensor, keyboard pressure sensor, voice acoustics (horns)

Spontaneous / Intelligent tutorial system for computer repair

Attentive and concentrated

23,072 instances

Video and EEG headband (Emotiv-EPOC). Extraction of Alpha and Beta waves.

Spontaneous / Psychological tests to induce mental states of attention and concentration

Pleasured, frustrated, confused, interested and bored

137 students

Video and the BROMP observation method (to detect the affective state of students by observers)

Spontaneous / Educational game Physics Playground

Interested

22 students

Video of the Kinect for windows and BIOPAC MP150 electrocardiogram signal acquisition system

Spontaneous / writing a summary

Own database of [39]

Interested

30 students

Video of Kinect V2

Spontaneous / excel course

Own database of [40]

Bored, interested and neutral

44 students

Typing analysis and video

Spontaneous / writing an essay

Interested, excited, confident and frustrated

two groups of 38 and 29 high school students

Physiological sensors (video camera, mouse, chair and bracelet)

Spontaneous / Multimedia adaptive tutoring system for geometry

Own database of [31]

Own database of [32]

Own database of [1], [33]

Frustrated, confused, interested, bored

Basic emotions and learningcentered emotions (19 emotions and a neutral state)

Frustrated, confused, interested, bored

29 students

Own database of [34]

Own database of [35], [36] y [37]

Own database of [38]

Own database of [41]

239

In contrast, the databases with spontaneous records show information with emotional content that does not belong to a single class, but a mixture of th em. In other cases, there are samples with a very light emotional charge, close to a neutral emotional state. In addition, databases with spontaneous emotions are usually recorded in noisy environments, such as classrooms, study rooms, entertainment areas, offices, factories or in phone conversations, which leads to the inclusion of noise. Finally, because of the very nature of the problem, it is not possible to ensure a balanced quantity of examples per class. Another challenge to be solved is the extraction and selection of a set of characteristics that allow recognizing emotions in the data captured spontaneously. Although progress in the area has been important, there is still much to be done in realistic contexts. Therefore, it is necessary to propose and explore other approaches that allow reaching a good performance of the emotion’s recognition in real world applications. An evident aspect to consider is the fact that the area of application has an important influence on the accuracy of emotion recognition, as well as the degree of intrusion of the tools used [8]. And these two aspects are crucial to consider in the creation of the database of physiological and behavioral signals. The problems that are usually faced in the capture of physiological and b ehavioral signals data in real contexts are: 1. Use of invasive technologies that cause discomfort in the capture of spontaneous expressions. 2. Synchronization in the different capture devices. 3. The environmental conditions where the capture takes place. 4. The correct sensing of the devices. 5. The previous emotional state of the participants. 6. The availability of the participants. 7. The physical appearance of the face (amount of hair or artifacts on the face). 8. The necessary capture time. Due to the above mentioned, the formal design of a protocol for the capture of data (either physiological, behavioral or both) is necessary. The protocol will guide the execution of a controlled experiment with the objective of avoiding errors in the acquisition of data, describing the technologies to be used, the capture process, the environmental conditions and the preparation of the participants. In the following sections, all the aspects to be considered in the design of a protocol for data collection will be defined.

Elements of the Protocol for the Capture of Physiological and Behavioral Data

In this section, we first describe the stages of research, analysis and selection of key elements for the creation of a spontaneous database captured in real educational environments through the execution of a controlled experiment. For which it is necessary the voluntary participation of students in a learning activity while recording their behavior and physiological signals through different data acquisition devices. In the second part, we propose legal, ethical and environmental (physical place) issues that should be considered in the design of a formal protocol. Both points to consider can be seen in the diagram shown in figure 1, in terms of key concepts, the physiolog ical

240

and behavioral signals to be captured must be defined, the data acquisition technologies that will be used to capture the signals, the model of emotions on which the identification of emotions will be based and the activity in which the student s will participate while their data is captured. Regarding physical aspects of the place where the experiment will be carried out, it is important to consider comfortable furniture and stable climate control. As well as the legal and ethical aspects dependent on the laws and customs of the country where the data capture will take place.

Figure 1. Elements of the protocol for the capture of physiological and behavioral data. 4.1

Research, Analysis and Selection of Key Concepts for the Design of the Data Capture Protocol

Physiological and Behavioral Signals. Physiological and behavioral responses occur during or immediately after the causal event and the signals can be quantitatively compared. There are many signals relevant to emotional responses that are physically measurable, especially by cameras, microphones, and sensors, the latter of which might be placed in physical contact with a person in a comfortable and non-invasive way. Consider what happens when we try to recognize somebody's emotion. First, our senses detect low-level signals: motion around their mouth and eyes, perhaps a hand gesture, a pitch change in their voice and, of course, verbal cues such as the words they are using. Signals are any detectable changes that carry information or a message. Sounds, gestures, and facial expressions are signals that are observable by natural human senses, while blood pressure, hormone levels, and neurotransmitter levels require special sensing equipment. Second, patterns of signals can be combined to

241

provide more reliable recognition. A combination of clenched hands and raised arms movements may be an angry gesture; a particular pattern of features extracted from an electromyogram, a skin conductivity sensor, and an acoustic pitch waveform, may indicate a state of distress. This medium-level pattern representation can often be used to make a decision about what emotion is present. At no point, however, do we directly observe the underlying emotional state. All that can be observed is a complex pattern of voluntary and involuntary signals, in physical (physiological signals) and behavioral forms. Deciding which signal to measure depends on the data we want to analyze and on the availability of the technologies for its capture. Technologies for Data Acquisition. The sensors measure physiological parameters related to autonomic nervous system changes. Nowadays these are already integrated into wearable human-machine interfaces that capture physiological signals that can help in the recognition of emotions, as mentioned in [8]. Brainwave diadems that send information to an electronic device in the form of an electroencephalogram (EEG) are found. Cardiovascular wristbands, which measure the heart rate and provide information in the form of an electrocardiogram (ECG). Electrothermal activity sensors, which measure the level o f conductivity of the skin through sweat on the hands measured with two small electrodes of silver chloride in which an imperceptibly small voltage is applied and then the conductance between the two electrodes is measured. Thermal cameras allow to measure the temperature change of the human body associated with the different emotional states. There are also devices that measure muscle electrical activity in response to nerve stimulation of muscles, in the form of electromyography (EMG). The electromyogram signal uses small electrodes to measure a small muscle tension, which indicates contraction [9]. Regarding devices related to the identification of people’s behavior -such as body postures and gestures - there are traditional video cameras, webcams or augmented reality devices that allow the recording of facial expressions and body movements, as well as eye tracking, important to the recognition of emotions. In this type of devices, there are also voice recorders, another medium used to identify emotions, as mentioned in [7] . The decision on which device to choose for data capture depends directly on the possibility of acquiring them and on the data that we are interested in gathering. Although there are varieties of them on the market, it is also true that, due to their cost, they are not available to any institution or individual. It is also important to consider how invasive they are to the human body, since, for example, in the case of acquiring student data in learning activities we should consider the comfort of the student; The emotions they show depend a lot on how comfortable and how much pleasure they feel. Something similar happens when they attend a face-to-face class at a classroom, the furniture or teacher makes them uncomfortable and therefore they do not care and get bored. Emotions Model. The term emotion refers to relations among external incentives, thoughts, and changes in internal feelings ; as weather is a remarkable term for the changing relations among wind velocity, humidity, temperature, barometric pressure, and form of

242

precipitation. Theorists have long discussed a small set of categories for describing emotional states. "Basic emotions" may be defined in many ways. Perhaps the most thorough definition has been given by Paul Ekman, who has linked basic emotions to those which have distinctive universal facial expressions associated with them, as well as eight other properties [3]. By these criteria, Ekman identified six basic emotions or discrete emotion categories mentioned in the introduction. Some authors have been less concerned with the existence of eight or so basic emotions and instead refer to continuous dimensions of emotion [10]. Three dimensions show up most commonly, although only the names of the first two a re widely agreed on. The two most common dimensions are "arousal" (calm/excited), and "valence" (negative/positive) [4]. The other dimension is called “activation” (active/passive) [9]. In general, two dimensions cannot be used to distinguish all the basic emotions; for example, intense fear and anger lie in the same region of h igh arousal and negative valence. However, these two dimensions do account for the most common descriptions of mood. The lack of a definition of emotion, and the lack of agreement on whether there are basic emotions or continuous spaces of emotions are obs tacles to the goals of computer-based recognition and synthesis. The question of whether to try to represent emotions with discrete categories or continuous dimensions can be considered a choice, as each representation has advantages in different applications, the best choice depends on what we are trying to explain [9]. Educational Tool. Educational software incorporates multimedia content and gives users a high level of interactivity. These two features differentiate them from traditional teaching practices. Multimedia content like pictures, graphics, and sound help engage the students in their lessons. Its objective is to help fill a need for more interactive and personalized educational experiences for students. Some examples of educational software are intelligent tutorial systems, educational games and MOOCs. An essential element of this software is the student's modeling, from which it must be possible to adapt the software automatically to the requirements of each student, such as their learning style, cognitive capacity, learning curve and emotional state. The choice of educational software to be used as a virtual interactive medium depends on the charact eristics of the students (age, school level, subjects of study, studies they carry out), the subject of interest and mainly learning-centered emotions that want to provoke during their interaction with it. 4.2

Considerations for the Design of the Protocol

Environment A learning environment is a "place" or a "space" where learning takes place, it is a set of physical places and the relationships that appear in it. It is a whole of objects, smells, shapes, colors, sounds, people who inhabit and relate in a phys ical frame [11]. The elements, components, dimensions and conditions that should be considered when planning and designing a quality learning environment are the dimensions of the space, the furniture, the lighting, temperature and ventilation, the decoration and

243

accessories, the computer equipment that will be used and the technologies for the acquisition of data and its physical disposition. Legal In Mexico, the Federal Law on the Protection of Personal Data in Possession of Individuals DOF: 05/07/2010, published in the Official Gazette of the Federation on July 5th, 2010, protects the privacy of individuals with respect to the treatment we give their personal information. Its provisions are applicable to all individuals or corporations, public and private sector, both at the federal and state level, which carry out the processing of personal data in the exercise of their activities, therefore, companies such as banks, insurers, hospitals, schools, telecommunications companies, religious associations, and professionals such as lawyers, doctors, among others, are obliged to comply with what is established by this law. According to article 3 section V of this Law, personal data is all that information that allows to identify a person. In the capture of physiological and behavioral data we must first obtain consent for its recording, this must be done formally through the drafting of an "informed consent letter". In this document the participant is explained what the experiment consists of, the devices that will be used for the capture, the use that will be given to their data, the confidentiality that will be saved from them and their signature is requested with which they authorize the recording. For any other country we must inform ourselves of the laws that protect the privacy of personal data and design our experiment being respectful of them. Ethics Emotions, perhaps more so than thoughts, are ultimately personal and private , providing information about the most intimate motivational factors and reactions. Any attempts to detect, recognize, not to mention manipulate, a user’s emotions thus constitute the ultimate breach of ethics and will never be acceptable to computer users. Attempts to endow computers with these abilities will lead to widespread rejection of such computer systems and will help promote an attitude of distrust to computers in general. Humans routinely detect, recognize, and respond to emotions or manipulate them in ways that most would consider highly ethical and desirable [12] . In the case of cheering up a student’s mood to take better advantage of a learning activity is perfectly acceptable, more, however, although the intention is goo d, he should be informed of this.

Discussion

Constructing a database with physiological and behavioral signals of humans performing activities in real environments , both controlled and uncontrolled, is a complex, laborious and time-consuming task, especially when the challenge is to integrate data from various devices to capture different signs. Due to this, the databases of expressions with emotional content that are publicly available correspond only to facial expressions and activities in which it is easier to place a video camera and record the emotional expressions. But while it is true, this is the least invasive way to capture

244

emotional signals, it is a challenge to be able to integrate more robust databases with emotional data, both physiological and behavioral. The advantage of this proposal is to be able to relate physiological signals with the physical expression of the behavior and corroborate their correspondence, that is, for example, if the facial expression indicates happiness, what facial temperature is related to this emotion or what changes in heart rate can be associated with this emotion. Defining the area in which the recognition is attempted and the physiological signals that are to be captured gives the pattern to the design of an experiment, formalized in a protocol for its execution. The protocol defines the details of the capture to perform a controlled experiment and is designed for a specific context. Therefore, the databases will contain data corresponding to the activity within t he context defined as those mentioned in Table 2 where we show examples of databases with data captured in learning environments. The design of the capture protocol requires prior investigation about the context in which it will be executed, the kind of participants, the activity that will be carried out, the data that will be captured, the devices that will be used, the time of the capture, the emotions model, among others that may vary depending on the use that will be given to the database. Other not less important aspects that must be considered are the physical conditions of the place, furniture, legal aspects referring to the confidentiality of the captured data and ethical aspects about the treatment that will be given to the data. All these elements are described in a document that supports both the capture of data and the use that is made of them. It is important for any type of research to be backed by a document that validates the data on which the research is based and that in many occasions is a requirement to be able to publish in academic and scientific dissemination events . This work is the result of research and activities that were carried out to capture physiological and behavioral signals in a controlled learning environment with the objective of creating a database that makes better contributions to the learning-centered emotions automatic recognition. Based on the considerations mentioned in this article, the formal protocol for data capture was designed that helped us solve the problems mentioned in section 3. Following the protocol, we have a control of the execution of the experiment. In this the steps to carry out the capture are listed, from the welcoming of the student, the explanation of the experiment in which they will participate and the definition of all the aforementioned elements. The sample population is college engineering students, the educational software they use is a MOOC of basic algebra, exploring only the introductory module to algebra. The devices to record selected b ehavior signals are a Logitech Full HD Web video camera and the Kinect 360 for Windows. The devices for capturing physiological signals are an ICI 9320P thermal camera (for capturing the facial temperature), a cardiac pulse sensor implemented with Arduino and the MUSE 2 headband (both for recording the heart rate). Once we have the data, the future work to be done in this investigation is the cleaning and selection of these with the aim of structuring the database. Subsequently, a prior analysis of the data should be carried out with the idea of finding relations (manual labeling) between changes in temperature or heart rate and emotion, detected by observation of the face also considering the emotional state identified by the same student through an emotional test. The manual labeling will give us the guidelines to do it automatically. Finally, the tagged data will be the input of machine learning algorithms that once trained auto-

245

matically identify emotions centered on learning, hoping to contribute both in the educational and computational area.

References 1. 2.

3. 4. 5. 6. 7.

8. 9.

10. 11.

12. 13.

14. 15.

S. D’M ello and A. Graesser, “Dynamics of affective states during complex learning,” Learn. Instr., vol. 22, no. 2, pp. 145–157, 2012. C. Fuentes, V. Herskovic, I. Rodríguez, C. Gerea, M . M arques, and P. O. Rossel, “A systematic literature review about technologies for self-reporting emotional information,” J. Ambient Intell. Humaniz. Comput., pp. 1–14, 2016. P. Ekman, Emotions Revealed. Recognizing Faces and Feelings to Improve Communication and Emotional Life, 1st ed. New York: Henrry Holt and Company, 2003. J. A. Russell, “A circumplex model of affect,” J. Pers. Soc. Psychol., vol. 39, no. 6, pp. 1161–1178, 1980. P. Ekman, W. Friesen, and J. Hager, Facial Action Coding System. The M anual. Salt Lake City, United States of America, 2002. M . Soleymani, S. M ember, and J. Lee, “DEAP: A Database for Emotion Analysis Using Physiological Signals,” vol. 3, no. 1, pp. 18–31, 2012. F. González-Hernández, R. Zatarain-Cabada, M . L. Barrón-Estrada, and H. Rodríguez-Rangel, “Recognition of learning-centered emotions using a convolutional neural network,” J. Intell. Fuzzy Syst., 2017. S. Steidl, “Automatic Classification of Emotion-Related User States in Spontaneous Children’s Speech,” Universität Erlangen-Nürnberg, 2009. S. R. Livingstone and F. A. Russo, “The ryerson audio-visual database of emotional speech and song (ravdess): A dynamic, multimodal set of facial and vocal expressions in north American english,” PLoS One, vol. 13, no. 5, pp. 14–18, 2018. A. Freitas-M agalhães, Facial Action Coding System 3.0: M anual of Scientific Codification of the Human Face (english edition). Porto: FEELab Science Books, 2018. P. Lucey, J. F. Cohn, T. Kanade, J. Saragih, and Z. Ambadar, “The extended CohnKanade dataset (CK+): a complete facial expression dataset for action unit and emotion-specified expression Conference on,” in Computer Vision and Pattern Recognition Workshops (CVPRW), 2010 IEEE Computer Society Conference on, 2010, no. July, pp. 94–101. M . J. Lyons, J. Gyoba, and M . Kamachi, Japanese Female Facial Expressions (JAFFE), Database of digital images. 1997. M . F. Valstar and M . Pantic, “Induced Disgust, Happiness and Surprise: an Addition to the M M I Facial Expression Database,” Seventh International Conference on Language Resources and Evaluation, 2010. I. Sneddon, M . M crorie, G. M ckeown, and J. Hanratty, “Belfast Induced Natural Emotion Database,” IEE Trans. Affect. Comput., vol. 3, no. 1, pp. 32–41, 2012. S. M . M avadati, S. M ember, M . H. M ahoor, K. Bartlett, P. Trinh, and J. F. Cohn, “DISFA: A Spontaneous Facial Action Intensity Database,” IEEE Trans. Affect. Comput., vol. 6, no. 1, pp. 1–13, 2013.

246 16.

17.

18.

19.

20.

21.

22. 23.

24.

25. 26. 27. 28.

29.

30.

N. Aifanti, C. Papachristou, and A. Delopoulos, “The M UG facial expression database,” 11th International Workshop on Image Analysis for M ultimedia Interactive Services WIAM IS 10, pp. 1–4. S. L. Happy, P. Patnaik, A. Routray, and R. Guha, “The Indian Spontaneous Expression Database for Emotion Recognition,” IEEE Trans. Affect. Comput., vol. 8, no. 1, pp. 131–142, 2017. O. Langner, R. Dotsch, G. Bijlstra, D. H. J. Wigboldus, S. T. Hawk, and A. van Knippenberg, “Presentation and validation of the radboud faces database,” Cogn. Emot., vol. 24, no. 8, pp. 1377–1388, 2010. Zhao, “Oulu-CASIA NIR&VIS facial expression database,” Center for M achine Vision and Signal Analysis. University of Oulu, Oulum, Yliopisto. [Online]. Available: http://www.cse.oulu.fi/wsgi/CM V/Downloads/Oulu-CASIA. [Accessed: 04-Apr2019]. D. Aneja, A. Colburn, G. Faigin, L. Shapiro, and B. M ones, “M odeling Stylized Character Expressions via Deep Learning,” Asian Conf. Comput. Vision. Springer, vol. 1, pp. 136–153, 2016. A. M ollahosseini, B. Hasani, and M . H. M ahoor, “AffectNet: A Database for Facial Expression, Valence, and Arousal Computing in the Wild,” IEEE Trans. Affect. Comput., vol. 10, no. 1, pp. 18–31, 2017. R. M ena-Chalco, Jesus M arcondes and L. Velho, “Banco de Dados de Faces 3D: IM PA-FACE3D,” Brasil, 2008. C. E. Thomaz, “FEI Face Database,” Department of Electrical Engineering. Centro Universitario da FEI, São Bernardo do Campo, São Paulo, Brazil, 2012. [Online]. Available: https://fei.edu.br/~cet/facedatabase.html. [Accessed: 04-Apr-2019]. S. Zafeiriou, D. Kollias, M . A. Nicolaou, A. Papaiooannou, and I. Kotsia, “Aff-Wild: Valence and Arousal ‘in-the-wild’ Challenge,” in Computer Vision and Pattern Recognition Workshops (CVPRW), 2017, pp. 1980–1987. R. el Kaliouby and R. W. Picard, “Affectiva Database,” M IT M edia Laboratory, 2019. [Online]. Available: https://www.affectiva.com. [Accessed: 08-Apr-2019]. B. Nye et al., “Analyzing Learner Affect in a Scenario-Based Intelligent Tutoring System,” Artif. Intell. Educ. Springer, pp. 544–547, 2017. X. Xiao, P. Pham, and J. Wang, “Dynamics of Affective States During M OOC Learning,” Artif. Intell. Educ. Springer, pp. 586–589, 2017. R. Zataraín, M . L. Barrón, F. González, and C. A. Reyes-García, “An Affective and Web 3.0-Based Learning Environment for a Programming Language,” Telemat. Informatics, 2017. R. Zatarain-Cabada, M . L. Barrón-Estrada, F. González-Hernández, R. OramasBustillos, G. Alor-Hernández, and C. A. Reyes-García, “Building a Corpus and a Local Binary Pattern Recognizer for Learning-Centered Emotions,” Adv. Comput. Intell., vol. II, pp. 524–535, 2017. R. Zatarain-Cabada, M . L. Barron-Estrada, F. González-Hernández, and H. Rodríguez-Rangel, “Building a Face Expression Recognizer and a Face Expression Database for an Intelligent Tutoring System,” Proc. - IEEE 17th Int. Conf. Adv. Learn. Technol. ICALT 2017, no. 2161–377X/17, pp. 391–393, 2017.

247 31. 32.

33. 34.

35. 36.

37. 38.

39.

40.

41.

42. 43. 44. 45. 46.

N. Bosch and S. D’M ello, “The Affective Experience of Novice Computer Programmers,” Int. J. Artif. Intell. Educ., vol. 27, no. 1, pp. 181–206, 2015. J. M . Harley, F. Bouchet, and R. Azebedo, “Aligning and Comparing Data on Emotions Experienced during Learning with M etaTutor.,” LNAI 7926 - Artif. Intell. Educ., no. July, pp. 61–70, 2013. A. C. Graesser and S. D’M ello, Emotions During the Learning of Difficult M aterial, vol. 57. 2012. J. Y. Arana-Llanes, G. González-Serna, R. Pineda-Tapia, V. Olivares-Peregrino, J. J. Ricarte-Trives, and J. M . Latorre-Postigo, “EEG lecture on recommended activities for the induction of attention and concentration mental states on e-learning students,” J. Intell. Fuzzy Syst., 2017. N. Bosch et al., “Detecting student emotions in computer-enabled classrooms,” IJCAI Int. Jt. Conf. Artif. Intell., vol. 2016–January, pp. 4125–4129, 2016. N. Bosch, S. K. D’mello, J. Ocumpaugh, R. S. Baker, and V. Shute, “Using Video to Automatically Detect Learner Affect in Computer-Enabled Classrooms,” ACM Trans. Interact. Intell. Syst., vol. 6, no. 2, pp. 1–26, 2016. N. Bosch et al., “Automatic Detection of Learning-Centered Affective States in the Wild,” Proc. 20th Int. Conf. Intell. User Interfaces - IUI ’15, pp. 379–388, 2015. H. M onkaresi, N. Bosch, R. A. Calvo, and S. K. D’M ello, “Automated Detection of Engagement using Video-Based Estimation of Facial Expressions and Heart Rate,” IEEE Trans. Affect. Comput., pp. 1–14, 2016. K. Almohammadi, H. Hagras, B. Yao, A. Alzahrani, D. Alghazzawi, and G. Aldabbagh, “A type-2 fuzzy logic recommendation system for adaptive teaching,” Soft Comput., vol. 21, no. 4, pp. 965–979, 2017. R. Bixel and S. D´M ello, “Towards Automated Detection and Regulation of Affective States During Academic Writing,” in LNAI 7926 - Artificial Intelligence in Education., 2013, no. July, pp. 904–907. I. Arroyo, D. G. Cooper, W. Burleson, B. P. Woolf, K. M uldner, and R. Christopherson, “Emotion sensors go to school,” Front. Artif. Intell. Appl., vol. 200, no. 1, pp. 17–24, 2009. R. Cowie et al., “Emotion recognition in human computer interaction,” IEEE Signal Process. M ag., vol. 18(1), no. January, pp. 32–80, 2001. R. W. Picard, Affective Computing. 1997. H. Schlosberg, “Three Dimensions Review,” J. Gen. Psychol., vol. 61, no. 2, pp. 81– 88, 1954. B. G. Wilson, “Constructivist learning environments: Case studies in instructional design,” 1996, p. 4. R. W. Picard, “Affective computing: Challenges,” Int. J. Hum. Comput. Stud., vol. 59, no. 1–2, pp. 55–64, 2003.

The RASE model as a means of applying communicative language teaching principles online María de los Milagros Cruz Ramos 1 , Juan Manuel González Calleros 2 , Luz Edith Herrera Díaz1 Centro de Idiomas Región Veracruz, Universidad Veracruzana, Calzada Juan Pablo II s/n. 18071, Veracruz, M éxico 2 Facultad de Ciencias de la Computación, Benemérita Universidad Autónoma de Puebla Avenida San Claudio 14, Puebla, M éxico mariacruz@uv.mx, jumagoca78@gmail.com, ehd63@hotmail.com 1

Abstract. Introduction: This paper reports on an ongoing project aimed at determining if the integration of language skills and interaction, under the principles of the RASE model and the communicative approach, would have an impact on the development students’ communicative competence in an online English course. The modifications to the instructional design of the experimental group’s online course are founded on the revised version of the Interaction Hypothesis, the study of communicative competence, the principles of the CLT approach, and those of the RASE model. Results are drawn from the application of communicative tests as part of a quantitative multiple time series quasiexperimental design. The tests are complemented with the use of a rubric specifically created to measure aspects related to the dimensions of the competence under study. The progress of those students working under the modified instructional design seems more evident that the progress of those in the control group, in particular when it comes to their oral communicative competence. The results in other areas of language and the other three language skills are not as clear. Two more tests are to be applied to collect data until the end of the implementation period, and statistical analysis needs to be carried out. Early results suggest that interaction and the use of integrated skills under a techno-pedagogical model do indeed aid the development of learners’ communicative competence. Unlike previous studies, this project collects data related to all three dimensions of said competence. Keywords: EFL, Techno-Pedagogical M odel; Communicative Competence

Introduction

Instructional design can be defined as a process aimed at the improvement and application of methods and models of instruction. The selection of instructional methods is necessary so as to achieve the desired changes in the knowledge and abilities of the students in a specific context and course [1]. In the case of courses that rely on the use of technology, such as those that involve blended or online learning, the principles of instructional design are translated into instructional design models [2]. These models

249 work around a learning theory which orients the planning and execution of instruction. Such models can also be called techno-pedagogical [3] due to the two dimensions they usually consider: 1) the technological, which deals with the tools and resources to be applied within the learning environment; 2) the pedagogical, concerned with the learners’ characteristics and needs, as well as with the learning objectives, skills and competences to be achieved, the materials and activities to be designed and implemented, among other factors. However, it has been argued that disciplinary differences are rarely taken into account when designing or research ing instruction in online courses [4]. This lack of disciplinary differences applied into instructional design seems evident when analyzing existent research around interaction and communication in a foreign language online. As matter of fact, EFL (English as a Foreign Language) online courses in general do not seem to take into account two essential disciplinary elements: EFL approaches to teaching and learning [5], and their implications for interaction and communication [6-7]. This turns out to be problematic to say the least, due to the fact that regardless of the modality of instruction, EFL courses tend to have an objective in common: the development of skills that allow students to communicate in the target language. These skills, referred to as communicative, involve the integrated use of four different language skills: speaking, writing, listening , and reading. A systematic literature review carried out last year [5] found that it was necessary to do further research on the development of communication skills, and the influence of techno-pedagogical strategies (product of instructional design) aimed at both the integration of language skills and learner interaction in the target language. Said literature review found that even those projects aimed at studying the effects that strategies, interaction and language skill integration have on students’ actual communicative competence [8-13] either do not obtain conclusive results, or focus exclusively on linguistic aspects which do not necessarily guarantee communication. In consequence, this paper reports on the quantitative methods involved in an ongoing project that seeks to determine if an instructional design modified to incorporate interaction and communication in the target language under the principles of a communicative approach influences the development of a group of online students’ communicative competence. This paper is organized as follows: Section 2 presents the literature review that serves as a theoretical framework to the modification s undergone by the instructional design; Section 3 presents the methodology followed, comprised by the implementation and data collection procedures ; Section 4 synthesizes the preliminary results obtained in the first half of the project; Section 5 discusses the results and their implications; and Section 6 presents the conclusions and further research to be carried out in the second half of the project.

Literature Review

Language is considered a social process. Thus, we can assume that interaction with other people is relevant in learning how to communicate. This section revolves around four main concepts and theories so as to explain how communication in a target lan-

250 guage can be achieved online under the principles of a teaching approach and a techno-pedagogical model. 2.1 Language and Interaction Thanks to interaction, students are exposed to language models that constitute input. The role this input plays in the development of language is often analyzed from three points of view: the behaviorist, focused on the linguistic environment, that is, the stimuli and feedback obtained from the medium; the nativist, focused on the internal way in which each individual processes the language; and the interactionist, which considers the two previous ones [14]. A revised version of the Interaction Hypothesis [15] proposes that three things are necessary to acquire a new language through interaction: 1) noticing, the individual perceives and is aware of the linguistic characteristics of the input; 2) comparison, the individual compares the characteristics of the input with that of his own output, or his own speech; 3) integration, the individual constructs his own linguistic knowledge, thanks to the two previous elements, and internalizes it. Regarding his revised version of the interaction hypothesis, [16] noted that there are additional aspects to be taken into account, such as student ability, and how prepared they are to negotiate meaning; yet he concludes that his version of the hypothesis has generated sufficient research in the area and demonstrated explanatory power. The interaction hypothesis is still one of the most coherent and consistent ones, since it does not seem to contain internal contradictions and the terminology used is clearly and widely explained [17]. Although the role of instruction is not taken into account, the interaction hypothesis is capable of explaining the learning process [18]. In this study as learning is defined as the process through which an individual receives input in the target language through the interaction with other individuals; input that constitutes models of language whose patterns the student is able to understand, reproduce and adapt in order to express meaning. 2.2 Communicati ve competence The term communicative competence, first emphasized the importance of learning not only what is grammatically correct, but also understanding what is appropriate [19]. Although the work of Hymes was not originally created in relation to foreign languages, it gave rise to the work of other linguists [20-21], who reevaluated the original definition, determined that this competence should be observable in communicative acts; and identified the need to look for ways to contribute to the development of the communicative competence and its assessment, since it is considered measurable. This concept has been defined and studied by different authors over four decades [20-24]. In this particular study, communicative competence is understood as the use that individuals give to their grammatical knowledge of morphology, phonology, and syntax, as well as to their social knowledge in order to emit, interpret and adapt the messages received during the interaction in the target language, English. The authors referenced in this section have laid the foundations for the development of new techniques, methods and approaches for teaching and learning languages, as well as for the assessment of communication.

251 2.3 Communicati ve approaches in EFL The TBLT (Task-Based Language Teaching) approach and the communicative approach, better known as CLT, stand out among the rest. These two approaches allow the integration of the four language skills and heighten the role of interaction between participants in EFL contexts [25]. Nevertheless, CLT has a unique view under which communication in the target language is not only the objective, but also the means to achieve it. Moreover, the TBLT approach is more likely to favor communication and interaction over linguistic forms, leaving aside other the elements of the so called communicative competence. The CLT approach, instead, focuses on students learning and using forms, meanings and linguistic functions that enable them to communicate in the target language. This approach also aims at helping students develop their communicative competence [25-26]. CLT instruction follows a series of principles [25], which are considered applicable in online synchronous instruction and interaction [27]: a) Authentic language should be used and introduced into the classroom whenever possible. b) A part of being communicatively competent is being able to understand the intentions of the interlocutor. c) The use of language is not only the objective; it is also the means to achieve it. d) Errors are a natural result of the development of communication. They must be observed and addressed without diverting attention from communication. e) The teacher must establish situations that can promote communication. f) The communicative interaction between the students gives them the opportunity to negotiate a meaning. g) The communicative context is essential to give meaning to expressions. h) It is important to learn to use language forms appropriately. As it was previously mentioned, a balance among communication and linguistic form is attainable through this approach, as made evident in principles d) and h) above. Applying CLT principles to instructional design, and online instruction itself is expected to help set communicative situations where the development of the communicative competence can be tracked online. 2.4 RASE as a techno-pedagogical model The theories, principles and approaches mentioned in three previous sections have important implications for instructional design in virtual environments at the system, course and lesson level [2]. The RASE model (acronym for Res ources, Activities, Support, Evaluation), contemplates the design and implementation of resources, activities and evaluation. Additionally, it highlights the importance of the support provided to students. Consequently, the four components of the model (Re sources, Activities, Support and Evaluation) must be integrated holistically into the practice of the teacher / instructional designer. This model contemplates multiple concepts based on the Theory of Activity proposed by Vygotsky and articulated more specifically by Engeström [28]. Its focus is primarily constructivist, because it focuses learning on the activities carried out by the students, which must be carried out in an environment that encourages the construc-

252 tion of knowledge, beyond its mere transmission. The RASE model can be approached from two perspectives: 1) instructional - it helps teachers to focus on the student while integrating educational technologies; and 2) learning - it helps students learn disciplinary content and develop new knowledge [29]. 2.5 Theoretical articulation The works of the authors cited in this section seem to suggest that in order to help the participants of an EFL English course to communicate competently, an essential element is necessary: interaction. In the particular case of courses that are carried out in a virtual environment, this will only be possible with the support of a techno pedagogical model that guides the instructional design according to the CLT principles.

Methodology

This ongoing study follows two groups of students at an undergraduate level in a public university in the southeast side of Mexico. Both groups receive online instruction for a compulsory English II course provided by their university. One of the groups (regarded as the control group) is currently taking the standard treatment in the form of the online course typically implemented, whereas the other group (considered an experimental group) is taking a version of the course that has been modified under the principles of the RASE techno-pedagogical model. The classes or groups were constituted naturally, and convenience sampling was used in selecting them, which means the groups are not guaranteed to be similar enough that a Pre -test can be omitted [30]. Accordingly, a quantitative quasi-experimental multiple time series design was chosen as a means to measure students’ communicative competence in different points throughout the course. This methodological design can be considered strong due to the certainty of interpretation it gains through the multiple measurements made [3031]. However, there are multiple threats to the internal validity of these studies, some of which have to do with the size of the sample, the loss of data, as well as subjects abandoning the implementation [32]. Some solutions, such involve: trying to keep a balance between the size of the experimental and control groups; the emulation random assignment by using a pairing technique (also known as matching); and the calculation of the size of the effect in relation with the s ize of the sample [32]. 3.1

Context and subjects

The study is being carried out within an English II online course at a public university, and the subjects belong to the two groups available in the region. There is a total of 46 students, from 21 different undergraduate programs, enrolled in the online course. About 92% of them are enrolled in the same region, with only four of them enrolled in other regions. During the initial phase 17 students from the experimental group (about 33% of the group’s population, and 71% of the participants currently active in that group) and 8 students from the control group (about 68% of the students enrolled in this group, and the totality of the ones currently active) took the pre-test. This pre-

253 test consisted in a communicative test comprised of two main components: a written one, involving three language skills; and an oral one, which constitutes an evaluation of the fourth language skill. Based on their pre-test scores, the students were paired mechanically according to the variable oral communicative competence. The course consists of ten thematic units delivered though the institutional platform, Eminus. The current design of the course shares aspects in common with Dick and Carey's techno-pedagogical model; said model allows for individual and/or group activities to be put into applications or virtual learning environments, and contemplates the use of asynchronous or synchronous resources [33]. These applications, environments, or platforms, in turn, host not only the materials and activities, but are the medium in which the instruction is carried throughout all its stages. In fact, in virtual environments, student participation and even feedback are carried out in the virtual environment itself [34]. The Dick and Carey techno pedagogical model defines instruction, as a set of examples, materials, and activities to be carried out by students , all of which are aimed at prompting a desired behavior or learning [34]. As we have previously established, this study defines learning as the process through which an individual receives information that constitutes input. Input from which the individual must discern, reproduce and adapt language models in order to express meaning. Since the current course design lacks opportunities for students to receive the aforementioned input in English, there seems to be a disparity between the instruction currently provided in those online English courses and the objective said courses pursue: communication. The study materials available on the institutional platform contain explanations in Spanish, and examples in English. Even though the explanations in Spanish could be considered practical or useful, they would not constitute input for the learning of the foreign language. Something similar happens with the activities students must perform asynchronously on the platform, some of which have instructions in Spanish, even when students are expected to answer them in English. Furthermore, the lack of oral and written interaction in the target language among individuals, whether synchronously or asynchronously, goes against the tenets of the CLT approach. 3.2 Changes to the instructional design The intervention aims to improve the four elements of the RASE model (resources, activities, support and evaluation), with the support of technological tools following the CLT principles listed by [25]. The creation of additional resources, in the form of multimedia presentations, would allow for language to be presented as it is used in real contexts. In addition, the creation of audios that contain a contextualized use of the language according to the objectives of each unit and recorded by speakers with a C2 level according to the CEFR can provide listening practice. This ability, listening, is not considered in the current instructional design. The audios are introduced in some forums as a set for the written interaction. This, in turn, constitutes a modification to the forum activities, in order to integrate language skills, and provide greater opportunities for contextualized practice of the language. Forums also serve as a space for students to exchange audio messages with the help of a free online voice recording tool called VoiceSpice.

254 Similarly, synchronous sessions on the web conferencing tool Zoom, provide students with the opportunity to interact with their instructor and classmates in the target language. The activities presented during synchronous sessions follow the principles of the CLT approach, as well as the recommendations provided to apply said principles during synchronous sessions [27]. The activities carried out during these sessions are mostly oriented to the interaction among participants in the target language, but the materials and situations that trigger the interaction are based on either of the two receptive skills, listening or reading. The sessions also serve as a space to clarify doubts not yet addressed by asynchronous means of support (email, Eminus messaging). Evaluation, the last element of the RASE model is also modified with the help of Zoom, since oral tests and final presentations will also take place online. 3.3 Data collection A quantitative instrument to assess students’ communicative competence was designed in order to complement the communicative standardized tests used within the university where the study takes place. The instrument consists in a rubric that assesses elements corresponding to each of the three dimensions of the variable communicative competence. There are two versions of the rubric, one to be used in oral int eractions, and another one to be used in written communicative tasks. The design process took into account the guidelines and examples offered in the CEFR [35], as well as the operationalization of communicative competence with all its dimensions and subskills. In order for the rubric to be as manageable as possible, the rubric includes six aspects that directly relate to the indicators established in the operationalization. The aspects included in the rubric were selected based on two conditions, how representative they were of each dimension, and how applicable they would be to the expected production of a basic level EFL leaner: a) Lexical and grammatical accuracy and precision b) Lexical and grammatical range c) Pronunciation d) Style and Register e) Fluency and coherency f) Contextual appropriateness. Since the third aspect, fluency, is only applicable to oral communication but not to written interaction, the second version of the rubric replaces it with orthography, which in turn can only be assessed in written communicative tasks. Along with the rubrics, a detailed set of scoring criteria was developed so as to let potential examiners know how to score each aspect. The rubric uses a scale between 1 and 5, with 5 being the highest possible score for each of the six aspects. This results in a total of 30 possible points overall. This instrument was validated both qualitative and quantitatively by two different groups of TEFL (Teaching English as a Foreign Language) experts. The first group of experts validated the design by emulating the operationalization process originally followed, whereas the second group of experts validated the relevance and coherence of each indicator to measu re its corresponding dimension. The rubric was then tested to determine if two different scorers would reach an acceptable agreement.

255 Since the quantitative design selected for this study is a quasi-experimental multiple time series, three measurement points or tests are planned: a pre-test the week before the start of the course; a control test, in the seventh week of the course; and a final test or post-test, the week after the end of the course. The dates of the tests were chosen to match the partial and final tests, according to the official course calendar. Both control and experimental group were required to take the written exams face-toface, whereas the oral exams were taken online in the case of the experimental group.

Preliminary Results

This section reports on the results obtained by both groups in the Pre-Test and the first Control-Test in terms of score means. It also reports on five pairs of learners’ (matched mechanically after the pre-test) performance. The reason only five pairs of students are reported is because participants needed to fulfill two conditions to be considered: to have taken the pre-test, and to be up-to-date with course work. It is important to mention this since some students have joined the course after the pre -test period, and thus would only be eligible if they take the remaining tests in the time series. Similarly, some students who did take the pre-tests are not active on the course anymore. Table 1. M ean scores obtained in the first two measurements with a standardized test. Experimental Group Area Vocabulary Grammar Listening Reading Writing

Control Group

Pre-Test

Control Test 1

Pre-Test

Control Test 1

4.48 5.97

7.39

6.33

8.24

5.64

7.45

5.45

6.29 8.86 5.78

6.92

8.16

7.87

8.32

8.37

7.47

5.10

7.62

5.38

Table 1, above, summarizes the mean scores obtained by each group wit h a standardized written test in two different moments, the pre-test before the beginning of the course, and a control test on week 7. The standardized written tests used within the institution cover five different areas: vocabulary and grammar (both of th em linguistic elements of the language); listening, reading and writing (all of which are communicative language skills). The global mean score obtained by the groups in each of the tests are shown next (see Fig. 1).

256

Fig. 1. The chart shows the mean scores obtained by each group in the pre-test and control test. A communicative test was applied to the students as part of the oral component of the communicative test. The students were assigned a score using the instrument described in the previous section. The mean scores obtained by each group are shown below (see Fig. 2).

Fig. 2. The chart shows the mean score obtained by each group in the pre-test and control test.

Five students from the control group were mechanically paired to five students from the experimental group who had obtained the same pre-test results according to the oral communicative competence in the pre-test. The following Table 2 shows the results each pair of students has obtained in the standardized written test in these first two moments in the time series.

257 Table 2. Pair scores in the first two measurements.

Area

Pair 1 Experimental Group (EGS7) Pre-Test Control Test 1

Control Group (CGS2) Pre-Test Control Test 1

Vocabulary

5.71

8.57

7.14

Grammar

6.09

5.91

7.83

4.55

Listening

8.57 10.00 5.00

6.15

7.14

5.38

10.00

5.71

4.29

6.67

5.00

Reading Writing

Area

Pair 2 Experimental Group (EGS8) Pre-Test Control Test 1

Control Group (CGS4) Pre-Test Control Test 1

Vocabulary

4.29

8.57

5.71

7.14

Grammar

8.26

6.82

5.65

5.45

Listening

10.00 10.00 10.00

8.46

8.57

10.00

8.57

10.00

5.00

8.33

5.00

Reading Writing

Pair 3 Experimental Group (EGS11) Area

Control Group (CGS12)

Pre-Test

Control Test 1

Pre-Test

Control Test 1

Vocabulary

1.43

10.00

7.14

10.00

Grammar

4.78

5.45

6.96

3.18

Listening

5.71 7.14 3.33

5.38

10.00

5.38

7.14

10.00

5.71

6.67

5.00

Reading Writing

5.00 Pair 4

Experimental Group (EGS4) Area

Control Group (CGS17)

Pre-Test

Control Test 1

Pre-Test

Control Test 1

Vocabulary

2.86

7.14

10.00

8.57

Grammar

6.96 8.57 10.00 8.33

5.45

9.13

9.09

7.69

10.00

8.46

7.14

10.00

5.00

10.00

8.33

Listening Reading Writing

258 Pair 5 Experimental Group (EGS12) Area

Control Group (CGS18)

Pre-Test

Control Test 1

Pre-Test

Control Test 1

Vocabulary

2.86

8.57

2.86

5.71

Grammar

6.09 8.57 10.00 8.33

6.36

5.65

2.73

8.46

4.29

6.92

10.00 6.67

7.14 3.33

4.29 3.33

Listening Reading Writing

The global written scores, as well as the oral communicative competence scores for each pair are presented below (see Fig. 3-4).

Fig. 3. The chart shows the global scores each p air of students obtained in the first two written tests.

Fig. 4. The chart shows the scores each pair of students obtained in the first two oral tests.

259

Discussion

Both individual and group scores show that, in average, the control group started ahead of the experimental group. This is made evident in the average scores obtained in the standardized written Pre-Test. Given that all students enrolled into English II courses credited the previous course, albeit in different modalities and temporalities, they share a similar starting point at least as far as their linguistic knowledge concerns. Still, the mean scores for the two areas related to linguistic knowledge (vocabulary and grammar) were almost two points lower for the experimental group. Something similar occurred with the language skills listening and writing. For reading, however, the experimental group scored three tenths higher. This difference in scores per section, is congruent with the global score for the same written tests. The experimental group’s overall score of 6.19, was almost a point and a half below the 7.40 score averaged by the students in the control group. The disparity in scores when measuring students’ oral communicative competence, even though not as steep as the one shown in written tests, was still present. Nevertheless, and in line with the findings of a previous quantitative study [7], oral communication was expected to be the aspect to improve the most, from the pretest to subsequent measurements, thanks to an intervention that fostered interaction in the target language among peers. This research implemented such a change through the modification of some course elements according to the principles of the RASE techno-pedagogical model [28-29]. Accordingly, when comparing the results obtained by both groups in the pre-test and the control test, the experimental group’s oral communicative competence rose, whereas the average scores for the control group decreased. An apparent cause for the latter’s oral score decrease could be the fact that English II presents a topic typically considered challenging for students during the first few weeks: the simple past. Keeping in mind that other authors have claimed that the contextualized practice of the contents aids the subsequent production of the language [25-26] and that interaction is key in providing students with models they can further modify and personalize [15], it would seem illogical to expect students lacking that contextualized interaction to produce contextualized language with the same results as students who have been exposed to those conditions. The scores for the areas covered in the standardized written tests, however, show mixed tendencies. The mean scores of the experimental group for vocabulary soared, whereas the score for grammar decreased moderately. The scores for the language skills showed three different behaviors: reading and writing by over half a point, while listening increased as much. Then again, the written task involved the use of the simple past, a typically challenging topic. The mean scores of the control group increased as well, yet not in as many points as the scores for the experimental group did. The control group’s scores for grammar decreased by two points. The scores for the language skills also decreased. The most affected skill was writing, which dropped over two points. Even before an in depth statistical analysis, the results of the individuals paired with a member of the other group, do not seem to show a clear tend ency. There have been individuals in each pair whose scores have increased and decreased in the different areas of the written test, while their counterpart’s either show a similar behavior or

260 do the opposite. In contrast, the score individuals obtained when measuring their oral communicative competence show that, within two of the five pairs, the members of the experimental group performed better in the control test than they had previously done in the pre-test; while in two other pairs their performance s tayed the same. The score of all members of the control group decreased, being outperformed by their experimental group counterparts . The comparison between the group’s global written scores seems to make things clearer. The global mean scores of the experimental group increased by three tenths, while the global means of the control group decreased by over seven tenths. Even so, the average written score of the experimental group is still below that of the control group. Nevertheless, what originally was a difference 1.21 points, has been reduced to 0.14 points. Moreover, if this tendency continues, the experimental group could surpass their counterparts by the end of the research. This could lead to the modification of the instructional design to be considered not only successful, but also overall beneficial.

Future work

There is still an additional moment to collect data as a part of the multiple time series quantitative design being implemented. The rest of the implementation, as well as the collecting of data with a final test are planned over the next couple of months. SPSS, a statistical analysis program, will be used to explore the causal relations among the different variables involved in this project. Due to the fact that ANOVA tests are usually run to verify if certain characteristics of two groups differ significantly at the end of an intervention [26], such a test will be considered to verify if the experimental and control group’s communicative competence differ significantly after the intervention has finished.

Conclusions

The body of work behind the idea that integrated skills and interaction are key in the development of communication is wide to say the least [7, 10, 14-16, 19-26]. As previous literature reviews have demonstrated [5], the use of synchronous and asynchronous communication technology in EFL courses interests researchers around the world. However, the correlation between instructional designs, strategies implemented and actual communicative results are not typically studied. The preliminary results shown in this paper seem promising, the time frame for the implementation and data collection have been quite limited, though. It is too soon to determine whether communicative competence will indeed keep improving for the experimental group. That is to say: until more data has been gathered to confirm the preliminary results, and a statistical analysis has been carried out, it would not be possible to assert whether the intervention has been beneficial or not. The extent of the benefits, given they exist, remains unknown. The outlook, however, is optimistic.

261

References 1. 2. 3.

8. 9. 10.

11. 12.

13.

14. 15. 16.

Reigeluth, C.: Instructional Design Theories and M odels. An Overview of their Current Status. 1st edn. Lawrence Erlbaum Associates, Hillside, New Jersey (1983). Universidad Veracruzana, http://www.uv.es/~bellochc/pedagogia/EVA4.pdf., last accessed 2016/09/15 Coll, C., M auri, T., Onrunbia, J.: Los entornos virtuales de aprendizaje basados en el análisis de casos y la resolución de problemas. In Psicología de la educación virtual. 1st end. M orata, Spain (2008). Smith, G., Torres-Ayala, A., Heindel, A.: Disciplinary Differences in E-learning Instructional Design: The Case of M athematics. Journal of Distance Education, 23(3), 63-88. (2019) Cruz-Ramos, M ., González-Calleros, J., & Herrera-Díaz, L.: Desarrollo de Habilidades Comunicativas en Cursos de Inglés en Línea: Una Aproximación desde el Estado del Conocimiento. 1st end. CIATA.org-UCLM ., Ciudad Real (2018). ProQuest Homepage, https://search.proquest.com.ezproxy.cdigital.uv.mx:8443/pqdtglobal/ docview/1708684020/2664C8A40C6345E1PQ/1?accountid=14833, last accessed 2017/06/12. ProQuest Homepage, https://search.proquest.com.ezproxy.cdigital.uv.mx:8443/pqdtglobal/ docview/1506973309/D548205F73BD4C5EPQ/1?accountid=14833, last accessed 2017/12/05. Arias, L.: Impacto de un curso mediado por la Web 2.0 en el desarrollo profesional de un grupo de futuros docentes de inglés. Folios 36, 51-76 (2012). Chang, H., Windeatt, S.: Developing collaborative learning practices in an online language course. Computer Assisted Language Learning, 29(8), 1-16 (2016). Cruz, M ., Sandoval, M .: Retroalimentación entre iguales en sesiones virtuales síncronas. Recopilación de Ponencias del Congreso Internacional de Investigación Academia Journals Celaya, Vol. 7, No. 4, pp. 1169-1174 (2015) Herrera Díaz, L., González M iy, D.: Developing the Oral Skill in Online English Courses Framed by the Community of Inquiry. PROFILE 16(1), 73-88 (2017). Rosales, B., Zárate, J., Lozano, A.: Desarrollo de la competencia comunicativa en el idioma inglés en una plataforma interactiva. Sinéctica. Revista Electrónica De Educación 41, 1-11 (2013). Vine, A., Ferreira, A.: M ejoramiento de la Competencia Comunicativa en Español como Lengua Extranjera a Través de la Videocomunicación. Revista De Lingüística Teórica Y Aplicada 50(1), 139-160(2012). Ellis, R.: Understanding Second Language Acquisition. 1st edn. Oxford University Press, Oxford (1985). Ellis, R.: The Interaction Hypothesis: A Critical Evaluation. Regional Language Centre Seminar, pp. 3-46. Singapore (1991). Ellis, R.: The study of second language acquisition. 2nd edn. Oxford University Press, Oxford (2008).

262 17. 18.

19. 20. 21.

22. 23. 24. 25. 26. 27.

28. 29. 30. 31. 32.

33.

34. 35.

Ghaemi, F., Salehi, N.: Interaction Hypothesis: A Comprehensive Theory of SLA? International Journal for Teachers of English 4(4), 23-33 (2014). Gass, S., M ackey, A.: Input, interaction, and output in second language acquisition. Theories in second language acquisition: An introduction. 1st edn. Lawrence Erlbaum, M ahwah, NJ (2007). Hymes, D.: On communicative competence. 1st edn. Penguin, London (1972). Canale, M ., Swain, M .: Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics 1, 1-47 (1980). Savignon, S.: Communicative Competence: Theory and Classroom Practice. Texts and Contexts in Second Language Learning. 1st edn. Addison-Wesley Publishing Company, Reading M assachusetts (1983) Widdowson, H.: Learning purpose and language use. 1st edn. Oxford University Press, Oxford (1983). Bachman, L.: Fundamental considerations in language testing. 1st edn. Oxford University Press, Oxford (2011). Pilleux, M .: Competencia comunicativa y análisis del discurso. Estudios Filológicos (36), 143–152 (2001). Larsen-Freeman, D., Anderson, M .: Techniques and principles in language teaching. 3rd edn. Oxford University Press, New York (2011). Professor Jack C. Richards, http://www.professorjackrichards.com, last accessed 2018/07/20. Sandoval, M ., Cruz, M .: The CLT Approach in Online Synchronous Sessions. In M emorias del Congreso Internacional de Investigación Academia Journals Celaya 2018 vol. 10, pp. 5096-5101. Academia Journal, Celaya (2018). Churchill, D., King, M ., Fox, R.: Learning design for science education in the 21st century. Zbornik Instituta Za Pedagoska Istrazivanja 45(2), 404-421 (2013). Fox, R.: The rise of open and blended learning. 1st edn. Open University of Hong Kong Press, Hong Kong (2015). Campbell, D., Stanley, J.: Experimental and quasi-experimental designs for research. 1st edn. R. M cNally, Chicago (1966). Fraenkel, J., Wallen, N., Hyun, H.: How to Design and Evaluate Research in Education. 8th edn. M cGraw-Hill, New York (2012). Fernández-García, P., Vallejo-Seco, G., Livacic-Rojas, P., Tuero-Herrero, E.: Validez Estructurada para una investigación cuasi-experimental de calidad: se cumplen 50 años de la presentación en sociedad de los diseños cuasi-experimentales. Anales de Psicología 30(2), 756-771 (2014). Chávez-Espinoza, J., Balderrama-Trápaga, J., Figueroa-Rodríguez, S.: The Dick and Carey Systems Approach M odel: Acercamiento y fundamentación. 1st edn. Creative Common, Ciudad de M éxico (2014). Dick, W., Carey, L., Carey, J.: Introduction to Instructional Design. 8th edn. Pearson, New Jersey (2015). Council of Europe: Common European framework of reference for languages: Learning, teaching, assessment. 1st edn. Cambridge University Press, Strasbourg, France (2018).

263 36.

Hernández, R., Fernández, C., Baptista, P.: M etodología de la investigación. 6th edn. M cGraw-Hill Education, M exico City (2014).

An Architectural Model for Collaborative Content Production to reduce Digital Divide Jaime Muñoz Arteaga 1 , José E. Guzmán Mendoza 2 , César Velázquez Amador1 1

Universidad Autónoma de Aguascalientes, M éxico Universidad Politécnica de Aguascalientes, M éxico jaime.munoz@edu.uaa.mx, jose.guzman@upa.edu.mx vace555@hotmail.com 2

Abstract. This work addresses the digital divide issue as a challenge to achieve the knowledge society. Nowadays, the digital divide in underdeveloped countries is not about the provision of infrastructure and Internet connectivity, is about innovation in models and strategies for digital literacy and digital content generation. This work proposes a collaborative content production model to generate content in order to mitigate the digital divide. For this, a modelcentered approach is required for massive content production to facilitate the integration of users with different skills. This innovative process should have as reference axis acquiring digital skills aimed at preparing people to XXI century digital citizenship. Keywords: User Interfaces, Digital Divide, Collaborative Work, Content Production,

Introduction

Currently, several countries have adopted changes that have led to what is known as "Knowledge Society". These transformations are driven and supported on new technologies to create, transmit and disseminate information, and direct product, which is knowledge. In this sense, it is desirable to achieve a knowledge society where inclusion of individuals in the generation of knowledge can be total, so knowledge societies are sources of development for all and not for a few. However, the difference in access and use of technology has created a new form of exclusion, called "digital divide", that it can impede the economic and human development, and expand the gap between regions and countries (international digital gap) and groups of citizens in a society (domestic digital divide) [1]. Digital divide is not limited only to a technological problem; it is a social and educational problem. The digital divide should be understood in three main concepts: access, ICT education and content [2]. On these concepts, Aguascalientes state in Mexico can be considered as one of state that has a better Internet connectivity infrastructure covering all its municipalities. However, there is still resolve the issues of ICT education -digital literacy- and content.

265

1.1

Digital Divide and Knowledge Society

The concept of “digital divide” appeared in conjunction with the rapid increase in Internet use in the late 1990s, and is typically defined in terms of physical access to computers. Over time the concept has been widened to include a number of issues, including the infrastructure needed to support the use of computers such as internet connections and electricity; physical barriers such as long distance to e.g. a library; economic barriers, such as being able to afford a computer; social and cultural barriers and patterns, including different social patterns in different social groups making for different uses of computers, e.g. for entertainment vs. for searching for information and taking part in politics; technical literacy, the ability to use a computer [3][4][5]. Sharing and creating information are a well know initiatives come from a knowledge society [6]. Ideally, the wealth generation engine in this society is knowledge, which constitutes the main factor of development and human welfare. Therefore, entering the knowledge society will allow any society access other levels of welfare and progress. 1.2

Digital Literacy and Collaborati ve Content Productions

In the knowledge society, a citizen has seen the need to adapt to an era of technological and social changes over the last century. This contemporary society intended to that citizens could be autonomous and own their own ideas, generators of transformations and knowledge to fostering progress in science, information technology, culture and other fields [7]. Digital literacy is a set of skills taught in order to identify and find the information needed to solve a problem [8]. In this sense, being digitally literate involves developing the knowledge and skills relating to ICT, and develop values and social attitudes and political in relation to technology [9]. One of the basic objectives pursued by the use of collaborative work is to promote proper training and proper job performance from the exchange of ideas and actions of the members involved in the management process of the organization, including the promotion of ideas, training activities and proposals for action [10]. In this work, we propose a collaborative content production model that allows collaborative content production, so that any user can collaborate and produce massively open content and use the content produced to help others users to develop basic digital skills and when these new users to appropriate these basic digital skills, they can be added to keep producing new content in order to mitigate their digital divide.

Collaborative content production model

Today, with the wide variety of tools available on the Web 2.0, allows a group of people, virtual communities, even educational institutions and companies to work collaboratively in order to create digital content on a massive scale. And once created these contents, may be available for different end users, such as older adults, Indigenous, professionals, children, etc. We believe that this collaboration can be used to produce digital content that help to reduce the digital divide.

266

People who join content collaboration, can work in both formal and informal environments. For example, formal collaboration can come from educational institutions, training centers, etc. Informal collaboration can occur from virtual communities and social networks. Then, it is necessary to combine these environments. For collaborative work in the production of content, this implies, using platforms that facilitate collaborative task for differentiated users. We use the term differentiated users to refer the task type that a user can performs based on the skills and competencies that provide to the content collaboration. That is, the role played by each user. A user can take the role as the coordinator who monitors and evaluates the quality of content, another user can perform the task of instructional designer, and another user can be the technical designer. In this sense, users must develop digital skills at least at a basic level to be able to perform a variety of tasks for the collaborative production of content. In other words, users should be able to handle different interfaces, devices and technology platforms. This situation creates another problem that relates to the need to provide training to these users, this is known as digital literacy. Digital literacy is introduced in the acquirement of some ICT technical skills. Technological skill refers to the knowledge about what the technology is, how it works, what it is for, and how it can be used to achieve specific objectives. Informational skill is the ability to recognize a need for particular information, and know how to locate, evaluate, select, summarize, and use the information effectively [11]. The collaborative learning environments allow to provide spaces which give developing individual and group skills from the discussion among users in exploring new concepts, each one responsible for their own learning [12]. Based on this premise, we propose the strategy of using the collaborative work focused o n the content production with a dual purpose. The first is to provide content designed according to certain digital skills, for example, a content that allows a person to learn to use a Web browser. And the second purpose is that during the development of content, users can apply the skills acquired and consequently strengthen these skills acquired to get the social appropriation of ICTs. Thus, we determined that the collaborative work focused on digital content production promotes the acquisition of basic digital skills for users (see figure 1).

Fig. 1. Collaborative production of digital content model

267

Servon [2] says that the problem of the digital divide can be understood as three problems: infrastructure problem, problem of digital literacy and content production problem. We're going to focus on solving the problem of content production from a collaborative approach. However, there are few methodologies for collaborative production of content that are supported by technology platforms [19][20]. Some systems provide collaborative platforms and are limited in relation to the tasks that users must perform, creating obstacles for collaborative work. We also want to mention that these systems require users to possess specific skills for its handling. So, our contribution to the reduction of the digital divide is in relation to design a model based on different technological platforms that allow support the collaborative work of a group of users (with different skills) for massive content production As we mentioned, many open platforms are available in Web 2.0 easy to use. These platforms encourage to perform the user tasks through interfaces for collaborative work environments. To exploit these open platforms, we propose the following architecture to support the collaborative content production:

Fig. 2. Architectural model for collaborative content p roduction

268

The objective of the proposed architecture is to provide support for the content production into collaborative platforms, so that users can adapt themselves depending on the environments learning. This architecture consists of four layers: 1) users, 2) Collaborative user interfaces and applications 3) Management for collaborative content production and 4) Resources. In following sections, we describe the operation of the layers. 2.1

Users: Roles and Tasks

In order to promote efficiency and allow people to take initiative in the content production it is important to have clearly defined roles in a multi-skilled team. In this section we propose a set of roles for collaborative content production. Table 1 shows the roles and functions or tasks that must be performed to content production. Table 1. Roles and user tasks for contents production. Role User Task Content This role represents a person or a group of persons that uses various tools Producer to produce content about a specific topic. Idea This person (or group of people) identifies the big picture of the project, Generator leads the others towards the main objective, and keeps the momentum of the project moving forward. Content This role is in charge of reviewing the contents about a subject. Then, this Reviewer person (or group) must work in close relation with the Content Organizer, proposing the appropriate modifications. Technical This role has the function of checking all the technical aspects of the conReviewer tent, for example, checks the consistency of the links, revise the formulas edition, etc. Pedagogical This person ensures that the organization of the content will make sense Reviewer from a pedagogical point of view, hence, works closely to the content organizer in order to notify missing contents, or to point out some elements that must be deeper investigated. Content It is a person able to guide the creation of a content, who has clear insights Organizer into the while providing swift feedback on all aspects of the content. This person proposes a table of contents where a collaborative production strategy will be defined in the group. Group This role is vital in order to organize the team that works in the project. Organizer The person in charge of this role should be characterized by various competencies such as project management, good social relationships, experience within the field of collaboration, and basic knowledge regarding the topic of the content.

269

2.2

Collaborati ve User Interfaces and applications

Collaborative user interfaces provide support for a group of users to perform content production tasks collaboratively. Collaborative UIs are responsible for providing dynamism to the production process and in addition, to establish a methodology facilitating the user tasks. Collaborative applications serves to combine different applications to provide for the users functionalities and allowing to make different collaboration tasks through the UIs. Also, this applications support different patterns for collaboration among the users: asynchronous, synchronous, multi-synchronous, local, distributed, etc. Collaborative applications performs its tasks in terms of user interfaces for content producers according to the role the users perform. The collaborative layer should provide user interfaces able to support a wide of tasks in collaborative environment for content production, and accessible through many possible devices. Many different applications are available to perform similar tasks and users choose those that are easier to understand and interact with and which, consequently, increase efficiency, productivity, and acceptance while reducing errors and the need for training [15]. When a user executes a task through collaborative UI, the UI sends the request to a collaborative application; this application can executes the task and returns an update to the collaborative UI so that other users can view updates (see figure 3).

Fig. 3. Collaborative graphical user interface model

Such collaborative layer can combine different applications and user interfaces for: communication, supporting different phases of process production, searching information, managing users, assessing content, etc., this facilitates the system to be available anywhere anytime.

270

2.3 Management for Collaborative Content Production The management for collaborative content production describes the workflow and user tasks for the content production using the defined collaborative UIs and applications provided by the collaborative layer. In this layer, the collaborative production of content is handled. This administration is determined by a series of tasks that forming a collaborative process. The process includes five tasks that are oriented to the production of content: 1) research, 2) ideas, 3) placement, 4) production and 5) publishing, and two tasks that are designed to control the quality of the content: 1) feedback and 2) Assessment. Thus, this manager allow for an easy generation/configuration of different collaborative application specific services for collaborative product/process design in international settings along the product/process development life cycle of a content, adapted to the specific needs of different users and their tasks. The platforms have to be open for different production content frameworks, i.e., they have to support different collaborative design approaches by combining different existing design tools with web services and LMS services . 2.4

Resources

Resources are on the backend which is composed of web services as a backbone for an entire architecture to operate and have the characteristics of cloud computing. Web services can provide services through Internet allowing to support formal and informal collaborative learning environments. These services include three characteristics: a) service by demand, b) Ubiquitous access to network and c) configuration elasticity [13]. Also, web services have characteristic that could be us ed as a repository to include different kind of content as Learning Objects (LO). LOs based learning environments are mostly used within web environments, especially through distance education. The LO can be designed to support exploration and investigation of ICT ideas and help conceptual understanding [14], thus, LOs allow the end users the acquisition of digital skills. With LMS, formal environments are integrated. Next table presents a list of some collaborative tools that can support collaborative tasks through collaborative UIs from open resources. Table 2. Tools to support UI for collaborative tasks Collaborative UI Research Ideas Placement Production Publishing

Collaborative Tools Google, Scoop.it, Slideshare, youtube, ZonaClic. M icrosoft Academic, M yScienceWork, Scientific Journal Finder. M indmeister, symbaloo, M iro, Google Calendar, Bubbl.us, IdeaBoardz Dropbox, Google Drive, Connexions, M iro, edocr, OneDrive, pCloud, Box, M ega M oodle, Educaplay, BookType, Connexions, Wix, Construcor, Cuadernia, Exelearning, BookStack, WordPress,Blogger Slideshare, M oodle, BookType, Wix, Scoop it, Scribd, WordPress, Blogger

271

FeedBack Assesment

M iro, M eetingBurner, skype, Wikispaces, M oodle, Join.me, Ekiga, Google Hangouts. M iro, Google Docs, OpenSchool ePorfolio, watermark.

With the proposed architecture, we intend to generate more flexible environments, which allow inclusion of differentiated users, with the understanding that a differentiated user can be a person or a community with different needs and characteristics. Through these flexible environments the users start as consumers of content, and after they grasp a high level of digital skills they can become content providers. We want to emphasize that formal education is inflexible because it excludes many people who do not have the skills, qualities, requirements or resources to join formal learning environments. Older adults, indigenous communities, people with low economic resources, even very small companies are different users who need to be literate in the use of ICT to join the society of knowledge, however, no one takes care of them. Informal environments can provide possible solutions to different users.

Case Study

The Government Plan of Aguascalientes State 2010-2016 considers within its main activities carried Aguascalientes toward a Knowledge Society as a strategy for improving the quality of life of the people of Aguascalientes. In order to illustrate our model this section presents a case study for the collaborative content production within the Fomix Project (FP) entitled “Integral Intervention to Reduce Digital Divide in Aguascalientes State”[16] and also poses a strategic way to develop basic digital skills through the creation of digital content collaboratively. This will contribute to the social appropriation of ICT in an effective and efficient to allow the entire population of Aguascalientes join the knowledge society. However, Aguascalientes is a state with very heterogeneous population, and to solve this situation we need to create different scenarios for integrating the various sectors of the population to the knowledge society. Within the FP involved several instructors that in the future they will be the digital facilitators, and through them, operate the FP directly to the Aguascalientes citizens. But first, the instructors must acquire the basic and intermediate digital skills, and then replicate. We follow the strategy that was raised and illustrated in figure 1. Adapting the general strategy to the FP context, is described as follows “Provide instructors the ability to build additional content to reduce the digital divide attached to basic and intermediate skills established in the FP”. So, Instructors can develop basic and intermediate digital skills (see figure 5) through the production of digital content collaboratively. This strategy allows one hand, creating collaborative work scenario and on the other, producing digital content in order to training Aguascalientes citizens. So, once citizens acquire digital competences, can become content producers. Figure 4 illustrates the application of the general strategy for the case study.

272

Fig. 4. General strategy applied to FP Project for collaborative content production.

As a case study we establish a categorizing digital skills into three levels that should be acquired by instructors and citizens in order to produce content through different online resources .

Fig. 5. Classification of digital skills

3.1

Creating a scenario for collaborative content production for case study.

To illustrate the operation of architecture to generate scenarios that promote collaborative production of content on a massive scale, we designed the following scenario for the FP case study.

273

Fig. 6. Scenario for collaborative content production

The scenario is developed for the Aguascalientes state context. It is based on virtual environments and distance to cover formal and informal learning, thus, different users can be integrated. Moodle and some cloud services (from table 2) are the backbone for providing this support. 3.2

Users: Roles and Tasks

Table 3 shows the different roles that can be performed by the users defined in the case study. Table 3. Users and roles. Users Instructors and Citizens

Role Content Producer, Content Reviewer, Content Organizer and Group Organizer.

Teacher

Content Reviewer, Technical Reviewer, Pedagogical Reviewer, Content Organizer and Group Organizer

Coordinator

Ideas, Generator and Content Reviewer.

274

3.3

Management for collaborative content production

To have a better understanding of the implementation of the management for collaborative content production (described in 2.3), the tasks to be performed by users of case study are illustrated in figure 7. We decided to use BPMN [17][18] for modeling the manager in terms of collaborative processes.

Fig. 7. Collaborative process for management of content production.

Independently that interfaces and platforms are used, figure 7 shows the processes that must be executed for the content production content by the users or roles. In this case, the processes are assigned to users, but, also be assigned to roles. A user can perform several roles in the process. 3.4

Collaborati ve User Interfaces and Applications

In this section we give an example of how the scenario manages and supports the processes in figure 8 on the side of the users group in relation to the role performed. We mention the production task in terms of its user interface.

275

Fig. 8. Graphical user interface for content production

Through the production UI: Teacher takes roles as content reviewer, technical reviewer and pedagogical reviewer. The teacher is responsible for periodically reviewing the production phase. Teacher valued pedagogical and technical aspects of content ensuring the quality of the content, generates a valuation report and provide feedback to the instructors. Instructors take role as content producer and some of them can also take the role of content reviewer. They use different applications to generate content according to the pedagogical and technical requirements. They are also based on the feedback of the teacher to make changes to the content. 3.5

Resources

According to the table 2, moodle, slideshare and scoop.it are resources that integrate tools that support the task of publication for the case study.

276

Fig. 9. Use of moodle, slideshare and scoop.it as a content repository

As shown in figure 9, moodle, slideshare and scoop.it are resources, they are used as content repositories. At the same time, these resources allow all the users to have access to these content. Also, these resources can support the task of feedback and allow users to assign a score to the content.

Conclusion

The digital divide must be understood as a process of social transformation, which requires the development of strategies, tools and public policies that contribute to wider access to and use of ICTs, through the development of digital skills needed to integrate Aguascalientes State into knowledge society. This involves creating environments where infrastructure, content and ICT education as a strategy to merge a solid intervention. We present a collaborative environment for content production as a proposal to reduce the problem of the digital divide in the state of Aguascalientes. Also, we find that the MSD are able to articulate collaborative strategies that allow digital content can be designed, developed and implemented within the same, and then generalized and expanded to citizenship. So, when citizens consume these contents, acquire basic digital skills Thus, it is intended that citizens are not content consumers only, in the moment they are integrating to knowledge society are able to generate new content and become content providers both individually and collectively. Moreover, it is important to the selection of technology resources. The resources should facilitate the collaborative content production scenarios, and ensure easy ac-

277

cess to contents. Our scenario described by the case study shows the ease in handling and performing tasks collaboratively by all roles. As future work, we will continue working on searching to strengthen the educational and technological features that should have digital content to be viable in other technology resources.

References 1. 2. 3. 4. 5.

6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16. 17. 18.

E. Tello Leal, «Las tecnologías de la información y comunicación (TIC) y la brecha digital: su impacto en la sociedad de M éxico», Rev. Univ. Soc. Conoc., 4(2) (2007). L. Servon, «Bridging the Digital divide: Technology, Community, and Public Policy», M alden M a Blackwell Pub (2002) R. K. Payne, «A Framework for Understanding Poverty.», Third Revis. Ed. Aha Process Inc Highl. TX (1996) M . Gurstein, «Effective Use: A community informatics strategy beyond the Digital Divide.», First M onday, 8 (12) (2003). M . A. Wimmer, H. J. Scholl, y A. Grönlund, «Inclusión in the E-Services Society Investigating Administrative Literacy Requirements for Using E-Services», SpringerVerl. Berl. Heidelb., pp. 216-227 (2007) L. Olivé, «El libro, la lectura y las bibliotecas en la sociedad del Conocimiento», Lect. Vida, 30(39), pp. 20-29 (2009). E. Álvarez, «Cinco ventajas del aprendizaje colaborativo online», Colombia Digital (2013) R. M onge y E. Frisicaro-Pawlowski, «redefining Information Literacy to Prepare Students for the 21st Century Workforce», Springer Sci. M edia New York (2013) M . Pinto, «Información, Acción, Conocimiento y Ciudadanía. La educación escolar como espacio de interrogación y de construcción de sentido», Gedisa Barc. (2008.) K. Jiménez G., «Propuesta estratégica y metodológica para la gestión en el trabajo colaborativo», Rev. Educ., 33(2), pp. 95-107 (2009) M . G. Badilla Quintana y M . Cortada Pujol, «Digital Literacy Development of Studens Involved in an ICT Educational Project», vol. 73, pp. 632-638 TechEducation (2010). M . M . Lucero, «Entre el Trabajo Colaborativo y el Aprendizaje Colaborativo», Rev. Iberoam. Educ. (2003) D. S. Linthicum, «Cloud Computing and SOA Convergence in Your Enterprise: A step-by-step guide», Addison-Wesley Prof. (2009) A. Baki y U. Cakroglu, «Learning Objects in high school mathematics classrooms: implementation and evaluation», Comput. Educ., vol. 55, pp. 1459-1469 (2010) G. M ori, F. Paterno, y C. Santoro, «CTTE: Support for Developing and Analyzing Task M odels for Interactive System Design», IEEE Trans. Softw. Eng., 28(8), (2002). Gobierno del Estado de Aguascalientes, «Plan sexenal del Gobierno 20102016»(2010). [S. A. White y D. M iers, «BPM N. Guía de Referencia y M odelado», Future Strat., (2009) OM G, «Business Process M odel and Notation (BPM N)V2», OM G Website (2011).

278 19.

20.

R. G. Ruiz, A. R. García, and M . del M . R. Rosell, “Educación en alfabetización mediática para una nueva ciudadanía prosumidora - M edia Literacy Education for a New Prosumer Citizenship,” Rev. Comun., 22 (43), pp.15–23 (2014). D. Contreras Guzmán, N. Vásquez Lara, and V. Salinas Silva, La Alfabetización Digital y el Desarrollo de Competencias Ciudadanas, Ediciones Universitarias de Valparaiso. Chile (2012).

Research and development of new technology Edited by Eugenia Erica Vera Cervantes Olga Leticia Fuchs Gómez Edited by this provision in htlm in the Computer Science Faculty page of the Benemérita Universidad Autónioma de Puebla (BUAP) http://www.dgie.buap.mx/dsae/index.php October 2019 File Weight 11.5 MB

View publication stats