Jmp 2016 28 issue 1

Volume 28 / Number 1 / 2016

Volume 28 / Number 1 / 2016

Journal of

Media Psychology Journal of Media Psychology

Editor-in-Chief Nicole Krämer

Theories, Methods, and Applications

Associate Editors Gary Bente Nick D. Bowman Christoph Klimmt Mary Beth Oliver Diana Rieger

zmp_55-0-0-0_58-60_positiv.indd 3

09.12.2015 11:47:57

Hogrefe OpenMind Open Access Publishing?

It’s Your Choice!

Your Road to Open Access Authors of papers accepted for publication in any Hogrefe journal can now choose to have their paper published as an open access article as part of the Hogrefe OpenMind program. This means that anyone, anywhere in the world will – without charge – be able to read, search, link, send, and use the article for noncommercial purposes, in accordance with the internationally recognized Creative Commons licensing standards.

The Choice is Yours 1. Open Access Publication: The final “version of record” of the article is published online with full open access. It is freely available online to anyone in electronic form. (It will also be published in the print version of the journal.)

2. Traditional Publishing Model: Your article is published in the traditional manner, available worldwide to journal subscribers online and in print and to anyone by “pay per view”.

Whichever you choose, your article will be peer-reviewed, professionally produced, and published both in print and in electronic versions of the journal. Every article will be given a DOI and registered with CrossRef.

www.hogrefe.com

How Does Hogrefe‘s Open Access Program Work?

After submission to the journal, your article will undergo exactly the same steps, no matter which publishing option you choose: peer-review, copy-editing, typesetting, data preparation, online reference linking, printing, hosting, and archiving. In the traditional publishing model, the publication process (including all the services that ensure the scientific and formal quality of your paper) is financed via subscriptions to the journal. Open access publication, by contrast, is financed by means of a one-time article fee (€ 2,500 or US $3,000) payable by you the author, or by your research institute or funding body. Once the article has been accepted for publication, it’s your choice – open access publication or the traditional model. We have an open mind!

Journal of

Media Psychology Theories, Methods, and Applications

Volume 28, No. 1, 2016

Editor-in-Chief

Editorial Assistant

Associate Editors

Nicole Kra¨mer, Social Psychology: Media and Communication, Department of Computer Science and Applied Cognitive Science, University Duisburg-Essen, Forsthausweg 2, 47057 Duisburg, Germany, Tel. +49 203 379-2482, Fax +49 203 379-3670, E-Mail nicole.kraemer@uni-due.de German Neubaum, Social Psychology: Media and Communication, University of Duisburg-Essen, Forsthausweg 2, Office LE 243, 47057 Duisburg, Germany, Tel. +49 203 379-2442, Fax +49 203 379-3670, Email german.neubaum@uni-due.de Gary Bente, University of Cologne, Cologne, Germany, E-mail bente@uni-koeln.de Nick D. Bowman, West Virginia University, Morgantown, VA, USA, E-mail Nicholas.Bowman@mail.wvu.edu Christoph Klimmt, Hanover University of Music, Drama and Communication Research, Germany, E-mail christoph.klimmt@ijk.hmtm-hannover.de Mary Beth Oliver, Penn State University, University Park, PA, USA, E-mail mbo@psu.edu Diana Rieger, University of Mannheim, Germany, E-mail diana.rieger@uni-mannheim.de

Annie Lang (Bloomington, IN, USA) Eun-Ju Lee (Seoul, South Korea) Jo¨rg Matthes (Vienna, Austria) Peter Nauroth (Marburg, Germany) Anne Oeldorf-Hirsch (Mansfield, CT, USA) Jochen Peter (Amsterdam, Netherlands) Daniel Pietschmann (Chemnitz, Germany) Robert F. Potter (Bloomington, IN, USA) Arthur A. Raney (Tallahassee, USA) Leonard Reinecke (Mainz, Germany) Meghan Sanders (Baton Rouge, LA, USA) Frank Schwab (Wu¨rzburg, Germany) Stephan Schwan (Tu¨bingen, Germany) Michael Slater (Columbus, OH, USA) Kaveri Subrahmanyam (Los Angeles, CA, USA) Ron Tamborini (East Lansing, MI, USA) Catalina Toma (Madison, WI, USA) Sabine Trepte (Hohenheim, Germany) Mina Tsay-Vogel (Boston, MA, USA) Dagmar Unz (Wu¨rzburg-Schweinfurt, Germany) Sonja Utz (Tu¨bingen, Germany) Sebastia´n Valenzuela (Santiago de Chile, Chile) Brandon van der Heide (East Lansing, MI, USA) Christian von Sikorski (Vienna, Austria) Peter Vorderer (Mannheim, Germany) Patrick Weber (Hohenheim, Germany) Rene´ Weber (Santa Barbara, CA, USA) Stephan Winter (Duisburg-Essen, Germany) Mike Yao (Hong Kong, ROC)

Editorial Board

Markus Appel (Koblenz-Landau, Germany) Florian Arendt (Mu¨nchen, Germany) Omotayo Banjo (Cincinnati, OH, USA) Anne Bartsch (Mu¨nchen, Germany) Paul Bolls (Columbia, MO, USA) Johannes Breuer (Cologne, Germany) Jennings Bryant (Tuscaloosa, AL, USA) Caleb Carr (Normal, IL, USA) Elizabeth Cohen (Morgantown, WV, USA) Enny Das (Nijmegen, The Netherlands) Kevin Durkin (Glasgow, UK) Allison Eden (Amsterdam, The Netherlands) Nicole Ellison (Ann Arbor, MI, USA) Malte Elson (Bochum, Germany) David Ewoldsen (Columbus, OH, USA) Christopher Ferguson (DeLand, FL, USA) Jesse Fox (Columbus, OH, USA) Sabine Glock (Wuppertal, Germany) Melanie Green (Chapel Hill, NC, USA) Matthew Grizzard (Buffalo, NY, USA) Dorothe´e Hefner (Hannover, Germany) Shirley S. Ho (Singapore) Matthias Hofer (Zurich, Switzerland) Juan Jose´ Igartua (Salamanca, Spain) Jimmy Ivory (Blacksburg, VA, USA) Jeroen Jansz (Rotterdam, The Netherlands) Julia Kneer (Rotterdam, The Netherlands) Elly Konijn (Amsterdam, The Netherlands) Maja Krakowiak (Colorado Springs, CO, USA)

Publisher

Hogrefe Publishing, Merkelstr. 3, 37085 Go¨ttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-425, E-mail publishing@hogrefe.com, Web www.hogrefe.com North America: Hogrefe Publishing, 38 Chauncy Street, Suite 1002, Boston, MA 02111, USA, Tel. (866) 823-4726, Fax (617) 354-6875, E-mail publishing@hogrefe.com Regina Pinks-Freybott, Hogrefe Publishing, Merkelstr. 3, 37085 Go¨ttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-425, E-mail production@hogrefe.com Hogrefe Publishing, Herbert-Quandt-Str. 4, 37081 Go¨ttingen, Germany, Tel. +49 551 50688-0, Fax +49 551 50688-24

Production Subscriptions Advertising/Inserts ISSN Copyright Information

Publication

Subscription Prices

Payment

Electronic Full Text Abstracting/Indexing Services

Melanie Beck, Hogrefe Publishing, Merkelstr. 3, 37085 Go¨ttingen, Germany, Tel. +49 551 99950-423, Fax +49 551 99950-425, E-mail marketing@hogrefe.com ISSN-L 1864-1105, ISSN-Print 1864-1105, ISSN-Online 2151-2388 Ó 2016 Hogrefe Publishing. This journal as well as the individual contributions and illustrations contained within it are protected under international copyright law. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without prior written permission from the publisher.

All rights, including translation rights, reserved.

Published in 4 issues per annual volume. The Journal of Media Psychology is the continuation of Zeitschrift fu¨r Medienpsychologie (ISSN 1617-6383), the last annual volume of which (Volume 19) was published in 2007.

Calendar year subscriptions only. Rates for 2016: Institutions: US $324.00 / 1238.00 / £190.00; Individuals: US $159.00 / 1114.00 / £91.00 (all plus US $16.00 / 112.00 / £10.00 postage & handling).

Single issue US $81.00 / 159.00 / £47.50 (plus postage & handling) Payment may be made by check, international money order, or credit card, to Hogrefe Publishing, Merkelstr. 3, 37085 Go¨ttingen, Germany. US and Canadian subscriptions can also be ordered from Hogrefe Publishing, 38 Chauncy Street, Suite 1002, Boston, MA 02111, USA. The full text of Journal of Media Psychology is available online at http://econtent.hogrefe.com/ and in PsycARTICLES. Abstracted/indexed in Current Contents/Social and Behavioral Sciences (CC/S&BS), Social Sciences Citation Index (SSCI), IBR, IBZ, PsycINFO, PsycLit, PSYNDEX, and Scopus. Impact Factor (2014): 0.882.

Journal of Media Psychology 2016; Vol. 28(1)

Ó 2016 Hogrefe Publishing

Contents Original Articles

The Spatial Presence Experience Scale (SPES): A Short Self-Report Measure for Diverse Media Settings Tilo Hartmann, Werner Wirth, Holger Schramm, Christoph Klimmt, Peter Vorderer, Andre´ Gysbers, Saskia Bo¨cking, Niklas Ravaja, Jari Laarni, Timo Saari, Feliz Gouveia, and Ana Maria Sacau

1

Social Categorization, Moral Disengagement, and Credibility of Ideological Group Websites Shane Connelly, Norah E. Dunbar, Matthew L. Jensen, Jennifer Griffith, William D. Taylor, Genevieve Johnson, Michael Hughes, and Michael D. Mumford

16

Violent Lyrics = Aggressive Listeners? Effects of Song Lyrics and Tempo on Cognition, Affect, and Self-Reported Arousal Stephanie Pieschl and Simon Fegers

32

Research Report

Effects of Light-Hearted and Serious Entertainment on Enjoyment of the First and Third Person Matthias Hofer

42

Erratum

Correction to Retzbach, Retzbach, Maier, Otto, & Rahnke, 2013

49

Meeting Calendar

Ó 2016 Hogrefe Publishing

50

Journal of Media Psychology 2016; Vol. 28(1)

Original Article

The Spatial Presence Experience Scale (SPES) A Short Self-Report Measure for Diverse Media Settings Tilo Hartmann,1 Werner Wirth,2 Holger Schramm,3 Christoph Klimmt,4 Peter Vorderer,5 André Gysbers,6 Saskia Böcking,7 Niklas Ravaja,8 Jari Laarni,9 Timo Saari,10 Feliz Gouveia,11 and Ana Maria Sacau12 1

Department of Communication Science, VU University Amsterdam, Amsterdam, The Netherlands, 2 Institute of Mass Communication and Media Research, University of Zurich, Zurich, Switzerland, 3 Department Human-Computer-Media, University of Würzburg, Würzburg, Germany, 4Hanover University of Music, Drama, and Media, Hanover, Germany, 5Department of Media and Communication Studies, University of Mannheim, Mannheim, Germany, 6Wink Stanzwerkzeuge, Neuenhaus, Germany, 7Energy Enterprise, Bern, Switzerland, 8Department of Social Research, University of Helsinki, Helsinki, Finland, 9VTT Technical Research Centre of Finland, Espoo, Finland, 10Department of Pervasive Computing, Tampere University of Technology, Tampere, Finland, 11Faculty of Science and Technology, University Fernando Pessoa, Porto, Portugal, 12Faculty of Human and Social Sciences, University Fernando Pessoa, Porto, Portugal Abstract. The study of spatial presence is currently receiving increased attention in both media psychology and communication research.

The present paper introduces the Spatial Presence Experience Scale (SPES), a short eight-item self-report measure. The SPES is derived from a process model of spatial presence (Wirth et al., 2007, Media Psychology, 9, 493–525), and assesses spatial presence as a two-dimensional construct that comprises a user’s self-location and perceived possible actions in a media environment. The SPES is shorter than many other available spatial presence scales, and can be conveniently applied to diverse media settings. Two studies are reported (N1 = 290, N2 = 395) that confirm sound psychometric qualities for the SPES. Keywords: spatial presence, measurement, self-report, immersion

Rapid advances in communication technologies have changed the way people use and experience media. New videoconferencing systems such as Cisco’s TelePresence SuiteTM, video games, 3D movies, and high-definition television are typical examples of such media change. A basic concept that has emerged with the growing body of research examining new media is spatial presence (Lombard & Ditton, 1997; also physical presence in Lee, 2004; telepresence in Draper, Kaber, & Usher, 1998).1 Spatial presence can be briefly defined as the user’s subjective feeling of ‘‘being there’’ in the space displayed by a medium (Riva, Davide, & IJsselsteijn, 2003; International Society 1

for Presence Research, ISPR, 2001; Sheridan, 1992; Slater & Wilbur, 1997). The concept emerged from the observation that users of virtual reality systems feel physically located in the mediated space (Slater & Steed, 2000; Steuer, 1992), but other research suggests that users can also feel spatially present while playing video games (Tamborini & Skalski, 2006), watching television (Bracken, 2005), and even reading books (Schubert & Crusius, 2002).

Spatial presence is considered an important concept in both psychology (Blascovich et al., 2002) and communication science (Lee, 2004). It has been linked to relevant constructs such as learning (e.g., Tichon, 2007), therapeutical

Spatial presence is a specific construct of a broader class of presence phenomena (Lombard & Ditton, 1997). Spatial presence focuses on ‘‘spatial illusions’’ and can be distinguished from social presence (‘‘the feeling of sharing a social situation with others’’; Lee, 2004) and transportation (‘‘the feeling of being part of a narrative’’; Green, Brock, & Kaufman, 2004). The way we use the term, spatial presence refers to the same user experience that others have addressed as physical presence (Lee, 2004) or telepresence (Draper, Kaber, & Usher, 1998).

2015 Hogrefe Publishing

Journal of Media Psychology 2016; Vol. 28(1):1–15 DOI: 10.1027/1864-1105/a000137

2

T. Hartmann et al.: Spatial Presence

issues (Robillard, Bouchard, Fournier, & Renaud, 2003), social behavior (Yee & Bailenson, 2007) including aggression (Eastin & Griffiths, 2006), enjoyment (e.g., Tamborini & Skalski, 2006), and the effectiveness of advertisement (Jin, 2010). Various measurements to assess spatial presence have been proposed in the past (Insko, 2003; Laarni et al., in press). Because spatial presence is considered a conscious experience or feeling (e.g., Biocca, 1997; Lombard & Ditton, 1997; Schubert, 2009; Sheridan, 1992; Witmer, Jerome, & Singer, 2005), subjective self-report measures represent an important type of measurement. Although several self-report measures of spatial presence exist, the ‘‘market’’ still offers a niche for a short, easily and flexibly applicable, and theoretically substantiated scale. In the present article, we introduce such a measure: the Spatial Presence Experience Scale (SPES).

The SPES rests on a theoretical model of spatial presence (Wirth et al., 2007), it is short (eight items), and it can be applied to diverse media settings. We report two studies (N1 = 290, N2 = 395) that confirm sound psychometric qualities for the SPES, and we discuss the scale’s practicability and applicability.

Review of Existing Self-Report Measures of Spatial Presence Researchers concerned with the study of spatial presence are able to choose from various self-report tools (for overviews, see IJsselsteijn, de Ridder, Freeman, & Avons, 2000; Laarni et al., in press; Schuemie, van der Straaten, Krijn, & van der Mast, 2001). According to our own experience, the most popular tools designed to assess spatial presence include the Presence Questionnaire (PQ; Witmer & Singer, 1998; Witmer et al., 2005), the Independent Television Commission–Sense of Presence Inventory (ITC-SOPI; Lessiter, Freeman, Keogh, & Davidoff, 2001), the Igroup Presence Questionnaire (IPQ; Schubert, Friedmann, & Regenbrecht, 2001), and the Temple Presence Inventory (TPI; Lombard, Ditton, & Weinstein, 2009). Subscales of these tools can also be used as short measures of spatial presence. In addition, scholars have proposed stand-alone short measures of spatial presence (Hendrix & Barfield, 1996; Kim & Biocca, 1997; Slater, Usoh, & Steed, 1995).

Existing Popular Scales of Spatial Presence Presence Questionnaire The Presence Questionnaire (PQ) was developed to measure spatial presence in immersive virtual environments that allow users to navigate through sceneries conveyed by highly immersive technology (e.g., head-mounted displays).

The PQ not only aims to assess the intensity of spatial presence as a state, but also strives to assess contributing factors (Witmer & Singer, 1998, p. 230). A second version of the PQ (Witmer & Singer, 1998) consisted of 17 items that Journal of Media Psychology 2016; Vol. 28(1):1–15

could be collapsed into a total presence score (a = .88).

Witmer and Singer (1998) report a significant negative correlation between the PQ total score and the Simulator Sickness Questionnaire (Kennedy, Lane, Berbaum, & Lilienthal, 1993), indicating the validity of the PQ. However, subsequent assessments of the validity of the PQ have yielded mixed results (Johns et al., 2000; Nystad & Sebok, 2004; Youngblut & Perrin, 2002). More recently, Witmer et al. (2005) proposed a revised third version of the PQ with several additional items. Data pertaining to 325 users of immersive virtual environments resulted in a 4-factor solution that included 29 items.

The 4 factors were labeled involvement (i.e., ‘‘focusing one’s mental energy and attention on the stimulus’’; Witmer et al., 2005, p. 308), sensory fidelity (i.e., the accuracy of sensory stimulation), adaptation/immersion (i.e., ‘‘perceiving oneself to be enveloped by, included in, and interacting with an environment’’; p. 308), and interface quality (i.e., the degree to which ‘‘display devices interfere/distract from task performance’’; p. 299).

The PQ is well suited to assess users’ experiences in interactive virtual reality systems, particularly if users perform a task. Potential problems associated with the PQ, however, may include its narrow scope (i.e., the wording of the questions is bound to interactive virtual environments; see Lessiter et al., 2001), a mix of states of spatial presence (e.g., feeling immersed) and determinants (e.g., identification of sounds), the absence of an explicated theoretical basis of spatial presence (Waller & Bachmann, 2006), and mixed findings regarding the validity of the measure.

Igroup Presence Questionnaire As with the PQ, the Igroup Presence Questionnaire (IPQ; Schubert, 2003; Schubert et al., 2001) aims to measure spatial presence as a sense ‘‘of being there’’ in interactive virtual environments. The 13-item IPQ was derived based on both explorative and confirmative factor analyses conducted in two survey studies (N = 246 and N = 296).

Although embedded in an intriguing conceptualization of spatial presence (‘‘embodied cognition framework’’; Schubert et al., 2001, p. 267; Schubert, Friedmann, & Regenbrecht, 1999), the IPQ’s initial item pool consisted of 75 items mainly sourced from existing presence questionnaires. The IPQ measures three potential subcomponents of presence (Schubert, 2003): spatial presence (‘‘being surrounded by the virtual environment, directly interacting in it. . ., and a sense of transportation to another place’’; Schubert, 2003, p. 70), involvement (‘‘focusing on the virtual environment instead of focusing on the real world’’; p. 70), and realness (‘‘how real the virtual environment is judged to be’’; p. 70).

The scales yielded acceptable internal consistency (all a > .62; Schubert, 2003), and the results of experimental tests support the validity of the IPQ (Regenbrecht & Schubert, 2002; Schubert, 2003). The IPQ builds on sound methodological steps, although the strong theoretical conceptualization is not fully reflected in the approach. Consequently, only one of the three IPQ subscales actually measures spatial presence, whereas involvement and 2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

realness may address closely related constructs or determinants rather than actual subdimensions of spatial presence.

Independent Television Commission–Sense of Presence Inventory The Independent Television Commission–Sense of Presence Inventory (ITC-SOPI; Lessiter et al., 2001) was designed to assess spatial presence across different types of media. The ITC-SOPI was developed based on an initial pool of 63 items (Lessiter et al., 2001) collected to indicate ‘‘possible manifestations of different content areas deemed relevant to presence’’ (Lessiter et al., 2001, p. 287), such as a sense of space, attention, and potential negative effects. In a study involving 604 participants and based on explorative factor analyses, a 4-factor solution (with a total of 43 items) was chosen ‘‘that made the most conceptual sense’’ (Lessiter et al., 2001, p. 290). The four subscales address a sense of physical space (physical placement in the mediated environment, and interaction with the environment), engagement (feeling psychologically involved and enjoying the content), ecological validity (perceiving the mediated environment as lifelike and real), and negative effects (‘‘adverse physiological reactions’’; Lessiter et al., 2001, p. 290).

Internal consistency is not reported for the final scales, but previous versions yielded a values > .76. Preliminary evidence reported by Lessiter et al. (2001) supports the validity of the scales. In summary, the ITC-SOPI is a frequently used spatial presence scale that can be applied to various types of media. With 43 items it is a rather lengthy measure, however. In addition, some of the four inductively derived factors (e.g., negative effects) may not resemble theoretically meaningful dimensions of spatial presence.

Existing Short Scales of Spatial Presence Hendrix and Barfield (1996) applied a two-item measure to assess spatial presence in virtual environments. One item asked participants to directly rate their sense of presence: ‘‘If your level in the real world is 100, and your level of presence is 1 if you have no presence, rate your level of presence in this virtual world’’ (Hendrix & Barfield, 1996). Another item asked participants on a 5-point scale how strongly they felt a sense of presence or ‘‘being there’’ in the virtual environment. Both items were applied in a smaller experimental study (N = 12) that tested effects of technological manipulations (e.g., monoscopic vs. stereoscopic display) on spatial presence.

These manipulations significantly affected both items.

In addition, items were correlated with a measure of realism. The two-item measure by Hendrix and Barfield is very short, and preliminary evidence supports its validity.

However, both items still need to be validated in bigger samples, and more extensive psychometric testing seems necessary. In addition, the wording of the first item is quite complicated and may be difficult to understand for some participants (Lessiter et al., 2001).

2015 Hogrefe Publishing

3

Kim and Biocca (1997) developed an eight-item measure of spatial presence, understood as a sense of being transported to the media world. The scale was applied in a 2 (presence of real-world visual stimuli: yes, no) · 3 (viewing angle: low, medium, high) between-subjects experiment on spatial presence in television exposure (N = 96). Unexpectedly, a factor analysis of the scale suggested 2 factors that distinguished positive from negative items. Accordingly, one factor was interpreted as arrival – or being in the media environment, the other as departure – or not being present in the media environment (Kim & Biocca, 1997). All but one item had only marginal crossloadings. Unfortunately, the measure did not respond to the experimental manipulation. The scale by Kim and Biocca follows an interesting approach, but contrary to expectations, it did not assess a unidimensional construct.

In addition, the nonsignificant findings of the experiment provide a serious challenge for the validity of the scale.

Slater, Usoh, and Steed (1995) developed another short measure of spatial presence: the Slater-Usoh-Steed (SUS) questionnaire. The SUS exists in an (older) three-item version (Slater, Usoh, & Steed, 1994, 1995) and in a more recent six-item version (Usoh, Catena, Arman, & Slater, 2000). The original three items assess on a 7-point scale users’ sense of being in a virtual environment, and the extent to which the virtual environment becomes the dominant reality and to which users feel like visiting somewhere rather than seeing pictures of something. In a small experiment by Slater et al. (1995), participants (N = 16) scored higher on the three-item SUS if they navigated through a virtual environment by naturally walking ‘‘in a place’’ in the lab than if navigating with a more artificial device – that is, by pressing a mouse button. The SUS has been applied in a range of other studies, usually with satisfactory results supporting its validity (e.g., Bormann, 2005). In a study (N = 20) that applied the six-item version of the SUS (Usoh et al., 2000), only two of the six items successfully discriminated a sense of spatial presence in the real world from a sense of spatial presence in a virtual environment. In addition, the six-item SUS was not significantly correlated with the PQ of Witmer and Singer (1998) among participants using a virtual environment.

Concluding Remarks About Existing Spatial Presence Scales Researchers have developed a couple of valuable measures in the past to assess spatial presence. All of these measures have contributed to progress in the field (Laarni et al., in press; Schuemie et al., 2001) and helped in the advance toward a standardized assessment of spatial presence. From a more critical point of view, a couple of caveats can be noted, however. Only a few of the published scale developments included an extensive testing of the scale’s psychometric qualities (e.g., IPQ or ITC-SOPI). Accordingly, it is difficult to evaluate the quality of the some of the existing measures. Furthermore, some of the existing scales, such as the PQ, IPQ, or ITC-SOPI, were derived based on an Journal of Media Psychology 2016; Vol. 28(1):1–15

4

T. Hartmann et al.: Spatial Presence

inductive approach, relying on factor analyses. Inductive approaches to scale development, however, have not been free of criticism (Clark & Watson, 1995; Cronbach & Meehl, 1955; Waller & Bachmann, 2006). The dimensions of some of the inductively derived measures may not necessarily represent the spatial presence concept, but may also be the result of methodological factors (Waller & Bachmann, 2006; Witmer et al., 2005). For example, little has been said in theory regarding ‘‘negative effects’’ (as assessed by the ITC-SOPI) as a dimension of spatial presence. And the realism subscale of the IPQ or the interface quality subscale of the PQ may capture determinants rather than dimensions of spatial presence.

The present article introduces the SPES as an alternative short measure of spatial presence experiences.2 In contrast to the ITC-SOPI and IPQ, the SPES was derived based on a deductive rather than inductive approach. In contrast to the SUS, which focuses on immersive virtual environments, the SPES is particularly designed and tested as a measure of spatial presence across diverse media settings.

Theoretical Foundations of the SPES The SPES builds on a process model of the formation of spatial presence experiences proposed by Wirth et al.

(2007; for empirical confirmations, see Hofer, Wirth, Kühne, Schramm, & Sacau, 2012), which was subsequently further enhanced by Schubert (2009) and by Wirth, Hofer, and Schramm (2012). Within the present approach, the model helps to achieve two things. First, it provides a theoretical view on the actual phenomenon – that is, spatial presence – and offers information about its dimensionality and characteristics. Second, the model explicates determinants of spatial presence. Accordingly, it embeds spatial presence within a nomological network (Cronbach & Meehl, 1955) of potentially correlated concepts. Hypotheses derived from this nomological network will allow us to test the validity of the SPES. Akin to other approaches (Herrera, Jordan, & Vera, 2006; Kim & Biocca, 1997; Schellenberg, 2007; Schubert, 2009; Slater, 2002), in the Wirth et al. (2007) model, spatial presence is understood as a user’s experience of being located within a space depicted by the media environment instead of the real environment. It is assumed that this shift in self-location also implies a shift in perceived action possibilities. Accordingly, if spatially present, users are assumed to perceive possible actions within the media environment rather than their real environment.

The Wirth et al. (2007) model conceptualizes a shifted self-location but also a shifted perception of action possibilities as dimensions of spatial presence. In this respect, the

2

Wirth et al. (2007) model resembles the embodied cognition–based approach toward spatial presence proposed by Schubert et al. (1999, 2001). According to this view, spatial presence arises if users focus on a media stimulus and develop a spatial mental model of the virtual environment and their perceived possible actions in it. Accordingly, users feel spatially present if they mentally represent actions of their own body in the virtual world (Schubert, 2009).

The crucial role of perceived action possibilities for spatial presence experiences is stressed in both approaches.

Determinants of Spatial Presence The Wirth et al. (2007) model argues that spatial presence emerges on the basis of two critical steps. First, users need to develop a mental model of the space depicted by a media offering. This spatial mental model is regarded as a necessary, albeit not sufficient, precondition of spatial presence.

Second, users may accept the spatial model as their own egocentric viewpoint. If they do, spatial presence is assumed to emerge. According to the Wirth et al. (2007) model, the process of accepting a spatial mental model is unconscious and automatic. The acceptance process is assumed to follow the mechanisms outlined in the context of perceptual hypothesis testing (Bruner & Postman, 1949). Accordingly, it is assumed that users automatically activate the most convincing – that is, consistent, error-free, evident – spatial model from existing alternatives to define their egocentric viewpoint. Spatial presence occurs if users accept the spatial model inferred from the media environment as their own egocentric viewpoint, and drop the model bound to the real environment. Accordingly, the model by Wirth et al. (2007) assumes that spatial presence increases, the more concise (consistent, error-free, evident) the spatial mental model that users develop (Hypothesis 1 [H1]).

If spatially present, users will feel as if they are physically located within the media environment and perceive their action possibilities within the mediated rather than the real environment. They will feel as if they could actually take part in the action of the media presentation, rather than merely observing it.

Across both steps, the Wirth et al. (2007) model conceptualizes various determinants of spatial presence. In line with the literature (e.g., Draper et al., 1998; Schubert et al., 2001), attention is considered the most basic determinant of spatial presence. According to the Wirth et al.

(2007) model, users’ attention to the media stimulus enhances the likelihood that they will develop a convincing spatial mental model based on the media input, and will be shielded from potentially conflicting spatial information from the real world. Accordingly, the model assumes that

The SPES evolved from the MEC Spatial-Presence-Questionnaire (MEC-SPQ; Vorderer et al., 2004). The MEC-SPQ was conceptualized to measure spatial presence and potential determinants of spatial presence. To date, the MEC-SPQ has not been officially published in a peer-reviewed journal. The SPES is simply a shortened and fine-tuned version of the two spatial presence–related scales of the MEC-SPQ that assess self-location and perceived possible actions. Due to more extensive analyses for the present approach, the items selected for the SPES and the self-location and perceived possible actions scales of the MEC-SPQ are not identical. We propose the SPES to replace both scales of the MEC-SPQ in the future. The MEC-SPQ remains a valuable tool to assess potential determinants of spatial presence.

Journal of Media Psychology 2016; Vol. 28(1):1–15

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

attention to a spatial media stimulus is positively associated with spatial presence (Hypothesis 2 [H2]). One reason users may direct their attention to a media stimulus is that they share a general interest in the topic depicted by the stimulus. According to the model, such a domain-specific interest positively affects (attention and thus) spatial presence (Hypothesis 3 [H3]).

In addition, the Wirth et al. (2007) model argues that users are more likely to develop convincing spatial mental models, the greater their skills to visually imagine spatial sceneries. Spatial imagery skills are an important part of a person’s general spatial ability (Hegarty & Waller, 2005). Spatial imagery skills ease the integration of retrieved spatial information and the filling in of incomplete spatial information. The ability to produce vivid spatial images should therefore support the development of convincing spatial models derived from the media. Accordingly, the Wirth et al. (2007) model hypothesizes that spatial imagery skills positively affect spatial presence (Hypothesis 4 [H4]).

Another determinant of spatial presence outlined in the model is cognitive involvement, which is among the typical factors of spatial presence considered by most researchers (e.g., Lessiter et al., 2001; Schubert, 2003; Witmer & Singer, 1998). Users are cognitively involved if they are preoccupied with the media stimulus and actively try to comprehend the depicted environment. When users are highly involved with media content, their mental capacity is primarily devoted to the media and not to reality.

Conversely, the majority of their information processes will be media-related and enrich their spatial mental model.

Accordingly, the Wirth et al. (2007) model assumes that cognitive involvement is a positive predictor of spatial presence (Hypothesis 5 [H5]).

Furthermore, the model assumes that users’ trait absorption positively influences (involvement and thus) spatial presence (Hypothesis 6 [H6]). Trait absorption refers to an individual’s motivation and skill in dealing with objects in an elaborate manner (Wild, Kuiken, & Schopflocher, 1995). High-absorption individuals are easily involved in things and ‘‘fascinated’’ without much effort. Trait absorption includes several abilities, of which synesthetic abilities may be most relevant for spatial presence. Synesthetic abilities could cause media stimulation of one sensory channel to trigger sensation across other sensory channels. This may strengthen the vividness of a media stimulus and users’ involvement in a media stimulus. Accordingly, higher trait absorption should result in a heightened cognitive involvement in a media offering and consequently in a more intense experience of spatial presence (Wirth, Hofer, & Schramm, 2012).

Spatial Presence in Diverse Media Settings In line with other conceptual accounts, the Wirth et al.

(2007) model argues that spatial presence can be experienced in varying degrees while using diverse media, ranging from highly immersive virtual reality systems (Steuer, 1992) to interactive audiovisual video game applications 2015 Hogrefe Publishing

5

(Tamborini & Skalski, 2006), noninteractive television (Bracken, 2005), and even books (Green, Brock, & Kaufman, 2004; Schubert & Crusius, 2002). Accordingly, the SPES was developed to measure spatial presence across diverse media settings.

Initial Item Pool Both the construction of the SPES items and the empirical test of the scale’s quality closely followed recommendations offered in the methodological literature (Bearden, Netemeyer, & Mobley, 1993; Clark & Watson, 1995; DeVellis, 2003).

Because the initial item pool exerts a strong influence on the validity of the developed instrument (Clark & Watson, 1995), the SPES builds on a pool of English-language items inspired by the conceptualization by Wirth et al. (2007).

Spatial presence was considered to be a narrow construct, with the two subdimensions (self-location and possible actions) covering only a few different facets. The goal was to develop the SPES as a short and convenient-to-apply scale consisting of just eight items: four items per subscale. The literature suggests starting with a number of items about twice as large as the anticipated length of the final scale (DeVellis, 2003), especially if the construct is narrow (Clark & Watson, 1995). Accordingly, in the present case, the initial item pool comprised 20 items. Ten items reflected users’ self-location (SL), and 10 items reflected their perceived possible actions (PA; see Table 1).

Self-location items included variants of users’ feelings of ‘‘being there’’ or ‘‘being physically present’’ in the media environment (SL1 to SL4). Self-location may also imply that users feel like they have departed from their real environment and feel like they have stepped into another place (SL5; Kim & Biocca, 1997). The experience of a shifted self-location has been linked to the feeling of being enveloped by or surrounded a media environment (e.g., Witmer & Singer, 1998). This was reflected in three additional items (SL6 to SL8: e.g., ‘‘I was convinced that things were actually happening around me’’). Items SL9 and SL10 rephrased the sensation of being enveloped by a medium, in a more specific way, by asking users to what extent they ‘‘experienced both the confined and open spaces in the presentation as though [they were] really there’’ and to what extent they were ‘‘convinced that the objects in the presentation were located on the various sides of [their] body.’’

Most possible action items dealt with users’ subjective impression that they would be able to carry out actions in the environment (PA2, PA3, and PA6) and to manipulate it (PA8 and PA9). A set of items specifically referred to the way users perceived action possibilities attached to objects presented in the environment (PA1, PA4, PA5, and PA7). For example, items captured users impression that they could actually touch objects (PA7) or use them as an utensil (PA1). Some items focused on movement and assessed the extent that users felt as if it was possible to move around in the environment (PA5 and PA10).

All items were phrased in such a way that they could be applied in a posttest after any kind of media exposure. For Journal of Media Psychology 2016; Vol. 28(1):1–15

6

T. Hartmann et al.: Spatial Presence

Table 1. The Spatial Presence Experience Scale (SPES): self-location (SL) and possible actions (PA) Sub SL

1

SL

2

SL

3

SL

4

SL

5

SL

6

SL

7

SL

8

SL

9

SL

10

PA

1

PA

2

PA

3

PA

4

PA

5

PA

6

PA

7

PA

8

PA

9

PA

10

Item

M

SD

p

F1–1

F1–2

ritc

F2–1

F2–2

SPES

I felt like I was actually there in the environment of the presentation.

It seemed as though I actually took part in the action of the presentation.

It was as though my true location had shifted into the environment in the presentation.

I felt as though I was physically present in the environment of the presentation.

I experienced the environment in the presentation as though I had stepped into a different place.

I was convinced that things were actually happening around me.

I had the feeling that I was in the middle of the action rather than merely observing.

I felt like the objects in the presentation surrounded me.

I experienced both the confined and open spaces in the presentation as though I was really there.

I was convinced that the objects in the presentation were located on the various sides of my body.

The objects in the presentation gave me the feeling that I could do things with them.

I had the impression that I could be active in the environment of the presentation.

I had the impression that I could act in the environment of the presentation.

I had the impression that I could reach for the objects in the presentation.

I felt like I could move around among the objects in the presentation.

I felt like I could jump into the action.

The objects in the presentation gave me the feeling that I could actually touch them.

It seemed to me that I could do whatever I wanted in the environment of the presentation.

It seemed to me that I could have some effect on things in the presentation, as I do in real life.

I felt that I could move freely in the environment of the presentation.

2.56

1.15

0.39

.853

.268

0.84

.845

.280

SL

2.33

1.08

0.33

.737

.389

0.78

.839

.313

SL

2.32

1.11

0.33

.848

.243

0.81

.855

.260

SL

2.09

1.07

0.27

.788

.274

0.77

.870

.234

SL

2.61

1.22

0.40

.599

.480

2.05

1.06

2.21

1.11

0.30

.543

.493

2.72

1.17

0.43

.653

.279

0.62

2.94

1.18

0.49

.698

.124

0.63

1.99

1.18

2.23

1.08

0.31

.272

.829

0.70

.243

.851

PA

2.43

1.11

0.36

.285

.848

0.69

.320

.812

PA

2.36

1.10

0.34

.270

.839

2.40

1.15

0.35

.708

.398

2.44

1.11

0.36

.368

.653

0.63

.361

.698

PA

2.40

1.12

0.35

.790

.271

2.43

1.14

0.36

.801

.221

2.16

1.11

0.29

.241

.495

0.51

.130

.655

PA

2.12

1.04

0.28

.061

.673

0.46

3.00

1.21

0.50

.413

.103

0.30

Notes. p indicates item difficulty. ritc indicates corrected item–total correlation. F1–1 indicates factor loading on Factor 1 of the first factor analysis conducted in Study 1. F2–1 indicates factor loading on Factor 1 of the second factor analysis conducted in Study 1.

Criteria formatted in bold represent the basis for the item being dropped.

Journal of Media Psychology 2016; Vol. 28(1):1–15

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

example, most items simply referred to ‘‘the environment of the presentation’’ in referring to the spatial scenery depicted by the medium. The wording was selected to be clear and unambiguous, and each item expressed only a single idea (Spector, 1992, p. 23). Lengthy items were avoided (DeVellis, 2003). All items were designed to be answered on a 5-point Likert scale ranging from 1 (= I do not agree at all) to 5 (= I fully agree). The chosen response format (i.e., degree of agreement) was preferred over alternatives such as frequency estimations or duration assessments, because posttest scales are unlikely to accurately assess the those alternative experiential aspects (Schwarz & Oyserman, 2001).

Study 1 Method Participants To choose a final set of items for the SPES and to assess the scale’s quality, we conducted four 1-factorial (distracted vs. nondistracted) between-subjects experiments involving a total of 290 participants. The sample size met the previously proposed recommendation to sample at least five times more participants than items tested (i.e., 20 items · 5 subjects/item = 100 subjects; Nunnally & Bernstein, 1994; Osborne & Costello, 2004). Sample size was also close to the general rule of thumb in scale developments to work with a sample of at least 300 subjects (Clark & Watson, 1995, p. 314). The four experiments were carried out at three different locations (Los Angeles, United States; Helsinki, Finland; Porto, Portugal) with either native English speakers or students screened for English proficiency. In Finland and Portugal, students were recruited at international schools and from English classes at local universities. The mean age of participants was 21.4 years (SD = 5.2; Min = 15 years; Max = 54 years). The majority of the participants were female (N = 212, 73.6%).

Each of the four experiments applied a specific medium: a text excerpt from a book or a film (Los Angeles), hypertext (Helsinki), or a virtual environment technology (Porto).3 This approach was taken to establish a basis for the cross-media applicability of the SPES (for a similar approach, see Lessiter et al., 2001). Of the 290 participants, 80 read the linear text, 81 watched the film, 80 read the hypertext, and 49 navigated through the virtual environment. The average ages of subjects in the hypertext sample 3

7

(M = 24.4 years, SD = 3.7) and the virtual environment sample (M = 22.9 years, SD = 9.5) were slightly higher than those in the text and film samples (text: M = 19.7, SD = 3.9; film: M = 19.5, SD = 1.4). Sex balance differed among the samples (female subjects: text, 88.5%; film, 84.0%; hypertext, 63.8%; virtual reality, 49.0%).

Experimental Manipulation Each of the four 1-factorial experiments involved a distraction-based manipulation of the subjects’ attention to the presented media stimulus. Because attention allocation was considered a crucial determinant of spatial presence, distracting participants seemed a plausible way to test the validity of the SPES (Draper et al., 1998; Schubert et al., 2001; Vorderer et al., 2004; Witmer & Singer, 1998).

In addition, a study by Brogni, Slater, and Steed (2003) employing an alternative presence measure found that awareness toward distracting real-world stimuli may disrupt or lower the experience of spatial presence. Accordingly, in line with H2, it was expected that spatial presence experiences would be stronger among nondistracted users compared with distracted users of a medium. Consequently, in all experiments, half of the participants, chosen at random, were distracted in the exposure situation to create a limit to their attention to the stimulus and, in turn, to decrease their feeling of spatial presence. In the other experimental condition, participants were not distracted.

Distraction was manipulated based on a dual-task procedure adapted from Bourke, Duncan, and Nimmo-Smith (1996). While using the presented medium, distracted participants had to perform a secondary task in the exposure situation. Audio signals were given at specific times. Participants were instructed to produce five ‘‘random numbers’’ of three digits each, as soon as the audio signal was heard.

The volume of the signal was chosen in such a way that it was clearly perceivable but did not obscure the audio of the media stimulus (if any). In the film and text study, distracted subjects were instructed to write their random numbers on a piece of paper. In the hypertext and virtual environment studies, in which distracted subjects had no free hand available, they were instructed to speak their random numbers aloud for the lab assistant to write down.

Procedure After being welcomed, subjects were placed behind tables situated about 2 m in front of the projection screen (film), or in front of a computer screen (hypertext and virtual

The text stimulus was an excerpt taken from the bestselling novel The Pillars of the Earth. The 12-page episode portrays how the protagonist, Jack, breaks into a cathedral, sets a fire, and attempts to escape from the rapidly spreading flames. While rushing through the different sections of the cathedral, the spatial environment is described in detail. As hypertext stimulus, we used ‘‘The Art of Singing’’ – a commercial multimedia production. Users navigated through rooms of an environment displayed as a series of screenshots on a desktop computer. The film stimulus showed a sequence taken from the movie The Boat – Director’s Cut. The movie tells the story of a German submarine crew during World War II. The selected episode was approximately 5 min long and portrayed the submarine’s assault on an allied convoy crossing the Atlantic Ocean. The movie received awards for outstanding recording of the atmosphere within a submarine.

In the interactive virtual environment stimulus, users navigated through the three-dimensional environment of a museum. The environment was displayed stereoscopically on a normal desktop computer. Users wore shutter-glasses to perceive the environment in three dimensions.

2015 Hogrefe Publishing

Journal of Media Psychology 2016; Vol. 28(1):1–15

8

T. Hartmann et al.: Spatial Presence

environment), or in front of printouts (book). In each of the virtual environment, hypertext, or film experiments, participants were seated at the same distance from the screen.

Participants in the distraction condition were instructed on how to deal with the audio signals during the exposure situation.

Upon receiving the instructions, participants spent a total of 10 min with the media stimulus. In the distraction condition, an audio CD containing the sequence of distracter signals was started simultaneously with the beginning of the reception of the media stimulus. Subsequent to the media exposure, all participants filled out the posttest questionnaire. Upon completing the questionnaire, participants were thanked and dismissed.

Measures In addition to the initial pool of spatial presence items, the posttest questionnaire also assessed all hypothesized determinants of spatial presence. Measures were taken from the precursor of the SPES, the MEC-SPQ (Vorderer et al., 2004). All scales ranged from 1 (= I do not agree at all) to 5 (= I totally agree): • domain-specific interest (eight items; example: ‘‘The [medium] corresponded very well with what I normally prefer’’; a = .93; M = 2.32; SD = 0.97); • spatial imagery skills (eight items; example: ‘‘When someone shows me a blueprint, I am able to imagine the space easily’’; a = .82; M = 3.5; SD = 0.71); • attention allocation (eight items; example: ‘‘My attention was caught by the [medium]’’; a = .93; M = 3.45; SD = 0.94); • conciseness of user’s spatial mental model (eight items; example: ‘‘Even now, I could still draw a plan of the spatial environment in the presentation’’; a = .9; M = 2.88; SD = 0.88); • cognitive involvement (eight items; example: ‘‘I thought about how much I know about the things in the presentation’’; a = .78; M = 2.85; SD = 0.76); and • trait absorption (nine items adapted from Tellegen & Atkinson, 1974; example: ‘‘I can be greatly moved by poetic language’’; a = .81; M = 3.55; SD = 0.75).

Selection of SPES Items Items were selected for the final SPES in three steps (e.g., Clark & Watson, 1995; DeVellis, 2003). First, the response distributions of all items were analyzed to exclude strongly skewed or difficult items, or those that showed little variance. Second, items were investigated using a varimaxrotated principal component analysis (PCA; forced 2-factor 4

solution; Kline, 1994). Third, standard reliability criteria were examined (corrected item – total correlation and Cronbach’s alpha). Table 1 provides a summary of the results obtained.

Step 1: Item Distribution and Item Difficulty Clark and Watson (1995) suggest identifying and eliminating items that have highly skewed and unbalanced distributions. One reason for this approach is that strongly unbalanced items convey little information. An analysis of the distributions obtained in the present study revealed that all items showed reasonable variance (1.03 < SD < 1.22). None of the item distributions revealed a strong ceiling effect or bottom effect (1.98 < M < 3.01).

However, none of the items showed a normal distribution (all K-S tests > 2.81; p < .01); instead, most of the items’ response distributions were skewed toward ‘‘no agreement’’ ( 0.23 < skewness < 1.06), especially items SL10 (skewness = 1.05) and SL6 (skewness = .8). Clearly, the use of the expression ‘‘I was convinced . . .’’ in these two items made it especially difficult for the subjects to agree with the statements. Both items were dropped from subsequent analyses.

Additional indices for item difficulty p were computed by dividing the item mean minus 1 by the theoretical maximum (5) minus 1, such that p had a range between 0 and 1.

It is preferable for items to yield p values between .2 and .8 (Fisseni, 1997). The values for all of the remaining 18 items fell within this recommended range (Table 1).

Step 2: Factor Structure and Factor Loadings All remaining items shared variance to a very high degree (Kaiser-Meyer-Olkin [KMO] = .95). The respondent to item ratio for the current sample was 16:1, clearly exceeding the recommendation of at least five respondents per item (Kline, 1994; MacCallum, Widaman, Zhang, & Hong, 1999).

In accordance with the deductive approach, we conducted a forced 2-factor PCA with varimax rotation.4 All remaining 18 items were entered; missing cases were deleted list-wise. The two factors obtained accounted for 60.05% of the variance (Factor 1: 36.33%; Factor 2: 24.72%). Factor loadings are listed in Table 1. Most of the self-location items loaded strongly on the first factor, whereas most of the possible actions items loaded strongly on the second factor. Accordingly, the first factor was regarded as reflecting a shift in the user’s self-location, whereas the second factor was thought to reflect the degree of perceived possible actions in a media environment. Items were retained if they loaded higher than 0.3 on their primary factor and if this primary loading was at least 0.2

Varimax was preferred over oblique rotation, although the enforced factors were expected to belong to spatial presence as a common factor. However, as suggested by Kline (1994), the observation and interpretation of factor loadings (a crucial aspect of the second step) are relatively hazardous with oblique factor rotation and relatively easy with orthogonal structures. In addition, an exploratory application of promax oblique rotation with kappa = 4 resulted in a similar solution to that obtained using varimax rotation.

Journal of Media Psychology 2016; Vol. 28(1):1–15

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

higher than any of their cross-loadings. Items SL5, SL7, PA4, PA6, and PA7 failed to meet this criterion and were subsequently dropped from consideration, leaving 13 remaining items.

Step 3: Internal Consistency PA2 and PA3 of the remaining items were almost identical in wording, with PA2 having slightly better psychometric qualities. To avoid ‘‘a scale with high internal consistency by writing the same items in different ways’’ (Bearden et al., 1993, p. 4), PA3 was dropped.

As a first test of internal consistency, we examined corrected item–total correlation ritc. The remaining six selflocation items and the remaining six possible actions items were analyzed separately. Items were required to have an acceptable ritc value of at least .5 (Fisseni, 1997). All items met this criterion (see Table 1) except for PA9 and PA10, which were dropped.

In a second step, we examined values for Cronbach’s alpha. The remaining four possible actions items (PA1, PA2, PA5, and PA8) yielded a satisfactory alpha value of .81. To obtain the most homogenous four-item subset among the remaining six self-location items, AlphaMax (Hayes, 2005) was applied, which checks alpha for all possible short forms of a k-item composite measure.

The four-item subset resulting in the highest alpha value (.92) consisted of SL1, SL2, SL3, and SL4.

Results

9

subscale compared with distracted participants, although the latter mean difference only approached significance (p = .057). The total SPES scale also responded to the distraction treatment, as expected, with nondistracted participants reporting significantly higher scores than did distracted participants. In summary, the results speak for the validity of the SPES.

Validity Test Based on Correlates of Spatial Presence According to the nomological network derived from the model by Wirth et al. (2007; see also Hofer et al., 2012), SPES was expected to be positively correlated with users’ domain-specific interest, spatial imagery skills, attention allocation, spatial mental model, cognitive involvement, and trait absorption (see H1 to H6). As shown in Table 3, all correlations obtained were highly significant and in the predicted direction. These results confirmed H1 to H6 and suggest a good validity of the SPES. As further evidence of the validity of the SPES, correlations with traits (domain-specific interest, spatial imagery skills, trait absorption) were considerably lower than those with immediate determinants of spatial presence (e.g., attention allocation, spatial mental model, cognitive involvement; Steiger’s Z for all rtrait vs. rprocess factor comparisons, p < .01). As a further indication of the good psychometric quality of the SPES, all correlations obtained between SPES scores and determinants were considerably weaker than the item-total correlations ritc of the SPES items (see Table 1, ninth column; Clark & Watson, 1995, p. 16).

Validity Test Based on Factor Analyses A varimax-rotated PCA of the finally selected eight SPES items suggested a 2-factor solution (eigenvalues 4.66, 1.15, .71, . . .). After rotation, Factor 1 accounted for 40.17% of the variance, and Factor 2 for another 32.48%.

All SPES items clearly loaded on their dedicated factors, and had only marginal cross-loadings (see Table 1, seventh and eighth columns).

Validity Test Based on Experimental Manipulation As expected, simple t tests (see Table 2) revealed that nondistracted participants attained higher scores on the SPES self-location subscale and on the SPES possible actions

Discussion Study 1 aimed to establish and validate the SPES as a short measure of spatial presence experiences. Based on the conceptual foundation laid out by Wirth et al. (2007), spatial presence was operationalized as a two-dimensional concept that includes a shift in users’ self-location and perceived action possibilities. Study 1 provided support for the reliability and validity of the SPES. The validity was confirmed in three approaches. First, factor analyses provided preliminary support of the two-dimensional structure of the SPES. Second, an experimental approach confirmed expectations and showed that total SPES scores were significantly higher in nondistracted than distracted participants.

Third, also in line with hypotheses, the SPES was

Table 2. Validation of the Spatial Presence Experience Scale (SPES) based on a distraction paradigm Distracted

SPES self-location SPES possible actions SPES total

Nondistracted

M

SD

M

SD

t(df)

2.16 2.22 2.19

0.97 0.92 0.85

2.49 2.41 2.45

0.98 0.83 0.79

2.89 (288)** 1.91 (288)+ 2.71 (288)**

Note. +p < .1, **p < .01.

2015 Hogrefe Publishing

Journal of Media Psychology 2016; Vol. 28(1):1–15

10

T. Hartmann et al.: Spatial Presence

Table 3. Validation of the Spatial Presence Experience Scale (SPES) based on zero-order correlations with potential correlates of spatial presence suggested by Wirth et al. (2007) 1 2 3 4 5 6 7 8 9

SPES self-location SPES possible actions SPES total Spatial mental model Attention Involvement Spatial imagery skills Domain–specific interest Trait absorption

1

2

3

4

5

6

7

8

9

– .60** .91** .47** .45** .38** .21** .23** .26**

– .88** .42** .41** .43** .24** .30** .23**

– .50** .48** .45** .25** .30** .27**

– .40** .49** .38** .24** .26**

– .51** .16** .26** .19**

– .32** .43** .37**

– .28** .38**

– .24**

–

Notes. Correlations with the SPES are formatted in bold. **p < .01.

significantly correlated with various determinants of spatial presence. Taken together, Study 1 resulted in a two-dimensional eight-item measure of spatial presence experiences with good psychometric properties.

screenshots of the virtual walk-through (n = 208), watched a film that showed a prerecorded walk through the museum (n = 82), or navigated through the museum in an interactive virtual environment that was stereoscopically displayed (n = 105). The sex balance was approximately equal in the sample (51.9% female), and the mean age was 23.87 years (SD = 4.72).5

Study 2 To examine the psychometric qualities of the SPES further, a second study was conducted. In Study 1, items were selected for the SPES partly based on exploratory factor analyses. An observed factor structure of a scale needs to be replicated in a sample different from the one in which it was initially obtained (e.g., Floyd & Widaman, 1995).

This was the main purpose of Study 2. More specifically, Study 2 aimed to replicate the two-dimensional structure of the SPES in the context of different media stimuli than the ones used in Study 1. A reconfirmation of the factor structure in a set of different media offerings would suggest that the SPES is a reliable and robust measure that can be applied to various media settings.

Results

Method

Replication of Factor Structure in CFA

The SPES was administered to a sample of 395 participants at four European universities: in Helsinki (Finland), Porto (Portugal), Hanover (Germany), and Zurich (Switzerland).

At each university, a researcher and a professional translator translated the SPES from English to Finnish, Portuguese, or German, as required. The items were then back-translated to English by a second professional translator, to check that the meaning of each item was retained (Cha, Kim, & Erlen, 2007).

The same procedure was applied as was employed in Study 1. All participants filled out the SPES after using a spatial media stimulus for at least 10 min. Participants were assigned to one of three different spatial media stimuli, all of which enabled the user to take a walk through a virtual Mozart museum. Subjects either used a hypertext with

After replacing missing values by multiple imputation (EM-A), the dimensionality of SPES was further examined in a confirmatory factor analysis (CFA) using the software LISREL. A test of multivariate normality confirmed that the variables were not normally distributed (v2 = 144.06; p < .001). Therefore, a robust maximum likelihood estimation was applied, and Satorra-Bentler scaled chi-square statistics were employed to evaluate model fit (Hoyle & Panter, 1995). Paths from the latent variable self-location to SL01 and from the latent variable possible actions to PA01 were set to 1. In line with the theoretical model, both latent constructs, self-location and possible actions, were allowed to covary. Fitting of the data to a congeneric model using robust maximum likelihood estimation revealed a good model fit according to Hu and Bentler (1999)

5

Replication of Factor Structure in PCA A forced 2-factor PCA with varimax rotation (KMO = .91) resulted in 2 factors that together explained 71.62% of the variance (Factor 1, 39.1%; Factor 2, 32.52%; eigenvalues 4.76, .97, . . .). All SPES items loaded strongly on their dedicated dimension (SPES SL: .87 to .77; SPES PA: .79 to .66). Cross-loadings were marginal. No item had a factor loading on the other dimension greater than .41. Results of the PCA confirm the factor structure of SPES obtained in Study 1.

In summary, Study 2 collected SPES data at four different locations, with three different media stimuli, in three different languages.

Journal of Media Psychology 2016; Vol. 28(1):1–15

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

11

Overall Discussion

Figure 1. Confirmatory factor analysis of the Spatial Presence Experience Scale (SPES) obtained in Study 2.

Standardized path coefficients are formatted in bold and italicized.

(Satorra-Bentler v2 = 27.6 (df = 18), p = .07; NFI = .99; root mean square error of approximation [RMSEA] = .037; 90% CI [.000, .063]; standardized root mean square residual [SRMR] = .03).6 All paths of the model were both substantial and significant (standardized path coefficients SPES SL: .91 to .76, p < .05; SPES PA: .70 to .64, p < .05; see Figure 1).

Internal Consistency Analyses of the two subscales using Cronbach’s alpha also confirmed a good internal consistency (aself-location = .91, apossible actions = .84). The CFA also provided support for the proposal that both self-location and possible actions are two dimensions of the joint construct spatial presence, because the two latent constructs showed a reasonable correlation (standardized path coefficient = .81, p < .05).

In fact, all items of the SPES could also be collapsed into one internally consistent (a = .89) total scale.7 In summary, Study 2 confirmed the structure of the SPES obtained in Study 1, and indicated a good validity of the measure.

6

7

Interest in the concept of spatial presence has grown rapidly in recent years (Lombard et al., in press). The present article introduces a new self-report measure (the SPES) to assess media users’ experience of spatial presence.

The present findings demonstrated that the eight-item scale reflects two dimensions of spatial presence (self-location and possible actions; Wirth et al., 2007) in a reliable and robust way. Validity tests showed that the experience of spatial presence – that is, scores on the SPES – increases with the amount of users’ attention allocation to the media stimulus, the conciseness of their spatial mental model, and their cognitive involvement in the media stimulus. In addition, in the present studies, spatial presence assessed with the SPES was positively correlated with users’ interest in the applied media stimuli, their visual imagery skills, and their absorption tendency. These observed relationships confirm the validity of the SPES. Strengths of the SPES include its derivation based on a deductive approach (Cronbach & Meehl, 1955) and its compactness. Items of the SPES are also designed to measure spatial presence in diverse media settings. The SPES is available in English, German, Portuguese, and Finnish. Depending on the interest of the researcher, the intensity of the spatial presence experience can be assessed by the total scores of the SPES, or in a more differentiated way by examining the two subdimensions of the SPES: self-location and possible actions.

Limitations and Outlook The tests performed as part of the present studies suggest a good psychometric quality for the SPES; however, additional tests may help to further illuminate the scale’s quality.

Study 1 revealed that the traits that represent distant determinants of spatial presence correlate less strongly with the SPES than with the immediate causes of the formation of spatial presence. However, the study did not include more extensive tests of discriminant validity. It would be interesting to further examine whether the SPES correlates with the assessment of constructs that are theoretically unrelated to spatial presence. Unlike broader concepts of nonmediation (e.g., transportation; Green et al., 2004), spatial presence is expected to be unrelated to, for example, users’ parasocial interaction with media characters. More extensive tests of convergent validity should also examine the degree to which the SPES correlates with alternative spatial presence scales. We assume that the SPES (and particularly the

An initial analysis suggested a slightly improved model fit if the error terms of SL03 and SL04 were allowed to covary, which they marginally did. We opted for letting the two error terms correlate, as we interpreted the correlation as resulting from the similar wording of the two items.

Are possible actions and self-location indeed two dimensions of spatial presence, or is one a determinant of the other? We argue that both are subdimensions of spatial presence that usually co-occur. Possible actions are strongly correlated (r = .6) with self-location, whereas zero-order correlations to theoretically proposed determinants are considerably weaker (all r < .47). Maybe more importantly, a joint varimax rotated factor analysis with all items entered both from the two subdimensions and from all suggested determinants yielded an 8-factor solution in which all SPES items jointly loaded on a single factor and had only marginal cross-loadings with other factors. Both results supported our view that both self-location and possible actions are dimensions of a joint underlying construct rather than a cause or consequence of each other.

2015 Hogrefe Publishing

Journal of Media Psychology 2016; Vol. 28(1):1–15

12

T. Hartmann et al.: Spatial Presence

self-location dimension) should positively correlate with the spatial presence subcomponent of the IPQ and the spatial presence factor of the ITC-SOPI. Given that we do not consider perceived realism an inherent dimension of spatial presence, we would expect considerably lower correlations of the SPES with subscales measuring perceived realism – for example, of the IPQ (e.g., in studies applying otherwise identical spatial scenarios in either a fictional/unrealistic or nonfictional/realistic setting). Finally, the use of retrospective assessment involves the possibility of having to deal with biased estimates or the memories of respondents (Schwarz & Oyserman, 2001). Accordingly, it would be valuable to further examine the convergent validity of the SPES with objective correlates of spatial presence experiences (e.g., Baumgartner, Valco, Esslen, & Jäncke, 2006; Freeman, Avons, Meddis, Pearson, & IJsselsteijn, 2000; Meehan, Insko, Whitton, & Brooks, 2001).

Participants were sampled in the present studies from different locations and were using different types of media.

On the one hand, the fact that our results confirm the expected dimensionality and a good psychometric quality for the SPES even in such heterogeneous samples may be considered evidence for the robustness of the scales.

On the other hand, additional exploratory tests of the latent factor structure across the subsamples collected in Study 2 failed to confirm measurement invariance. This finding is difficult to assess, as subsamples were small and differed across locations and applied media stimuli. However, measurement invariance of the SPES across media types and different samples deserves further scrutiny in future research.

Finally, in the present studies, SPES items were not normally distributed but skewed toward lower scores. This skewness may be a consequence of the media stimuli that we applied in the present approach, such as books or noninteractive films. Whereas we believe that even readers of books may feel spatially present, it seems more likely that the sensation is experienced by users of highly immersive virtual reality technology (Witmer & Singer, 1998).

However, in the present studies, the SPES has been only partially investigated in the context of such immersive systems.

The list of potential uses of the SPES is as long as the list of communication contexts in which spatial presence is relevant. The SPES offers another choice for researchers interested in assessing spatial presence. With the SPES, researchers can apply a theoretically plausible, validated, psychometrically sound, flexible, and particularly short self-report measure of spatial presence in addressing the challenges of current and future media technologies.

References Baumgartner, T., Valco, L., Esslen, M., & Jäncke, L. (2006).

Neural correlate of Spatial Presence in an arousing and noninteractive virtual reality: An EEG and psychophysiology study. Cyberpsychology & Behavior, 9, 30–45.

Journal of Media Psychology 2016; Vol. 28(1):1–15

Bearden, W. O., Netemeyer, R. G., & Mobley, M. F. (1993).

Handbook of marketing scales. Multi-item measures for marketing and consumer behavior research. Newbury Park, CA: Sage.

Biocca, F. (1997). The cyborg’s dilemma: Progressive embodiment in virtual environments. Journal of ComputerMediated Communication, 3(2), 1–18. Retrieved from http://jcmc.indiana.edu/vol3/issue2/biocca2.html Blascovich, J., Loomis, J., Beall, A., Swinth, K., Hoyt, C., & Bailenson, J. (2002). Immersive virtual environment technology as a methodological tool for social psychology.

Psychological Inquiry, 13, 103–124.

Bormann, K. (2005). Presence and the utility of audio spatialization. Presence: Teleoperators and Virtual Environments, 14, 278–297.

Bourke, P. A., Duncan, J., & Nimmo-Smith, I. (1996). A general factor in dual-task performance decrement. The Quarterly Journal of Experimental Psychology, 49, 525–545.

Bracken, C. C. (2005). Presence and image quality: The case of high-definition television. Media Psychology, 7, 191–205.

Brogni, A., Slater, M., & Steed, A. (2003). More breaks less presence. Proceedings of the 6th Annual International Workshop on Presence. Retrieved from http://www0.cs. ucl.ac.uk/staff/M.Slater/Papers/bipspres.pdf Bruner, J. S., & Postman, L. (1949). On the perception of incongruity: A paradigm. Journal of Personality, 18, 206–223.

Cha, E.-S., Kim, K. H., & Erlen, J. (2007). Translation of scales in cross-cultural research: issues and techniques. Journal of Advanced Nursing, 58, 386–395.

Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309–319.

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281–302.

DeVellis, R. F. (2003). Scale development: Theory and applications (2nd ed.). Thousand Oaks, CA: Sage.

Draper, J. V., Kaber, D. B., & Usher, J. M. (1998). Telepresence.

Human Factors, 40, 354–375.

Eastin, M. S., & Griffiths, R. P. (2006). Beyond the shooter game. Examining presence and hostile outcomes among male game players. Communication Research, 33, 448–466.

Fisseni, H. J. (1997). Lehrbuch der psychologischen Diagnostik [Handbook of psychological assessment] (2nd ed.).

Göttingen, Germany: Hogrefe.

Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7, 286–299.

Freeman, J., Avons, S., Meddis, R., Pearson, D., & IJsselsteijn, W. (2000). Using behavioural realism to estimate presence: A study of the utility of postural responses to motion stimuli. Presence: Teleoperators and Virtual Environments, 9, 149–164.

Green, M. C., Brock, T. C., & Kaufman, G. F. (2004). Understanding media enjoyment: The role of transportation into narrative worlds. Communication Theory, 14, 311–327.

Hayes, A. (2005). A computational tool for survey shortening applicable to composite attitude, opinion, and personality measurement scales. Paper presented at the meeting of the Midwestern Association for Public Opinion Research, Chicago, IL. Hegarty, M., & Waller, D. (2005). Individual differences in spatial abilities. In P. Shah & A. Miyake (Eds.), Handbook of visuospatial thinking (pp. 121–169). New York: Cambridge University Press.

Hendrix, C., & Barfield, W. (1996). Presence within virtual environments as a function of visual display parameters.

Presence: Teleoperators and Virtual Environments, 5, 274–289.

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

Herrera, G., Jordan, R., & Vera, L. (2006). Agency and presence: A common dependence on subjectivity? Presence: Teleoperators and Virtual Environments, 15, 539–552.

Hofer, M., Wirth, W., Kühne, R., Schramm, H., & Sacau, A. (2012). Structural equation modeling spatial presence: The influence of cognitive processes and traits on spatial presence. Media Psychology, 15, 373–395.

Hoyle, R. H., & Panter, A. T. (1995). Writing about structural equation models. In R. H. Hoyle (Ed.), Structural equation modeling (pp. 158–176). Thousand Oaks, CA: Sage.

Hu, L.-T., & Bentler, P. M. (1999). Cut-off criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.

IJsselsteijn, W. A., de Ridder, H., Freeman, J., & Avons, S. E. (2000). Presence: Concept, determinants and measurement.

Proceedings of the SPIE, 3959, 520–529.

Insko, B. E. (2003). Measuring presence: Subjective, behavioral and physiological methods. In G. Riva, F. Davide, & W. A. IJsselsteijn (Eds.), Being there: Concepts, effects and measurement of user presence in synthetic environments (pp. 109–119). Amsterdam, The Netherlands: Ios Press.

International Society for Presence Research (ISPR). (2001).

What is Presence? Retrieved from http://ispr.info/aboutpresence-2/about-presence/ Jin, S.-A. (2010). Effects of 3D virtual haptics force feedback on brand personality perception: The mediating role of physical presence in advergames. Cyberpsychology, Behavior, and Social Networking, 13, 307–311.

Johns, C. L., Nunez, D., Daya, M., Sellars, D., Casaneuva, J. S., & Blake, E. H. (2000). The interaction between individuals’ immersive tendencies and the sensation of presence in a virtual environment. In J. D. Mulder & R. van Liere (Eds.), Virtual environments 2000. Vienna, Austria: Springer.

Kennedy, R. S., Lane, N. E., Berbaum, K. S., & Lilienthal, M. G. (1993). Simulator sickness questionnaire: An enhanced method for quantifying simulator sickness. International Journal of Aviation Psychology, 3, 203–220.

Kim, T., & Biocca, F. (1997). Telepresence via television: Two dimensions of telepresence may have different connections to memory and persuasion. Journal of Computer-Mediated Communication, 3(2). Retrieved from http://jcmc.indiana. edu/vol3/issue2/kim.html Kline, P. (1994). An easy guide to factor analysis. New York: Routledge.

Laarni, J., Ravaja, N., Saari, T., Böcking, S., Hartmann, T., & Schramm, H. (in press). Ways to measure presence: Review and future directions. In M. Lombard, F. Biocca, J. Freeman, W. IJsselsteijn, & R. J. Schaevitz (Eds.), Immersed in media: Telepresence theory, measurement, & technology. New York: Springer.

Lee, K. M. (2004). Presence, Explicated. Communication Theory, 14, 27–50.

Lessiter, J., Freeman, J., Keogh, E., & Davidoff, J. (2001). A cross-media presence questionnaire: The ITC-Sense of Presence Inventory. Presence: Teleoperators and Virtual Environments, 10, 282–297.

Lombard, M., Biocca, F., Freeman, J., IJsselsteijn, W., & Schaevitz, R. J. (in press). Immersed in Media. Telepresence Theory, Measurement, & Technology. New York: Springer.

Lombard, M., & Ditton, T. (1997). At the heart of it all: The concept of presence. Journal of Computer-Mediated Communication, 3(2). Retrieved from www.ascusc.org/jcmc/ vol3/issue2/lombard.html Lombard, M., Ditton, T. B., & Weinstein, L. (2009). Measuring (tele)presence: The Temple Presence Inventory. Presentation given at the 12th International Workshop on Presence, Los Angeles, CA, USA, November.

2015 Hogrefe Publishing

13

MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological Methods, 4, 87–99.

Meehan, M., Insko, B., Whitton, M., & Brooks, F. P. (2001).

Physiological measures of presence in stressful virtual environments. Proceedings of the 4th Annual International Workshop on Presence. Retrieved from http://ispr.info/ presence-conferences/previous-conferences Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill.

Nystad, E., & Sebok, A. (2004). A comparison of two presence measures based on experimental results. In M. A. Raya & B. R. Solaz (Eds.), Proceedings of the 7th International Workshop on Presence (pp. 266–273). Valencia, Spain: Technical University of Valencia.

Osborne, J. W., & Costello, A. B. (2004). Sample size and subject to item ratio in principal components analysis.

Practical Assessment, Research & Evaluation, 9(11).

Retrieved from http://PAREonline.net/getvn.asp?v=9&n=11 Regenbrecht, H., & Schubert, T. (2002). Real and illusory interactions enhance presence in virtual environments.

Presence: Teleoperators and Virtual Environments, 11, 425–434.

Riva, G., Davide, F., & IJsselsteijn, W. A. (2003). Being there: The experience of presence in mediated environments. In G. Riva, F. Davide, & W. IJsselsteijn (Eds.), Being There: Concepts, effects and measurements of user presence in synthetic environments (pp. 4–14). Amsterdam, The Netherlands: Ios Press.

Robillard, G., Bouchard, S., Fournier, T., & Renaud, P. (2003).

Anxiety and Presence during VR immersion: A comparative study of the reactions of phobic and non-phobic participants in therapeutic virtual environments derived from computer games. CyberPsychology & Behavior, 6, 467–476.

Schellenberg, S. (2007). Action and self-location in perception.

Mind, 115(463), 603–632.

Schubert, T. (2009). A new conception of Spatial Presence: Once again, with feeling. Communication Theory, 19, 161–187.

Schubert, T. W. (2003). The sense of presence in virtual environments: A three-component scale measuring spatial presence, involvement, and realness. Journal of Media Psychology, 15, 69–71.

Schubert, T. W., & Crusius, J. (2002). Five theses on the book problem. Presence in books, film, and VR. In F. R. Gouveia & F. Biocca (Eds.), Proceedings of the 5th International Workshop on Presence (pp. 53–59). Porto, Portugal: Universidad Fernando Pessoa.

Schubert, T. W., Friedmann, F., & Regenbrecht, H. (1999).

Embodied presence in virtual environments. In R. Paton & I. Neilson (Eds.), Visual representations and interpretations (pp. 269–278). London: Springer.

Schubert, T. W., Friedmann, F., & Regenbrecht, H. (2001). The experience of presence: Factor analytic insights. Presence: Teleoperators and Virtual Environments, 10, 266–281.

Schuemie, M. J., van der Straaten, P., Krijn, M., & van der Mast, C. A. P. G. (2001). Research on presence in VR: A survey.

Cyberpsychology & Behavior, 4, 183–202.

Schwarz, N., & Oyserman, D. (2001). Asking questions about behavior: Cognition, communication, and questionnaire construction. American Journal of Evaluation, 22, 127–160.

Sheridan, T. B. (1992). Musings on telepresence and virtual presence. Presence: Teleoperators and Virtual Environments, 1, 120–126.

Slater, M. (2002). Presence and the sixth sense. Presence: Teleoperators and Virtual Environments, 11, 435–439.

Slater, M., & Steed, A. (2000). Avirtual presence counter. Presence: Teleoperators and Virtual Environments, 9, 413–434.

Journal of Media Psychology 2016; Vol. 28(1):1–15

14

T. Hartmann et al.: Spatial Presence

Slater, M., Usoh, M., & Steed, A. (1994). Depth of presence in virtual environments. Presence: Teleoperators and Virtual Environments, 3, 130–144.

Slater, M., Usoh, M., & Steed, A. (1995). Taking steps: The influence of a walking metaphor on presence in virtual reality. ACM Transactions on Computer Human Interaction, 2, 201–219.

Slater, M., & Wilbur, S. (1997). A framework for immersive virtual environments (FIVE): Speculations on the role of presence in virtual environments. Presence: Teleoperators and Virtual Environments, 6, 603–616.

Spector, P. E. (1992). Summated rating scale construction.

Newbury Park, CA: Sage.

Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal of Communication, 42(4), 73–93.

Tamborini, R., & Skalski, P. (2006). The role of presence in the experience of electronic games. In P. Vorderer & J. Bryant (Eds.), Playing video games: Motives, responses, and consequences (pp. 225–240). Mahwah, NJ: Erlbaum.

Tellegen, A., & Atkinson, G. (1974). Openness to absorbing and self-altering experiences (‘‘absorption’’), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology, 83, 268–277.

Tichon, J. G. (2007). Using Presence to improve a virtual training environment. Cyberpsychology & Behavior, 10, 781–788.

Usoh, M., Catena, E., Arman, S., & Slater, M. (2000). Using presence questionnaires in reality. Presence: Teleoperators and Virtual Environments, 9, 497–503.

Vorderer, P., Wirth, W., Saari, T., Gouveia, F. R., Biocca, F., Jäncke, L., . . . Jänke, P. (2004). Development of the MEC spatial presence questionnaire (MEC-SPQ) Unpublished project report. Presence: MEC (IST-2001-37661).

Waller, D., & Bachmann, E. R. (2006). The borderline of science: On the value of factor analysis for understanding presence. Presence: Teleoperators and Virtual Environments, 15, 235–244.

Wild, T. C., Kuiken, D., & Schopflocher, D. (1995). The role of absorption in experiential involvement. Journal of Personality and Social Psychology, 69, 569–579.

Wirth, W., Hartmann, T., Boecking, S., Vorderer, P., Klimmt, P., Schramm, H., . . . Jänke, P. (2007). A process model of the formation of spatial presence experiences. Media Psychology, 9, 493–525.

Wirth, W., Hofer, M., & Schramm, H. (2012). The role of emotional involvement and absorption in the formation of spatial presence. Media Psychology, 15, 19–43.

Witmer, B. G., Jerome, C. J., & Singer, M. J. (2005). The factor structure of the presence questionnaire. Presence: Teleoperators and Virtual Environments, 14, 298–312.

Witmer, B. G., & Singer, M. J. (1998). Measuring presence in virtual environments: A presence questionnaire. Presence: Teleoperators and Virtual Environments, 7, 225–240.

Yee, N., & Bailenson, J. (2007). The Proteus effect: The effect of transformed self-representation on behavior. Human Communication Research, 33, 271–290.

Youngblut, C., & Perrin, B. M. (2002). Investigating the relationship between presence and performance in virtual environments. Paper presented at the IMAGE 2002 conference, Scottsdale, Arizona, USA.

Tilo Hartmann (PhD) is associate professor in the Department of Communication Science at the VU University Amsterdam, The Netherlands. He acts as an editorial board member of the periodicals Journal of Communication, Human Communication Research, and Media Psychology.

He specializes in media-psychological research examining media users’ illusionary experiences, as well as media effects and media choice.

Tilo Hartmann Department of Communication Science VU University of Amsterdam 1081 HV Amsterdam The Netherlands Tel. +31 20 598-6899 E-mail t.harmann@vu.nl André Gysbers studied media management at Hanover University of Music, Drama and Media, from 1997 to 2002.

He received his PhD in communication in 2007, and is now working as marketing director at a leading supplier of the graphic and label industries.

Feliz Gouveia (PhD) is associate professor in the Faculty of Science and Technology of University Fernando Pessoa, Portugal.

His research interests include complex information and knowledge representation, and design of supportive systems for interaction.

Christoph Klimmt studied media management at the Department of Journalism and Communication Research (IJK) of Hannover University, 1996–2000. From 2007 to 2010, he served as assistant professor in the Department of Communication, University of Mainz. Since autumn 2010, he has been professor of communication science at IJK Hannover. His research interests include media effects and processes, entertainment, and digital games.

Date of acceptance: July 14, 2014 Published online: January 30, 2015

Journal of Media Psychology 2016; Vol. 28(1):1–15

2015 Hogrefe Publishing

T. Hartmann et al.: Spatial Presence

15

Saskia Boecking received her PhD in communication in 2007. Throughout recent years she has held various positions as market research manager in international market research companies.

Currently she is working as market insight manager in a Swiss energy enterprise.

Ana Maria Sacau is associate professor in the Faculty of Human and Social Sciences at the University Fernando Pessoa (Portugal). She received her PhD in psychology in 1998 at the University of Santiago de Compostela (Spain). Currently her research areas include legal psychology, children and law, and criminal behavior.

Jari Laarni (PhD) is principal scientist at the Systems Research Center of the VTT Research Centre of Finland. He specializes in the areas of visual perception, cognitive psychology, usability evaluation, and user experience analysis, and he has conducted research on the issues involved in visual attention and search, user interface evaluation, sense of presence in media environments, and operator work analysis.

Holger Schramm (PhD) is professor of communication at the University of WĂźrzburg (Germany) in the Department Human-Computer-Media. His research focuses on media processes and effects (parasocial interactions, mood and emotion, entertainment, presence, flow), music and media, sports communication, and advertising/brand communication.

Niklas Ravaja (PhD) is professor of the social psychology of information and communication technology at the University of Helsinki, and chair of the Social Interaction and Emotion group of the Helsinki Institute for Information Technology (HIIT). His research interests include emotion- and attention-related psychophysiological responses during media processing.

Peter Vorderer (PhD, TU Berlin, Germany) is professor of media and communication studies at the University of Mannheim, Germany. His previous affiliations include the University of Music and Theatre in Hannover, Germany, the Annenberg School for Communication at the University of Southern California, United States, and the Free University of Amsterdam, The Netherlands. He specializes in media use and media effects research, with a focus on media entertainment and digital games.

Timo Saari is professor of human-centric design in the Department of Pervasive Computing in Tampere University of Technology, Finland. During 2007â&#x20AC;&#x201C;2011, he was associate professor in the Department of Broadcasting, Telecommunications and Mass Media in Temple University, USAUnited States. He specializes in the user experience (emotion, cognition, well-being) of pervasive and mobile technologies as well as customized media and games.

2015 Hogrefe Publishing

Werner Wirth (MA, PhD) is full professor at the University of Zurich, Institute of Mass Communication and Media Research, and head of the section Media Psychology and Effects. His main research areas comprise media audiences and effects, entertainment and emotion research, political and commercial persuasion, online and mobile communication, and empirical methods.

Journal of Media Psychology 2016; Vol. 28(1):1â&#x20AC;&#x201C;15

Original Article

Social Categorization, Moral Disengagement, and Credibility of Ideological Group Websites Shane Connelly,1 Norah E. Dunbar,2 Matthew L. Jensen,1 Jennifer Griffith,3 William D. Taylor,1 Genevieve Johnson,1 Michael Hughes,4 and Michael D. Mumford1 1

Department of Psychology, University of Oklahoma, Norman, OK, USA, 2University of California, Santa Barbara, CA, USA, 3Alfred University, Alfred, NY, USA, 4Human Resources Research Organization, Alexandria, VA, USA

Abstract. The online presence of ideological groups has enabled the dissemination of group beliefs and ideas through a variety of new media outlets. Websites have offered a way for these groups to share aspects of their ideology and to create a sense of shared identity. While ideological groups have been of interest for decades, little empirical research has examined their online presence. The aims of this study were to compare nonviolent and violent ideological group websites with each other and with nonideological websites with respect to social categorization, moral disengagement, and website credibility, and to examine the relationships of psychological processes to website credibility. A content analysis approach was used to rate 105 websites (violent = 32, nonviolent = 36; nonideological = 37) for aspects of social categorization, outgrouping, moral disengagement, content features of credibility, and structural features of credibility. Violent ideological group websites manifest a greater degree of social categorization, outgrouping, and moral disengagement than nonviolent ideological and nonideological websites. Regression analysis shows that these three variables negatively predict content and structural website credibility for nonviolent ideological groups but do not significantly predict website credibility for violent groups or nonideological groups. Potential limitations include small sample, use of raters (vs. normative website visitors) to evaluate websites, inclusion of only English-language websites, limited number of psychological processes examined and the potential need for more specific website categories. Social identity processes on websites vary for different types of groups and impact perceived credibility of such groups in online environments.

Keywords: ideological groups, psychological processes, ideological websites, website credibility, social identity

The increasing Internet presence of ideological groups has generated interest in understanding more about what these groups do and how they exert influence in online environments. Broadly speaking, ideological groups have clear, persistent, strongly held values and beliefs that serve as a mental model or guiding framework for understanding and interpreting events, information, and the world in general (Mumford et al., 2008; Van Dijk, 2006). Ideological groups have a presence on social media websites such as Facebook and Twitter and develop their own websites designed to serve a variety of information dissemination, advocacy, and community-building goals (Dobratz & Shanks-Meile, 1997; Matusitz & O’Hair, 2008; Schafer, 2002; Stanton, 2002). Groups that create a website have the ability to present and promote beliefs, ideas, and opinions to anyone with access to the Internet – especially those who already share the ideology, others who may be exploring alternative ideologies and social identities, and still others who might oppose the ideology. The potential for promulgating ideological views is virtually unlimited.

Journal of Media Psychology 2016; Vol. 28(1):16–31 DOI: 10.1027/1864-1105/a000138

A number of nonviolent civic, religious, and political ideological groups such as the Sierra Club, the United Methodist Church, and the Jewish Voice for Peace have websites that offer a range of information about their ideologies, mission, history, and current events, and provide opportunities for dialogue with, and involvement in, the group. These groups tend to have prosocial ideologies as expressed in group mission statements like ‘‘promoting the responsible use of the earth’s ecosystems and resources’’ (http://www.sierraclub.org) or ‘‘working together for peace, social justice, and human rights . . . to support the aspirations for Israelis and Palestinians for security and self-determination’’ (http://www.jewishvoiceforpeace.org).

Unfortunately, there are analogous ideological groups such as the Animal Liberation Front, the Army of God, and the English Defence League that sanction violence and hate in service of their ideologies. These group ideologies and associated missions reflect intolerance and contempt.

It is critically important to understand the similarities and differences in the online presence and influence of Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

nonviolent ideological groups and violent groups that propagate violence, target youths, and provide information and resources that perpetuate hate (Schafer, 2002). First, online access has increased exponentially with the global reach of the Internet and with the high potential for mass replication and proliferation of novel, interesting, or extreme views through varied social media outlets. This may be especially true for younger generations of digital natives, who understand, seek out, and routinely use digital technology.

Second, ideological groups seek to influence or even convert neutral observers and do so through a number of means (Frischlich & Rieger, 2013). For example, they create strong online communities that are not immediately recognizable as extremist. Additionally, they may attract youths online, using this forum to promote offline activities such as concerts, volunteer opportunities, or other leisure activities to further socialize and indoctrinate them toward the ideology. At least one study has shown that propaganda from violent extremist groups has been rated as interesting and persuasive despite the fact that this material was also rated as unpleasant, one-sided, and aversive (Frischlich & Rieger, 2013), suggesting that these groups have surmounted an important first persuasive hurdle of getting people to pay attention. Youths whose identities are still forming and who are seeking to ‘‘belong’’ somewhere may be especially vulnerable to, and drawn in by, these kinds of online influences. Their limited life experiences and lower levels of skepticism may increase the likelihood that youths will accept ideological groups as legitimate and credible (Flanagin & Metzger, 2008).

Websites sponsored by ideological groups are of particular interest due to the degree of control groups have over the content and structure of their websites. Prior research has shown that the websites of violent ideological groups are more difficult to access, have more stringent rules and regulations on discussion boards, and offer less user control over online settings and content (Griffith et al., 2013; Jensen et al., 2013). Additionally, these websites may be subject to less monitoring and scrutiny than social media websites, such as Facebook, which have terms of service policies that forbid overt threats and hate speech (Heath & O’Hair, 2009; Facebook terms of service policy), enabling violent groups to advocate their views with less threat of being detected, blocked, or shut down. Websites also enable groups to attract individuals who would otherwise be unable or unwilling to attend functions or meetings but who still express an interest in the group’s ideology and desire to participate in the group (Lee & Leets, 2002).

Some important differences have been identified between violent and nonviolent group websites (Griffith et al., 2013) and discussion forums (Angie et al., 2011), but additional research is needed. Griffith et al. (2013) compared violent and nonviolent group websites from a communication and media-use perspective, examining information variety, types of media used, member characteristics (number of members, average member tenure, amount of active member participation), member control, and website functionality. Violent group websites contained more progroup information, displayed more emotionally evocative images, and had a more tightly controlled online Ó 2015 Hogrefe Publishing

17

presence than nonviolent groups. They also used multiple media types such as videos, flyers, newsletters, and music to share information. Online activities of ideological groups have also been studied with respect to psychological processes. In a study of discussion forum messages on a small number of violent (n = 9) and nonviolent ideological group websites (n = 10), Angie et al. (2011) showed that violent group discussions exhibited social identity formation (i.e., encouraging integration with the group through participation), moral superiority (i.e., conveying a sense of righteousness and superiority of the ideology), self-expression (i.e., discussing/displaying group symbols), dehumanization (i.e., discussing outgroup members in derogatory ways), ethnic outgrouping (i.e., demonizing ethnic minorities as the cause of current injustices or life conditions), and moral disengagement, to a greater extent than nonviolent group discussions. It is unclear whether these patterns would be similar in a larger set of violent group websites. Additionally, we do not know if these same psychological processes are present beyond the discussion boards and appear on the general areas of these group websites, or whether these processes influence perceptions of website credibility. Examination of group websites is important because this reflects the public face of the group. Furthermore, not all visitors to the website will view or participate in discussions on message boards. Accordingly, the present study had two main purposes. The first was to compare and contrast violent ideological, nonviolent ideological, and nonideological group websites with regard to the presence of several psychological processes, including social categorization, outgrouping, and moral disengagement mechanisms.

The second purpose was to evaluate how these psychological processes influence perceptions about the credibility of website content and structure for these groups.

Ideological Groups and Social Identity People are attracted to ideological groups for a variety of reasons. These groups provide sets of defined beliefs, structure and meaning in life, and predictable social interactions that reduce uncertainty and threat (Hogg, 2007; Hogg, Meehan, & Farquharson, 2010; Jost & Hunyady, 2005; Mumford et al., 2008). Social identity develops through group membership and emotional attachment to a group, with the knowledge that the group is distinct in positive ways from other groups (Hogg, 2003; Tajfel, 1982; Tajfel & Turner, 1979). Uncertainty-identity theory (Hogg, 2007) suggests that self-concept uncertainty motivates people to seek out social identities through group identification. Groups that have a clear sense of purpose, well-defined goals, and internal homogeneity provide unambiguous prescriptions for how members should think, feel, and behave. Individuals with uncertain selfconceptions are attracted to these types of unified and highly cohesive groups, including ideological groups, because they reduce uncertainty. In establishing and maintaining their social identities, ideological groups may encourage social categorization by highlighting the Journal of Media Psychology 2016; Vol. 28(1):16–31

18

S. Connelly et al.: Psychological Processes and Online Ideological Groups

uniqueness of their beliefs and values relative to other groups (Ellemers, Spears, & Doosje, 2002) or by proclaiming superiority over groups with competing views (Angie et al., 2011; Parker & Janoff-Bulman, 2013). The anonymity of the virtual community may perpetuate identification with the ideological group by giving individuals the opportunity to routinely socialize with other like-minded individuals and act on extremist views. Virtual communities have the ability to provide social identity, structure, and support for extremist views.

Violent and nonviolent ideological groups are likely to differ in their use of social categorization. Violent ideological groups not only define themselves through a specific ideology or worldview, but are also willing to take more extreme actions to promote, defend, and protect that worldview from outside threats. Hogg and colleagues (2010) suggest that this may be the appeal of extremist groups to individuals high in identity uncertainty or who feel their own self-concept is threatened. The willingness of these groups to aggressively defend and promote their agendas, sometimes through violent actions, provides a sense of security and can promote greater identification with and loyalty to these groups (Hogg, 2005, 2007; Hogg et al., 2010). Other research has suggested that different dimensions of social identity may differentiate violent and nonviolent groups (Ashmore, Deaux, & McLaughlin-Volpe, 2004; Roccas, Klar, & Liviatan, 2006). Subjective identification with, and attachment to, a group’s ideology is one aspect of social identity characterized by internal focus on the group of interest, something that fosters critical evaluation of the group’s actions (Leidner, Castano, Zaiser, & Giner-Sorolla, 2010), which may explain why many ideological groups never become violent. Alternatively, social identity can be rooted in glorifying the group, emphasizing the superiority of the group relative to other groups, and showing unquestioning deference and loyalty to the group’s ideology, leadership, rules, symbols, and actions (Roccas et al., 2006). This aspect of social identification is not conducive to critical internal evaluation, enabling a group to justify hate and other acts of violence toward others who do not share their views. In light of this, we proposed the following hypothesis: Hypothesis 1 (H1): Violent ideological group websites will display more extreme social categorization in the form of group differentiation, group superiority, and disagreement with dissenting views than nonviolent ideological and nonideological group websites.

Ideological Groups and Outgrouping Ingroup glorification is associated with other psychological processes also likely to differentiate violent and nonviolent ideological group websites, such as outgrouping. A number of theories identify potential threats that motivate outgrouping (Riek, Mania, & Gaertner, 2006). For example, realistic group conflict theory (RGCT) suggests that competition for Journal of Media Psychology 2016; Vol. 28(1):16–31

scarce resources results in bias toward outgroups (Bobo, 1983; Sherif & Sherif, 1969). Symbolic threats to group values and beliefs have also been shown to increase intergroup bias, negative stereotypes, and racism (Kinder & Sears, 1981; Stephan & Stephan, 1996, 2000). Threats to a group’s perceived value or to its distinctiveness as a group can also lead to denigration of the source of the threat, which is often an outside group that holds different or opposing views (Tajfel & Turner, 1979). This tendency toward positive evaluation of one’s own group and negative evaluation of outgroups, known as intergroup bias, has been linked to the self-enhancement and the self-esteem of individual group members (Aberson, Healy, & Romero, 2000; Hewstone, Rubin, & Willis, 2002). Ideological groups may exhibit intergroup bias more than nonideological groups because the beliefs and values at the very core of these groups’ existence are susceptible to symbolic and perceived group value threats, especially when the ideology falls outside of mainstream or normative societal views.

One form of outgrouping often seen in violent ideological groups is discrimination and hate toward ethnic minorities (Angie et al., 2011; Hogg, 2003; Moghaddam, 2005; Pittinsky, Shih, & Trahan, 2006). The ideologies of some hate groups are framed around ‘‘protecting’’ the ingroup from the threat of destruction by the enemy rival ethnic group. Violence in these groups is often viewed as a justifiable means to various ends such groups have in mind (Taylor & Moghaddam, 1994). These extreme forms of outgrouping are the consequence of decreased self-sanctioning or self-reprimanding associated with moral disengagement (Bandura, 1999).

Hypothesis 2 (H2): Violent ideological groups will display outgrouping, in the form of seeing outgroups as enemies, negatively comparing the outgroup to others, and distorting information about the outgroup, to a greater extent than nonviolent ideological and nonideological group websites.

Ideological Groups and Moral Disengagement Moral agency has both proactive and inhibitive components, in that individuals and groups can initiate actions that are right, just, and humane, or refrain from actions that are wrong, unjust, and inhumane (Bandura, 1999). Social cognitive theory (SCT) suggests that inhibitive or selfregulatory aspects of moral agency need to be activated in order for people to recognize and do what is right (Bandura, 1989). However, there are a number of ways people can bypass self-regulatory processes, resulting in selective disengagement from reflection on unethical behavior and self-sanctioning both before and after the behavior has occurred. First, reprehensible behavior and the consequences of such behavior can be cognitively restructured or framed in ways that make it seem acceptable or even morally correct. Cognitive restructuring mechanisms include moral justification (e.g., killing an enemy Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

that threatens one’s way of life or survival), euphemistic language (e.g., eliminate a target vs. kill a person), advantageous comparison of the reprehensible behavior with other more destructive or widespread atrocities, and disregard for or distortion of the consequences (e.g., a group uses hateful language without acknowledging the violence and discrimination that results from this). Second, moral agency can be weakened by lessening personal responsibility for one’s actions through displacing responsibility onto leaders or authorities (e.g., ‘‘I was just following orders’’), or diffusing responsibility through division of labor or group decision making such that no one person feels accountable.

Third, it is easier to commit violent or unjust acts against others who are made to seem less than human or when attributions of blame are assigned to these victims. A quote cited by Akins (2006) from a Ku Klux Klan website exemplifies many of these moral disengagement mechanisms: Enemies from within are destroying the United States of America. An unholy coalition of anti-White, antiChristian liberals, socialists, feminists, homosexuals, and militant minorities have managed to seize control of our government and mass media.. . . We shall liberate our nation from these savage criminals and restore law and order to America. (Akins, 2006, p. 129) Extremist ideological groups may manifest greater moral disengagement than more mainstream ideological groups and nonideological groups on their websites. First, the extreme beliefs and views held by these groups make it difficult for them to work through normal political and social channels to accomplish goals (Akins, 2006). Consequently, these groups may turn to coercive or violent actions to perpetuate their ideologies. For example, the Animal Liberation Front is an international underground group that engages in unlawful destruction of facilities and removal of animals from laboratories and other locations in support of their animal rights and liberation ideology. Website content can be framed in ways that enable group members to morally justify violent actions. Second, the ideologies of some extreme groups are grounded in hatred toward other groups. Identifying one or more groups of enemies that threaten the ideological values of the group provides fertile ground for dehumanizing others, blaming outgroups, and selectively interpreting or distorting outcomes of group actions.

The Ku Klux Klan’s White supremacy, anti-immigration ideology exemplifies these types of hate groups. Third, members of these groups who are highly identified with the group ideology are likely to be loyal to the group and its leaders (Angie et al., 2011; Hogg, 2003; Roccas et al., 2006). Calls to action on group websites make it easy to displace or diffuse responsibility for individual actions. In light of these observations, we proposed the following hypothesis: Hypothesis 3 (H3): Violent ideological group websites will display moral disengagement mechanisms to a greater extent than nonviolent ideological and nonideological group websites. Ó 2015 Hogrefe Publishing

19

While websites offer many opportunities for ideological groups to disseminate ideas and motivate members to support the group, little is known about how the psychological processes on these websites are related to aspects of website credibility.

Ideological Website Credibility and Psychological Processes Information credibility has been conceptualized as the believability of information and/or its source (Hovland, Janis, & Kelley, 1953), and it is assumed that credibility assessments involve both objective judgments and subjective receiver perceptions (Metzger, Flanagin, & Medders, 2010). Research on the credibility of websites suggests that at least two general categories of factors are important in understanding website credibility: content features and structural features (Metzger, 2007; Metzger et al., 2010; Teven & McCroskey, 1997). Initial research highlighted five content criteria potentially important for assessing the credibility of a website: information accuracy (free of errors), authority (credentials and/or qualifications of website author), objectivity (fact vs. opinion), currency (information is up-to-date), and coverage or scope (depth of information) (Alexander & Tate, 1999; Brandt, 1996).

Structural features of a website may also inform credibility judgments. Aesthetic qualities of the website, opportunities to examine or provide feedback, contact information, website architecture that facilitates navigation, and attractiveness and dynamism of a website have been shown to influence user assessments of credibility (Hong, 2006; Metzger, 2007; van Birgelen, Wetzels, & van Dolen, 2008).

Metzger’s review of online credibility research (Metzger, 2007) noted that several studies using a variety of samples demonstrated that Internet users rarely or occasionally engaged in fact-checking or verifying kinds of behavior to evaluate website credibility. More recent research has suggested that evaluation of credibility may be less tied to information accuracy than it is tied to perceived trustworthiness. Metzger, Flanagin, and Medders (2010) suggest that rather than evaluating the credibility of website information in a cognitively effortful manner, credibility assessments rely on social information and cognitive heuristics. Their focus group study findings showed that credibility assessments were influenced by social heuristics such as social information pooling (e.g., usergenerated comments, testimonials, reviews, reputation systems), social confirmation of personal opinion (e.g., users share similar interests, personalities, viewpoints), interpersonal recommendations (referrals from friends), cognitive heuristics such as endorsements of the website (expert opinions, ratings, links to other credible sites or groups), and confirmation of expectancies (e.g., appearance as expected, information consistent with own views).

In terms of structural website features, research by Fogg (2003) showed that people most frequently use site presentation criteria (graphics, readability, navigation ease, Journal of Media Psychology 2016; Vol. 28(1):16–31

20

S. Connelly et al.: Psychological Processes and Online Ideological Groups

functionality) to evaluate credibility, followed by content features (e.g., breadth, accuracy, bias), source motives (e.g., did source have something to gain), and source reputation (e.g., name recognition, third-party endorsements).

These findings were consistent with results of other studies (Eysenbach & Köhler, 2002; Rieh, 2002).

In considering credibility assessments of ideological websites, several points are important to bear in mind.

Given that some viewers of ideological websites are seeking to confirm or adopt a social identity, their evaluations of credibility may be particularly susceptible to social confirmation and cognitive expectancy biases. If ideological websites provide social confirmation of one’s personal views or are a good match with one’s beliefs and values, they are likely to be viewed as more credible (Metzger et al, 2010). Alternatively, if the content or structural features of an ideological website run counter to one’s expectations for that website, this expectancy violation is likely to reduce credibility. These biases have implications for how psychological processes on a website might relate to perceived ideological match or expectancy violation. For instance, the presence of social categorization, outgrouping, and moral disengagement on violent ideological group websites is likely to match the beliefs and values, and confirm the expectations, of individuals seeking to adopt or confirm this kind of extreme identity. However, individuals looking to identify with nonviolent ideologies may find the presence of social categorization, outgrouping, and moral disengagement processes on nonviolent websites incongruent with the beliefs and values highlighted on those sites, which could aversely influence credibility.

These psychological processes could be related to website credibility even for viewers not seeking identity or affiliation with the group. Violent ideological groups may recognize the need to include certain content and structural features that will make the group seem credible to a broader range of viewers in the face of the extreme ideas, beliefs, and values that comprise their ideologies. When outgrouping, social categorizing, and moral disengagement are infused within a violent ideology and are present on a group’s website, credibility features aimed at boosting goodwill, expertise, currency, and trustworthiness may also be more pervasive. This would not necessarily hold true for nonviolent ideological group websites which may not see as much need to appeal to viewers other than through their ideology.

While goodwill is likely to be part of nonviolent ideologies, features reflecting currency or expertise may not be highlighted to the extent they are on violent ideological websites.

Accordingly, we hypothesized the following: Hypothesis 4 (H4): Social categorization, outgrouping, and moral disengagement will correlate positively with website content credibility features for violent ideological groups and negatively for nonviolent ideological groups.

It is less clear how social categorization, outgrouping, and moral disengagement might relate to structural features of website credibility for different types of ideological Journal of Media Psychology 2016; Vol. 28(1):16–31

groups. Given that structural features influence credibility (Metzger et al., 2010) and that these features are particularly important when individuals first visit a website (Davern, Te’eni, & Moon, 2000), this may be important to examine. Here again, violent ideological groups may see the need to use structural website features to attract and retain visitors, such as how well the website is organized, suggesting a positive correlation of structural features and negative psychological processes. However, ideological groups in general may focus more on the ideas they are expressing and promoting and less on structural website features, so the relationship between psychological processes and structural features may be weak.

Research Question 1 (RQ1): Will psychological processes on ideological websites be related to structural website features related to credibility for violent and nonviolent ideological groups?

The present study used content analysis to compare violent and nonviolent ideological group websites with each other and with nonideological group websites, on three sets of key variables: social categorization, outgrouping, and moral disengagement. Additionally, this study sought to clarify whether and how these website characteristics relate to content and structural credibility website features and if these relationships differ across website types.

Method Website Identification and Classification Websites for actual groups represented at a national or international level were considered for inclusion in this study.

Community or regional groups were not included because the size and visibility of such groups is likely much smaller.

Websites eligible for consideration had to identify with a group that meets face-to-face or otherwise facilitates member connection and interactions through the computermediated contact. Three types of group websites were identified for this study: violent ideological, nonviolent ideological, and nonideological. A list of websites from prior research (Griffith et al., 2013) was used as a starting point, and other websites were identified through following links within these websites and through keyword Internet searches. Keyword search terms were based on the names of known ideological and nonideological groups (e.g., ‘‘the Ku Klux Klan’’ and ‘‘American Civil Liberties Union’’) and ideological issues (e.g., ‘‘animal rights’’).

An initial list of 120 websites was compiled for evaluation.

Information from the ‘‘About Us,’’ ‘‘Mission Statement,’’ or ‘‘Statement of Purpose’’ area(s) of each website were reviewed and rated on several criteria by three trained coders to determine whether a website was ideological or nonideological. These criteria included whether the website (1) articulates a mental model about the goals the group seeks to obtain that is rooted in events of the past; (2) ties Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

interpretation of events to the mental model; (3) focuses on a few, core transcendent goals largely centered on a return to a past idealized state; and (4) rejects beliefs that are not congruent with the mental model. An overall mean across raters and criteria was computed for each website, and means were transformed into z-scores. Groups with a z-score greater than 1.00 were classified as ideological, and groups with a z-score less than 1.00 were classified as nonideological.

Examination of z-scores showed that several groups did not cleanly fall into a single category, and therefore those groups were removed from consideration. These groups included Oxfam International, Amnesty International, Mothers Against Drunk Driving (MADD), Hare Krishna Society, Amitabha Buddhist Society, Students for a Democratic Society, Project Reason, La Leche League, Free Believers Network, Muslim Aid (UK), U.S. Sportsmen’s Alliance, and American Society for the Prevention of Cruelty to Animals (ASPCA). Additionally, face validity of group types was also taken into account to ensure that ratings matched overall impressions of the groups’ ideological/ nonideological standing. In several instances, groups received ratings that were incongruent with overall perceptions. These groups included the National Audubon Society and Save Our Wild Salmon. These groups showed ideological tendencies in some areas of the website while remaining nonideological in nature regarding other criteria. These groups, similar to those listed above, did not fit distinctly within one category and thus might have had an unintended influence on relationships between predictors and key outcomes. For that reason, these groups were also removed from the sample.

Ideological websites were identified and classified as violent if they met one or more of the criteria described

by Griffith et al. (2013): (1) the website sanctions violence, (2) the website is associated with one or more other groups known to sanction violence, (3) members of the group represented on the website have been linked to acts of violence, and (4) the website or group has been classified as a hate group and/or a violent group by a reputable third party source (e.g., the Southern Poverty Law Center). Overall, this process resulted in a sample of 105 groups, with 68 groups categorized as ideological (violent = 32, nonviolent = 36) and 37 groups categorized as nonideological.

The final list of selected websites was reviewed by three experts on ideological groups. The list is shown in Table 1.

Development of Content Rating Scales and Coding Process Concurrent with developing the criteria for categorizing websites and the website list, we reviewed the literature addressing the topics of social categorization, moral disengagement, ideological groups, website characteristics, and credibility. Based on this review, definitions and rating metrics for the content coding were developed for key variables in each of these categories. Rating metrics were based on

Ó 2015 Hogrefe Publishing

21

the literature or were borrowed or adapted from previously validated metrics (Angie et al., 2011). The format of each rating scale included the construct definition along with a 5-point rating scale with examples anchoring the low, middle, and high end of the scale. Example ratings scales are provided in the Appendix, including one aspect of social categorization (superiority), one aspect of outgrouping (information distraction) and one aspect of moral disengagement (misrepresenting consequences).

Once rating scales were developed, six doctoral student coders from the Departments of psychology and communication were divided into two groups of three. Each group was randomly assigned to rate either the predictors or the criteria for the coding assigned that week, in an effort to minimize same source method bias. Thus, predictors and criteria for a given website were always rated by different groups of coders. Coding took approximately 4 months.

Rater Training Raters were trained together in several ways prior to the coding. First, they were instructed to use a ‘‘surfing’’ perspective when scanning and reading the websites for information related to the variables of interest (Rains & Karmikel, 2009). Content analytic studies of websites often focus on the home page and other specific areas of the website to identify relevant information (e.g., Gerstenfeld, Grant, & Chiang, 2003; Zhou, Reid, Qin, Chen, & Lai, 2005). A study by Griffith et al. (2013) analyzed website content at both primary and secondary levels within websites, demonstrating the need for exploring beyond the home page. This and other evidence has suggested that website study designs should include at least a second-level analysis (Symonenko, 2006). Accordingly, raters were instructed to expand their search for information relevant to the rating scales to the second level of each group’s website and, if need be, to examine additional levels – for example, some of the structural credibility information such as contact information was found several levels into the website.

A second aspect of rater training involved frame-ofreference training (Bernardin & Buckley, 1981) to familiarize the coders with the variables of interest, the coding process, and rating errors that sometimes occur, such as central tendency and halo errors. Raters were provided with a complete set of rating scales for evaluating the predictors and criteria of interest. They practiced applying the rating scales on approximately eight ideological and nonideological websites not included in the present sample. After rating several websites independently, raters met to compare and discuss similarities and differences in how they applied each rating scale. When discrepancies existed, the relevant definition and scale anchors were reviewed and clarified, and discussion continued until the raters reached consensus regarding the application of that scale. Coding practice lasted for approximately 3 weeks until raters reached acceptable levels of interrater agreement.

Journal of Media Psychology 2016; Vol. 28(1):16–31

22

S. Connelly et al.: Psychological Processes and Online Ideological Groups

Table 1. Ideological groups by category Nonideological

Nonviolent ideological

Amateur Entomologists’ Society

American Baptist Church

American Association of Retired Persons American Astronomical Society

American Cause American Civil Liberties Union

American American American American

Americans United Center for Bioethical Reform Christian Exodus Coalition to Stop Gun Violence

Botanical Council Cancer Society Diabetes Association Fisheries Society

Friends of the Earth Hadassah Independent American Party IslamiCity Islamic Society of North America

Kingdom Identity Ministries Ku Klux Klan

Jewish Voice for Peace John Birch Society

League of the South National Alliance

Latter Day Saints (Mormon) Church Libertarian Party National Association for the Advancement of Colored People National Coalition for Men National Organization for Women National Rifle Association NOH8 Campaign

National Association for the Advancement of White People National Democratic Front National Socialist Movement

Coffee Party Council of Conservative Citizens Earth First!

American Sewing Guild

Federation for American Immigration Reform Freedom from Religion Foundation

Amnesty International Asian American Arts Alliance Association of Woodworking and Furnishing Suppliers Atomic Age Alliance Big Brothers/Big Sisters of America British Beatles Fan Club Children and Adults with Attention Deficit/Hyperactivity Disorder Doctors without Borders Habitat for Humanity Jenny Craig Lions Club Mensa Mustang Club National Association for Amateur Radio National Association for the Self-Employed National Association of Miniature Enthusiasts National Association of Rocketry National Street Rod Association Photographic Society of America Shriners International Society of Professional Journalists Special Olympics Teamsters US Tennis Association

Aggressive Christianity Missionary Training Corps Alpha 66 Americans for Truth About Homosexuality Anarchist Federation Animal Liberation Front Army of God Aryan Nations The Barnes Review Creativity Movement Earth Liberation Front English Defence League The Family International Ezzedeen Al-Qassam Brigade (Hamas) Heterosexuals Organized for a Moral Environment Imperial Klans of America Institute for Historical Review Jewish Defense League

American Heart Association American Meteorological Society American Red Cross

American Trucking Associations

Violent ideologicala

ONE Campaign

Negotiation is Over Operation Rescue Power of Prophecy People for the Ethical Treatment of Animals Prairie Fire Organizing Committee

ProLife Action League

Sovereign Citizens

Sierra Club Socialist Party USA Tea Party Nation

United for a Sovereign America Volksfront Westboro Baptist Church

Unitarian Universalist Association of Congregations United Methodist Church United Pentecostal Church International United States Conference of Catholic Bishops

Yellow Ribbon Club Note. aIt is not recommended to visit violent, ideological group websites without extensive antivirus protection software and encryption service enabled on your computer.

Journal of Media Psychology 2016; Vol. 28(1):16–31

Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

During the coding process, rater agreement was empirically evaluated on a weekly basis between coders in each group, and any agreement problems were addressed.

If agreement was low for a given week’s coding (e.g., more than 25% of the rating scales had agreement below .60) coders were retrained, and the websites were recoded. Interrater agreement was calculated with an r*wg to estimate the degree of interchangeability of raters (Lindell & Brandt, 1999). This measure of reliability refers to the degree to which ratings made by coders are nearly identical (Kozlowski & Hattrup, 1992). Raters typically had very similar ratings, and this lack of variance made r*wg more suitable for assessing reliability than intraclass correlation interrater reliability estimates. The latter estimate will be artificially suppressed when ratings across coders lack variability (Stemler & Tsai, 2008; Tinsley & Weiss, 1975).

Intercoder agreement was acceptable for all measures including site credibility (r*wg = .68), structural credibility (r*wg = .80), social categorization (r*wg = .76), outgrouping (r*wg = .75), and moral disengagement (r*wg = .95).

Psychological Processes Rating Scales Several psychological processes were measured by aggregating coder ratings on a number of metric rating scales.

The social categorization scale (a = .93) comprised four items: group differentiation (the degree to which the group compares and contrasts itself with others on its website), superiority (the extent to which the group expresses its own superiority and entitlement), disagreement with dissenting views (the degree to which the group rejects views differing from their core beliefs), and deindividuation (the degree to which group members are encouraged to view themselves as part of a group rather than individuals and to behave in ways consistent with group expectations).

The second psychological process scale, outgrouping, consisted of three items (a = .96). These were seeing outgroups as enemies (the extent to which the website portrays the world as black and white and that groups/individuals who do not share their views are enemies), information distortion (the degree to which the website misrepresents news and information about outgroups to serve ingroup beliefs and purposes), and negative social comparisons (the extent to which the website expresses negative social comparisons with the outgroup). The moral disengagement (a = .95) scale was based on four of Bandura’s (1999) dimensions: dehumanization (the degree to which the website deemphasizes human qualities and discourages compassion and sensitivity with regard to outgroups), misrepresentation of consequences (the extent to which the website distorts the relationships between the actions they advocate and the effects of those actions), euphemistic labeling (the extent to which language is sanitized to lessen the emotional intensity of the reality of the group’s message or actions promoted), and displacement and diffusion of responsibility (the extent to which the website assigns blame to other groups or avoids taking responsibility).

Ó 2015 Hogrefe Publishing

23

Website Credibility Rating Scales Website credibility was assessed on website content credibility and website structural credibility. Content credibility (a = .91) refers to the extent to which the website attempts to appear as if providing credible content (Flanagin & Metzger, 2007; McCroskey & Teven, 1999; McCroskey & Young, 1981). This scale contained five items: trustworthiness (extent to which the group website attempts to appear honest and reliable), fairness (the extent to which the group website attempts to appear objective and balanced), expertise (the extent to which the group website attempts to appear qualified and intelligent), goodwill (the extent to which the group website projects unselfishness and concern for others), and currency/recency (the extent to which information on the website is up-to-date). These were rated on a 5-point Likert scale (1 = not at all; 5 = to a great extent), and ratings were averaged across items and raters to form an overall score. Website structural credibility (a = .76) refers to the structural composition of websites (Eysenbach & Köhler, 2002; Fogg, 2003; Metzger, 2007; van Birgelen et al., 2008). This scale contained six items: website organization, website architecture, and overall cleanness, which were rated on 5-point scales, and presence of privacy policy (Yes = 1, No = 0), presence of contact information for the group (Yes = 1, No = 0), and whether the website included a third-party endorsement (Yes = 1, No = 0).

Item scores were averaged across items and coders to create an overall score for structural credibility features.

Covariates Two other variables were assessed as controls in this study.

To control for differences in the range of information provided on the websites about the groups, information diversity was assessed. This was a one-item rating of the degree to which the website provided a wide range of information about the group, ranging from 1 (very narrow scope of information), to 5 (contains information about many topics relevant to the groups and the group’s goals). The second control variable indicated whether or not (Yes = 1, No = 0) the website fostered an online community where visitors could participate in discussion forums, comment on articles, or comment on blog posts. A yes response was assigned if one or more of these kinds of community activities were available.

Analyses Basic descriptive statistics were calculated for the constructs of interest for each website category (nonideological, nonviolent ideological, and violent ideological).

Simple t tests were conducted to test the first three hypotheses, using unpooled estimates for the few instances when homogeneity of variance was violated. Correlations and

Journal of Media Psychology 2016; Vol. 28(1):16–31

24

S. Connelly et al.: Psychological Processes and Online Ideological Groups

Table 2. Means and standard deviations for psychological process and credibility by website type Nonideological (n = 37) Social categorization (overall)abc Group differentiationabc Superiorityabc Disagree with dissenting viewsabc Deindividuationabc Outgrouping (overall)abc Outgroups as enemiesabc Information distortionabc Negative social comparisonsac Moral disengagement (overall)abc Misrep. consequencesabc Displace/diffuse responsibilityabc Euphemistic labelingabc Dehumanizationabc Content credibility (overall)abc Trustworthinessabc Fairnessabc Expertiseabc Goodwillbc Currencybc Structure credibility (overall)bc Website organizationbc Website architecturebc Cleannessbc Contact information Privacy policy Third party endorsementb

Nonviolent, ideological (n = 36)

Violent, ideological (n = 32)

M

SD

M

SD

M

SD

1.45 1.47 1.47 1.59 1.38 1.20 1.17 1.17 1.47 1.16 1.18 1.12 1.23 1.14 3.90 4.08 3.51 3.96 4.04 4.04 2.65 4.05 3.41 3.71 1.66 1.59 1.47

0.51 0.80 0.59 0.77 0.62 0.54 0.67 0.34 0.80 0.29 0.39 0.30 0.30 0.51 0.60 0.64 0.56 0.85 0.82 0.87 0.41 0.71 1.02 0.75 0.15 0.23 0.29

3.06 3.10 2.71 3.67 2.74 2.62 2.50 2.70 3.10 2.22 2.19 2.35 2.40 1.95 3.48 3.50 2.75 3.30 3.51 4.34 2.59 3.93 3.39 3.72 1.64 1.53 1.45

0.70 1.12 0.89 0.91 0.85 0.99 1.20 0.90 1.12 0.76 0.96 0.97 0.78 0.79 0.70 0.85 0.84 0.80 1.05 0.55 0.37 0.59 1.02 0.74 0.15 0.24 0.33

4.13 4.38 4.15 4.54 3.47 4.16 4.20 4.06 4.38 3.43 3.35 3.36 3.56 3.42 2.33 2.46 1.64 2.52 2.03 3.02 2.03 2.98 2.26 2.59 1.57 1.46 1.35

0.49 0.56 0.85 0.53 0.71 0.55 0.71 0.76 0.56 0.59 0.69 0.75 0.76 0.80 0.59 0.72 0.58 0.79 0.80 1.15 0.43 0.95 0.89 0.98 0.26 0.29 0.15

Notes. aSignificant mean differences between nonideological and nonviolent ideological websites; bSignificant mean difference between nonideological and violent ideological websites; cSignificant mean difference between nonviolent and violent ideological websites; unpooled estimates for the t statistic were used when homogeneity of variance was violated.

hierarchical multiple regression analyses were used to test H4 and the research question.

Results Table 2 presents the means and standard deviations at the scale and item level for the psychological processes and credibility constructs. With regard to social categorization, t tests showed significant differences in the overall scale mean and item-level means. Consistent with H1, violent ideological group websites (M = 4.13, SD = 0.49) showed greater social categorization than nonviolent ideological (M = 3.06, SD = 0.70) websites, t(66) = 7.43, p < .001.

Both categories of ideological websites had higher mean levels of social categorization than nonideological group websites (M = 1.45, SD = 0.51), t(71) = 11.19, p < .001 (nonviolent), t(67) = 22.037, p < .001 (violent). The pattern of means for outgrouping and moral disengagement

Journal of Media Psychology 2016; Vol. 28(1):16â&#x20AC;&#x201C;31

was similar, with the violent websites showing the highest levels (Mo = 4.16, SD = 0.55; Mmd = 3.43, SD = 0.59) and the nonviolent ideological websites (Mo = 2.62, SD = 0.99; Mmd = 2.22, SD = 0.76) and nonideological websites (Mo = 1.20, SD = 0.54; Mmd = 1.16, SD = 0.29) showing significantly lower levels, t(66) = 7.99, p < .001 (violent vs. nonviolent outgrouping), t(67) = 22.45, p < .001 (violent vs. nonideological outgrouping), t(66) = 7.24, p < .001 (violent vs. nonviolent moral disengagement), t(67) = 19.76, p < .001 (violent vs. nonideological moral disengagement). These results supported H2 and H3. Interestingly, the nonviolent ideological websites evidenced modest amounts of social categorization and outgrouping, suggesting that ideological groups in general use these processes in forming and maintaining distinct social identities. Content and structure credibility were higher for the nonideological (Mc = 3.90, SD = 0.60; Ms = 2.65, SD = 0.41), t(67) = 10.92, p < .001 (content), t(66) = 5.98, p < .001 (structure) and nonviolent (Mc = 3.48, SD = 0.70; Ms = 2.59, SD = 0.37), t(66) = 7.28,

Ă&#x201C; 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

25

Table 3. Nonideological website correlations for psychological processes and website credibility 1 1 2 3 4 5 6 7

Social categorization Outgrouping Moral disengagement Information diversity Online community Website content credibility Website structural credibility

2

â&#x20AC;&#x201C; .86** .85** .27 .06 .05 .19

3

.92** .12 .07 .08 .03

4

.07 .02 .12 .09

5

.14 .60** .53**

6

.05 .18

7

.67**

Note. n = 37. *p < .05; **p < .001.

Table 4. Nonviolent and violent ideological website correlations for psychological processes and website credibility 1 1 2 3 4 5 6 7

Social categorization Outgrouping Moral disengagement Information diversity Online community Website content credibility Website structural credibility

2 .70**

.84** .77** .16 .15 .65** .45**

.82** .26 .27 .76** .65**

3 .52** .78** .17 .31 .59** .46**

4 .00 .10 .19 .11 .12 .22

5 .19 .28 .24 .45** .42** .30

6 .39* .25 .31 .41* .23

7 .43* .20 .10 .35 .37* .47**

.82**

Note. n = 36 for nonviolent ideological websites (below the diagonal); n = 32 for violent ideological websites (above the diagonal). *p < .05; **p < .001.

p < .001 (content), t(62) = 5.51, p < .001 (structure) websites compared with the violent ideological websites (Mc = 2.33, SD = 0.59; Ms = 2.04, SD = 0.43).

Correlations among the study variables by website category are shown in Tables 3 and 4. Social categorization, outgrouping, and moral disengagement showed large, positive correlations for nonviolent ideological websites (avg. r = .81) and nonideological websites (avg. r = .88). Correlations for the violent ideological websites were also fairly large (avg. r = .67), with social categorization and moral disengagement showing the lowest correlation (r = .52).

These psychological processes showed no relationship to website content credibility or structural credibility for the nonideological websites. However, they showed modest to large negative relationships with credibility for the nonviolent ideological websites, with an average correlation of .67 for content credibility and .37 for structural credibility. Comparatively, outgrouping and moral disengagement were not significantly related to either type of credibility for the violent ideological websites, while social categorization showed modest negative correlations with content credibility (r = .38) and structural credibility (r = .43).

A series of hierarchical multiple regression analyses were conducted to test H4 and RQ1, as shown in Table 5.

Separate regressions were conducted for website content credibility and website structural credibility for each website category, with covariates entered in Step 1 and the three psychological processes entered in Step 2. Consistent with the zero-order correlations and H4, standardized beta coefficients indicated that social categorization (b = .14), outgrouping (b = .78), and moral disengagement (b = .23) accounted for significant variance in website content Ă&#x201C; 2015 Hogrefe Publishing

credibility for nonviolent ideological groups, R2D = .48, F(5, 31) = 3.72, p < .01. The positive beta for moral disengagement was likely due to suppression effects given the high intercorrelations of these variables. With regard to the violent ideological websites, H4 suggested that these psychological processes would contribute positively to ratings of website content credibility. However, neither the covariates nor the psychological processes accounted for significant variance in website content credibility for the violent websites. Taken together, these findings partially supported H4 and suggest that these types of negative processes hurt content credibility perceptions more for nonviolent websites than for violent ones.

Regression analyses examining the relationships of psychological processes to website structural credibility showed a pattern of findings similar to the content credibility regressions. Covariates were not significantly related to ratings of structural credibility for either type of ideological website. However, as indicated by the standardized betas, social categorization (b = .14), outgrouping (b = .86), and moral disengagement (b = .14) accounted for significant variance in website structural content for the nonviolent ideological websites, R2D = .41, F(3, 32) = 7.37, p < .001. Again, the positive betas are likely due to suppression. The psychological processes did not account for significant variance in website structural credibility for the violent ideological websites.

With regard to the nonideological groups, standardized betas indicated that information diversity (b = .60) was the only variable that accounted for significant variance in website content credibility, R2D = .36, FD(2, 34) = 9.36, p < .001, and website structural credibility, b = .57, Journal of Media Psychology 2016; Vol. 28(1):16â&#x20AC;&#x201C;31

26

S. Connelly et al.: Psychological Processes and Online Ideological Groups

Table 5. Relationships of social categorization, outgrouping, and moral disengagement to website credibility across website type

Website content credibility Step 1 Information diversity Online community

Step 2 Social categorization Outgrouping Moral disengagement

Website structure credibility Step 1 Information diversity Online community

Step 2 Social categorization Outgrouping Moral disengagement

Nonideological websitesa (n = 37)

Non-violent ideological websitesb (n = 36)

Violent ideological websitesc (n = 32)

b .60** .03 R2D .36 R2 Adjusted .32 – – – R2 D ns R2 Adjusted ns

b – – – – .14 .78** .23 R2 D .48 R2 Adjusted .41

b – – – – – – – – –

.57** .26 R2D .35 R2 Adjusted .31 – – – R2 D ns R2 Adjusted ns

– – – – .14 .86** .14 R2 D .41 R2 Adjusted .35

– – – – – – – – –

Note. ns = not significant; F values for Website Content credibility steps 1 and 2: aFchange1 (2, 34) = 9.36, p < .001; Fchange2 (3, 31) = 3.87, p < .69; bFchange1 (2, 33) = 3.70, p < .04; Fchange2 (3, 30) = 11.54, p < .001; ccovariates and predictors did not significantly contribute to content credibility for violent websites. F values for Website Structural credibility steps 1 and 2: aFchange1 (2, 34) = 9.06, p < .001; Fchange2 (3, 31) = .46. p < .72; bFchange1 (2,33) = 2.19, p < .13; Fchange (3, 32) = 7.37, p < .001; ccovariates and predictors did not significantly contribute to structural credibility for violent websites. *p < .05, **p < .01.

R2D = .38, FD(2, 34) = 9.06, p < .001. Social categorization, outgrouping, and moral disengagement did not significantly predict website content credibility or website structure credibility after controlling for information diversity and the presence of an online community.

Discussion This study extends the research on the online presence of ideological groups in some important ways. First, it provides evidence that psychological processes important to social identity formation and ethical behavior are embedded within general areas (i.e., home pages, ‘‘About Us’’ pages) on websites developed by ideological groups. Ideological groups are important sources of social identity for some individuals. Their online public faces, as manifested through their websites, show the presence of social categorization (highlighting group distinctiveness and superiority) and outgrouping (diminishing outside groups) processes known to influence the development and maintenance of social identity (Hogg, 2007; Tajfel & Turner, 1979; Taylor & Moghaddam, 1994). Additionally, websites of ideological groups also included content reflecting a number of moral disengagement mechanisms – mechanisms that reduce self-sanctioning and can increase the probability of Journal of Media Psychology 2016; Vol. 28(1):16–31

unethical behavior. Our comparison of the degree to which these psychological processes were present on violent and nonviolent ideological group websites indicated substantially higher levels of social categorization, outgrouping, and moral disengagement on the violent sites. These findings are consistent with and extend prior findings with ideological group discussion forums and message boards (Angie et al., 2011). To some extent, all ideological groups in this sample appeared motivated to engage visitors and encourage group members online to initiate and enhance identification with the group. However, use of these processes was more pronounced on violent websites.

A second contribution of this research is that it highlights the relationship of psychological processes to website credibility and, by extension, group credibility for the different categories of websites examined. The psychological processes of social categorization, outgrouping, and moral disengagement were unrelated to the websites’ structural and content credibility for nonideological groups. This may be because the source of social identity for these nonideological groups comes not from a rigidly defined set of beliefs and values, but from shared activities, interests, and goals. Research on social identities in online communities has suggested that in addition to the social exclusion aspect of social identity, online groups have more positive ways of fostering a sense of cohesion such as promoting a shared repertoire of activities and artifacts and mutual engagement Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

of community members (Pape, Reinecke, Rohde, & Strauss, 2003).

Alternatively, we expected and saw that social categorization, outgrouping, and moral disengagement were negatively related to website content and structural credibility features for nonviolent ideological groups. Interestingly, these groups use psychological processes that likely increase social identification, but that may ultimately undermine their credibility, at least with the ratings used in this study. The very nature of these processes may contradict important beliefs and values held by nonviolent ideological groups, something that may be fairly noticeable to viewers of these websites. Thus, the means through which nonviolent groups communicate their ideological positions appear to be an important part of how the group is viewed. At the bivariate level, these processes were also negatively correlated with website content and structural credibility for the violent groups, but the relationships were not as strong. When used together to account for variance in website credibility, these processes did not significantly contribute to the credibility of violent websites.

Theoretical Implications Parker and Janoff-Bulman (2013) note that outgroups are fundamental to ingroup identification (i.e., social identity) for morality-based groups. While most ideological groups have some level of belief that their views, values, and beliefs are ‘‘right’’; the implied or stated consequences of not endorsing and following the ideology range from being labeled as socially unjust to condemning one’s soul for all eternity. Thus, there is a range of ‘‘morality’’ implied by different types of ideologies. Angie et al. (2011) found some evidence that violent ideological group discussion boards contained significantly more religious content than nonviolent ideological groups. This suggests that social identity formation is not the same for all types of groups. Some processes are emphasized much more than others in ideological groups, and those sanctioning violence are likely to engage in certain processes more than others. Additionally, social identity motivations of the individuals seeking to identify with ideological groups may influence credibility assessments (Hogg, 2007; Thoroughgood, Padilla, Hunter, & Tate, 2012). Hogg (2007) suggests that the need to reduce self-uncertainty (about one’s cognitions, perceptions, feelings, behavior, etc.) through social categorization is higher in some individuals. These individuals seek out ideological groups that provide meaning, strong sets of values and beliefs, and clear norms for behavior that reduce this uncertainty. Thoroughgood et al. (2012) suggest that there is a susceptible circle of followers who are more likely to seek out destructive leaders and their groups. One category of followers consistent with attraction to ideological groups are the ‘‘lost souls,’’ or those individuals who seek to identify with a leader because they lack self-concept clarity, have poor core self-evaluations, have basic needs that are unmet, and are experiencing distress in their personal life.

A second category of followers includes the ‘‘acolytes,’’

Ó 2015 Hogrefe Publishing

27

or individuals whose values and goals are congruent with the leader’s/group’s. They see the ideological leader as a source of expert power able to pursue those values and goals. We only explored three psychological processes here that might contribute to the social identity of such groups, but there are likely others that will also show differences across these types of groups.

Psychological processes associated with social identity formation in ideological groups may also impact perceived credibility of such groups in online environments. If the group’s ideology is inconsistent with the psychological processes used to convey it, its credibility may be harmed, which would reduce the group’s ability to attract new members. Additionally, persistent use of ideology-inconsistent processes could erode member support for the group because this could create conditions for deidentification.

Alternatively, group members could get used to these inconsistencies over time through moral disengagement processes, which justify the ethicality of exclusionary social categorization processes. Such disengagement may make the psychological processes seem more consistent with the ideology over time, and credibility doubts could decrease. Examining the credibility of violent ideological group websites over time from the perspective of group affiliates would be an interesting area for future research.

Practical Implications Ideological groups are undoubtedly important to the fabric of modern society, reflecting a wide variety of political, religious, and social ideals. Nonviolent ideological groups have the potential to promote prosocial values, provide meaning, and facilitate just ethical behavior. However, to the extent that the group uses psychological processes that undermine its credibility, members and potential members may remain less than fully committed to, and even skeptical about, the group. Nonviolent groups may want to capitalize on other ways to enhance social identity that do not involve negative comparisons with outside groups.

Findings in this study also suggest that the credibility of violent ideological groups can potentially be undermined when their use of negative psychological processes is highlighted publicly. For example, online media outlets, traditional media, watch agencies, and others have called attention to the violent nature of these types of groups, but have not yet focused on the type of psychological manipulation that can occur. Some violent ideological groups are moving toward a sanitized public face and Internet presence. Greater public awareness of this sanitization could embolden these groups, potentially increasing their credibility with group members and affiliates. Alternatively, public awareness of such facades could result in increased skepticism and vigilance of those outside the group. Additionally, identifying whether these changes are associated with a reduction in psychological and/or physical violence is also needed to understand these trends.

Reform for some violent groups may be possible through interventions aimed at increasing awareness of

Journal of Media Psychology 2016; Vol. 28(1):16–31

28

S. Connelly et al.: Psychological Processes and Online Ideological Groups

moral disengagement and providing support for greater self-sanctioning. Unfortunately, certain ideologies are grounded in violence against other people and groups, such that reducing or eliminating the violence necessarily would change the ideology. It remains to be seen whether these groups can adapt their ideological views so that violence is not seen as the only means to maintaining the ideology.

Limitations This study was not without limitations. First, while we analyzed three times as many groups as previous studies looking at ideological group websites, the number of groups within each website category was still relatively low, ranging from 32 to 37, reducing the power in some of the analyses. Nevertheless, these sample sizes are sufficient for means testing and bivariate correlations, and the regression results were fairly consistent with the correlations. Second, evaluations of credibility were made by coders rather than individuals seeking to explore ideological groups as a source of social identity. The relationship of psychological processes to credibility should be explored in this type of sample. Third, this study was limited to English-language websites, given that this was the language spoken by the researchers and coders involved in the study. It would be interesting to see if the patterns of results hold in nonEnglish-language ideological websites. A fourth limitation is that we only examined a few psychological processes for ideological websites, and we recognize that others may also be operating such as norm setting and enforcement, emotional appeals, and social conformity. These could be examined in future studies. Additional control variables would also have been desirable, such as website size, age, and location, but this information was not consistently available across all websites. Finally, given the sample sizes, we classified all nonviolent organizations into one category, without differentiating among the specific types of ideological organization or group the website represented. For example, websites reflected religious, environmental, political, and social groups. The same held true for the violent ideological websites, with there being a range of ideological types as well as levels of violence subsumed within our categorization. Future research should explore similarities and differences in distinct types of ideological groups with respect to psychological processes and website credibility.

Concluding Remarks In an age where social uses of the Internet are expanding exponentially, it is important to continue exploring the presence, nature, and impact of ideological groups that have an online presence. Such groups are capable of exerting substantial influence on attitudes, beliefs, and values, particularly for individuals seeking social support and social identities. While there are many prosocial ideological groups, there are also antisocial, violent ones that continue Journal of Media Psychology 2016; Vol. 28(1):16–31

to attract people. We hope this research stimulates new understanding and continued study in this domain.

References Aberson, C. L., Healy, M., & Romero, V. (2000). Ingroup bias and self-esteem: A meta-analysis. Personality and Social Psychology Review, 4(2), 157–173.

Akins, J. K. (2006). The Ku Klux Klan: America’s forgotten terrorists. Law Enforcement Executive Forum, 5, 127–144.

Alexander, J. E., & Tate, M. A. (1999). Web wisdom: How to evaluate and create information quality on the Web. New York: Routledge.

Angie, A. D., Davis, J. L., Allen, M. T., Byrne, C. L., Ruark, G., Cunningham, C. B., . . . Mumford, M. D. (2011). Studying ideological groups online: Identification and assessment of risk factors for violence. Journal of Applied Social Psychology, 41, 627–657.

Ashmore, R. D., Deaux, K., & McLaughlin-Volpe, T. (2004). An organizing framework for collective identity: articulation and significance of multidimensionality. Psychological Bulletin, 130(1), 80–114.

Bandura, A. (1989). Human agency in social cognitive theory.

American psychologist, 44(9), 1175-1184.

Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3(3), 193–209.

Bernardin, H. J., & Buckley, M. R. (1981). Strategies in rater training. Academy of Management Review, 6(2), 205–212.

Bobo, L. (1983). Whites’ opposition to busing: Symbolic racism or realistic group conflict? Journal of Personality and Social Psychology, 45, 1196–1210.

Brandt, D. S. (1996). Evaluating information on the Internet.

Computers in Libraries, 16(5), 44–46.

Davern, M. J., Te’eni, D., & Moon, J. Y. (2000, December).

Content versus structure in information environments: A longitudinal analysis of Website preferences. In Proceedings of the 21st International Conference on Information Systems (pp. 564–570). Atlanta, GA: Association for Information Systems.

Dobratz, B. A., & Shanks-Meile, S. L. (1997). White power, white pride! The white separatist movement in the United States. New York: Twayne.

Ellemers, N., Spears, R., & Doosje, B. (2002). Self and social identity. Annual Review of Psychology, 53(1), 161–186.

Eysenbach, G., & Köhler, C. (2002). How do consumers search for and appraise health information on the world wide web?

Qualitative study using focus groups, usability tests, and indepth interviews. BMJ: British Medical Journal, 324(7337), 573–577.

Flanagin, A. J., & Metzger, M. J. (2007). The role of site features, user attributes, and information verification behaviors on the perceived credibility of web-based information.

New Media & Society, 9(2), 319–342.

Flanagin, A. J., & Metzger, M. (2008). Digital media and youth: Unparalleled opportunity and unprecedented responsibility. In M. Metzger & A. J. Flanagin (Eds.), Digital media, youth, and credibility (pp. 5–28) The John D. and Catherine T. MacArthur Foundation Series on Digital Media and Learning. Cambridge, MA: MIT Press.

Fogg, B. J. (2003). Prominence-interpretation theory: explaining how people assess credibility online. In CHI’03 Extended Abstracts on Human Factors in Computing Systems (pp.

722–723) April. New York: ACM. Frischlich, L., & Rieger, D. (2013). Dealing with the dark side: Negative ingroups and the effects of right-wing and Islamic extremist propaganda videos. June. Paper presented at the Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

63th annual International Communication Association conference, London, UK. Gerstenfeld, P. B., Grant, D. R., & Chiang, C. P. (2003). Hate online: A content analysis of extremist Internet sites.

Analyses of Social Issues and Public Policy, 3(1), 29–44.

Griffith, J. A., Byrne, C. L., Nei, D. S., Barrett, J. D., Hughes, M. G., Davis, J. L., . . . Mumford, M. D. (2013). Online ideology: A comparison of website communication and media use. Journal of Computer-Mediated Communication, 18(2), 137–153.

Heath, R. L., & O’Hair, H. D. (2009). The significance of crisis and risk communication. In R. L. Heath & H. D. O’Hair (Eds.), Handbook of risk and crisis communication (pp. 5– 30). New York: Routledge.

Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias.

Annual Review of Psychology, 53(1), 575–604.

Hogg, M. A. (2003). Social identity. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (pp. 462– 479). New York: Guilford Press.

Hogg, M. A. (2005). Uncertainty, social identity, and ideology.

In S. R. Thye & E. J. Lawler (Eds.), Advances in group processes (Vol. 2, pp. 203–229). San Diego, CA: Elsevier.

Hogg, M. A. (2007). Uncertainty-identity theory. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 39, pp. 69–126). San Diego, CA: Academic Press.

Hogg, M. A., Meehan, C., & Farquharson, J. (2010). The solace of radicalism: Self-uncertainty and group identification in the face of threat. Journal of Experimental Social Psychology, 46, 1061–1066.

Hong, T. (2006). The influence of structural and message features on Web site credibility. Journal of the American Society for Information Science and Technology, 57(1), 114–127.

Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: Psychological studies of opinion change. New Haven, CT: Yale University Press.

Jensen, M. L., Dunbar, N., Connelly, S., Taylor, W., Adame, B., Rozzell, B., & Hughes, M. (2013). Social media on violent ideological group websites. Paper presented at the Americas Conference on Information Systems, Chicago, IL. Jost, J. T., & Hunyady, O. (2005). Antecedents and consequences of system-justifying ideologies. Current Directions in Psychological Science, 14(5), 260–265.

Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life.

Journal of Personality and Social Psychology, 40(3), 414– 431.

Kozlowski, S. W., & Hattrup, K. (1992). A disagreement about within-group agreement: Disentangling issues of consistency versus consensus. Journal of Applied Psychology, 77(2), 161–167.

Lee, E., & Leets, L. (2002). Persuasive storytelling by hate groups online examining its effects on adolescents. American Behavioral Scientist, 45(6), 927–957.

Leidner, B., Castano, E., Zaiser, E., & Giner-Sorolla, R. (2010).

Ingroup glorification, moral disengagement, and justice in the context of collective violence. Personality and Social Psychology Bulletin, 36(8), 1115–1129.

Lindell, M. K., & Brandt, C. J. (1999). Assessing interrater agreement on the job relevance of a test: A comparison of CVI, T, rWG(J), and r*WG(J). Journal of Applied Psychology, 84(4), 640–647.

Matusitz, J., & O’Hair, D. (2008). The role of the internet in terrorism. Terrorism: Communication and rhetorical perspectives, 383–407.

McCroskey, J. C., & Teven, J. J. (1999). Goodwill: A reexamination of the construct and its measurement. Communications Monographs, 66(1), 90–103.

Ó 2015 Hogrefe Publishing

29

McCroskey, J. C., & Young, T. J. (1981). Ethos and credibility: The construct and its measurement after three decades.

Communication Studies, 32(1), 24–34.

Metzger, M. J. (2007). Making sense of credibility on the Web: Models for evaluating online information and recommendations for future research. Journal of the American Society for Information Science and Technology, 58(13), 2078–2091.

Metzger, M. J., Flanagin, A. J., & Medders, R. B. (2010). Social and heuristic approaches to credibility evaluation online.

Journal of Communication, 60(3), 413–439.

Moghaddam, F. M. (2005). The staircase to terrorism: A psychological exploration. American Psychologist, 60(2), 161–169.

Mumford, M. D., Bedell-Avers, K. E., Hunter, S. T., Espejo, J., Eubanks, D., & Connelly, M. S. (2008). Violence in ideological and non-ideological groups: A quantitative analysis of qualitative data. Journal of Applied Social Psychology, 38(6), 1521–1561.

Pape, B., Reinecke, L., Rohde, M., & Strauss, M. (2003, November). E-community-building in wilnf-central. In Proceedings of the 2003 International ACM SIGGROUP Conference on Supporting Group Work (pp. 11–20).

New York: ACM. Parker, M. T., & Janoff-Bulman, R. (2013). Lessons from morality-based social identity: The power of outgroup ‘‘hate,’’ not just ingroup ‘‘love’’. Social Justice Research, 26, 81–96.

Pittinsky, T. L., Shih, M. J., & Trahan, A. (2006). Identity cues: Evidence from and for intra-individual perspectives on positive and negative stereotyping. Journal of Applied Social Psychology, 36(9), 2215–2239.

Rains, S. A., & Karmikel, C. D. (2009). Health informationseeking and perceptions of website credibility: Examining Web-use orientation, message characteristics, and structural features of websites. Computers in Human Behavior, 25(2), 544–553.

Rieh, S. Y. (2002). Judgment of information quality and cognitive authority in the Web. Journal of the American Society for Information Science and Technology, 53, 145–161.

Riek, B. M., Mania, E. W., & Gaertner, S. L. (2006). Intergroup threat and outgroup attitudes: A meta-analytic review.

Personality and Social Psychology Review, 10(4), 336–353.

Roccas, S., Klar, Y., & Liviatan, I. (2006). The paradox of group-based guilt: Modes of national identification, conflict vehemence, and reactions to the in-group’s moral violations.

Journal of Personality and Social Psychology, 91(4), 698–711.

Schafer, J. A. (2002). Spinning the web of hate: Web-based hate propagation by extremist organizations. Journal of Criminal Justice and Popular Culture, 9(2), 69–88.

Sherif, M., & Sherif, C. (1969). Ingroup and intergroup relations. Social psychology, 221–266.

Stanton, J. J. (2002). Terror in cyberspace: Terrorists will exploit and widen the gap between governing structures and the public. American Behavioral Scientist, 45(6), 1017–1032.

Stemler, S. E., & Tsai, J. (2008). Best practices in interrater reliability: Three common approaches. In J. W. Osborne (Ed.), Best practices in quantitative methods (pp. 29–49).

Thousand Oaks, CA: Sage.

Stephan, W. G., & Stephan, C. W. (1996). Intergroup relations.

Madison, WI: Brown & Benchmark.

Stephan, W. G., & Stephan, C. W. (2000). An integrated threat theory of prejudice. In S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 23–45). Mahwah, NJ: Erlbaum.

Symonenko, S. (2006). Exploring patterns in website content structure. Paper presented at the Information Architecture Summit (Learning, Doing, Selling). Vancouver, Canada.

Journal of Media Psychology 2016; Vol. 28(1):16–31

30

S. Connelly et al.: Psychological Processes and Online Ideological Groups

Retrieved from http://iasummit.org/2006/files/177_Presentation_ Desc.pdf Tajfel, H. (1982). Social psychology of intergroup relations.

Annual Review of Psychology, 33, 1–39.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47).

Monterey, CA: Brooks/Cole.

Taylor, D. M., & Moghaddam, F. M. (1994). Theories of intergroup relations: International social psychological perspectives. Westport, CT: Praeger.

Teven, J. J., & McCroskey, J. C. (1997). The relationship of perceived teacher caring with student learning and teacher evaluation. Communication Education, 46, 1–9.

Thoroughgood, C. N., Padilla, A., Hunter, S. T., & Tate, B. W. (2012). The susceptible circle: A taxonomy of followers associated with destructive leadership. The Leadership Quarterly, 23(5), 897–917.

Tinsley, H. E., & Weiss, D. J. (1975). Interrater reliability and agreement of subjective judgments. Journal of Counseling Psychology, 22(4), 358–376. van Birgelen, M. J., Wetzels, M. G., & van Dolen, W. M. (2008). Effectiveness of corporate employment web sites: How content and form influence intentions to apply.

International Journal of Manpower, 29(8), 731–751.

Van Dijk, T. A. (2006). Ideology and discourse analysis. Journal of Political Ideologies, 11(2), 115–140.

Zhou, Y., Reid, E., Qin, J., Chen, H., & Lai, G. (2005). US domestic extremist groups on the Web: link and content analysis. Intelligent Systems, IEEE, 20(5), 44–51.

Date of acceptance: August 5, 2014 Published online: January 30, 2015 Shane Connelly (PhD, George Mason University, 1995) is a Professor of Industrial and Organizational Psychology and Co-director of the Center for Applied Social Research at the University of Oklahoma. She has published over 70 journal articles in various outlets such as The Leadership Quarterly, International Journal of Conflict Management, Journal of Computer-Mediated Communication, Human Performance, and Ethics and Behavior. Her research interests include emotions in leadership and workplace contexts, research ethics and ethical decision-making, and communication in ideological groups.

Shane Connelly Department of Psychology University of Oklahoma 455 W. Lindsay St. Dale Hall Tower, Room 705 Norman, OK 73019 USA Tel. +1 405 325-4580 E-mail sconnelly@ou.edu

Journal of Media Psychology 2016; Vol. 28(1):16–31

Norah E. Dunbar (PhD, University of Arizona, 2000) is a Professor of Communication at University of California Santa Barbara. Her expertise is in nonverbal and interpersonal communication, with special emphasis on dominance and power relationships, interpersonal synchrony, and deception detection. She has published over 40 journal articles and book chapters including those in Journal of Computer-Mediated Communication, Journal of Nonverbal Behavior, Communication Research, Communication Monographs, and Journal of Social and Personal Relationships.

Matthew L. Jensen (PhD, 2007) is an assistant professor of management information systems and a codirector of the Center for Applied Social Research at the University of Oklahoma. His interests include computer-aided decision making, knowledge management, human-computer interaction, and computer-mediated communication.

Jennifer A. Griffith (Ph.D., University of Oklahoma) is an Assistant Professor of Management at Alfred University. Her research interests include leadership, emotions, and communication. She has published articles in a number of journals such as The Leadership Quarterly, International Journal of Conflict Management, and The European Journal of Work and Organizational Psychology.

William D. Taylor is a doctoral candidate in Industrial and Organizational Psychology at the University of Oklahoma. His current research interests include occupational health and safety, risk perception, leadership, and performance management.

Genevieve Johnson is a doctoral candidate in Industrial and Organizational Psychology at the University of Oklahoma. Her current research interests include emotions in the workplace, ethical decision making, leadership, interpersonal and organizational communication, and performance feedback. She has been published articles in various journals such as Human Relations, The Leadership Quarterly, and Journal of Computer-Mediated Communication.

Ó 2015 Hogrefe Publishing

S. Connelly et al.: Psychological Processes and Online Ideological Groups

Michael Hughes (PhD, University of Oklahoma) is an Industrial Organizational psychologist at Human Resources Research Organization, Alexandria, VA, whose primary areas of expertise include test development and validation, and analyses of high-stakes testing data. His research interests also include complex skill acquisition, training, and ideological groups.

31

Michael D. Mumford (Ph.D., University of Georgia) is a Professor of Industrial and Organizational Psychology at the University of Oklahoma and is a co-director of the Center for Applied Social Research.

He has achieved the highest research honor at OU, the George Lynn Cross Research Professor. He has published extensively in the areas of leadership, creativity, and ethics in outlets such as the Creativity Research Journal, The Leadership Quarterly, Journal of Organizational Behavior, and Ethics and Behavior.

Appendix Table A1. Example rating scales. Superiority/entitlement: Extent to which the website expresses group superiority or entitlement 1

2

3

4

5

No expression of superiority on the website

Some expression of superiority on the website

Pervasive expression of superiority on the website

Website views everybody (ingroup or outgroup) as equal and does NOT consider itself better than others

Includes a small number of statements expressing strong feelings of superiority or many statements making weak expressions of superiority

Website expresses or implies feelings of superiority (‘‘Our way is the only right way’’), entitlement (‘‘We deserve to have things our way’’)

Information distortion: Degree to which website misrepresents information relevant to outgroups 1

2

3

4

5

Low

Moderate

High

Website accurately reports information from newscasts, books, websites, etc. May even question the veracity of reports to ensure accuracy.

Some deliberate misrepresentation of information to support views.

Includes taking quotes out of context, putting misleading labels on stories, or spinning stories to seem more negative than they are.

Information distorted to the point of paranoia, where all information is twisted to fit with the ideology or instead is perceived as a challenge or slight.

Misrepresenting consequences: Extent to which the website distorts the relationship between the actions they advocate and the effects the actions cause 1

2

3

4

5

Website does not attempt to reframe or misrepresent consequences

Website somewhat attempts to reframe or misrepresent consequences

Website constantly attempts to reframe or misrepresent consequences

Website discusses consequences associated with group events accurately

Website may attempt to reframe some aspects of consequences associated with group events to make them seem more positive or less negative

Website pervasively misrepresents consequences associated with group events in order to minimize the harm caused or exaggerate positive outcomes (e.g., ‘‘We are liberating animals not destroying property’’)

Ó 2015 Hogrefe Publishing

Journal of Media Psychology 2016; Vol. 28(1):16–31

Original Article

Violent Lyrics = Aggressive Listeners?

Effects of Song Lyrics and Tempo on Cognition, Affect, and Self-Reported Arousal Stephanie Pieschl and Simon Fegers Institut für Psychologie, Westfälische Wilhelms-Universität Münster, Gemany Abstract. Research on music has had an impressive impact. For example, the semantic content of lyrics seems to cause associated short-term effects regarding cognition and affect. However, we argue that these effects might have been confounded by other musical parameters related to time, pitch, texture, or voice of the selected songs. This study overcame this methodological problem by using different versions of an experimentally manipulated song. In a 2 · 2 between-subjects design, 120 university students listened to four versions of a song with violent or prosocial lyrics presented in slow or fast tempo. As predicted by theories of priming, violent lyrics increased aggressive cognitions (word completion test) and aggressive affect (self-reported state anger) in comparison with prosocial lyrics. However, the reverse effects of prosocial lyrics on prosocial cognitions and prosocial affect could not be confirmed. Finally, the tempo of the song did not consistently increase selfreported arousal, and we did not find more extreme effects under conditions of fast tempo as predicted by the arousal-extremity model.

Keywords: music, aggression, cognition, affect, arousal

Nowadays media use is ubiquitous and allows for many benefits such as worldwide communication. However, media are also criticized regarding potential negative effects such as aggression or obesity. Empirically, a long research tradition has investigated effects of violent media – for example, of comic books, television, or video games.

Research regarding music is comparatively rare even though music is one of the most frequently used and most important media channels of adolescents (Medienpädagogischer Forschungsverbund Südwest, 2012). We present an empirical study that extended the current methodological repertoire of music research.

General Learning Model and Other Theories About the Effects of Music Historically, the general learning model (GLM; Buckley & Anderson, 2006) has been the default framework of media research. It integrates many theories to model positive and negative media effects. However, recently it has been criticized – for example, because of its causal assumption of media effects while ignoring alternative selection or catalyst explanations and because of its (over-)focus on (social-)cognitive variables instead of real-life behavior (Elson & Ferguson, 2014). We acknowledge these points Journal of Media Psychology 2016; Vol. 28(1):32–41 DOI: 10.1027/1864-1105/a000144

but nonetheless utilized the GLM for deducing testable hypotheses and selecting variables. The GLM is well-suited for these endeavors because we tested causal effects and focused exclusively on social-cognitive antecedents of aggression.

The GLM elaborates how media stimuli are processed on three levels that constitute a continuous loop. This study, starting from this description, focused exclusively on shortterm effects regarding the first two levels; the last behavioral (output) level was ignored. On the first input level, the starting situation is defined by persons in situations.

Personal factors include all characteristics that a person brings to a situation – for example, personality traits or genetic predispositions; situational factors include relevant features of a situation such as provocations or aggressive cues. In our study, musical parameters such as lyrics or tempo constituted situational variables, while aggressive or prosocial personality traits or preferences for different types of music constituted personal variables. On the second routes level, the listener’s internal state determines how this input information is processed via three interacting antecedents of aggression – namely, cognition, affect, and arousal.

The GLM explains the effects of media on cognition and affect based on the cognitive neoassociation model (Berkowitz, 1984): The human brain is assumed to resemble an associative network where knowledge is represented 2015 Hogrefe Publishing

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

in nodes that are linked by associations. These associations are built as a function of contiguity and semantic similarity.

If one node is activated, spreading activation causes simultaneous activation of associated nodes. Berkowitz (1990) proposed that not only are cognitions linked in such semantic networks but also emotions. One central principle of this model is priming (Huesmann & Taylor, 2006): External stimuli can activate specific nodes within the network which facilitate the activation of associated nodes.

Meta-analyses showed that aggressive primes were able to successfully activate aggressive cognitions (RoskosEwoldsen, Roskos-Ewoldsen, & Carpentier, 2002). According to the cognitive neoassociation model (Berkowitz, 1984) and the GLM (Buckley & Anderson, 2006), the semantic content of music can be considered as primes (Knobloch-Westerwick, Musto, & Shaw, 2008). Thus, the GLM predicts that violent or prosocial lyrics can cause violent or prosocial cognitions and affect by making associated cognitions and emotions more accessible by priming.

Arousal is assumed to be the third important facet of internal states in the GLM. Music effects can be predicted based on the excitation transfer theory (Zillmann & Bryant, 1974) and the arousal-extremity model (Hansen & Krygowski, 1994): Many concurrent sources can influence autonomic arousal, and cognitive adaptations to changing situations are faster than adaptations in physiological arousal. Therefore, residual arousal from multiple sources is additive and can transfer to later situations. Such enhanced arousal can be misinterpreted if attributed exclusively to the current situation, and consequently emotions, cognitions, and behavior might be artificially enhanced. For example, arousal was associated with more aggressive reactions to perceived provocations (Zillmann & Bryant, 1974) or more pronounced prosocial behavior (Mueller & Donnerstein, 1981). Thus, arousal determines the intensity of a reaction but not its direction. Arousal is also relevant for priming because more extreme schemas are predominantly associated with higher levels of arousal. For example, under conditions of high arousal, sexy videos primed the perception of a neutral video stronger than under conditions of low arousal (Hansen & Krygowski, 1994). Therefore, these theories predict that cognitive and affective priming effects should be more extreme under conditions of high arousal.

One musical parameter that is consistently associated with arousal is music tempo (see subsequent section).

Selective Review of Research Into the Effects of Music Research into the effects of music can be categorized into three traditions. Each tradition has advantages and disadvantages due to methodological trade-offs between ecological validity and experimental control.

Research within the first tradition investigates long-term effects of preferences for, and listening to, specific music genres. Results indicate, for example, that preferences for heavy metal and rap were correlated with behavioral problems (Allen et al., 2007; Hansen, 1995), low academic performance (Allen et al., 2007; Took & Weiss, 1994), 2015 Hogrefe Publishing

33

hostility (Rubin, West, & Mitchell, 2001), and verbal aggression (Atkin, Smith, Roberto, Fediuk, & Wagner, 2002). These studies have high ecological validity but do not allow for conclusions about causality because of their cross-sectional designs; effects may be due to self-selection or third variables (Allen et al., 2007).

These problems are addressed by the second tradition of experimental studies that have scrutinized the short-term effects of listening to different kinds of popular songs: Participants listen to songs – for example, with violent or neutral lyrics – and afterwards their cognitions, affect, arousal, or behaviors are assessed. Overall these studies show stronger effects than correlational studies but are less ecologically valid (Allen et al., 2007).

The accessibility of cognitions is measured by similarity judgments (Anderson, Carnagey, & Eubanks, 2003; Greitemeyer, 2011), lexical decision tasks (Anderson et al., 2003; Krahé & Bieneck, 2012), free associations (Fischer & Greitemeyer, 2006), word completion tasks (Anderson et al., 2003; Fischer & Greitemeyer, 2006; Greitemeyer, 2009b, 2011), or attitude measures (Greitemeyer, Hollingdale, & Traut-Mattausch, 2012; Sprankle, End, & Bretz, 2012). The majority of results are in line with the predictions of the GLM. Songs with violent (vs. neutral) lyrics increased aggressive cognitions (Anderson et al., 2003; Fischer & Greitemeyer, 2006), and songs with prosocial (vs. neutral) lyrics decreased aggressive cognitions (Greitemeyer, 2009a, 2009b, 2011); these effects persisted when controlling for trait aggression (Greitemeyer, 2011).

However, a few inconsistent results indicate that prosocial songs did not consistently change attitudes toward violence (Greitemeyer, 2011) or punishment (Greitemeyer et al., 2012).

Affect is measured by questionnaires about participants’ emotional states (Fischer & Greitemeyer, 2006; Greitemeyer, 2011), mood (Greitemeyer et al., 2012; Krahé & Bieneck, 2012), anger, anxiety (Ballard & Coates, 1995), or hostility (Anderson et al., 2003; Greitemeyer, 2011; Wanamaker & Reznikoff, 1989), or by confronting them with scenarios and measuring their empathy (Greitemeyer, 2009a, 2009b; Greitemeyer et al., 2012). The majority of results are in line with the predictions of the GLM. Violent, misogynic, or man-hating (vs. neutral) lyrics increased anger and overall aggressive affect, and decreased positive emotions (Anderson et al., 2003; Fischer & Greitemeyer, 2006; Greitemeyer, 2011); and prosocial or proequity (vs. neutral) lyrics increased affective empathy or good mood (Greitemeyer, 2009a, 2009b; Greitemeyer et al., 2012). Other results were less consistent: Even pleasant (vs. aversive) music without semantic content (lyrics) elicited differential effects regarding mood (Krahé & Bieneck, 2012); Ballard and Coates (1995) found no consistent effects of nonviolent, homicidal, and suicidal lyrics on anger, depression, anxiety, or suicidal tendencies; Greitemeyer (2011) found no effects of prosocial versus neutral lyrics on positive and negative emotions; Greitemeyer et al. (2012) in two out of three studies found no effects of proequity (vs. neutral) lyrics on mood; and Wanamaker and Reznikoff (1989) found no effects of nonaggressive versus aggressive music on hostility.

Journal of Media Psychology 2016; Vol. 28(1):32–41

34

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

The results of lyrics on arousal are even more mixed: Songs with violent, prosocial, and proequity lyrics did not elicit significantly different arousal from neutral songs (Anderson et al., 2003; Fischer & Greitemeyer, 2006; Greitemeyer, 2011; Greitemeyer et al., 2012). However, in one study, proequity lyrics elicited more arousal than neutral songs (Greitemeyer et al., 2012).

These studies also show that music has significant implications for behavior: Aggressive, misogynic, and man-hating music increased (sexually) aggressive behavior (Barongan & Hall, 1995; Fischer & Greitemeyer, 2006; Mast & McAndrew, 2011), and listening to prosocial or proequity music increased prosocial behavior – for example, donations (Greitemeyer, 2009a, 2009b), spontaneous helping behavior (Greitemeyer, 2009a), positive job-related evaluations (Greitemeyer et al., 2012), or tips in a restaurant (Jacob, Guéguen, & Boulbry, 2010) – and decreased aggressive behavior – for example, within the chili sauce paradigm (Greitemeyer, 2011). Mediation analyses imply that (prosocial) behavior might be mediated by the affective route rather than by the cognitive or arousal routes (Greitemeyer, 2009a, 2011).

Even though experimental research from this tradition shows impressive results for the short-term effects of music, it has a marked shortcoming because it uses existing popular songs. Thus, detected effects cannot be attributed to specific musical parameters such as lyrics but could be confounded by additional variables: Popular songs intrinsically differ in time (including speed/tempo, meter, rhythm, rhythmic articulation such as staccato vs. legato, accentuation, and duration), pitch (including tonality/key/mode such as major vs. minor, melody, and melodic direction), and texture (including harmony, harmonic complexity, timbre, percussiveness, orchestration, and genre) (Collier & Hubbard, 2001; Gomez & Danuser, 2007; Kellaris & Kent, 1993; Webster & Weir, 2005). For vocal music, additional variables such as tone and range of voice might play a role.

Furthermore, songs are differentially known and liked and can be played at different sound levels. These variables could also be responsible for the detected effects.

Researchers have tried to solve these problems by selecting songs of comparable lengths, genres, and artists for all experimental conditions or by running pilot studies ensuring matched songs regarding arousal and mood stimulating properties and regarding perceptions as neutral, violent, prosocial, or proequity (Greitemeyer, 2009b, 2011; Greitemeyer et al., 2012; Mast & McAndrew, 2011; Sprankle et al., 2012). Empirical results seem to point to lyrics as the driving force behind music effects: Detected effects are content-specific. Proequity lyrics increased positive attitudes toward women (Greitemeyer et al., 2012); misogynic lyrics increased negative cognitions of men toward women, while the reverse is true for man-hating lyrics and women (Fischer & Greitemeyer, 2006); and perceived prosocial content significantly predicted a reduction in aggressive behavior (Greitemeyer, 2011). Even though these efforts are noteworthy, the only way to unequivocally link the effects of music to specific music parameters is to use experimentally manipulated songs that vary only regarding the variable(s) of interest.

Journal of Media Psychology 2016; Vol. 28(1):32–41

A third tradition uses such experimental manipulations of musical parameters. However, these studies predominantly focus on parameters other than lyrics. For example, the tempo of music (in beats per minute [BPM]) can be manipulated, and associated emotional responses and arousal can be measured. Results show consistently that faster tempo is positively associated with arousal (Holbrook & Anand, 1990) whether measured objectively (galvanic skin response, heart rate, respiration rate, or blood pressure) or by self-reports (Gomez & Danuser, 2007; Hodges, 2009; Husain, Thompson, & Schellenberg, 2002). However, in some studies, tempo effects were enhanced for popular music in comparison with classical music (Kellaris & Kent, 1993), while in others they were diminished (Carpentier & Potter, 2007). Furthermore, fast tempo was also consistently associated with emotional responses such as happiness or brightness (Collier & Hubbard, 2001; Webster & Weir, 2005), pleasure (Kellaris & Kent, 1993), or overall emotional valance (Gomez & Danuser, 2007). There seem to be optimal tempo preferences that are associated with most positive emotions (Holbrook & Anand, 1990).

Up to now, song lyrics have not been manipulated, even though the first steps in this direction have been taken: Studies have compared the effects of instrumental-only and instrumental-plus-lyrics versions of songs. Wester, Crown, Quatman, and Heesacker (1997) found that gangsta rap lyrics increased adversarial sexual beliefs, and Brummert Lennings and Warburton (2011) found significant effects of violent lyrics on aggressive behavior (hot sauce paradigm). However, Wester et al. (1997) found no significant effects of gangsta rap lyrics on attitudes toward women, sexual conservatism, or sex role stereotyping; and Sprankle et al. (2012) found no significant effects of misogynic lyrics on attitudes or behavior. Thus, the results are inconclusive, and even in these experiments, variables other than lyrics might have contributed to these effects. The ultimate test for lyrics effects would be to create songs with different semantic content.

Study Rationale and Hypotheses This study constitutes the first exploration of the feasibility of such an approach. To minimize effects of personal preferences and genre, we utilized an unknown song of a genre not usually associated with violence or prosociality, and experimentally manipulated it according to a 2 · 2 design: Four versions of the song contained either violent or prosocial lyrics and were played in either slow or fast tempo.

To diagnose listeners’ internal states, we captured their (aggressive and prosocial) cognitions with word completion tasks and their affect (aggressive and prosocial) and arousal (alertness and agitation) with self-report questionnaires.

We investigated the following main hypotheses: Hypothesis 1 (H1): Listening to songs with violent lyrics will increase aggressive cognitions and aggressive affect (compared with listening to prosocial lyrics). We assumed that these lyrics main effects would be caused by aggressive priming.

2015 Hogrefe Publishing

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

Hypothesis 2 (H2): Listening to songs with prosocial lyrics will increase prosocial cognitions and prosocial affect (compared with listening to violent lyrics).

We assumed that these lyrics main effects would be caused by prosocial priming.

Hypothesis 3 (H3): All lyrics main effects (H1 and H2) will be stronger in the fast tempo conditions than in the slow tempo conditions. We assumed that these tempo-by-lyrics interactions would be caused by arousal-extremity. A prerequisite for this hypothesis was that fast tempo increases self-reported arousal (alertness and agitation).

We captured additional potential covariates pertaining to trait aggressiveness, trait prosociality, and music experiences and preferences. Furthermore, to test if our main hypotheses were specific (e.g., prosocial lyrics increased prosocial affect) we also measured ‘‘good mood.’’ We predicted that none of the experimental manipulations would significantly impact this more general measure of affect.

Method Sample Ninety-six female students and 24 male students (N = 120) aged between 19 and 41 years (M = 22.64, SD = 4.02) were recruited for this study. The majority (88%) were studying psychology, and the majority were in their first semester (75%). None of these students knew the experimental material beforehand, and all of them reported afterward that they were able to hear the song and lyrics normally. A chi-square test showed that gender distributions did not differ significantly between experimental conditions, X2(3) = .417, p = .937 (violent slow: 6 men, 24 women; violent fast: 5 men, 25 women; prosocial slow: 6 men, 24 women; prosocial fast: 7 men, 23 women). An ANOVA regarding age showed no significant main effects of the experimental factors lyrics, F(1, 116) = .245, p = .621, or tempo, F(1, 116) = .535, p = .466, but a significant interaction, F(1, 116) = 11.703, p = .001. Tukey post hoc contrasts indicated that the prosocial fast sample was significantly younger (M = 21.00, SD = 2.05) than the prosocial slow (M = 23.93, SD = 4.80) and the violent fast sample (M = 23.77, SD = 4.48); the violent slow (M = 21.87, SD = 3.55) sample did not differ significantly in age from any other sample. Nonetheless, we did not include age as a covariate in our main analyses, because this variable did not explain incremental variance beyond the experimental manipulations.

Design and Procedure Students were randomly assigned to one of the four experimental conditions of the between-subjects 2 · 2 design. In all conditions (each with n = 30), they had to 2015 Hogrefe Publishing

35

listen to one version of the experimentally manipulated song that differed according to lyrics (violent vs. prosocial) and tempo (slow vs. fast). Subsequently, their cognitions, affect, and arousal were measured via self-reports. Afterward, the following variables were captured: students’ tastes and prior experiences with music, trait aggressiveness, and trait agreeableness. Data are collected from up to 10 students at once, but students worked in individual cubicles and listened to the song individually with headphones on.

Each student received 30 min of credit as a trial subject.

Material Experimental Versions of the Song The German song Wilbur by the band Montague was selected as experimental material. This band as well as this song is unknown to the general public, but their indie music style is popular among students in general. Two new German lyrics versions were created, and both versions were re-recorded for this study; the singer from Montague sang both versions over the same instrumental soundtrack with the same melody and pronunciation. Afterward both versions were mixed to be played in two tempi.

More specifically, regarding the lyrics manipulation, the song described a person who wants to kill (violent version) or help (prosocial version) an old man. The text was exactly the same in both conditions except for 47 words that were interchanged – namely, eight adjectives, 13 verbs, and 26 nouns. These manipulated words were roughly matched regarding their length, frequency, valence, and associated arousal according to German norms (Grühn & Smith, 2008; Schwibbe, Räder, Schwibbe, Borchardt, & GeikenPophanken, 1994), ensuring that effects could not be attributed to differences in these variables. Here are some examples of matched words: ‘‘mean’’ and ‘‘kind’’ (German: ‘‘fies’’ vs. ‘‘nett’’), ‘‘cruel’’ and ‘‘caring’’ (German: ‘‘grausam’’ vs. ‘‘sozial’’), ‘‘threaten’’ and ‘‘help’’ (German: ‘‘drohen’’ vs. ‘‘helfen’’), ‘‘hate’’ and ‘‘love’’ (German: ‘‘hassen’’ vs. ‘‘lieben’’), and ‘‘destruction’’ and ‘‘trust’’ (German: ‘‘Zerstörung’’ vs. ‘‘Vertrauen’’).

Regarding the tempo manipulation, the software AtomixMP3 (version 1.1; Atomix Productions Inc., http:// atomix-productions.software.informer.com/) was used to mix both the violent and the prosocial versions of the song in two tempi: slow at 66.4 BPM resulting in a length of 4.18 min, and fast at 101.2 BPM resulting in a length of 2.50 min. We selected these specific tempi to achieve maximum tempo differences while simultaneously ensuring that both versions were believable, likable, and understandable (see Husain et al., 2002).

Cognitions: Word Completion Task Students’ cognitions were captured with a newly constructed Word Completion Task (WCT) similar to tasks used in previous studies (Anderson et al., 2003; Fischer & Greitemeyer, 2006). This WCT consisted of 32 word Journal of Media Psychology 2016; Vol. 28(1):32–41

36

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

stems that each could be completed by adding one letter.

Thirteen word stems could be completed into violent versus neutral words (four verbs, nine nouns), 13 word stems into prosocial versus neutral words (four verbs, nine nouns).

A further six items could only be completed into neutral words, and these served as distractors. For all experimental items, the word frequency of the neutral versus aggressive or neutral versus prosocial completions was similar in the German language (http://wortschatz.uni-leipzig.de/), ensuring that mere accessibility did not determine answers. For eight of the experimental aggressive or prosocial word stems, only one neutral and one experimental completion were possible; for the remaining five items more neutral completions were possible. Examples of the aggressive cognition items are ‘‘dr_hen’’ which could be completed into ‘‘drohen’’ (Engl. ‘‘to threaten’’) or ‘‘drehen’’ (Engl. ‘‘to spin’’), and ‘‘To_’’ which could be completed into ‘‘Tod’’ (Engl. ‘‘death’’) or ‘‘Ton’’ (Engl. ‘‘sound’’). Examples of the prosocial cognition items are ‘‘sch_tzen,’’ which could be completed into ‘‘schützen’’ (Engl. ‘‘to protect’’) or ‘‘schätzen’’ (Engl. ‘‘to estimate’’), and ‘‘_rost’’ which could be completed into ‘‘Trost’’ (Engl. ‘‘consolation’’) or ‘‘Frost’’ (Engl. ‘‘freeze’’). The average numbers of aggressive (WCT-A) and prosocial (WCT-P) word completions was computed as indices of cognition. Both scores range from 0 (no aggressive/prosocial completions, respectively) to 1 (all words were completed aggressively/prosocially, respectively). Thus, these scores can easily be transformed to percentages. For example, a WCT-A score of .37 corresponds to 37% completed aggressively.

adjectives (e.g., perky and unhappy). High values indicate positive mood (MDBF-GB: Cronbach’s a = .87).

Aggressiveness and Agreeableness Trait Questionnaires A translated version of the Buss Perry Aggression Questionnaire (Buss & Perry, 1992) with the subscales physical aggression (eight items; e.g., ‘‘Given enough provocation, I may hit another person’’), verbal aggression (five items; e.g., ‘‘I often find myself disagreeing with people’’), anger (six items; e.g., ‘‘I have trouble controlling my temper’’), and hostility (eight items; e.g., ‘‘I am suspicious of overly friendly strangers’’) was administered. Students indicated their agreement with these statements on 5-point scales (1 = extremely uncharacteristic of me, to 5 = extremely characteristic of me). The reliability of these subscales was only partly satisfactory (Cronbach’s a = .49 to .80).

Thus, the average score across all 27 items was used for further analyses; high values indicate high aggressiveness (trait-A: Cronbach’s a = .84).

As a proxy of a prosocial personality, the factor agreeableness from the new Five Factor Inventory (Costa & McCrae, 1985) was used. This scale consists of 12 items (e.g., ‘‘I try to be friendly with everyone I meet’’). To each of these items, students indicated their agreement on 5-point scales (1 = strong disagreement, to 5 = strong agreement). The average score was computed; a high value indicates high agreeableness (trait-P: Cronbach’s a = .68).

Affect and Arousal State Questionnaires Multiple facets of affect and arousal were measured via self-reports. In all cases, adjectives were used, and students had to indicate how well these adjectives described their current state on 5-point scales (1 = not at all, to 5 = very much). For each scale, average scores were computed.

Aggressive affect was captured by five items of the German version of the State Anger Expression Inventory (Schwenkmezger, Hodapp, & Spielberger, 1992). For these adjectives (e.g., angry and mad), high values indicate high anger (affect-A). We developed a further five adjectives to capture prosocial affect (e.g., friendly and sociable); high values indicate high prosocial affect (affect-P). Both scales were highly reliable in this sample: Cronbach’s a = .87 for affect-A and Cronbach’s a = .82 for affect-P. Two scales from a German internal state questionnaire (Mehrdimensionaler Befindlichkeitsfragebogen [MDBF]; Steyer, Schwenkmezger, Notz, & Eid, 1997) were used as indicators of self-reported arousal – namely, ‘‘alertness versus fatigue’’ (AF) with eight adjectives (e.g., rested or alert) and ‘‘calm versus agitation’’ (CA) with eight adjectives (e.g., calm and nervous). High values represent high alertness (MDBF-AF: Cronbach’s a = .91) and high levels of calmness (MDBF-CA: Cronbach’s a = .88). This questionnaire also includes a third general affect subscale that captures ‘‘good versus bad mood’’ (GB) with eight

Journal of Media Psychology 2016; Vol. 28(1):32–41

Music Questionnaire Two questions captured students’ musical experience – namely, ‘‘Do you play a musical instrument (if yes, which instrument(s) and how long did you play actively)?’’ and ‘‘How much time per day do you spend on average listening to music (answers: 1 = 0–1 hours to 5 = 5–6 hours)?’’ Furthermore, we translated and adapted existing questionnaires about students’ musical tastes and preferences: Students indicated their preferences toward the music genres of rock, classic, electro, soul/blues, pop, country, folk, religious, and soundtrack on 5-point scales from 1 = not at all to 5 = very (Litle & Zuckerman, 1986); we added the category of ‘‘indie’’ genre to this scale. Additionally, they indicated their preferences toward 12 characteristics of music on 5-point scales from 1 = not at all to 5 = very; examples are ‘‘romantic and dreamy,’’ ‘‘serious and reflective,’’ or ‘‘wild and violent’’ (Schwartz & Fouts, 2003); we deleted the item ‘‘played at fast tempo’’ from this scale. Further questions checked potential reasons for exclusion and captured participants’ liking the song: ‘‘Do you have impaired hearing?’’ ‘‘Do you know the band Montague?’’ ‘‘Did you know the song just presented?’’ ‘‘How much did you like the song just presented (liking the song; answers: 1 = not at all to 5 = very much)?’’

2015 Hogrefe Publishing

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

37

Table 1. Correlations between Measures of Aggressive and Prosocial Cognition and State and Trait Affect, Arousal, and Liking the Song WCT-A

WCT-P

affect-A

affect-P

WCT-A WCT-P affect-A affect-P MDBF-GB MDBF-AF MDBF-CA trait-A trait-P Liking the song

– .170 .004 .077 .118 .024 .134 .056 .086 .017

MDBF-GB MDBF-AF MDBF-CA

– .086 .056 .046 .018 .014 .095 .033 .017

– .541*** .657*** .304** .617*** .031 .043 .147

– .657*** .385*** .567*** .153 .267** .351***

– .524*** .765*** .062 .031 .201*

– .397*** .015 .074 .128

M (SD)

.50 (.15)

.51 (.12)

1.57 (.70)

3.58 (.67)

3.91 (.65)

3.22 (.83)

– .062 .028 .155 3.64 (.73)

trait-A

trait-P

– .642*** .021

– .119

2.15 (.42) 3.91 (.40)

Liking the song

– 2.55 (1.05)

Notes. WCT = Word Completion Task; MDBF = Mehrdimensionaler Befindlichkeitsfragebogen (German internal state questionnaire); -A = aggressive; -P = prosocial; -GB = good versus bad mood; -AF = alertness versus fatigue; -CA = calm versus agitation; WCT scores range from 0 = none of the words was completed aggressive or prosocial, respectively, to 1 = all of the words were completed aggressive/prosocial, respectively; scores of all other variables range from 1 to 5 with high scores indicating state anger or prosocial affect or good mood or alertness or calm or trait anger or trait agreeableness or liking the song, respectively. *p < .01, **p < .05, ***p < .001.

Results Descriptive Results Most students (67%) played musical instruments, most frequently piano (48%) or guitar (28%). On average, students spent 2.50 (SD = 1.07) hours a day listening to music. They preferred the music genres rock (M = 3.98, SD = 1.09) and soundtrack (M = 3.55, SD = 1.03) most strongly, while they preferred religious music least (M = 1.63, SD = 0.96).

They preferred the music characteristics ‘‘innocent and peaceful’’ (M = 3.90, SD = 0.86) and ‘‘romantic and dreamy’’ (M = 3.82, SD = 0.97) most, while they preferred ‘‘wild and violent’’ music (M = 2.13, SD = 1.19) least.

They moderately preferred indie music (M = 3.35, SD = 1.34).

Regarding the dependent variables (see Table 1), on average, participants completed approximately half of the words aggressively (WCT-A) and prosocially (WCT-P), while the other half was completed neutrally. On average, students self-reported an internal state characterized by good mood, calm, prosocial affect, and alertness, while they self-reported little aggressive affect. Aggressive and prosocial cognitions were not significantly interrelated or related to any other dependent variables (WCT-A and WCT-P).

All self-reported affective and arousal variables were significantly interrelated; all variables were negatively related to anger (affect-A) but positively interrelated (affect-P, MDBF-GB, MDBF-AF, and MDBF-CA).

Regarding potential covariates (see Table 1), on average, students self-reported high trait agreeableness (trait-P) and low trait anger (trait-A), and liked the experimentally manipulated song moderately well (liking song). We only included covariates in the main analyses that explained incremental variance beyond the experimental manipulations. To determine which covariates to include, we computed seven regressions, one for each dependent variable 2015 Hogrefe Publishing

(WCT-A, WCT-P, affect-A, affect-P, MDBF-GB, MDBFAF, and MDBF-CA). In each regression, we entered the experimental manipulations (lyrics and tempo) and their interaction in a first block and potential covariates (trait-A, trait-P, and liking song) in a second block. Only for affect-P, did the second block explain significant incremental variance, R2change = .153, Fchange(3, 113) = 7.082, pchange < .001; liking the song was the only significant predictor (b = .291**). Therefore, only this variable was included as a covariate in the corresponding ANCOVA for affect-P; for all other ANOVAs, no covariates were included (see subsequent section).

The Effects of Lyrics and Tempo on Internal States For hypotheses testing, ANOVAs (or ANCOVAs) were computed for each dependent variable separately (WCT-A, WCT-P, affect-A, affect-P, MDBF-GB, MDBFAF, and MDBF-CA) with lyrics (violent vs. prosocial) and tempo (slow vs. fast) as between-subject factors.

Cognitions The ANOVA for aggressive cognitions (WCT-A; see Figure 1, left) showed a significant effect of lyrics, F(1, 116) = 7.677, p = .007, g2p = .062, but no significant effect of tempo, F(1, 116) = .213, p = .646, and no significant interaction, F(1, 116) = .028, p = .867: Students who listened to violent lyrics completed significantly more words aggressively (M = 0.54, SD = 0.16) than those listening to prosocial lyrics (M = 0.46, SD = 0.12).

The ANOVA for prosocial cognitions (WCT-P) showed no significant effects of lyrics, F(1, 116) = .046, Journal of Media Psychology 2016; Vol. 28(1):32–41

38

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

Figure 1. Aggressive cognition (left panel, Word Completion Task for agression [WCT-A]) and aggressive affect (right panel, affect-A) by lyrics (x-axis: violent vs. prosocial) and tempo (bars: slow vs. fast); main effects of lyrics were significant. p = .831, tempo, F(1, 116) = 2.976, p = .087, or their interaction, F(1, 116) = 1.945, p = .166.

significant interaction, F(1, 116) = .760, p = .385: Students listening to prosocial lyrics (M = 3.79, SD = 0.72) self-reported being significantly calmer than students listening to violent lyrics (M = 3.49, SD = 0.71).

Affect The ANOVA for anger (affect-A; see Figure 1, right) showed a significant effect of lyrics, F(1, 116) = 4.697, p = .032, gp2 = .039, but no significant effect of tempo, F(1, 116) = .072, p = .789, and no significant interaction, F(1, 116) = .420, p = .518: Students who listened to violent lyrics reported significantly more anger (M = 1.71, SD = 0.73) than those listening to prosocial lyrics (M = 1.44, SD = 0.66). The ANCOVA for prosocial affect (affect-P) showed no significant effects of lyrics, F(1, 115) = 1.019, p = .315, tempo, F(1, 115) = .081, p = .777, or their interaction, F(1, 115) = .058, p = .811; however, the covariate liking the song elicited a significant main effect, F(1, 115) = 12.925, p < .001, gp2 = .101: The more students liked the song, the higher their reported prosocial affect. The experimental manipulations also had no significant effects on more general good mood (MDBFGB), namely we found no significant effects of lyrics, F(1, 116) = 1.149, p = .286; or tempo, F(1, 116) = .009, p = .923; and no significant lyrics-by-tempo interaction, F(1, 116) = .020, p = .887.

Self-Reported Arousal The ANOVA for alertness (MDBF-AF) showed a significant effect of tempo, F(1, 116) = 4.417, p = .038, g2p = .037, but no significant effect of lyrics, F(1, 116) = .735, p = .393, and no significant interaction, F(1, 116) = .093, p = .762: Students listening to the fast song self-reported significantly higher alertness (M = 3.38, SD = 0.82) than students listening to the slow song (M = 3.06, SD = 0.83). The ANOVA for agitation (MDBF-CA) showed a significant effect of lyrics, F(1, 116) = 5.385, p = .022, g2p = .044, but no significant effect of tempo, F(1, 116) = .030, p = .862, and no Journal of Media Psychology 2016; Vol. 28(1):32â&#x20AC;&#x201C;41

Discussion Summary and Discussion of Results Hypothesis 1 about aggressive priming was confirmed: Participants listening to songs with violent lyrics had significantly higher scores on the Word Completion Task for aggressive cognitions (WCT-A) and state anger (affect-A) than students listening to songs with prosocial lyrics. In line with the GLM (Buckley & Anderson, 2006), these effects can be explained by the psychological mechanism of priming that predicts that violent content of lyrics makes associated aggressive cognitions and affect more accessible (Berkowitz, 1984, 1990; Knobloch-Westerwick et al., 2008; Roskos-Ewoldsen, et al., 2002). Furthermore, these results are consistent with the majority of previous findings from correlational (Atkin et al., 2002) and experimental research traditions (Anderson et al., 2003; Fischer & Greitemeyer, 2006; Greitemeyer, 2011). Please note that this does not imply that violent lyrics caused highly aggressive cognitions or affect; these effects were only notable in comparison with prosocial lyrics conditions. On an absolute level, the resulting aggression was low: Violent lyrics triggered only moderate levels of aggressive cognitions (WCT-A) and relatively low state anger (affect-A scores were below the scale midpoint).

Hypothesis 2 about prosocial priming was refuted: Students listening to songs with violent or prosocial lyrics did not differ significantly in their scores on the Word Completion Task for prosocial cognitions (WCT-P) or in their prosocial affect (affect-P). Prosocial affect (affect-P) seemed more a matter of preference (liking the song). These findings contradict the GLM and the assumption of priming (see section General Learning Model and Other Theories About the Effects of Music). However, previous research 2015 Hogrefe Publishing

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

about the effects of prosocial lyrics is also partly inconsistent: Some experiments found significant positive effects on cognitions, affect, and behavior (Greitemeyer, 2009a, 2009b, 2011; Greitemeyer et al., 2012), while others detected inconsistent or nonexistent effects (Ballard & Coates, 1995; Greitemeyer, 2011; Greitemeyer et al., 2012; Wanamaker & Reznikoff, 1989).

Furthermore, the priming effects regarding violent and prosocial lyrics (see above) were specific and did not transfer to more general affective variables such as good versus bad mood (MDBF-GB); we found no significant effects of lyrics or tempo regarding this variable.

Potential explanations for the divergent pattern of results regarding violent and prosocial lyrics concern participants’ dominant states, traits, and preferences: This sample of predominantly female psychology students can be characterized by low aggression, high prosociality, and low preference for violent music. Therefore, listening to violent lyrics should have contradicted these dominant internal states substantially, and thus the violent semantic content should have been highly salient. On the other hand, listening to prosocial lyrics might have been in line with participants’ dominant internal states and might have been less salient.

Consequently, prosocial lyrics might not have triggered additional prosocial cognitions and affect, and violent lyrics might not have interfered with participants’ internal state enough to significantly lower their preexisting prosocial state. Further research with participants with other traits or music preferences (e.g., heavy metal) could explore this explanation.

Hypothesis 3 regarding arousal-extremity effects could not be tested conclusively: As prerequisite for testing this hypothesis we expected main effects of tempo on selfreported arousal. Faster tempo significantly increased selfreported alertness in comparison with slower tempo (MDBF-AF), but tempo did not significantly impact selfreported agitation (MDBF-CA) – instead students reported being significantly calmer for prosocial songs than for violent songs. The results for MDBF-AF are consistent with previous research (Gomez & Danuser, 2007; Hodges, 2009; Holbrook & Anand, 1990; Husain et al., 2002; Kellaris & Kent, 1993; van der Zwaag, Westerink, & van den Broek, 2011), but the results for MDBF-CA remain an enigma. Because the prerequisites for testing H3 were only partially fulfilled, it is not surprising that we also did not find the predicted tempo-by-lyrics interactions in the analyses regarding cognitions and affect (H1 and H2). Such interactions would have been consistent with the arousalextremity model, excitation transfer theory, and the GLM (Buckley & Anderson, 2006; Hansen & Krygowski, 1994; Zillmann & Bryant, 1974). However, previous research also did not find arousal-extremity effects consistently. For example, physiological arousal did not mediate the effects of music on cognitions or on behavior (Greitemeyer, 2011; Greitemeyer et al., 2012). Further research into this issue is necessary – for example, with more direct stimulations of arousal by physical exercise (see Hansen & Krygowski, 1994) or other measures of arousal (see Limitations).

2015 Hogrefe Publishing

39

Limitations This study constitutes the first conclusive test of lyrics effects enabled by systematic experimental manipulation of the semantic content of music (violent vs. prosocial lyrics). Because it was also the first exploration of the feasibility of this approach, we implemented a one-stimulus design (one song was manipulated) and utilized self-report instruments for the variables of interest (cognition, affect, and arousal). This approach has some inherent (methodological) limitations: On a general level, we maximized experimental control at the cost of ecological validity and generalizability. We used only one experimental song which is not from the Top 50 charts and not of a genre usually associated with violence; we have no information whether the detected effects transfer to other songs or genres, or not. Our convenience sample of predominantly female psychology students is not representative of the general population, and especially not representative of adolescents most at risk of negative music effects. These effects might not transfer to a population of heavy metal fans for whom violent lyrics might constitute the norm. Additionally, experimental manipulations of songs have inherent drawbacks: By manipulating tempo one simultaneously manipulates the length of the songs. Consequently, these different doses of exposure might impact cognitions, affect, or arousal. By manipulating lyrics, one might inadvertently create content–tempo mismatches that might not be believable (e.g., slow song with highly violent lyrics). However, in this study, we found no empirical indicators of this problem; students liked the prosocial song significantly more than the violent song independent of tempo. Furthermore, it is necessary to replicate these effects with alternative measures of cognition, affect, and arousal. For example, repeated before–after measures of affect would enable the detection of minute changes in affect, and physiological measures of arousal might be more sensitive and enable the detection of the predicted arousal-extremity effects. Furthermore, our conclusions are explicitly limited to shortterm effects and to internal states. We have no data about long-term effects or behavioral impact. Further research is needed to address these limitations.

Implications Nonetheless, this study is the first conclusive test of the hypothesis that lyrics are in fact the driving force behind the previously detected effects of violent and prosocial music. The two versions of the experimental song with violent and prosocial lyrics differed only in 47 words that were interchanged. The results indicate that violent lyrics increased aggressive cognitions and aggressive affect.

However, the resulting cognitions and affect were not very aggressive, the effect sizes were only small to medium, and the reverse effect for prosocial lyrics could not be confirmed. Therefore, this study points to relatively small negative short-term effects of violent lyrics. On a

Journal of Media Psychology 2016; Vol. 28(1):32–41

40

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

methodological level, this study shows that it is possible to experimentally manipulate musical parameters such as lyrics. Thus, it opens the door for considering other musical parameters as relevant variables. Further research could concentrate on the incremental contributions of different musical parameters related to time, pitch, texture, voice, popularity, and genre. The real-life effects of music can most likely be attributed to the complex interplay between all of these variables and the resulting holistic impression: A specific song might leave an overall impression of aggression and violence not only based on its lyrics, but on associations with specific genres, staccato rhythms, high rhythmic accentuations, percussiveness, and specific melodies and harmonies (or lack thereof). Furthermore, most likely, the enculturation into subcultures associated with specific genres is at least as relevant for aggressive or prosocial behavior as the lyrics of specific songs.

Acknowledgments The data were collected as part of S.F.’s Diploma thesis requirements. We wish to thank the band Montague for letting us use their song Wilbur and for their help in recording the two versions of the experimental song.

References Allen, M., Herrett-Skjellum, J., Jorgenson, J., Kramer, M. R., Ryan, D. J., & Timmerman, L. (2007). Effects of Music. In R. W. Preiss, B. Gayle, N. Burrell, M. Allen, & J. Bryant (Eds.), Mass media effects research: Advances through meta-analysis (pp. 263–279). Mahwah, NJ: Erlbaum.

Anderson, C. A., Carnagey, N. L., & Eubanks, J. (2003).

Exposure to violent media: The effects of songs with violent lyrics on aggressive thoughts and feelings. Journal of Personality and Social Psychology, 84, 960–971. doi: 10.1037/0022-3514.84.5.960 Atkin, C. K., Smith, S. W., Roberto, A. J., Fediuk, T., & Wagner, T. (2002). Correlates of verbally aggressive communication in adolescents. Journal of Applied Communication Research, 30, 251–268. doi: 10.1080/00909880216585 Ballard, M. E., & Coates, S. (1995). The immediate effects of homicidal, suicidal, and nonviolent heavy metal and rap songs on the moods of college students. Youth & Society, 27(2), 148–168. doi: 10.1177/0044118X95027002003 Barongan, C., & Hall, G. (1995). The influence of misogynous rap music on sexual aggression against women. Psychology of Women Quarterly, 19, 195–207. doi: 10.1111/j.14716402.1995.tb00287.x Berkowitz, L. (1984). Some effects of thoughts on antiand prosocial influences of media events: A cognitiveneoassociation analysis. Psychological Bulletin, 95, 410–427. doi: 10.1037/0033-2909.95.3.410 Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive neoassociationistic analysis.

American Psychologist, 45, 494–503. doi: 10.1037/0003066X.45.4.494 Brummert Lennings, H. I., & Warburton, W. A. (2011). The effect of auditory versus visual violent media exposure on aggressive behaviour: The role of song lyrics, video clips and musical tone. Journal of Experimental Social Psychology, 47, 794–799. doi: 10.1016/j.jesp.2011.02.006 Journal of Media Psychology 2016; Vol. 28(1):32–41

Buckley, K. E., & Anderson, C. A. (2006). A theoretical model of the effects and consequences of playing video games. In P. Vorderer & J. Bryant (Eds.), Playing video games: Motives, responses, and consequences (pp. 363–378). Mahwah, NJ: Erlbaum.

Buss, A. H., & Perry, M. (1992). The Aggression Questionnaire.

Journal of Personality and Social Psychology, 63, 452–459. doi: 10.1037/0022-3514.63.3.452 Carpentier, F., & Potter, R. F. (2007). Effects of music on physiological arousal: Explorations into tempo and genre.

Media Psychology, 10, 339–363. doi: 10.1080/ 15213260701533045 Collier, W. G., & Hubbard, T. L. (2001). Judgments of happiness, brightness, speed and tempo change of auditory stimuli varying in pitch and tempo. Psychomusicology, 17, 36–55.

Costa, P. T., & McCrae, R. R. (1985). The NEO personality inventory manual. Odessa, FL: Psychological Assessment Resources.

Elson, M., & Ferguson, C. J. (2014). Twenty-five years of research on violence in digital games and aggression: Empirical evidence, perspectives, and a debate gone astray.

European Psychologist, 19(1), 33–46. doi: 10.1027/10169040/a000147 Fischer, P., & Greitemeyer, T. (2006). Music and aggression: The impact of sexual-aggressive song lyrics on aggressionrelated thoughts, emotions, and behavior toward the same and the opposite sex. Personality and Social Psychology Bulletin, 32, 1165–1176. doi: 10.1177/0146167206288670 Gomez, P., & Danuser, B. (2007). Relationships between musical structure and psychophysiological measures of emotion. Emotion, 7, 377–387. doi: 10.1037/15283542.7.2.377 Greitemeyer, T. (2009a). Effects of songs with prosocial lyrics on prosocial behavior: Further evidence and a mediating mechanism. Personality and Social Psychology Bulletin, 35, 1500–1511. doi: 10.1177/0146167209341648 Greitemeyer, T. (2009b). Effects of songs with prosocial lyrics on prosocial thoughts, affect, and behavior. Journal of Experimental Social Psychology, 45(1), 186–190. doi: 10.1016/j.jesp.2008.08.003 Greitemeyer, T. (2011). Exposure to music with prosocial lyrics reduces aggression: First evidence and test of the underlying mechanism. Journal of Experimental Social Psychology, 47(1), 28–36. doi: 10.1016/j.jesp.2010.08.005 Greitemeyer, T., Hollingdale, J., & Traut-Mattausch, E. (2015).

Changing the track in music and misogyny: Listening to music with pro-equality lyrics improves attitudes and behavior toward women. Psychology of Popular Media Culture, 4, 56–67. doi: 10.1037/a0030689 Grühn, D., & Smith, J. (2008). Characteristics for 200 words rated by young and older adults: Age-Dependent Evaluations of German Adjectives (AGE). Behavior Research Methods, 40, 1088–1097. doi: 10.3758/BRM.40.4.1088 Hansen, C. (1995). Predicting cognitive and behavioral effects of gangsta rap. Basic and Applied Social Psychology, 16(1–2), 43–52. doi: 10.1207/s15324834basp1601&2_3 Hansen, C., & Krygowski, W. (1994). Arousal-augmented priming effects: Rock music videos and sex object schemas.

Communication Research, 21(1), 24–47. doi: 10.1177/ 009365094021001003 Hodges, D. (2009). Bodily responses to music. In S. Hallam, I. Cross, & M. Thaut (Eds.), Oxford handbook of music psychology (pp. 121–130). Oxford: Oxford University Press.

Holbrook, M. B., & Anand, P. (1990). Effects of tempo and situational arousal on the listener’s perceptual and affective responses to music. Psychology of Music, 18(2), 150–162. doi: 10.1177/0305735690182004 Huesmann, L., & Taylor, L. D. (2006). The role of media violence in violent behavior. Annual Review of Public 2015 Hogrefe Publishing

S. Pieschl & S. Fegers: Violent Lyrics = Aggressive Listeners?

Health, 27, 393–415. doi: 10.1146/annurev.publhealth.

26.021304.144640 Husain, G., Thompson, W., & Schellenberg, E. (2002). Effects of musical tempo and mode on arousal, mood, and spatial abilities. Music Perception, 20(2), 151–171. doi: 10.1525/ mp.2002.20.2.151 Jacob, C., Guéguen, N., & Boulbry, G. (2010). Effects of songs with prosocial lyrics on tipping behavior in a restaurant.

International Journal of Hospitality Management, 29, 761–763. doi: 10.1016/j.ijhm.2010.02.004 Kellaris, J. J., & Kent, R. J. (1993). An exploratory investigation of responses elicited by music varying in tempo, tonality, and texture. Journal of Consumer Psychology, 2, 381–401. doi: 10.1016/S1057-7408(08)80068-X Knobloch-Westerwick, S., Musto, P., & Shaw, K. (2008).

Rebellion in the top music charts: Defiant messages in rap/hip-hop and rock music 1993 and 2003. Journal of Media Psychology: Theories, Methods, and Applications, 20(1), 15–23. doi: 10.1027/1864-1105.20.1.15 Krahé, B., & Bieneck, S. (2012). The effect of music-induced mood on aggressive affect, cognition, and behavior. Journal of Applied Social Psychology, 42, 271–290. doi: 10.1111/ j.1559-1816.2011.00887.x Litle, P., & Zuckerman, M. (1986). Sensation seeking and music preferences. Personality and Individual Differences, 7, 575–577. doi: 10.1016/0191-8869(86)90136-4 Mast, J. F., & McAndrew, F. T. (2011). Violent lyrics in heavy metal music can increase aggression in males. North American Journal of Psychology, 13(1), 63–64.

Medienpädagogischer Forschungsverbund Südwest. (2012).

JIM-Studie 2012. Jugend, Information, (Multi-) Media.

Basisuntersuchung zum Medienumgang 12- bis 19-Jähriger [JIM study 2012: Youth, information, (multi-)media: Basic survey about the media use of 12- to 19-year-olds]. Stuttgart: Medienpädagogischer Forschungsverbund Südwest.

Mueller, C. W., & Donnerstein, E. (1981). Film-facilitated arousal and prosocial behavior. Journal of Experimental Social Psychology, 17(1), 31–41. doi: 10.1016/00221031(81)90004-4 Roskos-Ewoldsen, D. R., Roskos-Ewoldsen, B., & Carpentier, F. (2002). Media priming: A synthesis. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (2nd ed., pp. 97–120). Mahwah, NJ: Erlbaum.

Rubin, A. M., West, D. V., & Mitchell, W. S. (2001). Differences in aggression, attitudes toward women, and distrust as reflected in popular music preferences. Media Psychology, 3(1), 25–42. doi: 10.1207/S1532785XMEP0301_02 Schwartz, K. D., & Fouts, G. T. (2003). Music preferences, personality style, and developmental issues of adolescents.

Journal of Youth and Adolescence, 32, 205–213. doi: 10.1023/A:1022547520656 Schwenkmezger, P., Hodapp, V., & Spielberger, C. D. (1992).

Das State-Trait-Ärgerausdrucks Inventar (STAXI). Manual [The state trait anger expression inventory manual]. Bern: Verlag Hans Huber.

Schwibbe, M., Räder, K., Schwibbe, G., Borchardt, M., & Geiken-Pophanken, G. (1994). Zum emotionalen Gehalt von Substantiven, Adjektiven und Verben [About the emotional content of subjects, adjectives, and verbs]. In W. Hager & M. Hasselhorn (Eds.), Handbuch deutschsprachiger Wortnormen (pp. 272–284). Göttingen: Hogrefe.

Sprankle, E. L., End, C. M., & Bretz, M. N. (2012). Sexually degrading music videos and lyrics: Their effects on males’ aggression and endorsement of rape myths and sexual stereotypes. Journal of Media Psychology: Theories, Methods, and Applications, 24(1), 31–39. doi: 10.1027/18641105/a000060 Steyer, R., Schwenkmezger, P., Notz, P., & Eid, M. (1997). Der Mehrdimensionale Befindlichkeitsfragebogen (MDBF).

2015 Hogrefe Publishing

41

Manual [The multi-dimensional mental state questionnaire manual]. Göttingen: Hogrefe.

Took, K. J., & Weiss, D. S. (1994). The relationship between heavy metal and rap music and adolescent turmoil: Real or abstract? Adolescence, 29(115), 613–623. van der Zwaag, M. D., Westerink, J. M., & van den Broek, E. L. (2011). Emotional and psychophysiological responses to tempo, mode, and percussiveness. Musicae Scientiae, 15, 250–269. doi: 10.1177/1029864911403364 Wanamaker, C. E., & Reznikoff, M. (1989). Effects of aggressive and nonaggressive rock songs on projective and structured tests. Journal of Psychology: Interdisciplinary and Applied, 123, 561–570.

Webster, G. D., & Weir, C. G. (2005). Emotional responses to music: Interactive effects of mode, texture, and tempo.

Motivation and Emotion, 29(1), 19–39. doi: 10.1007/ s11031-005-4414-0 Wester, S. R., Crown, C. L., Quatman, G. L., & Heesacker, M. (1997). The influence of sexually violent rap music on attitudes of men with little prior exposure. Psychology of Women Quarterly, 21, 497–508. doi: 10.1111/j.14716402.1997.tb00127.x Zillmann, D., & Bryant, J. (1974). Effect of residual excitation on the emotional response to provocation and delayed aggressive behavior. Journal of Personality and Social Psychology, 30, 782–791. doi: 10.1037/h0037541

Date of acceptance: September 12, 2014 Published online: April 15, 2015

Stephanie Pieschl received her diploma in psychology in 2002 and her PhD in 2008, both from the Westfälische WilhelmsUniversität Münster. Currently she is a postdoctoral researcher in the department of educational psychology in Münster. Her research interests include learning and teaching with new media, and media literacy.

Stephanie Pieschl Institut für Psychologie Westfälische Wilhelms-Universität Münster Fliednerstr. 21 48149 Münster Germany Tel. +49 251 83-31386 Fax +49 251 83-39105 E-mail pieschl@uni-muenster.de Simon Fegers received his diploma in psychology in 2010 from the Westfälische Wilhelms-Universität Münster and his certification as psychotherapist in 2014 from the Academy of Behavioral Therapy (Akademie für Verhaltenstherapie) in Cologne. Currently he finishes his PhD at the Universität zu Köln, Department of Clinical Psychology and Psychotherapy.

Journal of Media Psychology 2016; Vol. 28(1):32–41

Research Report

Effects of Light-Hearted and Serious Entertainment on Enjoyment of the First and Third Person Matthias Hofer Institute of Mass Communication and Media Research, University of Zurich, Zurich, Switzerland Abstract. This was a study on the perceived enjoyment of different movie genres. In an online experiment, 176 students were randomly divided into two groups (n = 88) and asked to estimate how much they, their closest friends, and young people in general enjoyed either serious or lighthearted movies. These self–other differences in perceived enjoyment of serious or light-hearted movies were also assessed as a function of differing individual motivations underlying entertainment media consumption. The results showed a clear third-person effect for light-hearted movies and a first-person effect for serious movies. The third-person effect for light-hearted movies was moderated by level of hedonic motivation, as participants with high hedonic motivations did not perceive their own and others’ enjoyment of light-hearted films differently.

However, eudaimonic motivations did not moderate first-person perceptions in the case of serious films.

Keywords: third-person perception, light-hearted films, serious films, enjoyment, social distance, hedonic motivations, eudaimonic motivations

Movies can elicit pleasurable (hedonic) experiences, such as fun, excitement, or a thrill. Comedies, thrillers, or action movies (light-hearted movies) are regarded as the types of films that tend to provide these hedonic viewing experiences (Oliver & Bartsch, 2011) by focusing on short-term gratifications such as mood improvement or mood repair (Reinecke et al., 2012; Zillmann, 1988). In contrast, dramas, documentaries, or biographical films (serious movies) are regarded as having very different effects, providing viewers with meaningful (eudaimonic) experiences.

Serious movies can offer deep insights into the human condition, teaching viewers about morality or the value or meaning of human life. Serious movies can also lead to feelings of autonomy, competence, or personal growth, or can impart a sense that one’s own life has meaning (Oliver, Hartmann, & Woolley, 2012; Tamborini, Bowman, Eden, Grizzard, & Organ, 2010; Wirth, Hofer, & Schramm, 2012). While light-hearted films mainly produce egocentered short-term effects, Oliver et al. (2012) found that serious films promote the motivation to enact moral virtues, encouraging wisdom and virtue and possibly having altruistic consequences.

These two film genres are also associated with social implications about their worth for individuals or for society as a whole, and the use and display of different genres can express a certain social identity. Relying on semiotics and cultural theory (e.g., Gottdiener, 1985), Hall (2007, p. 260) introduces the concept of sign value to describe the social implications of different types of entertainment offerings. For example, showing off one’s consumption Journal of Media Psychology 2016; Vol. 28(1):42–48 DOI: 10.1027/1864-1105/a000150

and enjoyment of serious films can convey a positive image to others and thereby enhance one’s self-image, while viewing light-hearted movies is generally seen as socially undesirable (Peiser & Peter, 2000). As a result, users may downplay their involvement with certain media content to avoid negative social implications (Hall, 2007).

The third-person effect is a theoretical approach that is well-suited to grasping the social implications of the two film genres outlined above (Davison, 1983). This effect posits that individuals consider themselves less influenced by socially undesirable media than other people, a phenomenon known as third-person perception (TPP), and more influenced by socially desirable media than other people (first-person effect or first-person perception; FPP) (Davison, 1983; Tal-Or, Tsfati, & Gunther, 2009). Applied to the subject of the present paper, one could argue that an individual’s enjoyment of light-hearted or serious films may depend, at least in part, on what that person thinks others would experience while watching the same content.

Accordingly, Kim and Oliver (2012) hypothesized that serious films could trigger FPPs, while light-hearted films would be perceived by participants to have a greater effect on other viewers than on themselves (TPP). Although Kim and Oliver’s (2012) results regarding these hypotheses were not completely consistent, their study provides a good starting point for further investigation of perceptual gaps for light-hearted or serious films. Specifically, the authors recommended that future research examine the role that individuals’ motivation plays in their entertainment consumption. Therefore, the present study aimed to extend 2015 Hogrefe Publishing

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

Kim and Oliver’s (2012) findings by taking into account the possible moderating effects of hedonic and eudaimonic motivations, as introduced by Oliver and Raney (2011).

The Third-Person Effect The third-person effect of communication was first proposed by Davison (1983), who demonstrated that people tend to perceive media messages to have stronger effects on other people than on themselves. Cognitive and motivational explanations of TPP include unrealistic optimism (Gunther & Mundy, 1993), self-enhancement (Salwen & Dupagne, 2001), and impression management (Tal-Or & Drukman, 2010). All of these explanations are grounded in the idea that, to maintain or improve their self-esteem and to preserve a positive self-image, people tend to overestimate the effect of media on others, thereby presenting themselves as particularly intelligent (Tal-Or et al., 2009).

Several moderators have been found to either dampen or increase the perceptual gap between the self and others.

One such moderator is the perceived social distance between the self and other, with TPPs and FPPs tending to become larger as the comparison group is perceived as more socially distant (Gunther & Mundy, 1993; Hoffner et al., 2001). This social distance can either be conceptualized as a generality of the description of others or as a perceived similarity with the third person (Meirick, 2005).

Another known moderator of TPP and FPP is the social desirability of the media content, with many studies confirming clear TPP for socially undesirable messages such as pornography (Lee & Tamborini, 2005), misogynistic and violent rap lyrics (McLeod, Eveland, & Nathanson, 1997), and television violence (Hoffner et al., 2001; Salwen & Dupagne, 2001). Likewise, when messages were socially desirable (e.g., health-related messages; Gunther & Mundy, 1993), TPPs were diminished or even reversed.

Third-Person Perceptions and Entertainment Media The social desirability of the message, self-enhancement motivations, and impression management are also likely candidates for explaining TPP and FPP in response to light-hearted and serious entertainment, respectively.

Leone, Peek, and Bissell (2006), for instance, found that participants exhibited clear TPPs when they watched reality TV shows (a form of very light-hearted entertainment).

Combined with Baruh’s (2010) finding that watching reality TV can trigger feelings of guilt, it can be argued that, because comedies are widely regarded as shallow, silly, and socially undesirable, to be entertained by such lighthearted films offers little or no opportunity for selfenhancement and impression management (Oliver & Raney, 2011; Peiser & Peter, 2000). Asking people about their own and others’ enjoyment of light-hearted films

2015 Hogrefe Publishing

43

should therefore lead to a TPP, because one might seem to be a shallow person if he or she was entertained by a comedy or an action film, especially if he or she reported more entertainment than others. Serious films, in contrast, may foster deeper insights, feelings of empathy, or psychological maturation (Oliver & Bartsch, 2010; Wirth et al., 2012). Those who profess to be entertained by such films present themselves as being empathic, thoughtful, or altruistic, and as living in accordance with moral values. This can produce a positive self-image, as well as a positive image in the eyes of others (Hall, 2007). This, in turn, should lead to FPPs (Kim & Oliver, 2012).

In light of these observed effects, we propose that for lighthearted films, participants will rate their enjoyment as lower than that of their close friends and young people in general, whereas, for serious films, participants will rate their enjoyment as higher than that of their close friends and that of young people in general (H1). Second, we hypothesize that these effects will be stronger as the social distance between the first person and the third person increases (H2).

The Moderating Role of Hedonic and Eudaimonic Motivations To grasp preferences for different entertainment genres, Oliver and Raney (2011) introduced the concepts of hedonic and eudaimonic motivations. Hedonic motivations entail preferences for light-hearted films such as comedies or action adventures. Since holding such preferences would denote a favorable attitude toward such types of entertainment, enjoying light-hearted films would not be regarded as undesirable by persons with strong hedonic motivations.

These favorable attitudes would also lead to an increase in perceived effects on the self and, as a result, decrease TPPs. In other words, hedonic motivations should moderate the effects proposed in H1 and H2 in such a way as to dampen TPPs for light-hearted films (H3). Eudaimonic motivations, in contrast, go along with preferences for serious films depicting meaningful stories about moral virtues and the purpose of life. Motivations of this type would also logically have a moderating effect on perceptual gaps, as persons with high eudaimonic motivations enjoy serious movies more than persons with lower eudaimonic motivations. This difference in motivations and the resulting enjoyment would, in turn, increase FPPs between the first and the third person when watching serious films (H4).

Method Participants A total of 176 undergraduate communications students (women, n = 116) aged 17–29 years (Mage = 20.85,

Journal of Media Psychology 2016; Vol. 28(1):42–48

44

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

SDage = 1.88) took part in the present study. This study was conducted as an online experiment. Students received course credit for participation.

general. They then filled out the rest of the questionnaire, which included items assessing hedonic and eudaimonic motivations, as well as the demographic information of sex and age.

Design and Stimulus Material The study employed a 2 (light-hearted movies vs. serious movies) · 3 (social distance: self vs. close friends vs. young people) mixed factorial design with hedonic and eudaimonic motivations as continuous moderators.

A list of seven light-hearted movies (e.g., The Wedding Crashers, USA, 2005, Director: David Dobkin) or seven serious movies (e.g., La Vita è Bella, Italy, 1997, Director: Roberto Benigni) served as stimulus material (see the Appendix for the complete list). The films were chosen based on the results of a pre-study. This pre-study began with a list of 36 films (18 light-hearted and 18 serious films) generated according to the definitions of appreciation and enjoyment by Oliver and Bartsch (2010, 2011).

The labeled covers of these 36 films were shown to 161 undergraduate students (Mage = 20.53, SDage = 2.11), who were then asked to rate the light-heartedness or the seriousness of each film that they had seen on a 5-point Likert scale. Films were presented randomly to control for order effects. The subset of films used in the main study consisted of the films that had the highest ratings in one category along with the lowest ratings in the other (e.g., Dumb and Dumber, USA, 1994, Director: Peter Farrelly; Mlight-heartedness = 3.37, Mseriousness = 1.33). To be included in the main study, films also had to have been seen by more than 50% of the pre-study sample.

Procedure In the main study, participants first indicated their perceived similarity with both their close friends and young people in general (see Measures section). After this preliminary questionnaire, they were randomly assigned to one of the two experimental conditions (light-hearted movies: n = 88; serious movies: n = 88). Participants were given a list of seven films and asked to choose the film they knew best. Had any participants indicated that they were not familiar with any of these films they would have been excluded from the sample; however, all participants reported knowing at least one of the seven films. After selecting their film, they indicated their perceptions regarding the movie’s effect on themselves, their close friends, and young people in

Measures Manipulation Check The manipulation check for the experimental manipulation of different film genres consisted of five semantic differentials on which the chosen film had to be rated (e.g., 1 = meaningless, 5 = meaningful). The five differentials were then collapsed into a single index (a = .917, M = 3.34, SD = 1.22).

Perceived Similarity Two items were used to measure perceived similarity with either young people or close friends. The first item (‘‘How similar do you feel to the following group?’’) used a 5-point Likert-type scale ranging from 1 (= not at all similar) to 5 (= very similar) to measure perceived similarity with the third-person group of either young people or close friends. The second item (‘‘How strongly do you identify with the following group?’’) used a 5-point Likert-type scale ranging from 1 (= not at all) to 5 (= very strongly) to measure the degree of identification with the group in question (see Meirick, 2005). Since correlations between the two items in each group were strong, these two items were combined to form an index for each group: ryoung(174) = .75, p < .001, Myoung = 3.03, SDyoung = 1.02; rfriends(174) = .57, p < .001, Mfriends = 4.22, SDfriends = 0.75.

Enjoyment To measure perceived enjoyment of the selected film for oneself, close friends, and young people in general, the fun subscale by Oliver and Bartsch (2010) was used (e.g., ‘‘It was fun for me to watch this movie,’’ ‘‘It was fun for my close friends to watch this movie,’’ and ‘‘It was fun for young people to watch this movie’’). Together, the scales had good reliabilities (self: a = .842, M = 3.96, SD = 0.89; close friends: a = .865, M = 3.84, SD = 0.97; young people in general: a = .947, M = 3.72, SD = 1.10) (Table 1).

Table 1. Means, standard deviations, and zero-order correlations of Enjoyment, Hedonic, and Eudaimonic Motivations 1.

2.

3.

4.

5.

Enjoyment (first person) Enjoyment (close friends) Enjoyment (younger people in general) Eudaimonic motivations Hedonic motivations

M

SD

1.

3.96 3.83 3.72 3.91 3.24

0.89 0.97 1.10 0.62 0.76

– .78** .53** .14 .25**

2.

– .71** .01 .23**

3.

4.

5.

– .02 .18*

– .21

–

Note. *p < .05; **p < .01.

Journal of Media Psychology 2016; Vol. 28(1):42–48

2015 Hogrefe Publishing

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

45

Table 2. Multi-level model of the effects of experimental manipulation, social distance, and hedonic and eudaimonic motivations on perceived enjoyment Fixed Effects Intercept Social Distance Experimental Condition (= light-hearted films) Hedonic Motivations Eudaimonic Motivations Social Distance · Experimental Condition Social Distance · Hedonic Motivations Experimental Condition · Hedonic Motivations Social Distance · Experimental Condition · Hedonic Motivations Social Distance · Eudaimonic Motivations Experimental Condition · Eudaimonic Motivations Time · Experimental Condition · Eudaimonic Motivations

Hedonic Motivations The hedonic motivations subscale by Oliver and Raney (2011) was used to measure motivations of this type (e.g., ‘‘I like movies that may be considered ‘silly’ or ‘shallow’ if they can make me laugh and have a good time’’; a = .80, M = 3.24, SD = 0.76) (Table 1).

b

SE

4.13 .20 .42 .59 .15 .61 .22 .52 .24 .04 .32 .12

0.08 0.04 0.11 0.11 0.14 0.05 0.05 0.15 0.07 0.06 0.19 0.08

t Value 52.15 5.64 3.73 5.15 1.11 12.05 4.27 3.47 3.47 0.61 1.72 1.46

p Value .000 .000 .000 .000 .268 .000 .000 .001 .001 .545 .086 .146

young people, M = 3.03, SD = 1.02, t(175) = 25.54, p < .001. Results of a paired sample t test showed that perceived similarity to close friends differed significantly from perceived similarity to young people: t(175) = 15.07, p < .001.

Test of Hypotheses Eudaimonic Motivations Eudaimonic motivations were assessed with the eudaimonic motivations subscale by Oliver and Raney (2011) (e.g., ‘‘I like movies that make me more reflective’’; a = .79, M = 3.91, SD = 0.62) (Table 1).

Results Manipulation Check To examine whether the experimental manipulation was successful, an analysis of variance (ANOVA) was conducted with film genre (light-hearted vs. serious) as the factor and the manipulation check as the dependent variable.

Results showed that the experimental manipulation was successful, as light-hearted films were rated as less meaningful (M = 2.30, SD = 0.75) than serious films, M = 4.37, SD = 0.51, F(1, 174) = 459.27, p < .001, g2 = 73.

To examine the social distance between the self, close friends, and young people, perceived similarity with close friends and perceived similarity with young people were tested against the maximum value (this assumes that the self feels completely similar to himself or herself). Results showed significant differences between the maximum value of 5, perceived similarity to close friends, M = 4.22, SD = 0.75, t(175) = 13.95, p < .001, and perceived similarity to

2015 Hogrefe Publishing

To test the hypotheses (H1–H4), a linear mixed model was conducted using restricted maximum likelihood estimation (REML), with enjoyment as the dependent variable. The experimental condition was entered as a between-subjects factor; the three groups (self, close friends, and young people) were entered as a within-subject factor, and hedonic and eudaimonic motivations were used as continuous moderators. The final model was a random intercept model – that is, enjoyment of the first person varied significantly within the sample, var(u0j) = .39(.05), z = 7.32, p < .001.

Results for fixed effects are given in Table 2. In the case of serious movies, participants regarded themselves as more entertained than their close friends and than young people in general, an example of FPP. As depicted in Figure 1, the perceived effect increased as the third person became more socially distant from the first person. In the case of light-hearted movies, a clear TPP emerged, which also increased with social distance. Thus, all study data supported H1 and H2. However, the main effects, as well as the two-way interaction between the experimental condition and social distance, have to be interpreted in light of a significant three-way interaction between the experimental manipulation, social distance, and hedonic motivations: As can be seen in the right-hand panel of Figure 1, participants with higher hedonic motivations (i.e., 1 SD above the mean) regarded themselves, their close friends, and young people in general as equally entertained by light-hearted films, indicating a dampening of TPP and, therefore, supporting H3. However, there was no significant three-way

Journal of Media Psychology 2016; Vol. 28(1):42–48

46

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

Figure 1. Three-way interaction between experimental manipulation, social distance, and hedonic motivations on enjoyment. Left panel: hedonic motivations 1 SD below the mean; right panel: hedonic motivations 1 SD above the mean. interaction between the experimental manipulation, social distance, and eudaimonic motivations, meaning that H4 was not supported by the data.1

Discussion The results of the present study demonstrate that viewing and experiencing entertainment media has social implications. A person can present himself or herself to others as being too smart to be entertained by light-hearted films – as a person who is so smart, empathic, thoughtful, moral, or even altruistic as to require more complex and serious entertainment. However, with only the perceived shallowness or meaningfulness of the stimuli being evaluated, this interpretation remains speculative.

Our study indicated the presence of the proposed moderating effect of hedonic motivations, with the perceptual gap disappearing for participants who were highly hedonically motivated in their general film consumption. This would seem to denote that persons with high hedonic motivations do not regard light-hearted films as socially undesirable or guilt inducing, and therefore do not show TPPs. However, we did not find that eudaimonic motivations moderated FPPs. One possible explanation for this could be that persons with high eudaimonic motivations are generally more reflective than those with low eudaimonic motivations (Oliver & Raney, 2011). This reflectiveness could

1

then lower the need for self-enhancement through showing off one’s consumption of meaningful entertainment.

The results of the present study may have broader implications for entertainment-related theories such as mood management theory (Zillmann, 1988), which could, for instance, apply differently in social viewing situations such as movie theaters than when people view media alone.

More precisely, viewers with low hedonic motivations might not find much hedonic value in light-hearted films when watching such films with other people, as such viewers might downplay their enjoyment so as to not threaten their self-image. Future studies may therefore wish to to examine the effects of coviewing on enjoyment for persons with higher and lower hedonic motivations.

Finally, it should be noted that we only considered hedonic enjoyment in assessing the effects of light-hearted and serious films. One could, however, argue that concepts like fun or hedonic enjoyment might not be best suited to grasping the experiences associated with serious films. Future studies should, therefore, also include measures of appreciation so as to assess the entertainment experiences of serious movies more accurately (Oliver & Bartsch, 2010).

References Baruh, L. (2010). Mediated voyeurism and the guilty pleasure of consuming reality television. Media Psychology, 13, 201–221. doi: 10.1080/15213269.2010.502871

One can argue that the fact that not all participants responded to the same stimulus may have influenced the results. Therefore, we ran the same model with a dummy variable for each film. The results changed slightly: For example, the main effect of the experimental manipulation disappeared. However, the three-way interaction between the experimental manipulation, the social distance, and hedonic motivations remained significant.

Journal of Media Psychology 2016; Vol. 28(1):42–48

2015 Hogrefe Publishing

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

Davison, W. P. (1983). The third-person effect in communication. Public Opinion Quarterly, 47(1), 1–15. doi: 10.1086/ 268763 Gottdiener, M. (1985). Hegemony and mass culture: A semiotic approach. American Journal of Sociology, 90, 979. doi: 10.1086/228173 Gunther, A. C., & Mundy, P. (1993). Biased optimism and the third-person effect. Journalism Quarterly, 70, 58–67. doi: 10.1177/107769909307000107 Hall, A. (2007). The social implications of enjoyment of different types of music, movies, and television programming. Western Journal of Communication, 71, 259–271. doi: 10.1080/10570310701653794 Hoffner, C., Plotkin, R. S., Buchanan, M., Anderson, J. D., Kamigaki, S. K., Hubbs, L. A., . . . Pastorek, A. (2001).

The third-person effect in perceptions of the influence of television violence. Journal of Communication, 51, 283–299. doi: 10.1111/j.1460-2466.2001.tb02881.x Kim, K., & Oliver, M. B. (2012, May). Third-person perceptions and different movie genres. Paper presented at the International Communication Association, Boston, MA. Lee, B., & Tamborini, R. (2005). Third-person effect and internet pornography: The influence of collectivism and internet self-efficacy. Journal of Communication, 55, 292–310. doi: 10.1111/j.1460-2466.2005.tb02673.x Leone, R., Peek, W. C., & Bissell, K. L. (2006). Reality television and third-person perception. Journal of Broadcasting & Electronic Media, 50, 253–269. doi: 10.1207/ s15506878jobem5002_5 McLeod, D. M., Eveland, W. P., & Nathanson, A. I. (1997).

Support for censorship of violent and misogynic rap lyrics: An analysis of the third-person effect. Communication Research, 24, 153–174. doi: 10.1177/009365097024002003 Meirick, P. C. (2005). Rethinking the target corollary: The effects of social distance, perceived exposure, and perceived predispositions on first-person and third-person perceptions.

Communication Research, 32, 822–843. doi: 10.1177/ 0093650205281059 Oliver, M. B., & Bartsch, A. (2010). Appreciation as audience response: Exploring entertainment gratifications beyond hedonism. Human Communication Research, 36, 53–81. doi: 10.1111/j.1468-2958.2009.01368.x Oliver, M. B., & Bartsch, A. (2011). Appreciation of entertainment: The importance of meaningfulness via virtue and wisdom. Journal of Media Psychology, 23, 29–33. doi: 10.1027/1864-1105/a000029 Oliver, M. B., Hartmann, T., & Woolley, J. K. (2012). Elevation in response to entertainment portrayals of moral virtue.

Human Communication Research, 38, 360–378. doi: 10.1111/j.1468-2958.2012.01427.x Oliver, M. B., & Raney, A. A. (2011). Entertainment as pleasurable and meaningful: Identifying hedonic and eudaimonic motivations for entertainment consumption. Journal of Communication, 61, 984–1004. doi: 10.1111/j.14602466.2011.01585.x Peiser, W., & Peter, J. (2000). Third-person perception of television-viewing behavior. Journal of Communication, 50, 25–45. doi: 10.1111/j.1460-2466.2000.tb02832.x

2015 Hogrefe Publishing

47

Reinecke, L., Tamborini, R., Grizzard, M., Lewis, R., Eden, A., & David Bowman, N. (2012). Characterizing mood management as need satisfaction: The effects of intrinsic needs on selective exposure and mood repair. Journal of Communication, 62, 437–453. doi: 10.1111/j.1460-2466.

2012.01649.x Salwen, M. B., & Dupagne, M. (2001). Third-person perception of television violence: The role of self-perceived knowledge. Media Psychology, 3, 211–236. doi: 10.1207/ S1532785XMEP0303_01 Tal-Or, N., & Drukman, D. (2010). Third-person perception as an impression management tactic. Media Psychology, 13, 301–322. doi: 10.1080/15213269.2010.503516 Tal-Or, N., Tsfati, Y., & Gunther, A. C. (2009). The influence of presumed media influence: Origins and implications of the third-person perception. In R. L. Nabi & M. B. Oliver (Eds.), The Sage handbook of media processes and effects (pp. 99–112). Los Angeles, CA: Sage.

Tamborini, R., Bowman, N. D., Eden, A., Grizzard, M., & Organ, A. (2010). Defining media enjoyment as the satisfaction of intrinsic needs. Journal of Communication, 60(4), 758–777. doi: 10.1111/j.1460-2466.2010.01513.x Wirth, W., Hofer, M., & Schramm, H. (2012). Beyond pleasure: Exploring the eudaimonic entertainment experience. Human Communication Research, 38, 406–428. doi: 10.1111/ j.1468-2958.2012.01434.x Zillmann, D. (1988). Mood management through communication choices. American Behavioral Scientist, 31, 327–340. doi: 10.1177/000276488031003005

Date of acceptance: December 29, 2014 Published online: June 29, 2015

Matthias Hofer (PhD, University of Zurich) is a senior research and teaching associate at the Institute of Mass Communication and Media Research, University of Zurich, Switzerland. His main research areas are media audiences and effects, entertainment and emotion research, social capital in new media, and presence research. He also examines media effects through the lifespan.

Matthias Hofer Institute of Mass Communication and Media Research University of Zurich Andreasstrasse 15 8050 Zurich Switzerland Tel. +41 44 635-2063 Fax +41 44 634-4934 E-mail m.hofer@ipmz.uzh.ch

Journal of Media Psychology 2016; Vol. 28(1):42–48

48

M. Hofer: Enjoyment of Light-Hearted and Serious Movies

Appendix

Serious Films

Films Used in the Study

The Pianist (USA, 2002, Director: Roman Polan´ski) Seven Pounds (USA, 2008, Director: Gabriele Muccino) La Vita è Bella (Italy, 1997, Director: Roberto Benigni) Schindler’s List (USA, 1993, Director: Steven Spielberg) Into the Wild (USA, 2007, Director: Sean Penn) My Sister’s Keeper (USA, 2009, Director: Nick Cassavetes) The Green Mile (USA, 1999, Director: Frank Darabont)

Light-Hearted Films Hangover (USA, 2009, Director: Todd Phillips) Ted (USA, 2012, Director: Seth MacFarlane) Miss Congeniality (USA, 2000, Director: Donald Petrie) American Pie (USA, 1999, Directors: Paul Weitz & Chris Weitz)

Journal of Media Psychology 2016; Vol. 28(1):42–48

2015 Hogrefe Publishing

Erratum Correction to Retzbach, Retzbach, Maier, Otto, & Rahnke, 2013 The article ‘‘Effects of repeated exposure to science TV shows on beliefs about scientific evidence and interest in science’’ by Joachim Retzbach, Andrea Retzbach, Michaela Maier, Lukas Otto, & Marion Rahnke (Journal of Media Psychology, 2013, 25, 3–13, DOI: 10.1027/1864-1105/ a000073) contained an error on page 9, left column, 2nd paragraph.

Correctly, the p-value for Group 1 in the sentence should read p = .045: We found a significant decrease in USE-O scores in Group 1: certain treatment, t(166) = 2.02, p = .045, d = 0.15; a significant increase in Group 2: uncertain

Ó 2015 Hogrefe Publishing

treatment, t(170) = 4.69, p < .001, d = 0.36; and an increase in Group 3: t(155) = 3.10, p = .002, d = 0.20.

The authors regret any inconvenience or confusion this error may have caused.

Reference Retzbach, J., Retzbach, A., Maier, M., Otto, L., & Rahnke, M. (2013). Effects of repeated exposure to science TV shows on beliefs about scientific evidence and interest in science.

Journal of Media Psychology, 25, 3–13. doi: 10.1027/18641105/a000073

Journal of Media Psychology 2016; Vol. 28(1):49 DOI: 10.1027/1864-1105/a000178

Meeting Calendar March 2–4, 2016 18th General Online Research Conference Dresden, Germany Contact: Deutsche Gesellschaft f r Online-Forschung e.V. (DGOF)/ German Society for Online-Research, E-mail office@dgof.de, http:// www.gor.de/ March 21–23, 2016 TeaP 2016: 58th Conference of Experimental Psychologists Heidelberg, Germany Contact: Conference Management: Jan Rummel, Sabine Falke, University of Heidelberg, Germany, E-mail teap2016@psychologie.uni-heidelberg. de, http://www.teap2016.de/543/ index.hei?context=543&&card=heises.

0cf358d5cdda95644f88091e7729 c1b7 March 30–April 1, 2016 61st Annual Conference of the German Communication Association (DGPuK) Leipzig, Germany Contact: University of Leipzig, Institute of Communication and Media Research, Leipzig, Germany, E-mail dgpuk2016@uni-leipzig.de, http:// conference.uni-leipzig.de/ dgpuk2016/ May 7–12, 2016 CHI2016 San Jose, CA, USA Contact: Conference Chairs: Allison Druin (University of Maryland), Jofish Kaye (Yahoo Labs), E-mail generalchairs@chi2016.acm.org, http://chi2016.acm.org/wp/ May 10-12, 2016 WAPOR 69th Annual Conference Austin, TX, USA

Journal of Media Psychology 2016; Vol. 28(1):50 DOI: 10.1027/1864-1105/a000179

Contact: World Association for Public Opinion Research, Conference Co-Chairs: Bethany Albertson (E-mail balberts@austin.utexas.edu) and Talia Stroud (E-mail tstroud@austin.utexas.edu), http:// wapor.org/69th-annual-conference/ June 6–13, 2016 ICA – International Communication Association Annual Conference Fukuoka, Japan Contact: International Communication Association, Washington, DC, http://www.icahdq.org June 22–25, 2016 15th International Conference on Language and Social Psychology (ICLASP15) Bangkok, Thailand Contact: International Association of Language and Social Psychology, E-mail iclasplirod@gmail.com, http://ialsp.org/conferences/ July 24–29, 2016 ICP – 31st International Congress of Psychology Yokohama, Japan Contact: ICP2016 Office, c/o The Japanese Psychological Association, 5-23-13, Hongo, Bunkyo-ku, Tokyo, 113-0033 Japan, E-mail info@icp2016. jp, http://www.icp2016.jp July 27–31, 2016 IAMCR 2016 Conference Leicester, UK Contact: International Association for Media and Communication Research, University of Leicester, UK, E-mail iamcr2016@leicester.ac.uk, http:// iamcr.org/leicester2016

August 4–7, 2016 APA Psychology – 124th Annual Convention Denver, CO, USA Contact: American Psychological Association, 750 First St. NE, Washington, DC 20002-4242, http:// www.apa.org/convention/ September 18–22, 2016 50th Congress of DGPs (Deutsche Gesellschaft f r Psychologie) Leipzig, Germany Contact: University of Leipzig, Conference Management, E-mail info@dgpskongress.de, http://www. dgpskongress.de/frontend/index.php? page_id=120 October 13–14, 2016 18th International Conference on Intelligent Virtual Agents Bali, Indonesia Contact: World Academy of Science, Engineering and Technology, https:// www.waset.org/conference/2016/10/ bali/ICIVA November 9–12, 2016 ECREA 2016, 6th European Communication Conference Prague, Czech Republic Contact: Conference Secretariat, Prague, Czech Republic, E-mail info@ecrea2016prague.eu, http:// www.ecrea2016prague.eu/ July 11–14, 2017 ECP – 15th European Congress of Psychology Amsterdam, NL, The Netherlands Contact: ENIC Meetings and Events, Firenze, Italia, E-mail info@ecp2015.it, http://www.ecp2015.it/#sthash.

MESRCDSL.dpuf

Ó 2016 Hogrefe Publishing

Instructions to Authors Journal of Media Psychology (JMP) is committed to publishing original, high-quality papers which cover the broad range of media psychological research. This peer-reviewed journal focuses on how human beings select, use, and experience various media as well as how media (use) can affect their cognitions, emotions, and behaviors. It is also open to research from neighboring disciplines as far as this work ties in with psychological concepts of the uses and effects of the media. In particular, it publishes multidisciplinary papers that reflect a broader theoretical and methodological spectrum and comparative work, e.g., cross-media, cross-gender, or crosscultural. As JMP is intended to foster Open Science Practices, authors are offered to publish their data and materials (i.e., stimuli and surveys) as Electronic Supplementary Material on the publisher’s website at http://econtent@hogrefe. com. In line with the Peer Reviewers’ Openness Initiative, authors may be asked by reviewers to share their data and materials at any stage of the reviewing process.

Journal of Media Psychology publishes the following types of article: Original Articles, Theoretical Articles, Research Reports, Pre-Registered Reports.

Manuscript submission: All manuscripts should in the first instance be submitted electronically at http://www.editorialmanager.com/ jmp. Detailed instructions to authors are provided at http://www. hogrefe.com/periodicals/journal-of-media-psychology/advice-forauthors/ Copyright Agreement: By submitting an article, the author confirms and guarantees on behalf of him-/herself and any coauthors that the manuscript has not been submitted or published elsewhere, and that he or she holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches, and tables, and that the article and its contents do not infringe in any way on the rights of third parties. The author indemnifies and holds harmless the publisher from any third party claims. The author agrees, upon acceptance of the article for publication, to transfer to the publisher the exclusive right to reproduce and distribute the article and its contents, both physically and in nonphysical,

Ó 2016 Hogrefe Publishing

electronic, or other form, in the journal to which it has been submitted and in other independent publications, with no limitations on the number of copies or on the form or the extent of distribution. These rights are transferred for the duration of copyright as defined by international law. Furthermore, the author transfers to the publisher the following exclusive rights to the article and its contents: 1. The rights to produce advance copies, reprints, or offprints of the article, in full or in part, to undertake or allow translations into other languages, to distribute other forms or modified versions of the article, and to produce and distribute summaries or abstracts.

2. The rights to microfilm and microfiche editions or similar, to the use of the article and its contents in videotext, teletext, and similar systems, to recordings or reproduction using other media, digital or analog, including electronic, magnetic, and optical media, and in multimedia form, as well as for public broadcasting in radio, television, or other forms of broadcast.

3. The rights to store the article and its content in machinereadable or electronic form on all media (such as computer disks, compact disks, magnetic tape), to store the article and its contents in online databases belonging to the publisher or third parties for viewing or downloading by third parties, and to present or reproduce the article or its contents on visual display screens, monitors, and similar devices, either directly or via data transmission.

4. The rights to reproduce and distribute the article and its contents by all other means, including photomechanical and similar processes (such as photocopying or facsimile), and as part of so-called document delivery services.

5. The right to transfer any or all rights mentioned in this agreement, as well as rights retained by the relevant copyright clearing centers, including royalty rights to third parties.

Online Rights for Journal Articles: Guidelines on authors’ rights to archive electronic versions of their manuscripts online are given in the Advice for Authors on the journal’s web page at www.hogrefe.com.

February 2016

Journal of Media Psychology 2016; Vol. 28(1)

Using movies to help learn about mental illness

“I have been a fan of Movies and Mental Illness from the first edition.”

Steven Pritzker, PhD, psychology professor (Saybrook University) and former Hollywood script writer

Danny Wedding / Ryan M. Niemiec

Movies and Mental Illness

Using Films to Understand Psychopathology 4th edition 2014, xviii + 456 pp.

US $59.00 / € 42.95 ISBN 978-0-88937-461-4 Also available as an eBook Films can be a powerful aid to learning about mental illness and psychopathology – for students of psychology, psychiatry, social work, medicine, nursing, counselling, literature or media studies, and for anyone interested in mental health.

Movies and Mental Illness, written by experienced clinicians and teachers who are themselves movie aficionados, has established a great reputation as a uniquely enjoyable and highly memorable text for learning about psychopathology. The new edition has been fully updated to include DSM-5 and ICD-10 diagnoses.

www.hogrefe.com

The core clinical chapters each use a fabricated case history and Mini-Mental State Examination along with synopses and discussions about specific movies to explain, teach, and encourage discussion about all the most important mental health disorders.

Each chapter also includes: Critical Thinking Questions; “Authors’ Picks” (Top 10 Films); What To Read if You Only Have Time to Read One Book or Article; and Topics for Group Discussions.

Social Psychology Editor-in-Chief Christian Unkelbach Universität zu Köln, Germany New Editor-in-Chief starting April 1, 2016: Kai Epstude, Groningen, NL

nline free o issue le samp

Associate Editors Julia Becker, Osnabrück, Germany Malte Friese, Saarbrücken, Germany Michael Häfner, Berlin, Germany Hans J. IJzerman, Amsterdam, The Netherlands Markus Kemmelmeier, Reno, USA Ruth Mayo, Jerusalem, Israel Ulrich Kühnen, Bremen, Germany Michaela Wänke, Mannheim, Germany

Editorial Office Juliane Burghardt Universität zu Köln, Germany

ISSN-Print 1864-9335 ISSN-Online 2151-2590 ISSN-L 1864-9335 6 issues per annum (= 1 volume)

Subscription rates (2016) Libraries / Institutions US $464.00 / € 354.00 Individuals US $223.00 / € 159.00 Postage / Handling US $24.00 / € 18.00

www.hogrefe.com

About the Journal Social Psychology publishes innovative and methodologically sound research and serves as an international forum for scientific discussion and debate in the field of social psychology. Topics include all basic social psychological research themes, methodological advances in social psychology, as well as research in applied fields of social psychology. The journal focuses on original empirical contributions to social psychological research, but is open to theoretical articles, critical reviews, and replications of published research.

The journal was published until volume 38 (2007) as the Zeitschrift für Sozialpsychologie (ISSN 0044-3514). Drawing on over 30 years of experience and tradition in publishing high-quality, innovative science as the Zeitschrift für Sozialpsychologie, Social Psychology has an internationally renowned team of editors and consulting editors from all areas of basic and applied social psychology, thus ensuring that the highest international standards are maintained.

Manuscript Submissions All manuscripts should be submitted online at www.editorialmanager.com/sopsy, where full instructions to authors are also available.

Electronic Full Text The full text of the journal – current and past issues (from 1999 onward) – is available online at http://econtent.hogrefe.com/ loi/zsp (included in subscription price).

A free sample issue is also available here.

Abstracting Services The journal is abstracted/indexed in Current Contents/Social and Behavioral Sciences (CC/S&BS), Social Sciences Citation Index (SSCI), PsycINFO, PSYNDEX, ERIH, Scopus, and EMCare. Impact Factor (Journal Citation Reports®, Thomson Reuters): 2014 = 1.662

Alternatives to traditional self-reports in psychological assessment “A unique and timely guide to better psychological assessment.”

Rainer K. Silbereisen, Research Professor, Friedrich Schiller University Jena, Germany Past-President, International Union of Psychological Science

Tuulia Ortner / Fons J.R. van de Vijver (Editors)

Behavior-Based Assessment in Psychology Going Beyond Self-Report in the Personality, Affective, Motivation, and Social Domains Series: Psychological Assessment – Science and Practice – Vol. 1 2015, vi + 234 pp.

US $63.00 / € 44.95 ISBN 978-0-88937-437-9 Also available as an eBook Traditional self-reports can be an unsufficiant source of information about personality, attitudes, affect, and motivation. What are the alternatives? This first volume in the authoritative series Psychological Assessment – Science and Practice discusses the most influential, state-of-the-art forms of assessment that can take us beyond self-report. Leading scholars from various countries describe the

www.hogrefe.com

theoretical background and psychometric properties of alternatives to self-report, including behavior-based assessment, observational methods, innovative computerized procedures, indirect assessments, projective techniques, and narrative reports. They also look at the validity and practical application of such forms of assessment in domains as diverse as health, forensic, clinical, and consumer psychology.