Psi Chi Journal of Psychological Research - Winter 2023

ISSN: 2325-7342

Published by Psi Chi, The International Honor Society in Psychology ®

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH WINTER 2023 | VOLUME 28, NUMBER 4

EDITOR

STEVEN V. ROUSE, PhD

Pepperdine University

Telephone: (310) 506-7959

Email: steve.rouse@psichi.org

ASSOCIATE EDITORS

JENNIFER L. HUGHES, PhD

Agnes Scott College

STELLA LOPEZ, PhD

University of Texas at San Antonio

TAMMY LOWERY ZACCHILLI, PhD Saint Leo University

ALBEE MENDOZA, PhD

Delaware State University

KIMBERLI R. H. TREADWELL, PhD University of Connecticut

ROBERT R. WRIGHT, PhD

Brigham Young University-Idaho

EDITOR EMERITUS

DEBI BRANNAN, PhD

Western Oregon University

MANAGING EDITOR

BRADLEY CANNON

DESIGNER

JANET REISS

EDITORIAL ASSISTANT

EMMA SULLIVAN

ADVISORY EDITORIAL BOARD

GLENA ANDREWS, PhD

RAF Lakenheath USAF Medical Center

AZENETT A. GARZA CABALLERO, PhD Weber State University

MARTIN DOWNING, PhD Lehman College

HEATHER HAAS, PhD

University of Montana Western

ALLEN H. KENISTON, PhD

University of Wisconsin–Eau Claire

MARIANNE E. LLOYD, PhD Seton Hall University

DONELLE C. POSEY, PhD

Washington State University

LISA ROSEN, PhD

Texas Women's University

CHRISTINA SINISI, PhD

Charleston Southern University

PAUL SMITH, PhD

Alverno College

ABOUT PSI CHI

Psi Chi is the International Honor Society in Psychology, founded in 1929. Its mission: "recognizing and promoting excellence in the science and application of psychology." Membership is open to undergraduates, graduate students, faculty, and alumni making the study of psychology one of their major interests and who meet Psi Chi’s minimum qualifications. Psi Chi is a member of the Association of College Honor Societies (ACHS), and is an affiliate of the American Psychological Association (APA) and the Association for Psychological Science (APS). Psi Chi’s sister honor society is Psi Beta, the national honor society in psychology for community and junior colleges.

Psi Chi functions as a federation of chapters located at over 1,180 senior colleges and universities around the world. The Psi Chi Headquarters is located in Chattanooga, Tennessee. A Board of Directors, composed of psychology faculty who are Psi Chi members and who are elected by the chapters, guides the affairs of the Organization and sets policy with the approval of the chapters. Psi Chi membership provides two major opportunities. The first of these is academic recognition to all inductees by the mere fact of membership. The second is the opportunity of each of the Society’s local chapters to nourish and stimulate the professional growth of all members through fellowship and activities designed to augment and enhance the regular curriculum. In addition, the Organization provides programs to help achieve these goals including conventions, research awards and grants competitions, and publication opportunities.

JOURNAL PURPOSE STATEMENT

The twofold purpose of the Psi Chi Journal of Psychological Research is to foster and reward the scholarly efforts of Psi Chi members, whether students or faculty, as well as to provide them with a valuable learning experience. The articles published in the Journal represent the work of undergraduates, graduate students, and faculty; the Journal is dedicated to increasing its scope and relevance by accepting and involving diverse people of varied racial, ethnic, gender identity, sexual orientation, religious, and social class backgrounds, among many others. To further support authors and enhance Journal visibility, articles are now available in the PsycINFO®, EBSCO®, Crossref®, and Google Scholar databases. In 2016, the Journal also became open access (i.e., free online to all readers and authors) to broaden the dissemination of research across the psychological science community.

JOURNAL INFORMATION

The Psi Chi Journal of Psychological Research (ISSN 2325­7342) is published quarterly in one volume per year by Psi Chi, Inc., The International Honor Society in Psychology. For more information, contact Psi Chi Headquarters, Publication and Subscriptions, 651 East 4th Street, Suite 600, Chattanooga, TN 37403, (423) 756­2044. www.psichi.org; psichijournal@psichi.org

Statements of fact or opinion are the responsibility of the authors alone and do not imply an opinion on the part of the officers or members of Psi Chi.

Advertisements that appear in Psi Chi Journal do not represent endorsement by Psi Chi of the advertiser or the product. Psi Chi neither endorses nor is responsible for the content of third­party promotions. Learn about advertising with Psi Chi at http://www.psichi.org/Advertise

COPYRIGHT

Permission must be obtained from Psi Chi to reprint or adapt a table or figure; to reprint quotations exceeding the limits of fair use from one source, and/or to reprint any portion of poetry, prose, or song lyrics. All persons wishing to utilize any of the above materials must write to the publisher to request nonexclusive world rights in all languages to use copyrighted material in the present article and in future print and nonprint editions. All persons wishing to utilize any of the above materials are responsible for obtaining proper permission from copyright owners and are liable for any and all licensing fees required. All persons wishing to utilize any of the above materials must include copies of all permissions and credit lines with the article submission.

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

239 Bicyclist Helmet Use and Distractions: An Observational Study on a Designated Urban Shared-Use Path

Camryn H. Hutchins1 and Bryan E. Porter2*

1Department of Psychology, Knox College

2Department of Psychology and The Graduate School, Old Dominion University

247 Investigation of the Relationship Between Perceived Mental Workload and Chronic Pain

Kayli N. Colpitts, Jennifer L. Gibson Dias*, Thomas J. Faulkenberry* , and Amber L. Harris Bozer*

Department of Psychological Sciences, Tarleton State University

256 Ally See or Ally Do: Rewarding Corporate Social Responsibility Through Purchasing

Madison Will and Kevin R. Carriere*

Department of Psychology, Washington and Jefferson College

264 The Effects of Hormonal Contraception on Auditory Emotional Memory

Jessica Simonson, Courtney A. Durdle, and Michael B. Miller*

Department of Psychological and Brain Sciences, University of California, Santa Barbara

275 Social Media Use Motives: An Influential Factor in User Behavior and User Health Profiles

Carson R. Ewing, Christian Nienstedt, Robert R. Wright*, and Samuel Chambers

Department of Psychology, Brigham Young University–Idaho

287 Where’s the Party? How Clothing and Context Influence Perceptions of Women

Megan Sherman, Regan A. R. Gurung*, Callan Jackman, and Hannah Mather

School of Psychological Sciences, Oregon State University

296 Predicting Motivation and Learning Strategies in Community College Students

Karen A. Livesey*1, Alison K. Beatty2, Morrison F. Rubin2, and Niomi R. Kaiser3

1Department of Psychology, Northern Virginia Community College

2Department of Psychology, George Mason University

3 School of Neuroscience, Virginia Polytechnic Institute and State University

Bicyclist Helmet Use and Distractions: An Observational Study on a Designated Urban Shared-Use Path

Camryn H. Hutchins1 and Bryan E. Porter*2

1Department

of Psychology, Knox College

2Department of Psychology and The Graduate School, Old Dominion University

ABSTRACT. Distraction has been a behavior evaluated among drivers and more recently pedestrians but remains understudied among bicyclists. Demographic variables have often been used to differentiate those who are likely to be distracted. However, there are differing conclusions about demographic influences of distracted bicycling and even distraction’s relationship itself to bicycle safety. This study observed bicyclist distraction and safety behaviors, along with possible demographic factors that could predict these behaviors. The authors observed cyclists on the Elizabeth River Trail, a bicyclist/ pedestrian shared trail that covers 10.5 miles of an urban area. During data collection, observers collected data about the direction the cyclist was traveling, if they were with children, if they were wearing a helmet, if a distraction was present, and the type of distraction, if applicable (e.g., hand­held use of mobile phones, wearing headphones, eating, drinking). Observers also recorded participants’ perceived sex and estimated age. Frequency analyses revealed that 55.4% of 426 cyclists were not helmeted and 30.0% were distracted. The most common distraction was wearing headphones (19.5% of total cyclists observed). No significant relationship was found between helmet use and distraction. Younger cyclists were more likely than older cyclists to not wear a helmet and be distracted. These findings show a high prevalence of behaviors that may impact safety on designated cycling paths. The significant number of cyclists without a helmet and being distracted should create concern for potential injury risks to bicyclists on trails being built within urban areas if those bicyclists were to crash.

Keywords: bicyclist, helmet use, bicycle trail, distracted cycling

Bicycling has been increasingly encouraged as a more environmentally friendly, healthy, and fun mode of transportation (Maas et al., 2022; Wild & Woodward, 2021). As more people take advantage of bike trails, shared bike lanes, and bike sharing systems (Maas et al., 2022), focusing on their distractions becomes even more important. Specific distractions previously identified in observational studies involving bicyclists have included hand­held use of mobile phones, wearing headphones/

earbuds, eating, drinking, and smoking (Ethan et al., 2016; Huemer et al., 2022). The most common type of distraction identified for bicyclists are activities that divert mental resources from aural attention, such as listening to music or using headphones (Stavrinos et al., 2018). These types of distractions are often associated with delayed response times, which increases safety risks and possibility of crashes (Stavrinos et al., 2018). Yet, there does not appear to be a clear consensus regarding how

Diversity badge earned for conducting research focusing on aspects of diversity.

demographic factors, such as age and gender, influence the likelihood of a bicyclist being distracted. Age also seems to influence the likelihood of distraction; however, there has not been a consensus in terms of which age group was more likely to be distracted. One study completed in Germany found that younger cyclists were more likely to be distracted (Huemer et al., 2022), but one in Latin America found the opposite effect (Useche et al., 2019).

Age differences in distraction behaviors may be due to multitasking proficiency decreasing with age (Rumschlag et al., 2015). Therefore, younger vehicle operators may be more comfortable with being engaged in multiple tasks (e.g., bicycling and looking at their phones) than older people. Younger people in general are more comfortable with technology that could be considered distractions, such as headphones or cell phones. It is possible that this comfortability could lead to more young cyclists participating in risky bicycling behaviors, especially when technology is involved. Gender differences in transportation research seem to be more difficult to explain theoretically. Song et al. (2021) reported that increased driving experience predicted decreased risky driving behavior, regardless of gender. Whereas Rhodes and Pivik (2011) found that male drivers tended to have more positive affect related to risky behaviors, which increased their likelihood to engage in risky driving behaviors. It is important to evaluate risk factors that have been identified in driving research in other transportation contexts because predictors of decisions made when driving a vehicle are the same predictors that could relate to decisions made on a bicycle.

There also seems to be differing findings regarding how age and sex affect bicycle safety behaviors, such as helmet use. Researchers have shown that helmet use consistently reduces head injuries and death as a result of bicycle crashes, especially in pediatric populations (Howard et al., 2022). Helmet use seems to differ by gender and geographic location and seems to have a negative relationship with presence of distractions (Basch et al., 2014; Ethan et al., 2016; Huemer et al., 2022; Kilinc & Kartal, 2022; Osberg et al., 1998). Whereas Huemer et al. (2022) and Kilinc & Kartal (2022) both observed that male cyclists were less likely to wear a helmet, Ethan et al. (2016) and Basch et al. (2014) found that women were less likely to engage in that safety behavior.

Observational studies have also suggested that location plays a role, with big cities in the United States, such as Boston and New York City, having a higher percentage of helmet ­ wearers than European cities, such as Paris (Basch et al., 2014; Osberg et al., 1998). In countries with a more established cycling culture, such as European countries, there is more cycling infrastructure

and an increased number of cyclists in general who make the overall cycling experience less risky than it is in North America (Reynolds et al., 2009). This could be an explanation as to why it is less common for cyclists to wear safety gear, such as a helmet, in these locations. It is important to investigate possible predictors for helmet use (e.g., age, gender, and location) in order to structure possible public health initiatives to efficiently encourage safety behaviors in populations that need it the most. The current study served as a preliminary observation of cyclists on an urban shared­use trail to evaluate the prevalence of both distraction and safety behaviors, particularly helmet usage, auditory and tactile distractions (e.g., headphone use, eating, holding items in hand).

Safety researchers have mostly been focused on distracted driving, relative to other modes of transportation, and its relationship with demographic factors, risky behaviors, and crash risk. Observational and naturalistic driving studies evaluating distracted driving, particularly the use of mobile technology, have found that between 23–60% of drivers are distracted at one point during their journey (Calvo et al., 2020; Kidd & Chaudhary, 2019). The researchers of these studies have also identified that age might be an influence on the likelihood of drivers engaging in secondary tasks. Specifically, drivers under the age of 25 may be more likely to be distracted than those 65 and older (Calvo et al., 2020). Younger drivers are also more likely than older drivers to use mobile technology for a variety of tasks while driving, whereas older drivers tend to use their cell phones primarily for phone conversations (Dozza et al., 2015). Low­demand secondary tasks (e.g., cell phone use, holding a conversation) create significant safety risks while driving, such as increased erratic steering behavior, increased deviation of lateral position, increased lane departures, impaired reaction ability, and decreased situational awareness in simulated driving scenarios (Donmez et al., 2006; Irwin et al., 2014; Karthaus et al., 2019; Kass et al., 2007; Li et al., 2016; Rumschlag et al., 2015). By using a simulated driving scenario where participants were distracted by a simulated “hands­free” phone call, Kass et al. (2007) found that distracted participants were significantly more likely to get in collisions with other vehicles, strike more pedestrians, exceed the speed limit, and drive through more stop signs. These outcome behaviors demonstrate the potential detrimental effects of being distracted while operating a vehicle. However, there is a gap in the literature as to whether distracted cycling is as prevalent on designated cycling paths as distracted driving. Therefore, an exploratory study on cyclist distraction is warranted to gain understanding on whether these behaviors carry over to cycling as well as what predicts distraction behavior in order to mitigate it.

Hutchins and Porter | Bicyclist Helmet Use and Distractions

Bicyclist Helmet Use and Distractions | Hutchins and Porter

Stavrinos et al. (2018) found parallels between how secondary task engagement affects these three types of transportation in their review of distractions in adolescent pedestrians, drivers, and bicyclists. The use of mobile technology impacts both visual and cognitive processes, two systems that are necessary to maintain awareness of potential hazards on the road (Stavrinos et al., 2018). Even low­demand tasks, like conversing on the phone or listening to music, involve considerable risk because they divert mental resources away from the primary task: staying safe on the road (Stavrinos et al., 2018). The evidence of parallel effects of distraction among pedestrians, drivers, and bicyclists creates the need for further investigation into how secondary task engagement affects bicycling behavior and possible demographic factors that predict the likelihood of being distracted on a bicycle.

Much of the literature on distracted bicycling focuses on bicycle–motor vehicle interactions on urban streets, and not much is known about how bicycle safety and distraction operate on designated cycling paths in the United States (Akar et al., 2013; Basch et al., 2014). Even in settings with clear potential danger, including close interactions with motor vehicles, 31.2% of bicyclists in Boston and 20.3% of bicyclists in New York City were observed to be engaging in distraction behaviors (Ethan et al., 2016; Wolfe et al., 2016). Yet, designated trails for pedestrians and cyclists, which have limited interaction with motor vehicles, and therefore less overt potential danger, have remained understudied. There is diverging evidence as to whether crashes are more or less common on these off­road, designated bicycle paths (Hagel et al., 2015; Walker, 2011). Walker (2011) suggested that crashes are more common on off­road paths, whereas Hagel et al. (2015) suggested that cycling on off­road paths decreased

the risk of both crashes and injuries. Some studies have also suggested that distraction is more common on designated cycling paths, thus increasing the safety risk (Huemer et al., 2022). To determine whether urban spaces should be incorporating designated cycling paths into their infrastructure, it is important to first determine the prevalence of safety behaviors associated with environments that have a higher level of perceived safety.

The Elizabeth River Trail (ERT) is a designated, shared bicyclist/pedestrian path in Norfolk, Virginia, that stretches 10.5 miles with minimal motor vehicle interaction. This preliminary observation of cyclists on the ERT sought to evaluate the prevalence of both distraction and safety behaviors, particularly helmet usage, auditory and tactile distractions (e.g., headphone use, eating, holding items in hand). Given that most of the current bicycle and distraction literature offer differing conclusions about demographic factors (Ethan et al., 2016; Huemer et al., 2022), it was important to include data about age and gender, in particular, to assist in the debate. Therefore, this study focused on how age and gender may relate to bicycle safety and distraction on an urban bicycling path.

Method

Settings and Participants

The first author conducted a pilot study for location scouting one week prior to data collection. Seven prospective sites on the ERT were observed for 15 minutes each, recording cyclist volume. Locations were chosen primarily for the volume of cyclists recorded during this pilot study. Proximity to downtown Norfolk, Virginia, interaction with pedestrians/motor vehicles, and unique subpopulations of cyclists were also considered when

TABLE 1

Demographic Information by Site

Note. All values in the table are given as percentages by site.

selecting the sites. Four sites were selected: the Brambleton Bridge Exit, the Botetourt Footbridge, Jeff Robertson Park, and the Old Dominion University (ODU) campus. The times of day that seemed to yield the highest volume of cyclists were 12:00–3:00 p.m. on weekends and 3:00–6:00 p.m. on weekdays, also determined by the pilot study.

The sites chosen for observation were clustered primarily on the southern half of the ERT due to their increased cyclist volume and proximity to downtown Norfolk. Both the Brambleton Bridge Exit and the Botetourt Footbridge serve as popular routes to get to downtown Norfolk, possibly accounting for their moderate to high cyclist volume. Jeff Robertson Park was chosen for its potentially higher volume of family cyclist traffic, whereas the ODU campus provided a potentially higher number of college­aged and family bicyclists.

Observers recorded what they perceived to be the demographics of the participants, focusing on age and gender. Multiple age groups were observed: child(ages 9 and under), teen (ages 10–17), young adult (ages 18–35), middle­aged adult (ages 36–55), and senior (ages 56 and above). However, the estimated age of the participants observed in the study was spread unevenly (e.g., cyclists 17 years and younger made up 0.09% of the total sample). Therefore, for analysis purposes, the data were recategorized into two age groups: 35 years and younger and 36 years and older, making up 50.9% and 49.1% of the sample, respectively. Observed gender of the cyclists also provided a large disparity between categories, with 74.4% male and 25.6% female cyclists observed in the sample. Demographic factors did not significantly differ by location. Age and gender of observed cyclists by location are given in Table 1.

Measures

Data were recorded on paper datasheets designed to capture each case on one line of data. The datasheet recorded the direction the cyclist was traveling, if they were with children, if they were wearing a helmet, if a distraction was present, and the type of distraction (e.g., hand­held use of mobile phones, wearing headphones/ earbuds, eating, drinking, talking), if applicable. Demographic factors that were recorded included number of cyclists in a group (one, two, or three and above), type of clothing worn by the cyclist (whether designated cycling gear or not), perceived gender (male or female), and estimated age (9 and younger, 10–17, 18–35, 36–55, 56 and older).

Procedures

As previously described, the first author conducted a pilot study to determine observation sites and data collection times for the study. Prior data collection,

approval was received from the ODU Institutional Review Board. The work occurred over the course of five weeks, with three days of observation per week. Of those three days, two were weekdays and one was a day of the weekend. On weekdays, data were collected 3:00 to 6:00 p.m.; on weekends they were obtained from 12:00 to 3:00 p.m. Site order was reversed each collection day for counterbalancing. Each site was observed for 30 minutes, from a location on the edge of the trail path being observed. All observers traveled to and from each site by bicycle within 20 minutes in order to begin the next observation period. All data were recorded in real time, according to judgments made by each observer in the moment that the cyclist being observed passed them. In the case of groups of multiple cyclists, observers recorded information about the cyclist who passed by them first to represent the group.

Overall, 426 cyclists were observed over 15 collection days. Among key variables, helmet use, overall distraction, and demographics, there were fewer relationships found than expected. However, as will soon be presented, significant proportions of the observed cyclists were not helmeted, generally distracted, and distracted by headphones specifically.

Results

Reliability

Before presenting key results, information about reliability analyses is important to consider. Two observers worked together on six of the 15 collection days.

Interrater Reliability of Key Variables

Key Variables by Site

Hutchins and Porter | Bicyclist Helmet Use and Distractions

Bicyclist Helmet Use and Distractions | Hutchins and Porter

Although there was always a primary collector from whom data were used in main analyses during all 15 days, on those six days with pairs, one of the collector’s data were used to calculate interrater reliability. Percent agreement and Cohen’s κ estimates were calculated, with all key variables exceeding standard thresholds for reliability (percent agreements were all above 85% and Cohen’s κ values were above .75), as given in Table 2.

Helmet Use, Distractions, and Their Interrelationship With Demographics

A summary of frequency analyses for main variables is given in Table 3. Those analyses showed that 55.4% of 426 cyclists were not helmeted and 30.0% were distracted overall (by any type of distractor). However, helmet use and distraction were not significantly related, χ2(1, n = 426) = 0.76, p = .39, Ф = .04.

Analyses with demographic predictors had mixed results. There were no statistically significant relationships between sex and helmet use or distraction. (for helmet use: χ 2 (1, n = 426) = 2.72, p = .10, Ф = .08; for distraction: χ2(1, n = 426) = 2.67, p = .10, Ф = .08). The relationship between age and gender was also not statistically significant, χ2(1, n = 426) = 0.11, p = .74, Ф = .02. However, younger cyclists (50.9% of total cyclists) were more likely than older cyclists (49.1% of total cyclists) to not wear a helmet, χ2(1, n = 426) = 18.04, p < .001, Ф = .21, and be distracted, χ2(1, n = 426) = 11.15, p = .001, Ф = .16.

Types of Distractions and Location Differences

Although nearly one in three observed cyclists were distracted as mentioned above (30%), headphone use was by far the dominant type of distraction. Of those distracted cyclists, 64.8% of them were wearing headphones, either over­the­ear or in­ear buds. Other types of distraction that were observed but at much lower levels included: using a mobile phone (15.6%), listening to music with a speaker (7.8%), holding an object in their hands (7.0%), talking (6.3%), and drinking (3.1%).

Although the relationship between gender and overall distraction was not statistically significant, gender and headphone use were significantly related, χ2(1, n = 426) = 8.23, p = .004, Ф = .14. Male cyclists were more likely than female cyclists to wear headphones, whether over­the­ear or in­ear buds. Age was also significantly related to headphone use, χ2(1, n = 426) = 5.66, p = .02, Ф = .17. Younger cyclists were more likely than older cyclists to wear headphones.

Finally, helmet use significantly differed by location, χ 2 (3, n = 426) = 13.85, p = .003, Ф = .18. At the Brambleton Bridge Exit and Jeff Robertson Park, there were more cyclists with helmets than expected, whereas the Botetourt Footbridge and ODU campus

locations had fewer helmeted cyclists than expected. This effect was entirely driven by male participants, meaning that female participants had consistent helmet use across all four locations (for male participants: χ2(3, n = 317) = 13.40, p = .004, Ф = .21; for female participants: χ 2 (3, n = 109) = 1.71, p = .634, Ф = .13). However, location and overall distraction were not related, χ2(3, n = 426) = 3.27, p = .35, Ф = .09, nor were location and headphone use, χ2(3, n = 426) = 2.80, p = .42, Ф = .08. The relationship between age and location was not significant overall, χ2 (3, n = 426) = 1.91, p = .59, Ф = .07, or broken down by gender (for male participants: χ2(3, n = 317) = 3.22, p = .36, Ф = .10; for female participants: χ2(3, n = 109) = 0.11, p = .99, Ф = .03).

Discussion

In the current study, 426 cyclists were observed on an urban, shared ­ use trail. A significant proportion of these cyclists were not wearing a helmet (55.4%) and were observed to be in the presence of a distractor (30.0%). The most common distractor observed was headphones, with 64.8% of distracted cyclists wearing either over­the­ear or in­ear buds. Male cyclists were more likely than female cyclists to wear headphones. Chi square analyses revealed that younger cyclists (50.9% of total cyclists) were more likely than older cyclists (49.1% of total cyclists) to not wear a helmet and be distracted. Observation site predicted helmet use; at the Brambleton Bridge Exit and Jeff Robertson Park, there were more cyclists with helmets than expected, whereas the Botetourt Footbridge and ODU campus locations had fewer helmeted cyclists than expected. This relationship was entirely driven by male cyclists. However, there were no statistically significant relationships found between sex and helmet use or distraction when observing the trail for the current study.

Urban shared­use trails are becoming more common, and the setting of this study at the ERT is an exemplar. Trailheads, community art pieces, playgrounds, a glow­in­the­dark section in a public park, and other facilities are being added to the ERT with more to come. All these improvements encourage use of bicycles for recreation and, given that the trail connects business areas with residential neighborhoods, for work. However, bicycling is not without its safety risks. In 2021, the National Highway Traffic Safety Administration found that there were 966 bicyclist fatalities via traffic crashes in the United States. Additionally, Sarmiento et al. (2021) found that, from 2009–2018, there was an estimated 596,972 emergency department visits for bicycle­related traumatic brain injuries. The most common age group associated with both fatalities and emergency department visits were cyclists aged 10–14 years old. Understanding

risks for injury are important for developing parallel safety countermeasures with trail enhancements.

Many observational studies on distracted bicycling have shown the prevalence of these behaviors on roadways, intersections, and strictly urban environments (Ethan et al., 2016; Wolfe et al., 2017). Even in settings with clear potential danger and close interactions with motor vehicles, a significant portion of cyclists were observed to be engaging in distracting behaviors (Ethan et al., 2016; Wolfe et al., 2017). Yet, designated trails for pedestrians and cyclists, which have limited interaction with motor vehicles, and therefore less overt potential danger, have not been a priority in previous literature.

The observed cyclists represent the trail at this moment in time and also serve as good projections of what issues future policies may need to focus on for bicyclist safety. Clearly, the results show there is significant room for improvement with a large proportion of cyclists not wearing helmets and showing at least one distraction behavior.

Although the act of wearing a helmet in and of itself was unrelated to distractions, key bicyclist characteristics were. Being younger (under 35) was an important predictor of riding without a helmet and being distracted overall. Men and women did not differ overall, but for comparisons of wearing headphones, particularly men were more likely to do so than women. Location mattered as well. Observed sites in well­traveled areas had higher helmet use, particularly by male cyclists, regardless of participant age. Perhaps cyclists have a greater awareness for danger in these locations. Perhaps they are more likely to wear a helmet in those environments than in a location that seems more protected from dangerous situations like a college campus, or small pedestrian bridge close to a residential area. This discrepancy still has safety implications because the observation sites are located on the same trail, and it is unlikely that cyclists without safety gear in one area will put a helmet on in the more dangerous sections of the trail.

Very few observational studies have been conducted to evaluate bicycling behavior, and even fewer researchers have observed bicycling behavior on off­road, designated cycling paths. An observational study conducted by Huemer et al. (2022) focused on a comparison of the prevalence of safety and distraction behaviors of bicyclists and e­scooter users on a designated cycling/ pedestrian path, cycling street, and mixed traffic street in Germany. Huemer et al. (2022) observed over 4,000 riders, with 85.5% of them being conventional bicycle riders. In their sample, the researchers found that 17.8% of riders were helmeted and 12.2% of riders were distracted with the highest proportion of distracted riders on the designated cycling/pedestrian path, which is most

comparable to the environment of the ERT (Huemer et al., 2022). Similar to the current study, headphone use was the most common type of distraction, with 6.7% of the total sample wearing headphones or earbuds. Also similar to the current study, younger riders (aged 18–24) were more likely to be distracted and less likely to wear a helmet than older cyclists. Huemer et al. (2022) also found that riders engaging in risky behaviors were less likely to use safety equipment and more likely to be found at­fault in conflicts. Additionally, they identified an influence of gender in both distraction and safety behaviors: men were more likely to be distracted and not wear a helmet than women.

The results of the current study do not fully compare to Huemer et al. (2022), perhaps due to the much smaller sample size, less time spent observing, or the difference in how developed cycling culture is in the United States versus Germany. However, both studies identified a majority of cyclists in the sample being without a helmet (55.4% and 82.2%, respectively), a cause for concern due to how effective helmets are at reducing head injuries and death as a result of bicycling injuries, especially in pediatric populations (Howard et al., 2022). Both studies identified significant proportions of distracted riders (30.0% and 12.2%, respectively) with the most common distraction being the use of headphones/earbuds (19.5% and 6.7%, respectively). This proportion of distracted cyclists is also a concern because of distraction’s relationship with safety risks as shown in Stavrinos et al. (2018) and the findings of Huemer et al. (2022).

Limitations

Compared to other observational studies on bicycling, this study had a smaller sample size and less time spent observing. However, the findings of the current study were similar to findings observed in previous literature that had a larger sample size and spent more time observing.

As with all naturalistic studies, data collected were only observable actions, not how participants reported their decisions or motivations to act in ways that lead to distractions and risks while bicycling. These data reflect the possibility for cognitive distractions to be present, for as reviewed earlier, these behaviors have been demonstrated to have safety risks in driver simulation studies (Donmez et al., 2006; Irwin et al., 2014; Karthaus et al., 2019; Kass et al., 2007; Li et al., 2016; Rumschlag et al., 2015). Due to the nature of an observational study, the authors cannot claim that participants observed were actively being distracted by their observed distractors (e.g., a cyclist could have had headphones in, but was not playing music at the time of observation), but as with all such methodologies, observed behavior provides

Hutchins and Porter | Bicyclist Helmet Use and Distractions

Bicyclist Helmet Use and Distractions | Hutchins and Porter

prevalence of potential safety concerns. Further, this study is not alone in its methods. Previously established observational studies in the literature have used these behaviors (e.g., wearing headphones, holding a mobile phone, holding an object) as indicators of distraction prevalence (Ethan et al., 2016; Huemer et al., 2022; Ortiz et al., 2017; Rumschlag et al., 2015; Schwebel et al., 2022; Wolfe et al., 2016).

Future Directions

These findings provide evidence of possible predictors of unsafe cycling behaviors, which could help to structure policy or infrastructure that may need to be put in place for greater public health. Community programs such as using social marketing techniques to target groups that are at risk, like younger cyclist as suggested by the current study, could help to encourage more safety behaviors. Use of social media platforms or signage along the trail about the risk factors of distraction and not wearing a helmet could help to stimulate positive behavior change (Sarmiento et al., 2021; Wolfe et al., 2016).

Future research on designated cycling/pedestrian paths should focus on how the prevalence of distraction influences traffic rule violations, being at­fault in conflicts, and near­miss or actual crashes with pedestrians and any vehicular cross­traffic. Because the prevalence of distraction and safety behaviors seems to vary based on geographic location, observational studies should be replicated in more regions of Virginia particularly and the United States generally to identify geospatial influencers.

Conclusion

The authors conducted this exploratory observational study to gain a better understanding of what safety issues need to be monitored on designated cycling/ pedestrian paths. Because a majority of cyclists observed were not helmeted, and a meaningful proportion were distracted, possible public health initiatives to encourage safety behaviors should be considered in the advertising of these types of cycling paths. As more urban areas incorporate these types of paths into their infrastructure, it is important to make sure that these environments are well understood, safe, and ready for their users to enjoy.

References

Akar, G., Fischer, N., & Namgung, M. (2013). Bicycling choice and gender case study: The Ohio State University. International Journal of Sustainable Transportation, 7(5), 347–365. https://doi.org/10.1080/15568318.2012.673694

Basch, C. H., Zagnit, E. A., Rajan, S., Ethan, D., & Basch, C. E. (2014). Helmet use among cyclists in New York City. Journal of Community Health, 39(5), 956–958. https://doi.org/10.1007/s10900-014-9836-8

Calvo, J. A., Baldwin, C., & Philips, B. (2020). Effect of age and secondary task engagement on motor vehicle crashes in a naturalistic setting. Journal of Safety Research, 73, 297–302. https://doi.org/10.1016/j.jsr.2020.03.011

Donmez, B., Boyle, L. N., & Lee, J. D. (2006). The impact of distraction mitigation

strategies on driving performance. Human Factors, 48(4), 785–804. https://doi.org/10.1518/001872006779166

Dozza, M., Flannagan, C. A. C., & Sayer, J. R. (2015). Real-world effects of using a phone while driving on lateral and longitudinal control of vehicles. Journal of Safety Research, 55, 81–87. http://dx.doi.org/10.1016/j.jsr.2015.09.005

Ethan, D., Basch, C. H., Johnson, G. D., Hammond, R., Chow, C. M., & Varsos, V. (2016). An analysis of technology-related distracted biking behaviors and helmet use among cyclists in New York City. Journal of Community Health, 41, 138–145. https://doi.org/10.1007/s10900-015-0079-0

Hagel, B. E., Romanow, N. T. R., Enns, N., Williamson, J., & Rowe, B. H. (2015). Severe bicycling injury risk factors in children and adolescents: A case–control study. Accident Analysis and Prevention, 78, 165–172. http://dx.doi.org/10.1016/j.aap.2015.03.002

Howard, K. A., Griffin, S. F., Rolke, L. J., & Sease, K. K. (2021). Factors related to youth self-efficacy for injury prevention bicycle skills. Journal of Safety Research, 79, 94–99. https://doi.org/10.1016/j.jsr.2021.08.009

Huemer, A. K., Banach, E., Bolten, N., Helweg, S., Koch, A. K., & Martin, T. (2022). Secondary task engagement, risk-taking, and safety-related equipment use in German bicycle and e-scooter riders – An observation. Accident Analysis and Prevention, 172 https://doi.org/10.1016/j.aap.2022.106685

Irwin, C., Monement, S., & Desbrow, B. (2015). The influence of drinking, texting, and eating on simulated driving performance. Traffic Injury Prevention, 16(2), 116–123. https://doi.org/10.1080/15389588.2014.920953

Karthaus, M., Wascher, E., Falkenstein, M., & Getzmann, S. (2020). The ability of young, middle-aged and older drivers to inhibit visual and auditory distraction in a driving simulator task. Transportation Research Part F: Traffic Psychology and Behaviour, 68, 272–284. https://doi.org/10.1016/j.trf.2019.11.007

Kass, S. J., Cole, K. S., & Stanny, C. J. (2007). Effects of distraction and experience on situation awareness and simulated driving. Transportation Research Part F: Traffic Psychology and Behaviour, 10(4), 321–329. https://doi.org/10.1016/j.trf.2006.12.002

Kidd, D. G., & Chaudhary, N. K. (2019). Changes in the sources of distracted driving among Northern Virginia drivers in 2014 and 2018: A comparison of results from two roadside observation surveys. Journal of Safety Research, 68, 131–138. https://doi.org/10.1016/j.jsr.2018.12.004

Kilinc, E., & Kartal, A. (2022). The frequency of bicycle helmet use among college students and health beliefs for bicycle helmet attitude. Journal of American College Health https://doi.org/10.1080/07448481.2022.2032091

Li, X., Yan, X., Wu, J., Radwan, E., & Zhang, Y. (2016). A rear-end collision risk assessment model based on drivers’ collision avoidance process under influences of cell phone use and gender—A driving simulator based study. Accident Analysis and Prevention, 97, 1–18. https://doi.org/10.1016/j.aap.2016.08.021

Maas, S., Attard, M., & Caruana, M. A. (2022). Motivators and barriers for shared bicycle use in ‘starter’ cycling cities: Evidence from BSS user surveys in three Southern European Island cities. Transportation Research Procedia, 62, 67–74. https://doi.org/10.1016/j.trpro.2022.02.009

Ortiz, N. C., Ramnarayan, M., & Mizenko, K. (2017). Distraction and road user behavior: An observational pilot study across intersections in Washington, D.C. Road Danger Reduction, 7, 13–22. https://doi.org/10.1016/j.jth.2017.05.362

Osberg, J. S., Stiles, S. C., & Asare, O. K. (1998). Bicycle safety behavior in Paris and Boston. Accident Analysis and Prevention, 30(5), 679–687. https://doi.org/10.1016/s0001-4575(97)00097-3

Reynolds, C. C., Harris, M. A., Teschke, K., Cripton, P. A., & Winters, M. (2009). The impact of transportation infrastructure on bicycling injuries and crashes: A review of the literature. Environmental Health, 8(47). https://doi.org/10.1186/1476-069X-8-47

Rhodes, N., & Pivik, K. (2011). Age and gender differences in risky driving: The roles of positive affect and risk perception. Accident Analysis and Prevention, 43, 923–931. https://doi.org/10.1016/j.aap.2010.11.015

Rumschlag, G., Palumbo, T., Martin, A., Head, D., George, R., & Commissaris, R. L. (2015). The effects of texting on driving performance in a driving simulator: The influence of driver age. Accident Analysis & Prevention, 74, 145–149. https://doi.org/10.1016/j.aap.2014.10.009

Sarmiento, K., Haileyesus, T., Waltzman, D., & Daugherty, J. (2021, May 14). Emergency department visits for bicycle-related traumatic brain injuries among children and adults—United States, 2009–2018. Centers for Disease Control and Prevention.

Schwebel, D. C., Canter, M. F., Hasan, R., Griffin, R., White, T. R., & Johnston, A. (2022). Distracted pedestrian behavior: An observational study of risk by situational environments. Traffic Injury Prevention, 23(6), 346–351.

https://doi.org/10.1080/15389588.2022.2076282

Stavrinos, D., Pope, C. N., Shen, J., & Schwebel, D. C. (2018). Distracted walking, bicycling, and driving: Systematic review and meta-analysis of mobile technology and youth crash risk. Child Development, 89(1), 118–128. https://doi.org/10.1111/cdev.12827

Song, X., Yin, Y., Cao, H., Zhao, S., Li, M., & Yi, B. (2021). The mediating effect of driver characteristics on risky driving behaviors moderated by gender, and the classification model of driver’s driving risk. Accident Analysis & Prevention, 153, 106038. https://doi.org/10.1016/j.aap.2021.106038

Useche, S. A., Alonso, F., Sanmartin, J., Montoro, L. V., & Cendales, B. (2019). Wellbeing, behavioral patterns and cycling crashes of different age groups in Latin America: Are aging adults the safest cyclists? PLoS ONE, 14(8). https://doi.org/10.1371/journal.pone.0221864

Walker, I. (2011). Bicyclists. In B. Porter (Ed.), Handbook of traffic psychology (pp. 367–373). Elsevier Inc. https://doi.org/10.1016/C2009-0-01975-8

Wild, K., & Woodward, A. (2021). The bicycle as ‘constructive hope’: Children,

Hutchins and Porter | Bicyclist Helmet Use and Distractions

climate and active transport. Journal of Paediatrics and Child Health, 57(11), 1785–1788. https://doi.org/10.1111/jpc.15754

Wolfe, E. S., Arabian, S. S., Breeze, J. L., & Salzler, M. J. (2016). Distracted biking: An observational study. Journal of Trauma Nursing, 23(2), 65–70. https://doi.org/10.1097/JTN.0000000000000188

Author Note

Camryn H. Hutchins https://orcid.org/0000­0003­4492­6166

Bryan E. Porter https://orcid.org/0000­0002­3587­3624

This study was fully funded by the National Science Foundation [grant number: 1949760] for the Old Dominion University Research Experience for Undergraduates in Transportation Science. Correspondence concerning this article should be addressed to Bryan E. Porter, The Graduate School, Old Dominion University, 2102 Monarch Hall, Norfolk, VA, 23504, United States. Email: bporter@odu.edu

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

WINTER 2023

OF

Investigation of the Relationship Between Perceived Mental Workload and Chronic Pain

Kayli N. Colpitts, Jennifer L. Gibson Dias*, Thomas J. Faulkenberry*, and Amber L. Harris Bozer* Department of Psychological Sciences, Tarleton State University

ABSTRACT. Chronic pain is experienced by 1 in 5 adults in the United States and is often accompanied by fatigue, poor sleep quality, and psychological symptoms. These negative factors related to chronic pain are also associated with increased mental workload, particularly in the workplace. Mental workload refers to the amount of cognitive effort used by an individual to complete a task(s). The present study was designed to assess the relationship between chronic pain and perceived mental workload. Participants with and without chronic pain completed 4 variations of an N­back task (used to induce perceived mental workload at increasing levels). After completing each variation of the N­back task, participants completed the NASA­Task Load Index (TLX), a validated subjective measure of perceived workload. A mixed ­ repeated ­ measures ANOVA and Bayesian mixed­repeated­measures ANOVA were computed to assess the impact of chronic pain on perceived mental workload (as evaluated by NASA­TLX). There was an observed interaction effect of workload and chronic pain, F(3, 135) = 4.72, p = .004, η2 = .33, highlighting that potential increases in workload may affect individuals with chronic pain at a greater magnitude than individuals with no chronic pain. This relationship is important to understanding and mitigating the negative symptoms of chronic pain. Future studies should be completed to investigate further the relationship between chronic pain and mental workload, including electrophysiological measures (to assess workload more deeply and correlate with cortical activities) and measures of fatigue (to assess fatigue’s role in the relationship).

Keywords: chronic pain, mental workload, cognitive workload, NASA­Task Load Index, N­Back

Chronic pain, defined as pain lasting 12 weeks or longer, affects one in five U.S. adults (Dahlhamer et al., 2018; Treede et al., 2015). However, the presence of chronic pain does not exempt individuals from activities required for daily living. Individuals with chronic pain may experience strain to meet expectations at work, leaving about 72% of chronic pain sufferers unemployed (Dahlhamer et al., 2018). Approximately

20% of U.S. adults are estimated to experience chronic pain based on data from 2016 (Dahlhamer et al., 2018). Chronic pain brings forth various additional challenges that an individual must face. Fatigue, stress, and poor sleep quality have been reported in individuals experiencing chronic pain (Finan et al., 2013). These negative consequences can impact the day­to­day life of the individual, subsequently causing them to miss

Diversity badge earned for conducting research focusing on aspects of diversity.

work, miss time with family, spend additional money on treatments and medications, and experience an overall decrease in quality of life (Dahlhamer et al., 2018; de la Vega et al., 2018). Sleep quality has a bidirectional relationship with chronic pain (Bonvanie, 2016; de la Vega et al., 2018; Finan et al., 2013). Poor sleep quality and sleep issues significantly predict chronic pain onset as well as produce a long­term increase in pain severity (Bonvanie, 2016). Bonvanie (2016) observed the relationship between chronic pain and sleep issues to be mediated by fatigue and psychological symptoms, such as depression and anxiety.

In addition to the relationship with sleep, pain also has a heavily investigated relationship with fatigue (Aaron & Buchwald, 2003; de la Vega et al., 2018). Fatigue refers to the mental and physical tiredness often experienced by individuals, sometimes due to a particularly extraneous activity and sometimes with no specific identifiable cause (Aaron & Buchwald, 2003). As their names suggest, physical fatigue is caused by a high physical demand (e.g., exercise or completing a task quickly), and mental fatigue is caused by a high mental demand (e.g., complex arithmetic or quick recall). Although different in origin, mental and physical fatigue effects often produce similar results, such as slow reaction times and increased errors (Torres et al., 2016). Overall, fatigue is also associated with increases in sleep issues, physical disability, and depressive symptoms, all of which are symptoms of chronic pain (de la Vega et al., 2018).

In addition to fatigue, chronic pain may impact cognitive workload, which refers to the amount of mental effort an individual uses to complete a task or tasks (Gao et al., 2013; Xie & Salvendy, 2000). Understanding challenges related to workload is crucial in many contexts, such as driving (Castro et al., 2019) and piloting (Blanco et al., 2018). Cognitive workload is a measure often used in workplace settings to assess whether individuals are working at a level that is efficient and maximizes productivity (Blanco et al., 2018; Castro et al., 2019). A higher workload level indicates that individuals may be overloaded and thus perform tasks in a less than optimal state, which may produce errors (Castro et al., 2019). Workload can be affected by several factors, including sleep, stress, and fatigue (Aaron & Buchwald, 2003; Bonvanie, 2016; de la Vega et al., 2018; Ericsson et al., 2002; Finan et al., 2013; Gao et al., 2013; Naughton et al., 2007; Torres et al., 2016). Poor sleep quality and the presence of fatigue results in many errors and low productivity, indicating less than optimal workload levels (Lerman et al., 2012).

Mental workload is mediated by sleep quality in individuals with musculoskeletal disorders (Heidarimoghadam et al., 2019). Occupational fatigue

has been found to have a positive correlation with mental workload within an administrative work setting (Sartang et al., 2018). Furthermore, workload is one of the most crucial factors resulting in fatigue (Lerman et al., 2012). It has been suggested that fatigue may be responsible for workers losing concentration, making mistakes, falling asleep, and having a decreased reaction time (Torres et al., 2016). The evidence would suggest a clearly bidirectional relationship between workload levels and fatigue.

To assess workload levels, four techniques are typically used: primary task measures, secondary task measures, subjective measures, and physiological measures (Meshkati et al., 1995). Primary tasks are simply the task completed by the individual. For instance, when assessing the workload of an individual completing a written test, the score of the written test should be utilized (Hicks & Wierwille, 1979; Meshkati et al., 1995).). Secondary tasks are tasks an individual is asked to complete while maintaining success in a primary task (Meshkati et al., 1995). If the primary task is maintained while the secondary task is also maintained, the primary task is then assumed to require a lower workload. If the primary task becomes too difficult to maintain due to the presence of the secondary task, the primary task is then assumed to require a higher workload (Knowles, 1963; Meshkati et al., 1995). Subjective measures of the task refer to the individual completing the task’s subjective review about the difficulty of the task. Physiological measures elucidate changes in how the body is functioning, such as those measured with the electroencephalogram and galvanic skin conductance responses (Charles & Nixon, 2019; Meshkati et al., 1995).

It seems that the deleterious effects of chronic pain are known to affect cognitive workload, however, little research has investigated the direct relationship between chronic pain and workload. Although these two concepts have been heavily studied independently, a literature search conducted using the APA PsycINFO Database conducted in 2023 revealed the novelty of studying them in conjunction with one another. Although the term “chronic pain” yielded 27,512 hits, and workload also yielded a large number of hits (“mental workload” = 1,214 hits & “cognitive workload” = 578 hits), the combination of the terms “chronic pain” and “workload” only yielded 457 hits. Few other studies have detailed experimental designs specifically addressing the relationship between chronic pain and mental workload. Many of the studies referred to workload of physicians (patient load) or instructors (course load/ student enrollment), which are fundamentally different workload constructs.

One study did investigate the effects of stressinduced workloads on individuals with and without

Colpitts, Dias, Faulkenberry, and Bozer | Mental Workload and Chronic Pain

Mental Workload and Chronic Pain | Colpitts, Dias, Faulkenberry, and Bozer

musculoskeletal pain (Wang, 2012). Specifically, Wang sought to determine differences in performance, psychological, and physiological changes in individuals with and without neck pain during computer tasks with varying mental and physical workload requirements. The data showed the pain group to have demonstrated increased heart rate variability and increased feelings of anxiety and discomfort in response to tasks. The neck pain group also had overall poorer performance on tasks as measured by correct and total trials as well as response times. Overall, Wang (2012) provided evidence for a relationship between the presence of pain and workload. For the current study, we aimed to provide support and a basis for future research directly linking the effects of chronic pain to outcomes on workload. Further understanding the relationship between chronic pain and workload may lead to better mitigation of issues faced by individuals with chronic pain. To that end, we designed the current study to investigate whether individuals with chronic pain maintain an equivalent perceived workload level as those without chronic pain when completing cognitive tasks. We hypothesized that the group with chronic pain would demonstrate increased levels of perceived workload (as measured by the NASA­Task Load Index; NASA­TLX) in comparison to the group without chronic pain when completing cognitive tasks (N­Back and Dual N­Back). The directional hypothesis that chronic pain individuals would display specifically increased workload levels was based on the logic of chronic pain providing additional sensory experience for the individuals to process. This is based on previous research suggesting a relationship between factors that accompany chronic pain and workload (e.g., fatigue, sleep issues, and psychological symptoms).

Method

13 indicated at least once that they did not update settings and/or complete a task. Participants were between the ages of 18–35 years of age. Optional stopping was utilized to determine the final halt to data collection (as described in Hendriksen et al., 2020). A summary of descriptive statistics is found in Table 1. No identifiable information about participants was collected.

Measures

Participants completed the Short­Form McGill Pain Questionnaire (SF­MPQ) to ensure that the chronic pain group reported significantly different pain scores than the control group (Melzack, 1987). The SF­MPQ has been previously validated (Melzack, 1975). Participants were asked to rate 11 sensory words (throbbing, shooting, stabbing, sharp, cramping, gnawing, hot­burning, aching, heavy, tender, and splitting) and four affective words (tiring­exhausting, sickening, fearful, and punishing­cruel) on a 4­point scale labeled none, mild, moderate, and severe (Melzack, 1987). The SF­MPQ has a question focused on the present pain intensity (PPI) index; participants rate their current pain on a 6­point scale labeled no pain, mild, discomforting, distressing, horrible, and excruciating. The ratings on the entire SF­MPQ were summed to compute a composite score for each participant, in which a higher score indicates more severe pain (Melzack, 1987).

The N­Back task, introduced by Kirchner (1958), can be administered in several formats; however, the concept of the task remains the same (Muele, 2017). A stream of information is presented to the participant (e.g., positions of a marker in a grid, shapes, numbers, letters), and the participant is then asked to report when the target piece of information matches the target presented N trials ago (Kirchner, 1958). In the current study, individuals were asked to identify if the position of a marker in the grid they were viewing matches the one presented N (1 or 2) trials ago.

Participants

Institutional review board approval was obtained prior to data collection at Tarleton State University. A total of 47 participants were included in this study (Mage = 24.45 years, SD = 4.24 years); 21 individuals comprised the group of individuals reporting chronic pain, and 26 individuals were included in the group without chronic pain group. Participants were recruited using a participant pool enrollment management system, email, flyers, and word of mouth. Individuals self­reported the presence of ongoing pain lasting 12 weeks or more (Treede et al., 2015) and completed the Short­Form ­ McGill Pain Questionnaire (Melzack, 1987). Participants were asked between iterations of the N­Back and Dual N­Back tasks if they had correctly updated the task settings and completed the task, and if participants responded no, they were excluded from the study. Of 60 total respondents,

The dual N­Back task, a modified version of the N­Back proposed by Jaeggi et al. (2003), is another way to induce workload. This task follows the same structure as the N­Back task, except participants were presented with a stream of two types of information (e.g., positions of a marker in a grid and shape). The original dual N­Back utilized audio­visual cues, specifically audio recordings of consonants and various random shapes (Jaeggi et al., 2003). Participants in the present study identified if the marker was in the same position on the grid as it was N (1 or 2) trials ago and if the shape of the marker matches the shape presented N (1 or 2) trials ago. The decision to use two varying visual cues was made to eliminate the need for speakers and audio calibration as respondents completed the task on a personally accessed computer. Studies have indicated that the processes involved with the tasks are largely robust regardless of

the material used (type of stimuli; Jaeggi et al., 2010).

Specifically, one group sought to define activation differences when using differing stimuli for the N­Back task (Nystrom et al., 2000). It was shown that, whether letters, spatial locations, or shapes were used, reaction times, correct response data, and fMRI activation did not significantly vary. Although additional stimuli specific brain regions may become activated based on stimuli type, the areas thought to be specifically involved in working memory remain the same regardless of the stimuli presented in the N­Back and dual N­Back (Nystrom et al., 2000). Although the support for variations of the task not changing the effects induced by the task are predominantly discussed in the single N­Back task, we only utilized the N­Back and dual N­Back task in the present study to induce increasing workload. To facilitate replicability, screenshots of settings and example trials for both the N­Back and Dual N­Back can be found on Open Science Framework (Colpitts, 2023).

The NASA­Task Load Index (TLX), proposed by Hart and Staveland(1988), is a common and validated subjective measure for assessing perceived workload (Gao et al., 2013; Sartang et al., 2018; Torres et al., 2016). The instrument asks participants to respond to 15 pairwise comparisons of the six different subscales: Mental Demand, Physical Demand, Temporal Demand, Performance, Effort, and Frustration. The pairwise comparison simply asks which of the two presented factors (e.g., Mental Demand versus Physical Demand) was more important to their experience (Sartang et al., 2018). The pairwise comparisons were collected only on the initial administration of the NASA­TLX. Next, participants responded via a scale (labeled either low to high or poor to good) as to what level of each subscale best represents their experience completing the tasks (Sartang et al., 2018).

Procedure

The study was administered via Google Forms. After providing informed consent, participants completed the

demographic questions, the pain self­report question, and the SF­MPQ (Melzack, 1987). All N­Back trials were completed on BrainScale.net; trials were preceded with specific instructions as to how to complete the task (Dual N­Back Training). The initial N­Back block consisted of 21 practice trials of the N­Back task with N set equal to 1, 24 practice trials of the N­Back task with N set equal to 2, 21 practice trials of the dual N ­ Back task with N set equal to 1, and finally, 24 practice trials of the dual N­Back with N set equal to 2. The purpose of the practice trials was to ensure participants could understand the expectations of the task before beginning the task. Participants began the N­Back block after completing the practice block of the N­Back tasks. After the first 21 trials of the N­Back task with N set equal to 1, the participants completed the initial NASA­TLX, including the pairwise comparisons. The next block of the N­Back task consisted of 24 trials, however, with N set equal to 2. Once again, the participants completed the NASA­TLX, but this time without pairwise comparisons. After completing the different levels of the N­Back and associated NASA­TLX administrations, the participants completed 21 trials of the dual N­Back task with N set equal to 1, followed by the NASA­TLX. Then, participants completed 24 trials of the last dual N­Back task with N set equal to 2, followed by a final NASATLX administration. A summary of the procedures is available in Figure 1.

Data Analyses

G*Power software was used to compute a power analysis with the following parameters: .95 power, an expected effect size of 0.25, and an alpha criterion level of .05 (Cohen, 1988; Faul et al., 2007; Lakens, 2013) for the largest statistical model, yielding a suggestion of a minimum of 36 participants. Data are reported as mean ± SD. All responses were exported into a GoogleSheet where they were scored. All data analyses were completed with JASP software (JASP Team, 2020) and SPSS software (IBM Corp., 2017). We first conducted a battery

Figure 1

Flow Chart of Methods

Note. NASA-TLX = NASA-Task Load Index

of frequentist tests. An independent­samples t test was used to assess the difference in age, and chi ­ square tests were computed to test for demographic differences (gender and race) between the groups reporting chronic pain or no pain. An independent ­ samples t test was used to assess the difference in SF­MPQ scores between the chronic pain and no chronic pain groups. A mixed­repeated­measures ANOVA was computed to examine the impact of chronic pain and experimentally induced workload on perceived workload as measured by the NASA­TLX. We used Bonferroni post hoc tests to further examine the pairwise differences in perceived workload between groups (chronic pain and no chronic pain) for each level of the N­Back task. For all frequentist tests, the alpha criterion was set at α=.05. .

In addition to the frequentist tests, we supplemented our analyses with Bayesian hypothesis tests. Compared to frequentist testing, Bayesian testing offers many advantages, including the ability to quantify evidence on a continuous scale (e.g., by the Bayes factor; Kass & Raftery, 1995) as well as the ability to coherently assess the presence of (and quantify the evidence for) null effects (Faulkenberry et al., 2020). Specifically, we used the JASP Summary Statistics module (Ly et al., 2018) to compute a Bayes factor from the summary statistics obtained in each of our independent­samples t tests. The Bayes factor expresses the extent to which

the observed data are more likely under the alternative hypothesis compared to the null hypothesis (or vice versa). Further, we used a Bayesian mixed­repeatedmeasures ANOVA (Rouder et al., 2012) to assess the impact of chronic pain and experimentally induced workload on participants’ perceived mental workload as measured by the NASA­TLX. Bayesian mixed­repeatedmeasures ANOVA works by building competing models, which contain factorial combinations of each main effect. Posterior probabilities are then computed for each of the competing models. For our results to mirror those of conventional factorial analysis of variance, we used Bayesian model averaging (Hinne et al., 2020) to compute an inclusion Bayes factor for each main effect and interaction. Inclusion Bayes factors indicate the evidence for including each main effect and interaction by computing the multiplicative change in odds from prior to posterior when considering all models including the effect compared to all models not including the effect. Additionally, we performed optional stopping (Hendriksen et al., 2020; Rouder, 2014), where we monitored evidence by computing an inclusion Bayes factor for each effect after each participant, giving us the

SF-MPQ Data by Group

Note *significantly different at p < . 05.

Note. SF-MPQ = short-form McGill Pain Questionnaire.

NASA-TLX Ratings Data by Group

Note. NASA-TLX = NASA-Task Load Index. The ANOVA revealed a main effect of N-back task level, F(3, 135) = 6.47, p < .001, ηp2 = .17, no main effect of pain group, F(1, 45) = 0.12, p = .73, ηp2 = .002, and a significant interaction effect between N-back task level and pain group, F(3, 135) = 4.72, p = .004, ηp2 = .10. The Bonferroni post hoc analysis revealed that individuals with chronic pain demonstrated a significantly higher perceived workload on the 2-back than 1-back.

ability to monitor the accumulating evidence for each effect over time.

Results

The purpose of the study was to assess the relationship, if any, between chronic pain and mental workload. The hypothesis was that individuals reporting chronic pain would report higher subjective cognitive workload (as measured by the NASA­TLX) on the progressively more difficult variations of the N­Back task when compared to individuals without chronic pain.

First, we tested for differences in demographic variables (age, gender, and race) between the groups with chronic pain and without pain. For age, an independentsamples t test revealed a significant difference between groups, t(45) = 2.62, p = .01, d = 0.31. Participants in the chronic pain group (M = 26.1 years) were on average 3.0 years older than participants in the no pain group (M = 23.1 years). A Bayesian independent­samples t test confirmed that these data were BF10 = 4.24 times more likely under the alternative hypothesis. A chi­squared test indicated no difference between the groups with respect to gender, χ2 (1, N = 47) = 0.69, p = .41, or race, χ2 (5, N = 47) = 3.20, p = .67. Demographics of participants split by group (chronic pain or no chronic pain) are shown in Table 1.

The independent ­ samples t test to compare SF­MPQ scores by group revealed that individuals with chronic pain (13.52 ± 7.21) scored significantly higher than the no chronic pain group (5.04 ± 7.71), t(45) = 3.86, p < .001, d = 0.33. These data were highly evidential for the alternative hypothesis; we obtained a Bayes factor of BF10 = 73.33, indicating that the observed data were 73 times more likely under the alternative than the null. Descriptive statistics of the NASA­TLX scores as well as SF­MPQ scores are shown in Table 2.

The mixed­repeated­measures ANOVA was computed to compare the workload ratings as a function of N­Back task level (four levels manipulated within subjects) and pain group (two levels between subjects). The ANOVA revealed a main effect of N­Back task level, F(3, 135) = 6.47, p < .001, ηp2 = .17, but no main effect of pain group, F(1, 45) = 0.12, p = .73, ηp2 = .002. Critically, there was a significant interaction effect between N­Back task level and pain group, F(3, 135) = 4.72, p = .004, ηp2 = .10, such that the chronic pain groups subjectively rated the 2­back, dual 1­back, and dual 2­back to require more workload (see Table 3 and Figure 2).

The evidence for these effects on perceived workload was indexed with a Bayesian mixed­repeated­measures ANOVA. The Bayesian ANOVA (Faulkenberry et al., 2020; Rouder et al., 2012) builds five competing models for the observed data: (a) a null model, (b) a model

including only a main effect of pain group, (c) a model including only a main effect of N­Back task level, (d) an additive model containing both main effects, and (e) an interactive model containing both main effects and their interaction. After observing data, the most likely of these models was Model 5 (the interactive model), with a posterior probability of Bayesian model averaging revealed that the data were approximately 25 times more likely under models including the main effect of N­Back task level than under any models not including the main effect of N­Back task level (BFinclusion = 25.30). The data were approximately three times more likely under models not including a main effect of chronic pain group than under models including the main effect of chronic pain (BFexclusion = 2.98). Last, the data were approximately nine times more likely under models including an interaction between N­Back task level and chronic pain group than under models not including this interaction (BFinclusion = 8.95).

Discussion

The purpose of this study was to investigate any relationship that may exist between mental workload and chronic pain. Understanding this relationship may lead to better mitigation of the effects of chronic pain experienced by individuals in the workplace. We hypothesized that the group with chronic pain would demonstrate increased levels of perceived workload (as measured by the NASATLX) compared to the group without chronic pain when

Note. NASA-TLX = NASA-Task Load Index. Error bars represent standard deviation. The ANOVA revealed a main effect of N-back task level, F(3, 135) = 6.47, p < .001, ηp2 = .17, no main effect of pain group, between N-back task level and pain group, F(3, 135) = 4.72, p=.004, ηp2 = .10. The Bonferroni post hoc analysis revealed that individuals with chronic pain demonstrated a significantly higher perceived workload on the 2-back than 1-back.

Colpitts, Dias, Faulkenberry, and Bozer | Mental Workload and Chronic Pain

Mental Workload and Chronic Pain | Colpitts, Dias, Faulkenberry, and Bozer

completing cognitive tasks (N­Back and Dual N­Back). The data demonstrate that all participants did find the variations of the N­Back task to require progressively higher workload. The only exception is that the dual 1­back was not reported to be more difficult than the 2­back. This demonstrates the utility of the N­Back to induce increasing workload among participants (Jaeggi et al., 2003; Jaeggi et al., 2010; Kirchner, 1958; Muele, 2017). There was not a main effect of chronic pain observed, however, the interaction effect of workload and chronic pain highlights that potential increases in workload may affect individuals with chronic pain at a greater magnitude than individuals with no chronic pain.

The chronic pain and no chronic pain groups did not have any significant differences for gender or race. We observed that the chronic pain group was approximately 3 years older than the no pain group. However, we believe that combined with the invariance in the other two demographic variables, there is no cause for concern that the group of individuals reporting chronic pain and those reporting no chronic pain were significantly different on demographics. The significance of the t test comparing SF­MPQ scores across groups demonstrates that the chronic pain group did experience higher levels of pain than the no chronic pain group. This is vital to the study because it was imperative that the individuals in the chronic pain group would score higher on a validated pain scale to confirm they had more pain than the no chronic pain group (Melzack, 1987). This allows for the relationship between chronic pain and workload to be assessed.

The variations of the N­Back were utilized in the study to provide an experimentally manipulated induction of workload at four increasing levels of difficulty, and the observed main effect of N­Back level in the mixed ANOVA confirmed that the perceived workload levels were indeed different (Jaeggi et al., 2003; Jaeggi et al., 2010; Kirchner, 1958; Muele, 2017). Critically, the mixed ANOVA also demonstrated an interaction between chronic pain and induced workload level during the four N ­ Back tasks. However, post hoc tests are needed to understand the specifics of the relationship. The Bonferroni post hoc analysis revealed that the group of individuals who did not report chronic pain had no significant differences in workload ratings between the different N­Back levels.

In contrast, the analysis revealed differences in workload ratings between N ­ Back levels within the chronic pain group. Specifically, the chronic pain group reported lower levels of perceived workload (as measured by the NASA­TLX) on the 1­back level than on the 2­back, dual 1­back, or dual 2­back. Individuals with chronic pain demonstrated a significantly higher

perceived workload on the 2­back than 1­back, but considerably lower levels on the dual 1­back when compared to the 2­back. This is likely because, although there were now two streams of information to recall (shape and position), the participants were only being asked to recall targets from one trial ago on the dual 1­back as opposed to only having to recall position from two trials ago. Lastly, individuals with chronic pain reported the highest perceived workload levels on the dual 2­back than any level of the N ­ Back task. These results demonstrate that individuals with chronic pain found the levels of N­Back to require more perceived mental workload, thus providing support for our hypothesis. Individuals in the group not reporting chronic pain did not display significantly higher perceived workload levels on any N­Back level over another.

One unique component of our analysis was the use of Bayesian analysis of variance to assess the evidence for each effect that we reported from the traditional (frequentist) analyses. A Bayesian ANOVA compares several models and compares each model to the best model. For example, a model containing only a factor for N­Back task level provides insight to how likely the data are under a model that only accounts for the level of the N­Back task. Likewise, a chronic pain only model considers how likely the data are under a model that only accounts for the presence, or lack thereof, of chronic pain. An interactive model considers how likely the data are under a model that accounts for both chronic pain as well as N­Back level, along with the interaction between the two. The Bayesian mixed ANOVA suggests the data are most likely under the interactive model compared to all other models. We then used Bayesian model averaging to separately index the evidence for each possible main effect and interaction. We found that the observed data were substantially more likely under models containing a main effect of N­Back task level and the interaction between this level and the chronic pain group, but no separate main effect of chronic pain. The Bayesian mixed ANOVA aligns well with the reports of the traditional mixed ANOVA and provides evidence to support the hypothesis that individuals with chronic pain present an increased perceived mental workload.

Previous literature has shown a gap in research regarding the relationship between chronic pain and mental workload. Chronic pain symptoms commonly include poor sleep quality, stress, as well as fatigue (Aaron & Buchwald, 2003; Bonvanie, 2016; de la Vega et al., 2018; Finan et al., 2013). Fatigue is also common when individuals are mentally overloaded (Gao et al., 2013). Results generally support the hypothesis because the chronic pain group reported higher workload levels than the no chronic pain group. Additionally, the task

was subjectively rated by the chronic pain group to require significantly more workload as the N ­ Back level increased. This suggests that there is indeed a relationship between chronic pain and mental workload. Furthermore, it is important to the study of chronic pain as it cannot be properly mitigated without understanding the full scope of its effects.

An additional limitation of the current study concerning the demographic data was the higher number of women in the study compared to men. This may be because the data were collected primarily in a Department of Psychological Sciences, and psychology students are predominantly women. This is considered a limitation as the data may not be representative of the larger population. However, it should be noted that women are overrepresented in the clinical chronic pain population compared to men (Dahlhamer et al., 2018; Rickard et al., 2023). Also, pertaining to demographics, the age range of this study was restricted because the study was originally designed to be implemented with electroencephalogram (EEG) components, which normally limit the age range (Hultsch, 2000). It has been shown in the EEG literature that functional connectivity varies with age (Knyazev et al., 2015). The study was approved prior to, yet conducted at the beginning of, the COVID­19 pandemic, so EEG components could not be implemented due to social distancing. A future study will incorporate psychophysiological measures, including EEG and galvanic skin response measures (GSR), and these age restrictions will allow for comparisons across those studies.

Future studies will include electrophysiological measures, including EEG and GSR. EEG literature on mental workload has provided a formula for calculating workload from raw EEG signals. Wang et al. (2019) provided a calculation of workload based on the Holm et al. (2009) study. Averages from all frontal theta power divided by averages from all parietal alpha power equate to a workload score (Holm et al., 2009; Wang et al., 2019). Additionally, including a measure of fatigue may be useful to examine whether fatigue moderates the relationship between chronic pain and mental workload (Aaron & Buchwald, 2003; de la Vega et al., 2018; Lerman et al., 2012; Sartang et al., 2018).

In conclusion, this study provided data that supports some relationship between chronic pain and cognitive workload. The effect observed is such that chronic pain alone does not necessarily make a task more difficult, but instead that increasing the difficulty of a task may affect an individual with chronic pain on a greater scale than individuals without chronic pain. As individuals with chronic pain are rarely, if ever, excused from day­to­day activities and responsibilities due to

their pain, this relationship is important to understand so that aids to mitigate the effect can be implemented.

References

Aaron, L. A., & Buchwald, D. (2003). Chronic diffuse musculoskeletal pain, fibromyalgia and co-morbid unexplained clinical conditions. Best Practice & Research Clinical Rheumatology, 17(4), 563–574. https://doi.org/10.1016/S1521-6942(03)00033-0

Blanco, J. A., Johnson, M. K., Jaquess, K. J., Oh, H., Lo, L. C., Gentili, R. J., & Hatfield, B. D. (2018). Quantifying cognitive workload in simulated flight using passive, dry EEG measurements. IEEE Transactions on Cognitive and Developmental Systems, 10(2), 373–383. https://doi.org/10.1109/TCDS.2016.2628702

Bonvanie, I. J. (2016). Functional somatic symptoms in adolescence and young adulthood: Personal vulnerabilities and external stressors. [Groningen]: Rijksuniversiteit Groningen.

Castro, S. C., Strayer, D. L., Matzke, D., & Heathcote, A. (2019). Cognitive workload measurement and modeling under divided attention. Journal of Experimental Psychology: Human Perception and Performance, 45(6), 826–839. https://doi.org/10.1037/xhp0000638

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates.

Colpitts, K. (2023, June 15). Investigation of the relationship between perceived mental workload and chronic pain. https://.osf.io/rsfm9

Charles, R. L., & Nixon, J. (2019). Measuring mental workload using physiological measures: A systematic review. Applied Ergonomics, 74 221–232. https://doi.org/10.1016/j.apergo.2018.08.028

Dahlhamer, J., Lucas, J., Zelaya, C., Nahin, R., Mackey, S., DeBar, L., Kerns, R., Von Korff, M., Porter, L., & Helmick, C. (2018). Prevalence of chronic pain and high-impact chronic pain among adults — United States, 2016. MMWR. Morbidity and Mortality Weekly Report (Vol. 67). https://doi.org/10.15585/mmwr.mm6736a2

de la Vega, R., Racine, M., Castarlenas, E., Sole, E., Roy, R., Jensen, M. P., Miro, J., & Cane, D. (2018). The role of sleep quality and fatigue on the benefits of an interdisciplinary treatment for adults with chronic pain. Pain Practice, 19(4), 354–362. https://doi.org/10.1111/papr.12746

Dual N-Back Training. (n.d.). BrainScale.net Retrieved June 4, 2020, from https://brainscale.net/dual-N-Back/training

Ericsson, M., Poston, W. S. C., Linder, J., Taylor, J. E., Haddock, C. K., & Foreyt, J. P. (2002). Depression predicts disability in long-term chronic pain patients. Disability and Rehabilitation, 24(6), 334–340. https://doi.org/10.1080/09638280110096241

Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. https://doi.org/10.3758/bf03193146

Faulkenberry, T. J., Ly, A., & Wagenmakers, E.-J. (2020). Bayesian inference in numerical cognition: A tutorial using JASP. Journal of Numerical Cognition, 6(2), 231–259. https://doi.org/10.5964/jnc.v6i2.288

Finan, P. H., Goodin, B. R., & Smith, M. T. (2013). The association of sleep and pain: An update and a path forward. Journal of Pain, 14(12), 1539–1552. https://doi.org/10.1016/j.jpain.2013.08.007

Gao, Q., Wang, Y., Song, F., Li, Z., & Dong, X. (2013). Mental workload measurement for emergency operating procedures in digital nuclear power plants. Ergonomics, 56(7), 1070–1085. https://doi.org/10.1080/00140139.2013.790483

Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. Advances in Psychology, 52, 139–183. https://doi.org/10.1016/S0166-4115(08)62386-9

Heidarimoghadam, R., Saidnia, H., Joudaki, J., Mohammadi, Y., & Babamiri, M. (2019). Does mental workload can lead to musculoskeletal disorders in office workers? Suggest and investigate a path. Cogent Psychology, 6(1), 1664245. https://doi.org/10.1080/23311908.2019.1664205

Hendriksen, A., de Heide, R., & Grunwald, P. (2020). Optional stopping with Bayes Factors: A categorization and extension of folklore results, with an application to invariant situations. Bayesian Analysis, 16(3), 961–989. https://doi.org/10.1214/20-BA1234

Hicks, T. G., & Wierville, W. W. (1979). Comparison of five mental workload assessment procedures in a moving-base driving stimulator. Human Factors, 21(2), 129–143. https://doi.org/10.1177/001872087902100201

Colpitts, Dias, Faulkenberry, and Bozer | Mental Workload and Chronic Pain

WINTER 2023

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

255

Mental Workload and Chronic Pain | Colpitts, Dias, Faulkenberry, and Bozer

Hinne, M., Gronau, Q. F., van den Bergh, D., & Wagenmakers, E.-J. (2020). A conceptual introduction to Bayesian model averaging. Advances in Methods and Practices in Psychological Science, 3(2), 200–215. https://doi.org/10.1177/2515245919898657

Holm, A., Lukander, K., Korpela, J., Sallinen, M., & Müller, K. M. I. (2009). Estimating brain load from the EEG. The Scientific World Journal, 9, 639–651. https://doi.org/10.1100/tsw.2009.83

Hultsch, D. F., MacDonald, S. W., Hunter, M. A., Levy-Bencheton, J., & Strauss, E. (2000). Intraindividual variability in cognitive performance in older adults: Comparison of adults with mild dementia, adults with arthritis, and healthy adults. Neuropsychology, 14(4), 588–598. https://doi.org/10.1037//0894-4105.14.4.588

IBM Corp. (2017). IBM SPSS Statistics for Windows, Version 25.0. IBM Corp.

Jaeggi, S. M., Buschkuehl, M., Perring, W. J., & Meier, B. (2010). The concurrent validity of the N-Back task as a working memory measure. Memory, 18(4), 394–412. https://doi.org/10.1080/09658211003702171

Jaeggi, S. M., Seewer, R., Nirkko, A. C., Eckstein, D., Schroth, G., Groner, R., & Gutbrod, K. (2003). Does excessive memory load attenuate activation in the prefrontal cortex? Load-dependent processing in single and dual tasks: Functional magnetic resonance imaging study. NeuroImage, 19(2), 210–225. https://doi.org/10.1016/S1053-8119(03)00098-3

JASP Team. (2020). JASP (Version 0.12.2) [Computer software].

Kass, R. E., & Raftery, A. E. (1995). Bayes factors. Journal of the American Statistical Association, 90(430), 773–795. https://doi.org/10.1080/01621459.1995.10476572

Kirchner, W. K. (1958). Age differences in short-term retention of rapidly changing information. Journal of Experimental Psychology, 55(4), 352–358. https://doi.org/10.1037/h0043688

Knowles, W. B. (1963). Operator Loading Tasks. Human Factors, 5(2), 109–171.

Knyazev, G. G., Volf, N. V., & Belousova, L. V. (2015). Age-related differences in electroencephalogram connectivity and network topology. Neurobiology of Aging, 36(5), 1849–1859. https://doi.org/10.1016/j.neurobiolaging.2015.02.007

Lakens, D. (2013). Calculating and reporting effect sizes to facilitate cumulative science: A practical primer for t-tests and ANOVAs. Frontiers in Psychology, 4(NOV), 1–12. https://doi.org/10.3389/fpsyg.2013.00863

Lerman, S. E., Eskin, E., Flower, D. J., George, E. C., Gerson, B., Hartenbaum, N., Hursh, S. R., & Moore-Ede M. (2012). Fatigue risk management in the workplace. Journal of Occupational and Environmental Medicine, 54(2), 231–258. https://doi.org/10.1097/JOM.0b013e318247a3b0

Ly, A., Raj, A., Marsman, M., Etz, A., & Wagenmakers, E. J. (2018). Bayesian reanalyses from summary statistics: A guide for academic consumers . Advances in Methods and Practices in Psychological Science, 3, 367–374. https://doi.org/10.1177/2515245918779348

Melzack R. (1975). The McGill Pain Questionnaire: Major properties and scoring methods. Pain, 1(3), 277–299. https://doi.org/10.1016/0304-3959(75)90044-5

Melzack R. (1987). The short-form McGill Pain Questionnaire. Pain, 30(2), 191–197. https://doi.org/10.1016/0304-3959(87)91074-8

Meshkati, N., Hancock, P. A., Rahimi, M., & Dawes, S. A. (1995). Techniques in mental workload assessment. In J. R. Wilson & E. G. Corlett (Eds.), Evaluation of human work: A practical ergonomics methodology. (2nd ed., pp. 749–782). CRC Press.

Muele, A. (2017). Reporting and interpreting working memory performance in

N-Back tasks. Frontiers in Psychology, 8 https://doi.org/10.3389/fpsyg.2017.00352

Naughton, F., Ashworth, P., & Skevington, S. M. (2007). Does sleep quality predict pain-related disability in chronic pain patients? The mediating roles of depression and pain severity. Pain, 127(3), 243–252. https://doi.org/10.1016/j.pain.2006.08.019

Nystrom, L. E., Braver, T. S., Sabb, F. R., Delgado, M. R., Noll, D. C., & Cohen, J. D. (2000). Working memory for letters, shapes, and locations: fMRI evidence against stimulus-based regional organization in human prefrontal cortex. NeuroImage, 11(5), 424–446. https://doi.org/10.1006/nimg.2000.0572

Rickard, S. M., Strahan, A. E., Schmit, K. M., & Guy Jr., G. P. (2023). Chronic pain among adults – United States, 2019–2021. MMWR Morbidity and Mortality Weekly Report, 72(15), 379–385. https://doi.org/10.15585/mmwr.mm7215a1

Rouder, J. N. (2014). Optional stopping: No problem for Bayesians. Psychonomic Bulletin & Review, 21, 301–308. https://doi.org/10.3758/s13423-014-0595-4

Rouder, J. N., Morey, R. D., Speckman, P. L., & Province, J. M. (2012). Default Bayes factors for ANOVA designs. Journal of Mathematical Psychology, 56(5), 356–374. https://doi.org/10.1016/j.jmp.2012.08.001

Sartang, A. G., Haghshanas, B., Habibi, E., & Abedi, M. (2018). The investigation relationship between mental workload and occupational fatigue in the administrative staffs of a communications service company. Iranian Journal of Health, Safety & Environment, 6(1), 1221–1225.

Torres, Y., Nadeau, S., & Morency, F. (2016). Study of fatigue and workload among aircraft de-icing technicians. Occupational Ergonomics, 13(2), 79–90. https://doi.org/10.3233/OER-160240

Treede, R., Rief, W., Barke, A., Aziz, Q., Bennett, M. I., Benoliel, R., Cohen, M., Evers, S., Finnerup, N. B., First, M.B., Giamberadino, M. A., Kaasa, S., Kosek, E., Lavand’homme, P., Nicholas, M., Perrot, S., Scholz, J., Schug, S., Smith, B. H., . . . Wang, S. (2015). A classification of chronic pain for ICD-11. Pain, 156(6), 1003–1007. https://doi.org/10.1097/j.pain.0000000000000160

Wang, X., Li, D., Menassa, C. C., & Kamat, V. R. (2019). Investigating the effect of indoor thermal environment on occupants’ mental workload and task performance using electroencephalogram. Building and Environment, 158, 120–132. https://doi.org/10.1016/j.buildenv.2019.05.012

Wang, Y. (2012). An interactive physio-psychological model on understanding stress among individuals performing sedentary computer tasks (Publication No. 3570047) [Doctoral dissertation, The Hong Kong Polytechnic University]. ProQuest Dissertations Publishing.

Xie, B., & Salvendy, G. (2000). Review and reappraisal of modeling and predicting mental workload in single- and multi-task environments. Work & Stress, 14(1), 74–99. https://doi.org/10.1080/026783700417249

Author Note

Thomas J. Faulkenberry https://orcid.org/0000­0002­8997­3794

Amber L. Harris Bozer https://orcid.org/0000­0003­0487­5237

Kayli N. Colpitts is now at the Texas A&M Institute of Neuroscience, and Jennifer L. Gibson Dias is now at the Department of Psychology at Texas Christian University.

Correspondence concerning this article should be addressed to Kayli N. Colpitts at colpittskayli@gmail.com

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

Ally See or Ally Do: Rewarding Corporate Social Responsibility Through Purchasing

Madison Will and Kevin R. Carriere* Department of Psychology, Washington and Jefferson College

ABSTRACT. Corporate social responsibility is an organization’s obligation to be aware of its impact on different aspects of society (e.g., social, environmental, economic). Many companies utilize hashtags, trends, and popular social movements to signal allyship, yet this may be a marketing ploy. Do consumers acknowledge the difference between performative activism and substantial activism? In this study, 198 participants (M = 38.8 years old, SD = 11.5 years) rated companies’ perceived support of social issues and their willingness to shop based on different levels of performative and substantive activism. Results showed that consumers were most willing to shop at companies exhibiting substantial activism over and above performative activism (d = 0.39) due to perceiving them as supportive of the issue (d = 0.98), finding evidence for full mediation (indirect CI [.24, .63]). These results were independent of the type of topic and personal beliefs of the participants. This research helps understand how individuals make purchasing decisions and how corporate social responsibility has become more crucial for organizations to elevate their relationships with customers.

Keywords: corporate social responsibility, performative activism, purchasing behavior, decision­making process, substantial activism

In 2020, George Floyd’s murder shocked the nation, amplifying the Black Lives Matter Movement. June 2, 2020, the Tuesday following his death, became known as “Black Out Tuesday,” when users flooded social media with black photos. Companies also joined in, changing their logos to black squares to raise awareness for police brutality and racial justice (Blair, 2021). This social media campaign on social issues is not unique to Black Lives Matter. For example, Target releases its annual “Pride Month Collection” in June, which usually features a myriad of rainbows and other stereotypical designs, sometimes called “rainbow washing” (Champlin & Li, 2020).

Companies also respond to social issues by making structural changes. In 2020, Nike introduced several corporate social responsibility (CSR) initiatives, such as the Juneteenth Learning Program and Unconscious Bias Awareness Training, to promote understanding social change and racial inequality (NIKE, 2020). Throughout the year, Nike also developed a five­year workforce plan to create a more diverse and inclusive community; by 2025, Nike wants their workforce to be 50% women and 35% racial/ethnic minorities. Furthermore, their Supplier

Climate Action Program ensures they are committed to reaching carbon neutrality by 2025 (NIKE, 2020).

Both of these forms of activism represent different modes of CSR, which came to prominence in business practices after a rise in social movements in the 1960s. These voluntary actions that benefit society instead of the company run in stark contrast to the capitalistic mindset that a corporation’s only responsibility is to its shareholders. CSR initiatives can be beneficial and rewarding for many organizations; nevertheless, they still bring actual costs (Sprinkle & Maines, 2010).

On Nike’s 30th Anniversary, Colin Kaepernick was the face of their signature “Just Do It” campaign, championing a profound message, “[believe] in something, [even] if it means sacrificing everything,” in their Dream Crazy Advertisement (Carroll, 2018). Numerous customers’ initial reactions claimed they were boycotting Nike products, with some even cutting off logos and burning them (Carroll, 2018), particularly for White individuals (Intravia et al., 2020). However, despite the initial backlash, the campaign was successful with most of their target audience, as their sales increased by 10% in the quarter following the release of their ad (Carroll, 2018).

In comparison, Pepsi tried to accomplish something similar with their “Live for Now – Moments’’ advertisement campaign starring Kendall Jenner, in which she walks up to a police officer with a Pepsi can amidst a protest, hoping to send a unifying message (Tillman, 2019). Criticism followed immediately, describing the advertisement as tone­deaf and trivializing the issue of police brutality. Less than 48 hours later, Pepsi removed the commercial and released an apology. Pepsi’s target audience is younger people, and in a study following the release of the advertisement, the number of millennials considering buying Pepsi fell by 10% (Marzilli, 2018). Thus, real­life evidence is mixed. Do consumers respond favorably to these actions, and if so, what kind of actions are the best to be taken? Our work presented seeks to answer that dilemma.

Literature Review

Notably, corporations may engage in CSR for many reasons. In a review of CSR’s positive and negative outcomes, Sprinkle and Maines (2010) listed five main reasons for CSR: altruistic benefits to society, appeasing stakeholders, recruiting employees, increasing sales, and reducing production costs. Each of these reasons comes with its costs and benefits, and firms must always consider how the consumer will react to their actions. A qualitative study found these five reasons emerge in interviews with managers (Öberseder et al., 2013), along with additional explanations of the altruistic benefits and stakeholder perceptions, noting concerns for the community, consumers, and the environment.

Engaging in CSR is not straightforward. Companies that engage in environmental CSR, such as promoting a brand­recycling program, were viewed as less authentic and received lower ratings when the company was a disposable plastic company (Childs et al., 2019). Other factors in determining whether participants believe in the sincerity and authenticity of firms were the cost of the advertising expense in informing the public of their actions and the individual differences of the consumer. For example, if a company were to donate $2,000,000 to a cause but spent $20,000,000 in advertising said cause, participants evaluated the company worse than if they had not acted at all (Yoon et al., 2006). In light of Nike’s ads in support of #BlackLivesMatter, White and conservative participants held negative attitudes towards Nike and felt they were overstepping their role as a company (Intravia et al., 2020). Other individual difference variables, such as social dominance orientation (one’s support for maintaining the existing status quo hierarchy; Pratto et al., 1994), may also influence how one reacts to CSR, as those high on social dominance orientation were found to be opposed to fair­trade consumption (Rios et al., 2015) and corporate intergroup responsibility (Halevy et al., 2020)

Therefore, firms are left in a double­bind situation. For example, after the September 11th attacks, Honda donated cash, all­terrain vehicles, and generators for first responders. Ultimately, they decided to keep their donations secret despite public polls claiming they were “anti­American” because their activism was not broadly advertised (Alsop, 2002). Firms must wrangle with the diverse perspectives of their consumers, placing firms in a damned­if­you­do, damned­if­you­don’t situation. Although Yoon et al.(2006) suggests that companies may not want to advertise their actions, there are consequences for not advertising their social justice motives, especially if deemed insincere.

Perceived insincerity could occur due to many factors, such as the relevance of the CSR initiative to the company (Sen & Bhattacharya, 2001) and the consumer’s trust in the company (Osterhus, 1997) . However, one factor that may be particularly telling is the firm’s investment in the social issue. If firms change their logos or put up signs, this may not be a significant enough investment in the social cause to elicit a strong positive consumer reaction. Instead, such actions may be seen as performative. Performative allyship and activism involve publicly taking a stand on a given social issue but without substantial actionable investment. The theoretical construct of performative activism is relatively new, with the most recent research surrounding the construct from media studies (Blair, 2021; Hesford, 2021; Wellman, 2022). These critical scholars describe the actions of performative activism as a potentially damaging act that brings about no change to the social issue at hand.

Some research has shown that performative activism may have benefits, also known as “slacktivism” (Vie, 2014). Individuals view those who use ‘image filters’ on their social media profiles as greater activists compared to those who do not use image filters (Matsick et al., 2020). The argument made in favor of these seemingly empty gestures is that even something as simple as using an image filter can increase the awareness of an issue and may lead to future engagements if the individual is engaging due to prosocial motivations compared to image­enhancing motivations (Moussaoui et al., 2022).

Yet, if a company is perceived to be sincere, consumers should reward the firm for taking a stand or acting towards a specific cause with which they agree. Although CSR cannot guarantee strong consumer relationships, building consumer identification can positively impact consumers’ commitment to the brand (Sen et al., 2009), leading to greater purchase likelihood and brand loyalty (Du et al., 2007) and even to customers also increasing their donations to nonprofits (Lichtenstein et al., 2004) We posit that more substantial activism should display more sincerity and signal a deeper commitment to the social issue.

Unfortunately, research has not adequately addressed whether consumers acknowledge the difference between performative and substantial activism. Drawing on the results of Yoon et al. (2006) and others (Blair, 2021; Hesford, 2021; Holmes IV, 2020), we hypothesized that firms that engage in performative activism will not be rewarded for their activism by consumers with less intention to purchase their goods due to a perceived insincerity in the cause.

In this study, participants viewed businesses that were engaged with various issues, including Black Lives Matter (BLM), LGBTQ+, and fast fashion (cheap, mass­produced clothes that generally utilize sweatshop labor along with the use of unsustainable materials; Niinimäki et al., 2020; Shi et al., 2023). Each company performed one of four levels of activism: no activism (our control), performative activism (small changes with little cost to the business and no meaningful action to the social justice problem), substantial activism (larger changes with a potential cost to the business and an authentic action towards the social justice problem), or harmful activism (putting up signs in favor but actively taking a stand to perpetuate the social justice issue). We hypothesized that consumers would be less likely to shop at those businesses that portray performative activism than substantial activism and that consumers would perceive the substantial activism shops as caring more about social issues than all others.

Method

Participants were asked to rate a series of companies, randomly presented in order, who engaged in four combinations of performative and substantive activism: performative activism, negative activism, substantial activism, or no activism. Other stores were also presented as distractor firms that took actions not of interest to this study and served to hide the independent variable levels of interest. Firms that engaged in performative activism were described to have posted a sign at their storefront but made no changes to hiring practices or promotion policies. Firms that engaged in negative activism put up a sign in support of the social issue but changed hiring practices and promotion policies to discriminate against minorities. Firms engaged in substantial activism put up signs and changed hiring practices and promotion policies to benefit minorities. Finally, the no activism condition described a store that put up no sign and took no action. After the presentation of each store, participants were asked to rate their willingness to shop at the store and their perceived support of the social issue of each store. After reading all scenarios, participants then indicated which store would be their top choice to shop at.

Participants

Before conducting the study, the study was approved via expedited review by the Washington and Jefferson College Institutional Review Board (PSY­2021SS­285). One hundred ninety­eight individuals were recruited internationally from Prolific Academic and participated in a mixed­effects design, averaging 10 minutes and 43 seconds to complete the survey. Participants were randomly assigned to one of the following social issue conditions: LGBTQ+, BLM, and fast fashion. The participants in the study identified as White (78.57%), Black (12.75%), Asian (7.14%), Native American (1.02%), and Native Hawaiian (0.51%), with 88% identifying as Non­Hispanic or Latinx. Three participants (1.53%) identified as transgender or nonbinary, and the rest of the sample identified as cisgender (98.5%), with 59.69% identifying as men and 38.28% identifying as women. There was a diverse range of ages (M = 38.8, SD = 11.5) and political views (M = 4.94, SD = 2.16, where 1 = socialist and 9 = fascist).

Procedure

Participants were invited to participate in a survey to understand their general beliefs about topics.

Between subjects, we randomly assigned participants to read about stores that focused their activism on either Black Lives Matter (N = 68), LGBTQ+ (N = 62), or fast fashion ( N = 66). For the Black Lives Matter conditions, to demonstrate substantial activism, the companies changed their hiring and promotional policies to benefit minorities. Similarly, for LGBTQ+ conditions, companies donated to pro/anti­LGBTQ+ organizations or did not donate at all. Last, the fast fashion condition focused on environmental impacts and price, with companies who were acting to introduce changes to their company to reduce/increase pollution.

Materials and Measures Personal Beliefs on Fast Fashion

We measured participant’s existing feelings about fast fashion with four self­generated questions (e.g., “You are concerned about the environmental impact of disposing of unwanted clothing items”) on a 7­point scale, anchored from strongly disagree to strongly agree (α = .85, M = 4.72, SD = 1.31).

Personal Beliefs on the Black Lives Matter Movement

We measured participant’s attitudes toward the BLM Movement using a six ­ item, 5 ­ point scale (Holt & Sweitzer, 2018). The anchors in the scale change depending on the question (e.g., My personal attitude about the Black Lives Matter movement is I... [dislike a great deal to like a great deal]) (α = .97, M = 3.34, SD = 1.23).

Personal Beliefs on the LGBTQ+ Community

We measured participant’s personal attitudes toward the LGBTQ+ community using a developed scale of 19 items on a 7­point scale from strongly disagree to strongly agree (α = .96, M = 5.07, SD = 1.49; Ellis et al., 2003). An example item is: “Just as other species, homosexuality is a natural expression of sexuality in humans.” Questions that required reverse coding were reversed before any analysis.

Social Dominance Orientation

Participants were asked to rate their feelings on a social dominance orientation scale (Pratto et al., 1994). The scale consisted of 15 items, and questions were anchored on a 7­point scale from strongly disagree to strongly agree (α = .95, M = 2.78, SD = 1.40). An example item is: “It is sometimes necessary to use force against other groups to get what you want.” Questions that required reverse coding were reversed before any analysis.

Willingness to Shop

Participants were asked a single question for each shop on “How likely are you to shop at Company __”? (M = 3.18, SD = 1.17) on a 5­point scale, from 1 (extremely unlikely) to 5 (extremely likely).

Support of Social Issue

Participants were asked a single question for each shop on “How much do you believe the company supports [Condition: LGBTQ+/Black Lives Matter/being environmentally friendly]?” (M = 2.86, SD = 1.33) on a 5­point scale, from 1 (not at all) to 5 (a great deal).

Analytic Strategy

Within­subject ANOVAs applying Greenhouse­Gauser sphericity corrections when needed (Haverkamp & Beauducel, 2017) were used to test to see if there are group differences between the type of shop presented to the participant and both the perceived support of the issue as well as the participant’s willingness to shop at the store. We use Shapiro’s test of normality to test for normality. Estimated marginal mean comparisons using Tukey’s adjustment and Kenward­Roger estimated degrees of freedom (Kenward & Roger, 2009) was used to test group differences if the omnibus ANOVA main effect was significant. Finally, we used the PROCESS macro in SPSS to test if there was a significant mediation effect of perceived support of the social issue on willingness to shop at the store.

Results

There was no significant main effect from the type of social issue for perceived support, F (2, 193) = 3.02, p = .05, η2p = .03, or likelihood to shop, F(2, 193) = 1.62, p = .20, η2p= .02. Therefore, all further analyses have collapsed conditions. Likewise, adding personal beliefs as covariates, χ2(4) = 5.50, p = .24, or moderating terms on the condition, χ2(12) = 20.09, p = .06, did not improve model fit for perceived support or for likelihood to shop, χ2(4) = 2.19, p = .70; χ2(12) = 9.95, p = .27, so we do not include these controls in the proceeding analyses. Two repeated measures (within­subjects) ANOVA analyses were used to assess participants’ willingness to shop and perceived support of the shop. No significant outliers were identified by condition for either willingness to shop or perceived support. All conditions violated the assumption of normality based on Shapiro’s test of normality; however, researchers have argued that ANOVAs are robust to this violation (Schmider et al., 2010) and that Shapiro’s test, while the best test for normality, still has significant issues with power (Razali & Wah, 2011).

Perceived Support

Perceived support for the social issue by type of

Note. Participants' mean willingness to shop at each store

activism was significant, F(2.45, 477.23) = 76.79, p < .001, η2p = .28. Post­hoc analyses were run to examine where the differences between the groups were. The posthoc analyses in showed that the no activism condition was significantly lower than performative ( b diff = –0.49, SE = 0.10, t = –4.77, p < . 001, d = 0.48), substantial (bdiff = –1.49, SE = 0.10, t = –14.47, p < .001, d = 1.46), and negative activism (bdiff = –0.36, SE = 0.10, t = –3.48, p < .001, d = 0.35). Additionally, performative activism companies were rated significantly lower than substantial activism companies (bdiff = –1.00, SE = 0.10, t = –9.70, p < .001, d = 0.98), but no different than negative companies ( b diff = –0.13, SE = 0.10, t = 1.29, p = .57, d = 0.13). Substantial action companies scored significantly higher than negative activism companies (bdiff = 1.13, SE = 0.10, t = 10.99, p = .003, d = 1.11). This result is illustrated in Figure 1, with no activism (M = 2.28, SD = 1.32), performative (M = 2.77, SD = 1.16), substantial (M = 3.77, SD = 1.10), and negative activism (M = 2.64, SD = 1.26).

Willingness to Shop

Next, we tested to see if participants’ willingness to shop at each store was different based on the type of activism, which it was, F(2.64, 515.73) = 11.88, p < .001, η2p= .06. This is shown in Figure 2, with the following means and standard deviations of each group: no activism (M = 3.16, SD = 1.14), performative (M = 3.12, SD = 1.11), substantial ( M = 3.51, SD = 1.16), and negative activism (M = 2.92, SD = 1.20). The posthoc analysis in Table 2 revealed that when participants read about the shop that engaged in substantial activism, they indicated much higher willingness to shop than shops that exhibited negative activism ( b diff = 0.59, SE = 0.10, t = 5.86, p < .001, d = 0.59), performative activism (bdiff = –0.39, SE = 0.10, t = –3.87, p < .001, d = 0.39), and no activism; (bdiff = –0.35, SE = 0.10, t = –3.52, p = .003, d = 0.36; see Figure 2). On the other hand, performative was not significantly different than no activism (bdiff = –0.04, SE = 0.10, t = .36, p = .98, d = 0.04) or negative activism (bdiff = 0.20, SE = 0.10, t = 1.99, p = .19, d = 0.20), and negative activism was not significantly different from no activism (bdiff = 0.23, SE = 0.10, t = 2.35, p = .09, d = 0.23).

A mediation analysis of 5,000 bootstrapped samples was run using the MEMORE SPSS extension (Montoya & Hayes, 2017) to examine the indirect effect of willingness to shop through the perceived support of the shop. To run the mediation analysis, we compared the substantial activism shop against all others because a multilevel, within­subjects mediation requires pairwise comparisons (Hayes & Preacher, 2014; Montoya & Hayes, 2017). In separate regressions, both substantial activism (b = 0.44, SE = 0.09, t = 4.82, p < .001) and

perceived support of the shop ( b = 1.39, SE = 0.03, t = 14.19, p < .001) predicted willingness to shop. Likewise, there was a significant difference between the kind of activism and perceived support (b = 1.20, SE = 0.10, t = 11.41, p < .001). When both were included in the same model, though, the effect of Substantial

Participants' Mean Willingness to Shop at Each Store

TABLE 2

FIGURE 2

FIGURE 3

Mediation Analysis of Action Through Perceived Support to Willingness to Shop

activism condition became non­significant (b = 0.01, SE = 0.11, t = .10, p = .92), but the effect of perceived support remained significant ( b = 0.36, SE = 0.06, t = 6.28, p < .001). A bootstrap test produced a confidence interval for the overall indirect effect that did not include zero, b = 0.43, SE = 0.10, CI = [0.24, 0.63] (see Figure 3), suggesting full mediation.

Lastly, in the end, participants were asked to decide which store would be their top choice to shop at; 29.3% chose the substantial action shop, 22.2% picked the no activism shop, 17.7% named the negative activism shop, and 4.5% selected the performative activism shop, but 26.3% opted not to answer the question1. Using a chi­square test of independence, each company was expected to receive 18.4%, showing a significant difference than expected, χ2(146) = 35.0, p < .001, Cramer’s V = 0.28.

Discussion

Using a mixed ­ effects design, we tested to see if participants would treat companies who engage in performative activism differently than those who engage in substantive change to their policies and if this effect differed by social issue. We found a difference in how participants perceived the type of activism the companies took. The results showed that consumers were most willing to shop at the substantial activism shop, which exhibited substantial activism. Beyond that, participants identified these substantial activism companies as caring more about the social issue overall. This perceived care of the issue fully mediated the relationship between the type of action and one’s willingness to shop.

We found no difference in the type of issue; participants treated companies that engaged in BLM, LGBTQ+, and fast fashion corporate social responsibility the same. This result suggests that it is not necessarily what companies are standing for but standing for something and engaging in corporate social responsibility. Firms looking to enter this space can feel confident knowing they need only take action in what they care about and not feel the need to target one cause over another.

Our study was intentionally vague on what kinds of shops were being presented to participants. Prior research showed that public support for CSR initiatives depends on congruence between the product the firm sells and the issue being addressed (Ellen et al., 2000; Sen & Bhattacharya, 2001). By leaving details out of what the shop sells, participants had to focus primarily on the company’s action without knowing if there was an incongruence between the issue and the product. At the same time, it is unclear what store would match some 1A logistic regression predicting opting out of this question on all personality scales and demographic characteristics yielded no significant predictor at our alpha level of .05. As such, we tentatively claim this may be missing at random.

issues, such as BLM and LGBTQ+, except for stores in highly demographically homogeneous areas.

Interestingly, we found no significant effects of any individual difference variable. Neither social dominance orientation nor feelings on the three presented social issues significantly predicted how participants responded to perceived support or willingness to shop at the store. This finding goes against published research (Holt & Sweitzer, 2018; Rios et al., 2015). It may have been the case that some of these factors were still important in influencing the participants’ reactions but differed across within­subjects experimental conditions. More extensive studies with higher power should test for these possible moderating effects.

Each study design comes with its limitations, and ours is no different. For example, our study did not include a manipulation check, so we cannot be entirely sure if the within­subjects manipulation was successful. However, this is of minimal concern because the mediating variable of interest, support of the social cause, could be considered a type of manipulation check. If our manipulation were unsuccessful, then there would be no significant difference in perception of the shop’s stance towards the social justice issue. On the other hand, because our manipulation was successful, we saw that different shops were rated differently consistently and significantly. Other concerns with manipulation checks have also been discussed within the literature (c.f. Hauser et al., 2018). However, the within­subjects manipulation does lead to other concerns, notably, the tiredness of the participant. The fact that 26.3% of individuals opted out of answering our final question suggests that individuals were tired over the length of the study, and further research may want to consider a between­subjects manipulation with a more overt manipulation check.

Our manipulation itself was varied. BLM shops acted in a way focused on hiring practices, but LGBTQ+ shops focused on donations, and fast fashion shops focused on change in pollution­emission policies. Our choice to do this was directed at the inability to “hire” for fast fashion and the variety of ways shops may take social justice action. However, it still leads to a limitation in which we cannot be sure if our manipulation of the action of the shop was due to the action of the shop precisely or the fact that an action was being taken more generally. Here, too, our insignificant effects suggest that participants treated each condition equally, but a more controlled manipulation may shed additional light on this non­significant finding.

A more substantive limitation is our inability to test actual behavior. Although we asked participants how likely they were to shop at a given store, asking a

participant’s intent to shop is not necessarily the same as behaviorally putting the question to the participant (Sheeran & Webb, 2016). This ties in with concerns that our study may be subject to social desirability bias, which has been shown to have a small but significant effect in environmental justice studies (Vesely & Klöckner, 2020). Although our sample was anonymous to reduce those concerns, it is still unclear how much of our effect is due to this bias, and future research could consider a social desirability measure to account for this.

Finally, we should note the question of diversity within our sample. Using Prolific Academic, we achieved similar rates of racial identity compared to the United States population ( QuickFacts of the United States , 2022). However, with small samples, proportional equivalence in sample characteristics does not equate to full representation, and the weights of the oversampled White population may undersell important characteristics of other racial groups. In studying a topic about how corporations take action on social issues, it may be increasingly important to consider how minority groups, who may stand to benefit the most from said activism, perceive the action, which this research cannot adequately describe or examine (see Intravia et al., 2020 for an example that does investigate this). Beyond this, although our racial demographics were representative of the country’s population at large, we cannot claim that those who opt ­ in to take surveys online are representative of the population. Significant differences between these groups may exist that this research cannot account for. On the other hand, one may argue that because the backlash against this corporate activism has been primarily led by the white­majority group (Genter, 2022), understanding the majority group’s psychological reactions to corporate social responsibility may be particularly interesting to research.

In keeping details of the shop low, many potential avenues for future research arise. For instance, our study did not examine whether there would be a difference between a “Mom­and­Pop” local business putting up the sign and an international conglomerate such as Walmart having the exact change in sign. Some research suggests that national issues are more acceptable for giant corporations than small stores (Sung et al., 2022). These more significant, well­known companies may also carry with them internal biases. Participants may be more skeptical to find out that Chick­fil­A supports an LGBTQ+ issue, given their national attention and support for anti­LGBTQ+ causes. At the same time, highlighting that a single franchise from the larger chain is taking an opposing stand may lead participants to reward these stores even more. Future research should tease apart these manipulations.

Finally, our manipulations were all more typical of liberal ideals, and we had not considered companies that engage in more conservative movements, both within the United States (#WhiteLivesMatter, presidential preferences, fears against ‘wokeness’ or critical race theory) or abroad (refugees, Brexit, and others). Future research should examine whether or not responses to companies that take more conservative positions find the same results.

Consumers’ perception of corporate activism is vital to understanding the decision­making process and consumer purchasing behavior. Consumers are becoming more vocal about the importance of equality, protecting the environment, and economic factors. Can companies get away with simply changing their logo for a month? In our study, the conclusion is a potential no. In the eyes of the consumer, performative activism does not have the same recognition as substantial activism (see Figure 2). If companies want to receive the benefits of corporate activism, consumers are more responsive and rewarding of substantial activism. Our results indicate that relying solely on performative activism is ineffective; consumers discern the difference and prefer companies engaging in substantive change.

References

Alsop, R. (2002, January 16). Perils of corporate philanthropy: Touting good works offends the public, but reticence is perceived as inaction. The Wall Street Journal, B1. Blair, K. (2021). Empty gestures: Performative utterances and allyship. Journal of Dramatic Theory and Criticism, 35(2), 53–73. https://doi.org/10.1353/dtc.2021.0005

Carroll, C. (2018, December 22). Report: Nike sales rise despite Colin Kaepernick ad. Sports Illustrated

https://www.si.com/nfl/2018/12/22/colin-kaepernick-nike-ad-campaign-sales

Champlin, S., & Li, M. (2020). Communicating support in pride collection advertising: The impact of gender expression and contribution amount. International Journal of Strategic Communication, 14(3), 160–178. https://doi.org/10.1080/1553118X.2020.1750017

Childs, M., Woo, H., & Kim, S. (2019). Sincerity or ploy? An investigation of corporate social responsibility campaigns. Journal of Product & Brand Management, 28(4), 489–501. https://doi.org/10.1108/JPBM-07-2018-1953

Du, S., Bhattacharya, C. B., & Sen, S. (2007). Reaping relational rewards from corporate social responsibility: The role of competitive positioning. International Journal of Research in Marketing, 24(3), 224–241.

https://doi.org/10.1016/j.ijresmar.2007.01.001

Ellen, P. S., Mohr, L. A., & Webb, D. J. (2000). Charitable programs and the retailer: Do they mix? Journal of Retailing, 76(3), 393–406. https://doi.org/10.1016/S0022-4359(00)00032-4

Ellis, S. J., Kitzinger, C., & Wilkinson, S. (2003). Attitudes towards lesbians and gay men and support for lesbian and gay human rights among psychology students. Journal of Homosexuality, 44(1), 121–138.

https://doi.org/10.1300/J082v44n01_07

Genter, S. (2022). Collective racial resentment and emotion: Toward an understanding of Whites’ reactions to demands for racial equity. Sociology Compass, 16(10). https://doi.org/10.1111/soc4.13026

Halevy, N., Jun, S., & Chou, E. Y. (2020). Intergroup conflict is our business: CEOs’ ethical intergroup leadership fuels stakeholder support for corporate intergroup responsibility. Journal of Business Ethics, 162(1), 229–246.

https://doi.org/10.1007/s10551-018-4013-0

Hauser, D. J., Ellsworth, P. C., & Gonzalez, R. (2018). Are manipulation checks necessary? Frontiers in Psychology, 9, 998.

263

https://doi.org/10.3389/fpsyg.2018.00998

Haverkamp, N., & Beauducel, A. (2017). Violation of the sphericity assumption and its effect on type-I error rates in repeated measures ANOVA and multilevel linear models (MLM). Frontiers in Psychology, 8, 1841. https://doi.org/10.3389/fpsyg.2017.01841

Hayes, A. F., & Preacher, K. J. (2014). Statistical mediation analysis with a multicategorical independent variable. British Journal of Mathematical and Statistical Psychology, 67(3), 451–470. https://doi.org/10.1111/bmsp.12028

Hesford, W. S. (2021). Reading the signs: Performative white allyship. Quarterly Journal of Speech, 107(2), 239–244. https://doi.org/10.1080/00335630.2021.1915447

Holmes IV, O. (2020). Police brutality and four other ways racism kills Black people. Equality, Diversity and Inclusion: An International Journal, 39(7), 803–809. https://doi.org/10.1108/EDI-06-2020-0151

Holt, L. F., & Sweitzer, M. D. (2018). More than a black and white issue: Ethnic identity, social dominance orientation, and support for the black lives matter movement. Self and Identity, 19(1), 16–31.

https://doi.org/10.1080/15298868.2018.1524788

Intravia, J., Piquero, A. R., Leeper Piquero, N., & Byers, B. (2020). Just do it? An examination of race on attitudes associated with Nike’s advertisement featuring Colin Kaepernick. Deviant Behavior, 41(10), 1221–1231.

https://doi.org/10.1080/01639625.2019.1604299

Kenward, M. G., & Roger, J. H. (2009). An improved approximation to the precision of fixed effects from restricted maximum likelihood. Computational Statistics & Data Analysis, 53(7), 2583–2595.

https://doi.org/10.1016/j.csda.2008.12.013

Li, J.-Y., Kim, J. K., & Alharbi, K. (2022). Exploring the role of issue involvement and brand attachment in shaping consumer response toward corporate social advocacy (CSA) initiatives: The case of Nike’s Colin Kaepernick campaign. International Journal of Advertising, 41(2), 233–257.

https://doi.org/10.1080/02650487.2020.1857111

Lichtenstein, D. R., Drumwright, M. E., & Braig, B. M. (2004). The effect of corporate social responsibility on customer donations to corporatesupported nonprofits. Journal of Marketing, 68(4), 16–32.

https://doi.org/10.1509/jmkg.68.4.16.42726

Marzilli, T. (2018). Pepsi recovers but purchase consideration hasn’t. YouGov. Matsick, J. L., Kim, L. M., & Kruk, M. (2020). Facebook LGBTQ pictivism: The effects of women’s rainbow profile filters on sexual prejudice and online belonging. Psychology of Women Quarterly, 44(3), 342–361.

https://doi.org/10.1177/0361684320930566

Montoya, A. K., & Hayes, A. F. (2017). Two-condition within-participant statistical mediation analysis: A path-analytic framework. Psychological Methods, 22(1), 6–27. https://doi.org/10.1037/met0000086

Moussaoui, L. S., Blondé, J., Phung, T., Tschopp, K. M., & Desrichard, O. (2022). Does a low-cost act of support produce slacktivism or commitment? Prosocial and impression-management motives as moderators. Frontiers in Psychology, 13, 783995. https://doi.org/10.3389/fpsyg.2022.783995

Niinimäki, K., Peters, G., Dahlbo, H., Perry, P., Rissanen, T., & Gwilt, A. (2020). The environmental price of fast fashion. Nature Reviews Earth & Environment, 1(4), 189–200. https://doi.org/10.1038/s43017-020-0039-9

NIKE. (2020). 2025 Targets (FY20 NIKE Inc. Impact Report). NIKE, Inc.

Öberseder, M., Schlegelmilch, B. B., & Murphy, P. E. (2013). CSR practices and consumer perceptions. Journal of Business Research, 66(10), 1839–1851.

https://doi.org/10.1016/j.jbusres.2013.02.005

Osterhus, T. L. (1997). Pro-social consumer influence strategies: When and how do they work? Journal of Marketing, 61(4), 16–29.

https://doi.org/10.1177/002224299706100402

Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F. (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67(4), 741–763.

https://doi.org/10.1037/0022-3514.67.4.741

QuickFacts of the United States. (2022). United States Census Bureau.

https://www.census.gov/quickfacts/fact/table/US/RHI125221

Razali, N. M., & Wah, Y. B. (2011). Power comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and Anderson-Darling tests. Journal of Statistical Modeling and Analytics, 2(1), 21–33.

Rios, K., Finkelstein, S. R., & Landa, J. (2015). Is there a “fair” in fair-trade? Social dominance orientation influences perceptions of and preferences for fairtrade products. Journal of Business Ethics, 130(1), 171–180.

https://doi.org/10.1007/s10551-014-2221-9

Schmider, E., Ziegler, M., Danay, E., Beyer, L., & Bühner, M. (2010). Is it really robust?: Reinvestigating the robustness of ANOVA against violations of the normal distribution assumption. Methodology, 6(4), 147–151.

https://doi.org/10.1027/1614-2241/a000016

Sen, S., & Bhattacharya, C. B. (2001). Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. Journal of Marketing Research, 38(2), 225–243.

https://doi.org/10.1509/jmkr.38.2.225.18838

Sen, S., Shuili, D., & Bhattacharaya, C. B. (2009). Building brand relationships through corporate social responsibility. In D. J. MacInnis, C. W. Park, & J. W. Priester (Eds.), Handbook of brand relationships (pp. 195–211). Routledge. https://doi.org/10.4324/9781315703886

Sheeran, P., & Webb, T. L. (2016). The intention-behavior gap. Social and Personality Psychology Compass, 10(9), 503–518. https://doi.org/10.1111/spc3.12265

Shi, Z. (June), Liu, X., Lee, D., & Srinivasan, K. (2023). EXPRESS: How do fast fashion copycats affect the popularity of premium brands? Evidence from social media. Journal of Marketing Research, 002224372311644.

https://doi.org/10.1177/00222437231164403

Sprinkle, G. B., & Maines, L. A. (2010). The benefits and costs of corporate social responsibility. Business Horizons, 53(5), 445–453. https://doi.org/10.1016/j.bushor.2010.05.006

Sung, Y. H., Lim, R. E., & Lee, W.-N. (2022). Does company size matter in corporate social responsibility? An examination of the impact of company size and cause proximity fit on consumer response. International Journal of Advertising, 41(2), 284–308. https://doi.org/10.1080/02650487.2020.1850997

Tillman, L. (2019, June 17). Case study: Pepsico & Kendall Jenner’s controversial commercial. Astute. Astute.co.

https://astute.co/pepsi-kendall-jenner-commercial/ Vesely, S., & Klöckner, C. A. (2020). Social desirability in environmental psychology research: Three meta-analyses. Frontiers in Psychology, 11, 1395. https://doi.org/10.3389/fpsyg.2020.01395

Vie, S. (2014). In defense of “slacktivism”: The human rights campaign Facebook logo as digital activism. First Monday https://doi.org/10.5210/fm.v19i4.4961

Wellman, M. L. (2022). Black squares for Black lives? Performative allyship as credibility maintenance for social media influencers on Instagram. Social Media + Society, 8(1), 205630512210804.

https://doi.org/10.1177/20563051221080473

Yoon, Y., Gürhan-Canli, Z., & Schwarz, N. (2006). The effect of corporate social responsibility (CSR) activities on companies with bad reputations. Journal of Consumer Psychology, 16(4), 377–390.

https://doi.org/10.1207/s15327663jcp1604_9

Author Note

Kevin R. Carriere https://orcid.org/0000­0001­6032­286X

Kevin R. Carriere is now at Stonehill College, Department of Psychology.

The authors do not have any conflicts of interest to disclose. Correspondence concerning this article should be addressed to Kevin R. Carriere, Stonehill College, 230 Washington Street, North Easton, MA 02356. Email: kcarriere@stonehill.edu

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

The Effects of Hormonal Contraception on Auditory Emotional Memory

Jessica Simonson, Courtney A. Durdle, and Michael B. Miller* Department of Psychological and Brain Sciences, University of California, Santa Barbara

ABSTRACT. Emotional episodic memory is an important cognitive mechanism that has been extensively studied, however, auditory emotional memory in particular has yet to be thoroughly understood. In addition, sex hormones have been found to affect brain structure and regulate regions of the brain that support higher order cognitive functions. Considering the global usage of oral hormonal contraceptive pills, it is vitally important to investigate the effects of oral contraceptives on executive function, including memory. The aim of the present study was to investigate the extent to which oral contraceptives influence recall for an emotional auditory episodic memory compared to a neutral memory. Participants (N = 90; 45 on an oral contraceptive, 45 naturally cycling) performed a free recall task for an emotional and a neutral auditory story, and their recalls were categorized into gist and detail elements and rated for accuracy. Recall accuracy for an emotional or neutral auditory story was not different between women on oral hormonal contraceptives and women who were naturally cycling, however, both groups of women recalled more information regarding the neutral story compared to the emotional story. These findings inform how the use of hormonal contraceptive pills, combined with high emotional valence, may impact the content and accuracy of recalled episodic events.

Keywords: emotional memory, auditory memory, neuroendocrinology, hormonal contraceptives, women’s health

Episodic memory is a form of long­term memory that allows an individual to remember distinct events from their past personal experiences (Tulving, 1993). It differs from other types of memory due to its autonoetic component, in which individuals essentially relive a prior experience through episodic memory recall, often including sensory components (Tulving, 2005). Episodic memory is essential for many aspects of one’s life, including learning, forming identity, and decision­making (Tulving, 2005). Previous research has established that sex hormones have a large impact on various cognitive processes (Boss et al., 2014), including beneficial effects of estradiol on episodic memory (Rentz

Diversity badge earned for conducting research focusing on aspects of diversity. Open Data and Open Materials badges earned for transparent research practices. Data and materials are available at https://osf.io/k7h8c/

et al., 2017). Given the widespread usage of hormonal contraception, it is, therefore, pertinent to explore how disruptions to the endocrine system directly impact episodic memory.

Emotion and Memory

The impact of emotional events on episodic memory recall has been thoroughly studied, especially when looking at varying stimulus types and valence (degree of pleasantness). Previous research has demonstrated that negative emotional events tend to be more richly and accurately remembered compared to neutral events (Cahill & McGaugh, 1995; Dunsmoor et al.,

Hormonal Contraception and Memory | Simonson, Durdle, and Miller

2015; Kensinger, 2009). Stimuli with a negative valence are often more easily and accurately recalled than neutral or positive stimuli (Kensinger & Corkin, 2003). Although emotional arousal and stress responses may be a moderator in episodic memory recall (Buchanan & Tranel, 2008), evidence has suggested that the valence of a story alone may still have these enhancing effects (Bowen et al., 2018). Researchers have also found that the hippocampus and amygdala show increased activity when recalling an emotional memory, which coincides with the perception of a more vivid autonoetic reliving of the event and enhanced memory for it (Bowen et al., 2018). Improved memory for events with emotional salience is beneficial, as it aids in decision­making for future events. For example, one may not recall a typical drive to work, however, a drive in which they narrowly avoid an automobile accident would likely form a stronger memory. This may inform them to increase their attention while driving, or to consider alternative routes. Emotions and memory being intertwined serves an evolutionary benefit in increasing chances of survival. Although there is a large amount of research on emotional memory, more research focusing specifically on auditory emotional memory is needed. The subset of emotional memory relating to auditory stimuli alone is very relevant to our everyday lives, such as through storytelling and relaying information using conversation alone. Previous research regarding auditory memory has typically utilized sounds or single words, whereas recall for narrated stories has been less thoroughly explored. Auditory stories may retain higher ecological validity, as they accurately represent the means in which people transmit information in their day­to­day lives through conversations (Baldassano et al., 2017). As social beings, humans are constantly engaging with others through conversation, and the ability to recall information that was transmitted verbally is pertinent in guiding future decisions. A study examining the effects of gender and age on different domains of episodic memory demonstrated that women may outperform men on auditory episodic memory paradigms, whereas men may perform better on visual episodic memory tasks (Pauls et al., 2013). This reveals potential differences in episodic memory depending on the sensory domain and should be more thoroughly investigated.

Hormonal Contraceptives and Memory

Although currently over 100 million women worldwide use hormonal oral contraceptives as their preferred method of birth control, the literature surrounding the short and long­term effects of the drug on women’s cognitive health is relatively limited (Christin­Maitre, 2013; Cooper et al., 2022). Previous studies have found

that birth control can have widespread effects on areas such as brain structure (gray matter volume; Pletzer et al., 2010), emotion (Lewis et al., 2019), and improved verbal fluency (Griksiene & Ruksenas, 2011) among other cognitive domains (Warren et al., 2014). Given that the endocrine system is an essential component of cognitive function, it is important to evaluate how the pill impacts endocrine structures and hormones. The oral combination pill is made up of varying levels of synthetic forms of both estrogen and progesterone (Cooper & Mahdy, 2020), and affects sex steroid hormone receptors, mimicking the negative feedback effects of estrogen and progesterone. This reduces the secretion of gonadotropin releasing hormone, which in turn reduces the pituitary gland’s production of luteinizing hormone and follicle­stimulating hormone (Taylor et al., 2021). This mechanism prevents pregnancy in two ways: the reduction of follicle­stimulating hormone prevents follicle growth, whereas the reduction of luteinizing hormone prevents ovulation that typically occurs following a luteinizing hormone surge (Rivera et al., 1999). Depending on the specific chemical makeup of the pill, this mechanism leads to constant levels of either progesterone, estrogen, or both in women on a combined hormonal contraceptive (HC), contrasting the large fluctuations in sex hormone levels seen in naturally cycling (NC) women (Fleischman et al., 2010).

HCs have the potential to affect cognition in a variety of ways, such as by impacting different types of memory (Pletzer & Kerschbaum, 2014), including working memory (Griksiene & Ruksenas, 2011; Herrera et al., 2020) and verbal memory (Mordecai et al., 2008). Yet, few studies have expanded on the effects of HCs on memory for an emotional stimulus using narration alone, a stimulus that is quite relevant to our everyday lives. Previous studies on the effects of hormones on emotional memory have primarily utilized visual stimuli in combination with narration (Nielsen et al., 2011; Presuss et al., 2009) or videos (Wegerer et al., 2014). In Nielsen et al. (2011), NC and participants on oral HCs listened to a recording of either an arousing or a neutral short story while viewing a corresponding slideshow of images. Participants recalled each slide image and associated storyline, then recall accuracy was assessed based on the participants’ ability to report gist and detail elements of the stimulus. Gist is the main idea of the story, and when changed or omitted, it alters the entire storyline, whereas details are the smaller, more specific aspects (Cahill & van Stegeren, 2003). The researchers found that there was a difference in memory for emotionally arousing stimuli in women on a combined HC pill compared to NC women. NC women recalled more detail elements of the emotional stimuli compared

to the neutral, but showed no difference for gist recall, whereas HC women recalled more gist elements for the emotional content condition with no difference for detail recall (Nielsen et al., 2011). This may be due to HCs’ blunting effects on amygdala reactivity, as explored by Petersen and Cahill (2015). These findings regarding emotional memory, combined with research on the effects of HCs on auditory memory, suggest that the oral contraceptive pill has the capability to impact memory in a variety of domains. This research is important in advising individuals about potential long­term effects of HCs, and should be utilized to make well­informed decisions about using this medication.

Verbal memory in particular has also been shown to be affected by HC use. A study by Mordecai et al. (2008) demonstrated a difference in recall for neutral auditory stimuli (single words administered through the California Verbal Learning Task) for women on an oral contraceptive when compared to NC women. They found that HC women had improved verbal memory during their active pill phase, whereas NC women showed no noticeable differences throughout their menstrual cycle. These findings suggest that exogenous sex hormones, such as the oral contraceptive pill, contribute more to differences in verbal memory than endogenous sex hormones.

The Present Study

Previous studies have demonstrated numerous effects of hormonal contraception on cognition and memory, yet the drug’s impact on auditory emotional memory has not been fully explored. The aim of the present study was to explore how HCs affect recall of the gist and details of an emotional auditory story. We hypothesized that recall accuracy would be higher for the emotional content condition, regardless of birth control status (NC or HC) or recall content (gist or details), analogous to previous work regarding emotional memory. In accordance with the previous studies on visual memory, we hypothesized that the use of hormonal contraception would significantly impact the content of recall for an emotional story, such that: (a) NC women would remember more details of an auditory emotional content story compared to (b) those on hormonal birth control, who would remember more gist elements. We predicted (c) that no significant differences would be found between NC women and women on an HC for gist or detail recall for a neutral story. We did not predict (d) a main effect of contraceptive status on recall accuracy. Lastly, we predicted that (e) participants, regardless of contraceptive status, would recall more information from the emotional story than the neutral story.

To assess these predictions, four 2 (story condition:

emotional or neutral) x 2 (contraceptive status: HC or NC) mixed ANOVAs were used. Specifically, we predicted that there would be a significant interaction effect in all four analyses, such that for prediction (a) NC hit rate (HR) would be significantly greater than HC HR for emotional detail information and that NC false alarm rate (FAR) would be significantly less than HC FAR for emotional detail information. For prediction (b), we predicted that HC HR would be significantly greater than NC HR for emotional gist information, and HC FAR would be significantly less than NC FAR for emotional gist information. We anticipated the interaction effects to be driven by the emotional story, with no significant differences for HR and FAR between HC and NC groups in the neutral gist or neutral detail conditions (c). There was also a predicted main effect (e) of story condition on FAR and HR in all conditions, such that HR for the emotional condition would be significantly greater than the neutral condition, regardless of contraceptive status and gist versus detail information. FAR were predicted to be significantly less for the emotional condition than the neutral condition in all group comparisons.

The findings of this study were anticipated to further elucidate the need to specify a participant’s HC usage when including them in a research study regarding emotional content memory, as women may perform differently depending on their respective contraceptive status. Additionally, the results of this study were

1

Demographics –

Simonson, Durdle, and Miller | Hormonal Contraception and Memory

Hormonal Contraception and Memory | Simonson, Durdle, and Miller

expected to increase the understanding of the impacts that this widely utilized medication may have on specific cognitive domains.

Methods

Subjects

Ninety participants, all biologically female and identifying as women, between the ages of 18 and 33 (M = 20.5, SD = 2.4) were recruited through the online participant database (SONA) at University of California, Santa Barbara and received a $5 Amazon gift card for completing the study. Participants completed an online prescreening questionnaire to validate that they were either NC (i.e., not on an HC and for at least three months prior) or on an oral combination pill for at least three months (HC). In total, there were 45 NC participants and 45 participants on a combination oral birth control

pill. Participants were primarily White, Asian, or Multiple Ethnicities (see Table 1). A chi­square test of independence showed that there was a significant association between ethnicity and contraceptive status, χ2 (5, N = 90) = 13.2, V = .38, p = .01, with disproportionately high representation of Asian participants in the NC group and White participants in the HC group (see Table 1). There was also a significant association between year in school and contraceptive status, χ2 (4, N = 90) = 9.6, V = .33, p = .04, such that first year participants had a disproportionately high representation in the NC group (see Table 2) Additionally, an independent sample t­test showed that the number of participants in each contraceptive group (HC and NC) differed significantly by age, t(73) = 2.00, p < .05, d = 0.42.

Measures and Materials Birth Control Eligibility Questionnaire

To determine if a person was eligible to participate in the experiment, participants completed a birth control eligibility questionnaire. To participate in the study, participants for the HC group needed to indicate that they were on an oral HC pill, and had been for at least three months prior, and were asked about the brand and dosage of their oral combination pill. NC participants indicated that they were not on any HC and had not taken one for at least three months prior. They were also asked for information about the first day of their last menstrual cycle and how confident they were that the reported date was correct. All components of this questionnaire, including specific oral contraceptive usage, can be found in Appendix A and Table 3.

Positive and Negative Affect Schedule

Birth Control Information

The Positive and Negative Affect Schedule (PANAS) is a 5­point scale questionnaire that assesses positive and negative affect. It was originally developed to determine state and trait­based affect. It consists of 20 questions regarding how well a participant’s current mood matches various emotion words, ranging from 1 (very slightly or not at all) to 5 (extremely). For the purposes of this study, the state focused scale was used to assess both baseline and any changes in participants’ emotional state based on the story narrations. The scale has been shown to have both validity evidence and high reliability (Watson et al., 1988), including through online uses (Díaz­García et al., 2020).

Auditory Stimuli

The stimuli for this study consisted of prerecorded narrations of two different stories of a similar length. One story was used as a neutral condition, whereas the other was used with the aim of evoking an emotional

response. The neutral story focused on two women completing mundane tasks, such as going to a local mall, browsing stores, having dinner, and returning home. The negative emotional content story included distressing medical situations due to an unfortunate accident. A boy was hit by a car on his way to visit his father at work and was rushed to the hospital where he underwent many medical tests and surgery (see Appendix B for full transcript). The emotional story was based on the emotionally arousing stimuli used in Nielsen et al.’s (2011) study, which was originally adapted from Cahill et al. (1994). Each story was narrated and recorded by the author in a monotonous (limited in variation of tone or pitch) voice to avoid any vocal fluctuations so that participants’ recall would only be impacted by the contents of the story.

Because this study was completed remotely, the protocol asked participants to ensure proper administration of the auditory stimuli. Participants underwent a short audio functional test to check that they could properly hear the recordings on their respective devices. Participants noted what device they would be listening with (computer speakers, iPad/tablet/cellphone speakers, headphones, or other), played a short clip of a single tone, and then marked whether they could hear the noise. Just prior to the stimuli, participants had to mark two boxes signifying that their audio was working and that they were “ready and paying attention.” When they proceeded to the auditory stimuli, the stimuli only played once and would not allow for repeat.

Demographics

Between the administration of the two stimuli, participants were asked a series of questions regarding their demographics, including age, gender, sex at birth, year in school, and race/ethnicity (see Table 1). This served to not only provide more information regarding their demographics, but also as a small break between listening portions. The aim of this break was to allow time for participants to return to the baseline affective state if it had changed due to the previous stimulus.

Procedures

The study protocol was reviewed and approved by the University’s Institutional Review Board prior to data collection. Due to the COVID­19 pandemic, all research was required to be remote, thus participants completed the experiment through an online Qualtrics survey. They filled out an initial PANAS questionnaire (Watson et al., 1988) to assess their baseline emotional state. They then completed a computer audio functionality test to ensure that they would be able to hear the audio recordings. Next, participants were asked if they were ready to listen

Durdle, and Miller | Hormonal Contraception and Memory

and confirmed that they were free of distractions before proceeding to the story narration task. Participants then listened to a recorded narration of either a story with negative emotional content or neutral content (randomized order for each participant; see Appendix B). Another PANAS was administered to assess any changes in emotional states following the story. They were next instructed to type a summary recalling the story they previously heard. Participants were instructed to take at least two minutes to input their responses and needed to have a minimum of 75 words, otherwise, they would not be permitted to move forward in the study. After the first recall portion, participants were asked a series of demographic questions, and then were instructed to listen to the story narration they had not heard. Upon completion of listening to the second story, participants were tasked with a final PANAS questionnaire to reassess their emotional state. They then were given the same instructions to write a summary of what they recalled from the narrated story. Therefore, this experiment utilized a 2 (within­subjects story condition: emotional and neutral) x 2 (within­subjects recall content: gist and neutral) x 2 (between­subjects contraceptive use: HC or NC) design on free recall for auditory stories.

Scoring Free Recall

The present study categorized participants’ responses to ambiguous stimuli into one of two categories: hits (H; correctly identifying a stimulus that was present) or false alarms (FA; identifying a stimulus that was not present). From this, scores can be formulated into rates for each category. The hit rate (HR) is obtained by dividing participants’ total hits by the total possible correct categories whereas the false alarm rate (FAR) is the total false alarms divided by the total possible incorrect categories (see equations below):

HR = (number of H) / (number of total possible H)

FAR = (number of FA) / (number of total possible FA)

A recall was considered more accurate if it had a higher HR and lower FAR. This scoring method was implemented in conjunction with the gist and detail categorization rules used by Nielsen et al. (2011). In the current study, participants’ free­recall responses were scored by two independent judges, who qualitatively scored using the categories Correct Gist Information, Correct Detail Information, Incorrect Gist Information, and Incorrect Detail Information. In line with free recall scoring guidelines used by Diamond et al. (2020), an item category for both correct and incorrect information was created if two or more participants mentioned the item. If a portion of the participant’s response contained an aspect belonging to a particular category, it was

Simonson,

Hormonal Contraception and Memory | Simonson, Durdle, and Miller

marked with a 1, whereas the absence of an aspect was marked with a 0. Disagreements were settled by a third judge. Interrater reliability between the two judges was assessed using Cohen’s kappa, κ = .832, which is considered near­perfect agreement (McHugh, 2012)

Both the neutral and emotional content story consisted of 13 Correct Gist categories. The neutral story contained 26 Correct Detail categories, whereas the emotional story contained 31. The neutral story had six Incorrect Gist categories, whereas the emotional story

4 Hit Rates

Note. Descriptive Statistics for hit rates across conditions, given as Mean (Standard Deviation).

was determined to have eight. In terms of Incorrect Detail categories, the neutral story contained 19, whereas the emotional content story had 28. Because detail and gist categories have unequal amounts, they were analyzed separately.

Results

Note. Descriptive Statistics for false alarm rates across conditions, given as Mean (Standard Deviation).

FIGURE 1 Hit Rates

Recall Accuracy

To assess the effects of story content and birth control status on gist HR, a 2 (birth control status: HC or NC) x 2 (emotional content condition: neutral or emotional) mixed ANOVA with a within­subjects factor of story content was used. There was a main effect of story content, such that gist HR was higher for the neutral condition (M = 0.71, SD = 0.17) than for the emotional content condition (M = 0.59, SD = 0.22), F(1, 88) = 31.19, p < .001, η2 = .09 (see Table 4 and Figure 1). There was no significant main effect of birth control status on gist HR, indicating that HC women (M = 0.64, SD = 0.20) and NC women (M = 0.66, SD = 0.21) did not differ significantly in their recall for gist elements, F(1, 88) = 0.13, p = .72, η2 < .01. The interaction effect between birth control status and story content was also not significant, F(1, 88) = 0.69, p = .41, η2 < .01. To assess the effects of story content and birth control status on gist FAR, a 2 (birth control status: HC or NC) x 2 (emotional content condition: neutral or emotional) mixed ANOVA was used. There was a near­ significant main effect of story content on gist FAR (Emotional: M = 0.04, SD = 0.10; Neutral: M = 0.07, SD = 0.12), F(1, 88) = 3.01, p = .09, η 2 = .02, and there was no significant main effect of birth control status on gist FAR (HC: M = 0.06, SD = 0.11; NC: M = 0.06, SD = 0.11), F(1, 88) = 0.11, p = .74, η2 < .01. Additionally, the interaction between birth control status and story content on gist FAR was not significant, F(1, 88) = 1.08, p = .30, η2 < .01 (see Table 5 and Figure 2).

To assess the effects of story content and birth control status on detail HR, a 2 (birth control status: HC or NC) x 2 (emotional content condition: neutral or emotional) mixed ANOVA with a within­subjects factor of story content was used. There was a main effect of story content, where detail HR was higher for the neutral condition (M = 0.43, SD = 0.21) than for the emotional content condition (M = 0.35, SD = 0.16), F (1, 88) = 18.22, p < .001, η 2 = .05 (see Table 4 and Figure 1). There was not a significant main effect of birth control status on detail HR, indicating that HC women (M = 0.38, SD = 0.18) and NC women (M = 0.40, SD = 0.20) did not differ significantly in their recall for detail elements, F(1, 88) = 0.37, p = .55, η2 < .01. The interaction effect between birth control status and story content was also not significant, F(1, 88) = 0.33, p = .57,

Note. Mean hit rates for hormonal contraceptive versus naturally cycling women’s free recall in the emotional and neutral conditions. Error bars represent standard error.

η2 < .01. Finally, to assess the effects of story content and birth control status on detail FAR, a 2 (birth control status: HC or NC) x 2 (emotional content condition: neutral or emotional) mixed ANOVA was used. There was no significant main effect of story content on detail FAR (Emotional: M = 0.03, SD = 0.04; Neutral: M = 0.04, SD = 0.05), F(1, 88) = 2.36, p = .13, η2 = .01, nor was there a significant main effect of birth control status on detail FAR (HC: M = 0.04, SD = 0.04; NC: M = 0.03, SD = 0.04), F(1, 88) = 1.02, p = .32, η2 < .01. Additionally, the interaction between birth control status and story content on detail FAR was not significant, F(1, 88) = 2.75, p = .10, η2 = .01.

Assessing Current Emotional State

It is important to consider the possibility of current emotional states as a moderator in this experiment. To see if participants were significantly emotionally aroused by our emotional content auditory story, and not emotionally aroused at baseline or by the neutral stimulus, they completed a PANAS at three different time points: (a) before hearing any stimuli, (b) after the first randomized stimulus, and (c) after the final auditory story stimulus. To assess this, a 2 (birth control status: NC or HC) x 2 (emotional content condition: baseline, postemotional, or postneutral) mixed model ANOVA was used. There was a significant main effect of the story condition (Baseline: M = 17.9, SD = 6.93; Emotional: M = 16.90, SD = 7.31; Neutral: M = 14.00, SD = 5.79) on negative PANAS scores, F(1, 176) = 24.79, p < .001, η2 = .06), however participants showed lower negative affect scores following either narration than at baseline (see Table 6). Post hoc analyses with Bonferroni correction were then run to determine which pairwise comparisons within the emotional content condition factor were driving the main effect. Dependent­samples t tests revealed that this effect was driven by significantly lower scores following the neutral story than at baseline, t(88) = 7.45, p < .001, d = 0.79, as well as significantly lower scores following the neutral story than following the emotional story, t(88) = 4.58, p < .001, d = 0.59. There was no significant difference between negative PANAS scores following the emotional content story and at baseline, t(88) = 1.82, p = .22, d = 0.19. There was no main effect of birth control status on negative arousal, nor was there a significant interaction between birth control group and stimulus administration. This shows that participants, regardless of contraceptive status, were not measurably negatively aroused by either story, however, they might have been significantly calmed by administration of the neutral narration.

To prevent any carry ­ over emotional arousal effects from one condition to another, demographic

information was collected between recall of the first narration and administration of the second narration. Additionally, the order of stimuli was counterbalanced to further prevent order effects. To further confirm that there was no effect of order of stimulus administration, an independent­sample t­test was conducted between negative PANAS difference scores (postneutral score subtracted from postemotional score) and order of stimuli. There was no significant difference in these scores for participants who heard the emotional content story first ( M = 3.77, SD = 6.62) and those who heard the neutral story first (M = 2.11, SD = 5.41), t(88) = 1.30, p = .20.

Discussion

The current study sought to investigate whether oral hormonal contraception has a significant effect on recall accuracy for a narrated story with emotional content

Note. Descriptive Statistics for negative PANAS scores, given as Mean (Standard Deviation). Higher scores indicate more negative affect.

TABLE 6

Negative PANAS Scores

FIGURE 2

False Alarm Rates

Note. Mean false alarm rates for hormonal contraceptive versus naturally cycling women’s free recall in the emotional and neutral conditions. Error bars represent standard error.

Hormonal Contraception and Memory | Simonson, Durdle, and Miller

compared to a neutral story. Contrary to previous literature, HC and NC women did not differ significantly in their recalls of emotional or neutral episodic events, regardless of gist or detail content. However, all participants, regardless of contraceptive status, recalled more gist information (regardless of story valence) and recalled more information for the neutral story than the emotional story (regardless of gist or detail components).

Given that the story containing emotional content was intended to be negatively arousing, negative affect scores from the PANAS were anticipated to be highest following the emotional stimulus, with lower scores at baseline and following the neutral stimulus. Negative PANAS scores indicated that participants were not significantly aroused by the emotional stimulus. However, inferences can be made on the effects of an auditory story containing emotional content in general. We found one area of significant change in negative PANAS scores following stimulus administration, though it was following the neutral story. Although participants did not appear to be significantly aroused by the emotional story, NC women appear to have been significantly calmed by both stories. This may indicate that all participants had a particularly high level of negative affect at baseline, and neither story significantly elevated that level further.

The lack of significant differences between baseline and post­negative story PANAS scores may indicate that the story did not induce a stress response. This could possibly explain the discrepancy between our findings and others’ regarding HCs and emotional memory, as memory for emotional stories often differ depending on whether a cortisol response was evoked (Buchanan &Tranel, 2008). This factor may compound with the HC variable, as birth control can affect the influence of cortisol on memory retrieval for the free recall of words and numbers (Kuhlmann & Wolf, 2005), and therefore HCs might not have had as strong of an effect due to the lack of a cortisol response. However, future research will need to be conducted to directly assess this biological explanation.

The results of this study do not indicate a significant difference in recall for emotional auditory stimuli when comparing women on an HC and NC women, regardless of recall content type. NC women and women on a HC demonstrated similar recall accuracy for both stories, even when considering gist versus detail elements. Regardless of contraceptive use, participants had higher HR for the neutral condition than the emotional content condition, in contrast to prior research.

Analyses did reveal interesting differences in accuracy rates between the emotional content story and neutral story. Contrary to previous studies, participants in this study, regardless of contraceptive status, had

higher HR when recalling a neutral auditory story than for an emotional auditory story. Therefore, the hypothesis that participants’ recalls would be more accurate for the emotional content condition was incorrect, because HR was higher for the neutral condition than the emotional content condition. However, FAR was also higher for the neutral condition, indicating that participants might have been more liberal in their recall responses for the neutral story. Both of the hypotheses pertaining to the interaction between birth control status and the emotional content condition were not supported, because there was no significant difference between NC and HC women in the emotional content condition. The hypothesis that NC and HC women would not significantly differ in the neutral condition was supported. Although the findings of the current study did not match the hypothesized results, a few important implications can be drawn. The finding that participants had significantly higher HR for neutral content recall than for emotional content conflicts with some of the previous literature, including the findings of Nielsen and colleagues (2011). One important distinction between the current study and previous research regarding emotional memory lies in the current study’s analysis of participants’ criterion as it pertains to recalling an emotional versus a neutral story. Previous studies have found a difference between groups in recall for gist and detail elements of a story (Nielsen, 2011), though they did not account for FAR. This could be a crucial distinction in whether a difference actually occurs for the memory itself, or if the difference lies in a participant’s criterion for reporting a memory. Furthermore, a memory cannot necessarily be considered more accurate from correct responses alone, as the amount of incorrect information recalled must also be considered.

One limitation of this study would be its remote administration due to the COVID ­ 19 pandemic. Participants completed the study remotely on their own personal devices in an environment that could not be controlled completely, and thus participants might not have paid full attention to the audio recordings. Another limitation may lie in the auditory stimuli themselves. To control for the effects of tonality on emotional arousal, both the neutral and emotional stimuli were recorded with a very neutral, nonfluctuating tone. However, perhaps a large aspect of what is necessary to be aroused by auditory stimuli alone is within the delivery of the story itself. Removing vocal inflections from the auditory story may be directly removing an important component of emotional storytelling and could potentially explain the lack of significance in negative arousal scores following the emotional stimulus. Although the neutral story was created to mimic an everyday scene, individual

participants’ personal preferences and experiences might have led to a positive interpretation of the story as opposed to neutral. Nevertheless, the neutral story evoked less of a negative response than the emotional content story, as was intended. Additionally, the use of pre­existing groups (i.e., not assigning participants to the HC or NC condition) does introduce potential confounding variables. Although various demographic information was collected, we did not control for demographic differences that might have impacted performance on the task, such as age, religiosity, and education. Given that there was an unequal distribution of age, year in school, and ethnicities into HC and NC groups (see Tables 1 and 2), it is possible that any of these may have factored into the experimental results. However, the contribution of demographic variables into recall accuracy is beyond the scope of this experiment.

Future iterations of the study should incorporate various changes to the design. First, in­person administration of the experiment would allow for better control of the participants’ environment, which would, in turn, lead to fewer confounds. Second, allowing for natural vocal intonations in the recorded narrative may allow for larger emotional arousal by the emotional content story, as previous research has demonstrated limbic activation (in particular the amygdala) associated with emotional responses to changes in affective prosody (Brück et al., 2011). A control group that reads the stories as opposed to listening to them would also allow for further conclusions on the direct effects of an auditory component, and assessments of auditory processing abilities would aid in clarifying whether all participants were adequately absorbing the auditory material. Finally, validating the level of valence and arousal for the auditory stimuli is an important step for future iterations of the current study. Using a separate sample of participants to assess their momentary affective state at baseline and also poststimulus would be valuable in guaranteeing that the stimuli properly arouse participants as intended. A longer delay between story administrations may also have resulted in PANAS scores that more closely resembled the anticipated results.

Information was collected regarding HC participants’ specific birth control pills, including brand, dosage, and approximately how long they had been on their current form of birth control (see Table 2). For NC participants, data collection included menstrual health information including the first day of their last menstrual period and their confidence that the reported date is accurate. These questions were meant to approximate which phase of menstruation they may be in at the time of the study. Due to large individual variabilities in menstrual cycles, we were not able to

draw definitive conclusions on the cycle stage based on these self ­ reports. However, previous research demonstrating the impact of the menstrual cycle stage on memory has indicated that this might have been an important factor to consider, and future research should incorporate a reliable measure of cycle stage. The current study focused primarily on negative PANAS scores, because the emotional stimulus was intended to be negatively arousing. However, future analyses should include positive PANAS scores to see if there were any changes in momentary positive affect following either the emotional or the neutral stimulus. Finally, a deeper look into the contributions of the correct and incorrect categories may elucidate whether there was a difference between the number of categories that HC and NC women contributed. The current analyses created categories if the element reported was mentioned by two or more participants; however, this was not broken down into HC and NC participants. Perhaps there would have been a significant difference between groups that would be clarified by further analyses. Last, due to time constraints, there was not a significant time delay between the administration of the story and free recall, as well as between the stories themselves. Some evidence has suggested that the effects emotions have on memory are most apparent for memories for events following a significant delay (Yonelinas & Ritchey, 2015), and therefore more significant effects might be observed with a longer duration between study elements.

The present study’s findings reveal a potential difference in HCs’ impacts on emotional memory that is domain­specific. While previous research has demonstrated differences in recall for emotional stories with a visual component, the current study’s findings allude to an alternative effect when stimuli are exclusively auditory. These results add to the knowledge base of an incredibly complex relationship between hormonal contraception and cognition. However complicated, this area of research is paramount in understanding the short and long­term effects of a drug utilized by hundreds of millions of women worldwide.

References

Baldassano, C., Chen, J., Zadbood, A., Pillow, J. W., Hasson, U., & Norman, K. A. (2017). Discovering event structure in continuous narrative perception and memory. Neuron, 95(3), 709–721. http://doi.org/10.1016/j.neuron.2017.06.041

Buchanan, T. W., & Tranel, D.x (2008). Stress and emotional memory retrieval: Effects of sex and cortisol response. Neurobiology of Learning and Memory, 89(2), 134–141. https://doi.org/10.1016/j.nlm.2007.07.003

Bowen, H. J., Kark, S. M. & Kensinger, E. A. (2018). NEVER forget: Negative emotional valence enhances recapitulation. Psychonomic Bulletin & Review, 25, 870–891. https://doi.org/10.3758/s13423-017-1313-9

Boss, L., Kang, D. H., Marcus, M., & Bergstrom, N. (2014). Endogenous sex hormones and cognitive function in older adults: A systematic review. Western Journal of Nursing Research, 36(3), 388–426. https://doi.org/10.1177/0193945913500566

Brück, C., Kreifelts, B., & Wildgruber, D. (2011). Emotional voices in

Simonson, Durdle, and Miller | Hormonal Contraception and Memory

WINTER 2023

PSI

Hormonal Contraception and Memory | Simonson, Durdle, and Miller

context: A neurobiological model of multimodal affective information processing. Physics of Life Reviews, 8(4), 383–403. https://doi.org/10.1016/j.plrev.2011.10.002

Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). β-Adrenergic activation and memory for emotional events. Nature, 371, 702–704.

Cahill, L., & McGaugh, J. L. (1995). A novel demonstration of enhanced memory associated with emotional arousal. Consciousness and Cognition, 4(4), 410–421. https://doi.org/10.1006/ccog.1995.1048

Cahill, L., & van Stegeren, A. (2003). Sex-related impairment of memory for emotional events with β-adrenergic blockade. Neurobiology of Learning and Memory, 79(1), 81–88. https://doi.org/10.1016/s1074-7427(02)00019-9

Christin-Maitre, S. (2013). History of oral contraceptive drugs and their use worldwide. Best Practice & Research Clinical Endocrinology & Metabolism, 27(1), 3–12. https://doi.org/10.1016/j.beem.2012.11.004

Cooper, D. B., Patel, P., & Mahdy, H. (2022). Oral contraceptive pills. In StatPearls [Internet]. StatPearls Publishing.

Diamond, N. B., Armson, M. J., & Levine, B. (2020). The truth is out there: Accuracy in recall of verifiable real-world events. Psychological Science, 31(12), 1544–1556. https://doi.org/10.1177/0956797620954812

Díaz-García, A., González-Robles, A., Mor, S., Mira, A., Quero, S., García-Palacios, A., Baños, R. M., & Botella, C. (2020). Positive and Negative Affect Schedule (PANAS): Psychometric properties of the online Spanish version in a clinical sample with emotional disorders. BMC Psychiatry, 20(1). https://doi.org/10.1186/s12888-020-2472-1

Dunsmoor, J. E., Murty, V. P., Davachi, L., & Phelps, E. A. (2015). Emotional learning selectively and retroactively strengthens memories for related events. Nature, 520(7547), 345–348. https://doi.org/10.1038/nature14106

Fleischman, D. S., Navarrete, C. D., & Fessler, D. M. (2010). Oral contraceptives suppress ovarian hormone production. Psychological Science, 21(5), 750–752. https://doi.org/10.1177/0956797610368062

Pauls, F., Petermann, F., & Lepach, A. C. (2013). Gender differences in episodic memory and visual working memory including the effects of age. Memory, 21(7), 857–874. https://doi.org/10.1080/09658211.2013.765892

Griksiene, R., & Ruksenas, O. (2011). Effects of hormonal contraceptives on mental rotation and verbal fluency. Psychoneuroendocrinology, 36(8), 1239–1248. https://doi.org/10.1016/j.psyneuen.2011.03.001

Herrera, A. Y., Velasco, R., Faude, S., White, J. D., Opitz, P. C., Huang, R., Tu, K., & Mather, M. (2020). Brain activity during a post-stress working memory task differs between the hormone-present and hormone-absent phase of hormonal contraception. Neurobiology of Stress, 13, 100248. https://doi.org/10.1016/j.ynstr.2020.100248

Kensinger, E. A., & Corkin, S. (2003). Effect of negative emotional content on working memory and long-term memory. Emotion, 3(4), 378–393. https://doi.org/10.1037/1528-3542.3.4.378

Kensinger, E. A. (2009). Remembering the details: Effects of emotion. Emotion Review, 1(2), 99–113 https://doi.org/10.1177/1754073908100432

Kuhlmann, S., & Wolf, O. T. (2005). Cortisol and memory retrieval in women: Influence of menstrual cycle and oral contraceptives. Psychopharmacology, 183(1), 65–71. https://doi.org/10.1007/s00213-005-0143-z

Lewis, C. A., Kimmig, A. C. S., Zsido, R. G., Jank, A., Derntl, B., & Sacher, J. (2019). Effects of hormonal contraceptives on mood: A focus on emotion recognition and reactivity, reward processing, and stress response. Current Psychiatry Reports, 21(11). https://doi.org/10.1007/s11920-019-1095-z

McHugh, M. L. (2012). Interrater reliability: The kappa statistic. Biochemia Medica, 22(3), 276–282. https://doi.org/10.11613/bm.2012.031

Mordecai, K. L., Rubin, L. H., & Maki, P. M. (2008). Effects of menstrual cycle phase and oral contraceptive use on verbal memory. Hormones and Behavior, 54(2), 286–293. https://doi.org/10.1016/j.yhbeh.2008.03.006

Nielsen, S. E., Ertman, N., Lakhani, Y. S., & Cahill, L. (2011). Hormonal contraception usage is associated with altered memory for an emotional story. Neurobiology of Learning and Memory, 96(2), 378–384. https://doi.org/10.1016/j.nlm.2011.06.013

Petersen, N., & Cahill, L. (2015). Amygdala reactivity to negative stimuli is influenced by oral contraceptive use. Social Cognitive and Affective Neuroscience, 10(9), 1266–1272. https://doi.org/10.1093/scan/nsv010

Pletzer, B. A., & Kerschbaum, H. H. (2014). 50 years of hormonal contraception— Time to find out, what it does to our brain. Frontiers in Neuroscience, 8 https://doi.org/10.3389/fnins.2014.00256

Pletzer, B., Kronbichler, M., Aichhorn, M., Bergmann, J., Ladurner, G., & Kerschbaum, H. H. (2010). Menstrual cycle and hormonal contraceptive use modulate human brain structure. Brain Research, 1348, 55–62.

https://doi.org/10.1016/j.brainres.2010.06.019

Presuss, D., Schoofs, D., & Wolf, O. T. (2009). Associations between endogenous cortisol levels and emotional memory in young women: Influence of encoding instructions. Stress, 12(5), 379–387.

https://doi.org/10.1080/10253890802524592

Rentz, D. M., Weiss, B. K., Jacobs, E. G., Cherkerzian, S., Klibanski, A., Remington, A., Aizley, H., & Goldstein, J. M. (2017). Sex differences in episodic memory in early midlife: Impact of reproductive aging. Menopause, 24(4), 400–408. https://doi.org/10.1097/GME.0000000000000771

Rivera, R., Yacobson, I., & Grimes, D. (1999). The mechanism of action of hormonal contraceptives and intrauterine contraceptive devices. American Journal of Obstetrics and Gynecology, 181(5), 1263–1269.

https://doi.org/10.101https://doi.org/10.1016/S0002-9378(99)70120-16/S00029378(99)70120-1yonelina

Taylor, C. M., Pritschet, L., & Jacobs, E. G. (2021). The scientific body of knowledge–Whose body does it serve? A spotlight on oral contraceptives and women’s health factors in neuroimaging. Frontiers in Neuroendocrinology, 60, 100874. https://doi.org/10.1016/j.yfrne.2020.100874

Tulving, E. (1993). What is episodic memory? Current Directions in Psychological Science, 2(3), 67–70. https://doi.org/10.1111/1467-8721.ep10770899

Tulving, E. (2005). Episodic memory and autonoesis: Uniquely human? In H. S. Terrace & J. Metcalfe (Eds.), The missing link in cognition: Origins of selfreflective consciousness (pp. 3–56). Oxford University Press. https://doi.org/10.1093/acprof :oso/9780195161564.003.0001

Warren, A. M., Gurvich, C., Worsley, R., & Kulkarni, J. (2014). A systematic review of the impact of oral contraceptives on cognition. Contraception, 90(2), 111–116. https://doi.org/10.1016/j.contraception.2014.03.015

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. https://doi.org/10.1037/0022-3514.54.6.1063

Wegerer, M., Kerschbaum, H., Blechert, J., & Wilhelm, F. H. (2014). Low levels of estradiol are associated with elevated conditioned responding during fear extinction and with intrusive memories in daily life. Neurobiology of Learning and Memory, 116, 145–154. https://doi.org/10.1016/j.nlm.2014.10.001

Yonelinas, A. P., & Ritchey, M. (2015). The slow forgetting of emotional episodic memories: An emotional binding account. Trends in Cognitive Sciences, 19(5), 259–267. https://doi.org/10.1016/j.tics.2015.02.009

Author’s Note

Jessica M. Simonson http://orcid.org/0000­0003­0173­8042

Courtney A. Durdle https://orcid.org/0000­0002­8582­0977

The authors thank Courtney Durdle for her indispensable help on this project and Dr. Michael Miller for his guidance. Additionally, the authors would like to acknowledge two undergraduate research assistants in the Miller Memory Lab, Ziyuan Chen and Natasha Pansare, for qualitatively coding the participant responses. Special thanks to Dr. Tyler Santander for his help with the analysis of this project. Finally, the authors thank their peers, mentors, friends, and family for their constant support throughout this project.

Materials and data for this study can be accessed at https://osf.io/k7h8c/. The authors have no known conflict of interest to disclose. Research was sponsored by the U.S. Army Research Office and accomplished under cooperative agreement W911NF­19­2­0026 for the Institute for Collaborative Biotechnologies.

Correspondence concerning this article should be addressed to Jessica Simonson. Email: jessicasimonson@ucsb.edu

273

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

Eligibility Questionnaire

1. Are you currently on any form of hormonal contraceptive?

• No

• Yes, I am on an oral contraceptive/birth control pill

• Yes, I am on another form of contraceptive other than the pill (if they select this option they are deemed ineligible for the current study)

If they answer “No” to question 1:

2. Have you been on any type of hormonal contraceptive in the past 3 months?

• Yes (if they select this option they are deemed ineligible for the current study)

• No

• Unsure (if they select this option they are deemed ineligible for the current study)

If they answer “Yes” to question 1:

3. Have you been on this same type of hormonal contraceptive for at least 3 months?

• No, I have been on this hormonal contraceptive for less than 3 months (if they select this option they are deemed ineligible for the current study)

• Yes, I have been on this hormonal contraceptive for 3 months or more

Hormonal contraceptive pathway:

If they answer “Yes” to question 3:

4. How long have you been on this same type of hormonal contraceptive in months? (Please give your best guess)

5. Is your oral contraceptive a combined (estrogen and progestin) or progestin-only pill?

• Combined pill (estrogen and progestin)

• Progestin-only pill (minipill)

• Unsure (All selected options remain eligible to continue, this is just to collect further data)

6. What type of oral contraceptive are you on? Please specify the brand and dosage if possible.

No hormonal contraceptive pathway:

If they answer “No” to question 2:

3. When was the first date of your last menstrual period? (Please look at a calendar if you need)

4. On a scale of 1 to 10, how confident are you that you gave an accurate estimate of the first date of your last menstrual period? (Slider function from 1–10 displayed here for them to toggle)

APPENDIX B

Story Scripts

Arousal Version

Robin and her son, Teddy, are leaving their New York apartment on a cold December morning. She is taking him to visit his father’s workplace, which is a 20-minute walk from their home. The father is a laboratory technician at Victory Memorial Hospital.

They look both ways before crossing Broadway, a busy street.

While crossing the road, Teddy is struck by a white Mercedes, which critically injures him. A man at a nearby bus stop calls 911 for an ambulance.

At the hospital, the staff prepare the emergency room, to which Teddy is rushed.

An image from a brain scan machine shows severe bleeding in the left side of Teddy’s brain. All morning long, Dr. Smith’s surgical team struggled to save Teddy’s life.

Specialized surgeons were able to stop the bleeding in his brain, though he did sustain blindness in his right eye. After the surgery, while the father stayed with Teddy, the mother left to phone her other child, Samantha’s, preschool. Feeling distraught, she phones the preschool to tell them she will pick up Samantha as soon as possible. Heading to pick up her child, she hails a taxi at the number nine bus stop.

Neutral Version

Two women decide to go to the mall after they get off work on Sunday afternoon. Susan picked up Kate in her red BMW at 4:00.

They headed to Westfield mall in Placerville.

Once they got to the mall, they headed straight for Dottie’s clothing store for a big sale. At the store they tried on lots of different sweaters, but neither of them bought anything. After that, they decided to go to the mall’s candy shop because Kate had a gift certificate for $10.00. They looked around the shop for a while and tried different samples.

They each decided on buying an assortment of chocolate and split a pound of saltwater taffy. Next, they decided to get some dinner at a local diner located down the street from the mall. Once there, Kate ordered a bacon cheeseburger and fries, and Susan ordered a garden salad. While enjoying their meals they talked about work.

After they finished, Susan dropped Kate off at her house before heading home herself.

Social Media Use Motives: An Influential Factor in User Behavior and User Health Profiles

Carson R. Ewing, Christian Nienstedt, Robert R. Wright*, and Samuel Chambers Department of Psychology, Brigham Young University–Idaho

ABSTRACT. Social media use is an increasingly popular behavior and has been differentially correlated with both positive and negative user characteristics and health indicators. However, the current literature has yet to fully explore the role of social media user motivations in the relationship between social media use, user characteristics, and user health. Aiming to address this gap, the current cross­sectional study examined data gathered from 1,547 online undergraduate students who completed an online survey regarding their social media motives (entertainment, information seeking, personal utility, convenience), user demographics, behavior, and health indicators (behavioral, mental, physical, social). Results demonstrated unique differences in social media motives according to social media platform, as well as user demographic characteristics, electronic media use, and health indicators, especially for the entertainment motive. Although the entertainment motive was not significantly different between women and men, t(1545) = 1.78, p = .07, d = 0.10, it was notably different across class level, F(4, 1542) = 8.03, p < .001, η2 = .03, and relationship status, F(4, 1542) = 15.63, p < .001, η2 = .04, and it was the motive most strongly related to user behavior and health indicators. Additionally, entertainment motivation was moderately correlated with problematic smartphone use (r = .45, p < .001) and had a stronger correlation with study variables than any other social media motive. These findings suggest that motives, especially entertainment, are important for understanding social media use in user interface, behavior, and health.

Keywords: social media, motivation, uses and gratification, mood management theory, problematic smartphone use

In recent years, social media has become increasingly popular, as it is widely available and offers numerous benefits to consumers. Indeed, a recent estimate suggested that more than 80% of Americans are regular users of social media (Auxier & Anderson, 2021), and these numbers likely increased following the onset of the COVID­19 pandemic (Vargo et al., 2021; Wright et al. 2022). Moreover, use of social media has been consistently associated with a large variety of negative mental, physical, and social health behaviors (Naslund et al., 2020; Woods & Scott, 2016; Wright et al., 2021), including addictive tendencies in the use of cell phones and social media platforms (Hou et al., 2019; Zivnuska et al., 2019). Although most of these relationships suggest health detriments to social media users, especially over prolonged periods, there are some discrepancies in the literature that may be indicative of moderating factors

such as platform­specific effects (Wright et al., 2020, 2021) or social media use motivations (Al­Menayes, 2015). Social media use motivations may explain some of these discrepant findings. However, this remains a relatively underexplored area in the literature, especially regarding how motivations may shape user technology behavior, and influence user wellness. Thus, the current study seeks to address this gap in the literature.

Social Media Use: Two Theories of Motivation

It has been well­documented that social media user motivations vary and are related to individual user experiences and outcomes (e.g., Campisi et al., 2015). Although motivations for using social media both across and within persons fluctuate, it is important to note that there are some consistencies observed in the literature. For instance, many studies have identified entertainment

and information seeking as two primary motivations for using social media (Jarman et al., 2021; Meng et al., 2020; Stockdale & Coyne 2020;). In a large­scale study on motivations for social media use, Al­Menayes (2015) identified four motivations, including entertainment, information seeking, personal utility, and convenience. Moreover, in that same study, Al­Menayes (2015) noted that contextualizing these motivations is particularly important to understanding their role in social media use. Indeed, two prominent models of motivation provide useful theoretical frameworks for this contextualization.

First, the uses and gratification (U&G) theory of media use posits that media users seek out media among a large variety of options that will meet their needs and lead to the greatest amount of gratification (Whiting & Williams, 2013). For example, those looking for something to distract themselves might watch videos or play games, but those looking to find a good recipe for dinner might peruse various blogs or cooking websites to find the information they are seeking. As individuals find activities that gratify their needs, they are more likely to repeat these same activities. U&G holds that gratification taken from one form of media can vary based on the different purposes for which the individual chose to partake in that media (Pelletier et al., 2020). From this viewpoint, simply studying participants’ time spent using social media would not provide complete details regarding how social media use may impact users. Moreover, U&G theory espouses four key assumptions: (a) media use is goal­driven; (b) people use media to meet needs and desires; (c) social and physical factors mediate the use of media; and (d) the use of media and interpersonal communication are related (Kircaburun et al., 2020). As such, media users might not always achieve the gratification they seek, which could yield differential outcomes depending on their motivation for using social media, the outcomes of that use, and how well those two matched. Therefore, from the U&G perspective, examining specific motivations for using social media may provide a more accurate understanding of differential user outcomes, particularly when the original intended motivation is clear (e.g., entertainment, information seeking).

Second, mood management theory (MMT) proposes that individuals rearrange their stimulus environment to eliminate negative mood states and to maintain heightened positive mood states (Carpentier, 2020). According to this approach, the use of social media is just one possible means for stimulus control (Reinecke, 2017). Indeed, before many forms of media existed, stimulus rearrangement required entire changes of scenery, but the advent of modern technology has made this nearly instantaneous by simply looking at a small,

portable handheld device. Furthermore, MMT assumes that the mood management motivation for media use is largely an unconsciously learned process (Reinecke, 2017) such that an individual using social media to manage their own mood might have different results depending on their original purpose. For instance, a bored individual may implicitly seek out social media as a form of entertainment for mood improvement and to relieve their boredom. In comparison, a different individual with a specific problem at hand might use social media to find information but, due to mood management, end up spending more time on social media than intended and thus gain a negative mental effect. As such, the MMT perspective accounts for instances when social media use motivations may be more subtle and implicit (e.g., entertainment, convenience).

Social Media Use, User Health, and Motivations

When examining user behavior and health outcomes tied to social media use, most of the literature has focused on time spent on social media in general and its relationship with poor mental health (e.g., Cunningham et al., 2021; Ivie et al., 2020). These findings span many mental health variables including life satisfaction (Kross et al., 2013), depression and anxiety (Woods & Scott, 2016), body image (Fardouly & Vartanian, 2016; Rodgers et al. 2020), mood, and perceived stress (Wright et al., 2021). In a recent meta­analysis on the topic, Huang (2020) identified clear correlative relationships between increased social media use and lower reported self­esteem and life satisfaction alongside higher reports of loneliness and depressive symptoms. Other studies have uncovered a positive relationship between social media use and increases in overall time spent watching television, daily screen time, and number of social media platforms (e.g., Wright et al., 2022), behaviors that are themselves strongly related to problematic health profiles.

Not all the research in this area is conclusive, however. For instance, an eight­year longitudinal study of 500 adolescents conducted by Coyne et al. (2020) found no clear association between social media use and mental health. Additionally, Berryman et al. (2018) reported similar null results among a sample of over 450 young adults. Going one step further, some research has identified evidence of possible user health benefits of using social media. In fact, Wright et al. (2020) found that certain social media platforms when used for video chatting were associated with increased feelings of well­being, suggesting some social media may elicit beneficial health outcomes. Similarly, Campisi et al. (2015) reported that those using social media to stay in touch with friends reported a higher quality of life, but those using it for dating purposes had a lower

Examining Social Media Motives | Ewing, Nienstedt, Wright, and Chambers

quality of life. Such stark contrasts between results may be indicative of other moderating factors, such as the motivation behind social media use (Chen et al., 2015; Kim & Lee 2016).

Two recent studies have highlighted unique health profiles associated with different social media platforms that further imply motives may be involved in social media use correlations. First, among more than 600 college students, Wright et al., (2020) studied user health profiles associated with prominent social media including Facebook, Instagram, and Snapchat. They found that those who used video­based social media had better health profiles (e.g., less loneliness, depressive symptoms, anxiety, life satisfaction) than those who used more image and text based social media sites, suggesting a differential association. As a follow­up to this study, Wright et al. (2021) found similar results among a more diverse sample of MTurk users and an expanded array of social media platforms and health variables. Although these studies point to a connection between the specific qualities of the platforms with unique outcomes, the researchers suggested that motivations may play a decisive role in these observed differences.

In further support, some recent studies have highlighted the relationship between social media motivations and user demographic and behavior characteristics. First, in a study by Twenge and Martin (2020), adolescent boys and girls differed in their online media activities, time spent on media, and the psychological impact of media used, suggesting gender differences. Moreover, another study by Wright et al. (2018) identified strong support for differences in terms of perceived loneliness and social media use based on gender, year of college education, and relationship status, suggesting that social media motivations may also differ. Relatedly, the literature has suggested that the development of problematic smartphone use is associated with time spent using social media, especially for young people (Naslund et al., 2020) and that this relationship may be getting stronger over time (Wright et al., 2022).

Finally, a few studies provide some preliminary evidence of how certain motives may be related to specific health outcomes. In a large study among more than 8,000 participants, Meng et al. (2020) found a positive relationship between the motive of entertainment and problematic smartphone use. However, there was no such relationship for the motive of learning, suggesting that social media motives can influence other potentially problematic user behavior. In a longitudinal study conducted among adolescents, Stockdale and Coyne (2020) found that boredom (i.e., entertainment) was the most common motivation, and this motive was positively correlated with problematic social media use,

financial stress, and anxiety. Interestingly, using social media for social connection was also associated with higher levels of problematic media use, anxiety, and delinquency, although seeking information was not associated with any specific negative health indicators. Moreover, examining social media motives among a large sample of adolescents, Jarman et al. (2021) reported that the motive of passing time (i.e., entertainment) was inversely associated with well­being. However, once again, it is important to note that some studies in this area of research have identified no clear relationship between problematic social media use, well­being, and motivations such as entertainment and information seeking (Arness & Ollis, 2022). Therefore, the literature could benefit from research that clarifies the nature of these relationships.

The Current Study and Research Questions

Building on the existing literature, the current study used cross­sectional data collected at multiple times across the years of 2020–2022 to explore the unique relationships between social media use motives, user characteristics, behavior, and health among U.S. college students.

We had four primary research questions:

1. How are social media motives different among college students according to basic demographic characteristics (i.e., age, gender, education level, relationship status)?

2. What is the typical social media motives profile among the most popular social media platforms?

3. How do social media motives correlate with other electronic media use behaviors (i.e., social media time, number of social media platforms, screen time, TV watching, smartphone use)?

4. How are health variables related to different social media motives (i.e., behavioral, mental, physical, social)?

Methods Research Design and Procedure

After receiving Brigham Young University Idaho IRB approval, we administered a questionnaire to a convenience sample of online enrolled students completing an introductory psychology course from Brigham Young University Idaho regarding their social media use, motivations, and well ­ being. Potential student participants were solicited to participate by email invitation that contained a link to the online questionnaire (via Qualtrics) and those who participated received class credit. Our research design consisted of an exploratory correlational analysis of data collected

from this cross­sectional questionnaire. Participants were kept anonymous after data was collected, and each participant was required to sign an agreement form. Data were collected across seven semesters spanning spring 2020 to spring 2022. After narrowing the sample to include only those who provided consent for their data to be used for publication and who responded that they completed the survey accurately (attention check), our final sample size was 1,547.

To examine our first research question regarding how social media motives are related to demographic differences, we examined correlations for age, independent­samples t tests for gender, and ANOVA for each motive (information, entertainment, personal utility, and convenience) within each college year and relationship status. Next, we addressed research question two by examining average levels of each motive within each social media platform that had at least 10% representation in our sample, which resulted in our top 10 platforms in prevalence. For our third and fourth research questions, we calculated Pearson’s r correlation coefficients to examine the relationship between the four social media motives and electronic media behaviors as well as health indicators.

Participants

Average age of the sample was 26.38 (SD = 10.76), most were women (69.5%) with the remaining 30.5% indicating men, and ethnicity was mostly White/European American (75.2%) with Hispanic/Latino (12.0%), Asian American (3.5%), and Black/African American (3.0%) representing the next largest categories. Nearly half the sample were single (46.7%) with a third married (34.9%) and 16.5% engaged to be married or in a committed relationship. Although every college year was represented in our sample, the most common was first­year students (44.7%) with most of these being first­semester students (33.6% of entire sample). The remainder of our sample consisted of 28.1% sophomores, 16.4% juniors, 9.5% seniors, and 1.3% post baccalaureate students. A total of 41.5% were not employed, although 38.2% and 20.3% were employed part­ and full­time, respectively. Students were enrolled in an average of 10.52 credits (SD = 3.84), which is just below the 12 credits considered full­time. Finally, participants took an average of 86.38 minutes (SD = 407.02) to complete the online questionnaire, although a more accurate portrayal may be a median time of 33.93 minutes.

Measures Technology Variables

Technology Use . Daily time spent on social media during the past month was assessed with a single item where participants reported their estimate of time spent

across all social media on a typical day in minutes. Time spent on television (TV) was assessed using 3 questions that queried time spent during the past month in front of a TV (i.e., watching television, watching videos/shows on the computer, playing video games) on a 9­point frequency scale (0 = none , 8 = 6 hours or more ). To measure problematic smartphone use we combined the 8­item, 5­point agreement scale from Van Deursen et al. (2015) with 2 items constructed for this study (“The first thing I see in the morning when I wake is my mobile phone,” “The last thing I see before sleeping at night is my mobile phone”). This scale had acceptable internal consistency ( α = .87). As a measure of social media platform accounts, participants indicated whether they had an account for 15 different social media platforms (e.g., Facebook, Snapchat, TikTok), and these responses were then used to compute a sum score. Daily screen time was assessed by taking the average number of hours spent on screens daily across 4 items created for this study and then a sum score was calculated of minutes spent on a typical day during the past month on a smartphone (or other phone with a screen), tablet (or other small device with a screen), computer, and television.

Social Media Motivations. For the measurement of social media motivations, we used Al­Menayes’ (2015) 16­item, 5­point scale (1 = not at all, 5 = very much). These items queried reasons for using social media in general including 5 items for entertainment (e.g., “I use social media when I have nothing else to do/because it entertains me,” α = .87), 4 items for information seeking (e.g., “I use social media to search for information,” α = .81), 4 items for personal utility (e.g., “I use social media to listen to other’s opinions,” α = .81), and 3 items for convenience (e.g., “I use social media because it is easier than meeting in person,” α = .74).

Behavioral Variables

Frequency of consumption of drinks with added sugar (e.g., regular soda, sports drinks, coffee, iced tea, lemonade, or fruit punch) and fast food over the previous month were assessed using one item for each on a 10 ­ point serving frequency scale (0 = never to 10 = 5 or more times a day; Wright et al., 2017). Sleep quality was measured using a single item (“During the past month, how would you rate your sleep quality overall?”) on a 5­point scale ranging from very good to very bad. Lastly, to represent sedentary behavior, we used a 10­item adapted version of the Sedentary Behavior Questionnaire from Rosenberg et al. (2010) on a 9­point frequency scale (0 = none, 9 = 6 hours or more) in which participants indicated how much time, in general, they spend in a week on a variety of activities

WINTER 2023

Examining Social Media Motives | Ewing, Nienstedt, Wright, and Chambers

TABLE 1

Variables and Motive Correlates

279

Note * p < .05. Bonferroni correction was applied to each correlation test relevant to the research question, such that the significance level of .05 was converted by dividing .05 by 31 for these analyses presented here, which was .0016.

TABLE 2

Health Behavior Variables and Motive Correlates

Note * p < .05. Bonferroni correction was applied to each correlation test relevant to the research question, such that the significance level of .05 was converted by dividing

for these analyses presented here, which was .0022. TABLE

3

Mental Health Indicators and Motive Correlates

by

Note * p < .05. Bonferroni correction was applied to each correlation test relevant to the research question, such that the significance level of .05 was converted by dividing .05 by 50 for these analyses presented here, which was .001.

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

that involve little to no physical exertion (e.g., watching television, playing video games, sitting listening to music, sitting while reading a book, riding in a car, and doing homework)

Mental Health Variables

The mental health variables we assessed included anxiety, perceived stress, life satisfaction, mood, depressive symptoms, and body appreciation. First, anxiety during the last 3 months was captured using a 5­point frequency scale (1 = never to 5 = very often; α = .83) with the four items in the scale from Butz et al. (2011). Perceived stress was captured using seven items from the Perceived Stress Scale (Cohen et al., 1983; α = .84) on the same 5­point frequency scale (e.g., “In the past 3 months, how often have you found that you could not cope with all the things that you had to do?”). Life satisfaction was measured using the 5­item Satisfaction With Life Scale (Diener et al., 1985; α = .87) on a 5­point agreement scale (e.g., In most ways, my life is close to my ideal”). Positive and negative mood were assessed using an 8­item scale where participants indicated how much each of four positive mood adjectives (happy, enthusiastic, relaxed, alert; α = .66) and four negative mood adjectives (sad, irritable, bored, nervous; α = .70) described their mood over the last 30 days (Wright et al., 2017). Depressive symptomology was captured using the 5­item, 4­point frequency shortened CES­D (Bohannon et al., 2003; α = .77) measure (e.g., “I could not ‘get going’ ”). Finally, participants’ body appreciation was queried on a 7­point scale (1 = not at all true, 7 = very true) using the 13­item Body Appreciation Scale (e.g., “I respect my body”; Avalos et al., 2005; α = .95).

Physical Health Variables

Subjective overall health was measured using a single item on a scale of 0 (worst physical health) to 100 (best physical health). We measured the presence of common physical symptoms using a dichotomous 18­item scale derived from Spector and Jex (1998), which asked participants if they had experienced a variety of symptoms such as fever or headache, which was then summed with possible scores ranging from 0 to 18.

Social Health Variables

To access frequency of perceived loneliness during the past 3 months, we included the 3 ­ item, 5 ­ point loneliness scale (e.g., “How often do you feel that you lack companionship?”; Hughes et al., 2008; α = .86). Interpersonal conflict (Wright et al., 2017; α = .89) with others was measured with a 6­item, 5­point frequency scale (e.g., “In the past 3 months, how often have you felt like you were treated unfairly by other people?”). Finally, in­person social interaction was measured using

a 6­item, 4­point frequency scale (e.g., “How often do you get together with friends informally”; Twenge et al., 2017; α = .80).

Data Analysis

To reduce the risk of Type 1 error, a Bonferroni correction was applied to all correlation analyses by dividing

Note * p < .05.Bonferroni correction was applied to each correlation test relevant to the research question, such that the significance level of .05 was converted by dividing .05 by 16 for these analyses presented here, which was .0032.

Social Media Platforms and Corresponding Motives

the total number of correlations run by an original alpha level of .05. Correlations not relevant to the research question (i.e., how each social media motive is correlated to one another) were not included in these calculations.

Results

Descriptive statistics, including correlations, means, and standard deviations for study variables are displayed in Tables 1 through 5. Participants reported daily average time spent on screens as 3.5 hours (SD = 2.3) or 210 minutes (SD = 138), and the average number of social media platforms used was 3.9 (SD = 2.7). Each social media motive was significantly (p < .001) correlated to each other such that entertainment was related to personal utility (r = .25), information seeking (r = .24), and convenience (r = .39). Personal utility was related to information seeking (r = .46) and convenience (r = .44), but the largest relationship observed was between information seeking and convenience (r = .50). Regarding specific social media platforms, Instagram (75.4%) and Facebook (73.3%) were the most popular (see Table 6).

Research Question 1

First, in our examination of Research Question 1, age was significantly ( p < .01) correlated with both the entertainment ( r = –.33) and personal utility (r = –.08) social media motives (not information seeking

TABLE 7

Motivation Scores Between Genders

or convenience), suggesting both motives decrease for those who are older. Next, the between­gender analysis using independent samples t tests yielded significantly higher levels of convenience, information seeking, and personal utility motivation among women than men participants (p < .01), though they were rather small effect sizes (see Table 7). However, there was not a significant difference for entertainment motives, suggesting that although women and men differ in many social media use motivations, they are similar in their levels of entertainment motivation.

Next, our findings suggest that entertainment motivation is stronger for the younger class levels (first­year students, sophomore) and lower for those who are married. ANOVA tests examining each of the four motivations between class levels yielded significant results only for the entertainment motive, F(4, 1542) = 8.03, p < .001, η2 = .03 (see Table 8). Subsequent post hoc analyses yielded significantly higher entertainment motivation for first ­ year students ( p < .001, 95% CI [­0.12, 0.19]) and sophomores (p < .001, 95% CI [0.15, 0.64]) compared to seniors, who reported the lowest average entertainment motive. Third, significant differences in entertainment motivations were also observed between different relationship statuses, F(4, 1542) = 15.63, p < .001, η2 = .04; see Table 9). Specifically, married participants reported significantly less entertainment motivation than participants who were engaged (p = .01, 95% CI [­0.87, ­0.06]), in a committed relationship (p < .001, 95% CI [ ­0.55, ­0.14]), and single (p < .001, 95% CI [ ­0.44, ­0.15]). Participants reporting to be divorced or separated had the lowest average entertainment score, being significantly lower than participants reportedly married (p = .02, 95% CI [ ­0.03, ­0.98]), engaged (p < .001, 95% CI [­0.37, ­1.58]), in a committed relationship (p < .001, 95% CI [­0.35, 1.34]), or single (p < .001, 95% CI [0.33, 1.28]). All other ANOVA models on the other motives were nonsignificant for both class level and relationship status.

8

Level and Entertainment Scores

Research Question 2

TABLE 9

Relationship Status and Entertainment Scores

Our profiles of social media motives by most prevalent social media platform (Research Question 2) revealed some interesting patterns (see Table 6) that indicate notable similarities and differences between the 10 most prevalent social media platforms. For instance, across all platforms, on average, the entertainment motive was the highest of all four motives (M = 3.26, SD = 0.92), and personal utility was the lowest (M = 2.33, SD = 0.92). Within the entertainment motive, TikTok (M = 3.64, SD = 0.75), Twitter (M = 3.63, SD = 0.84), and Snapchat (M = 3.55, SD = 0.78) had the highest values and these were all significantly (ps < .001, ds > 0.60) greater than

the lowest entertainment motive platform, LinkedIn (M = 3.01, SD = 0.88). Personal utility was lowest for Facebook (M = 2.36, SD = 0.89) and highest for Twitter (M = 2.75, SD = 1.01), which was statistically significant (p < .001, d = 0.41). Also statistically significant (p < .001, d = 0.40), information seeking was lowest for Marco Polo (M = 2.87, SD = 0.88) and highest for Twitter (M = 3.24, SD = 0.96). Finally, convenience motive was rated as lowest for Facebook (M = 2.84, SD = 0.99) and highest for TikTok (M = 3.00, SD = 0.96) and Twitter (M = 3.00, SD = 1.00), though this relationship was significantly different only for the Facebook­TikTok comparison (p = .02, d = 0.16). Collectively, this suggests there were unique patterns between every prevalent social media platform.

Research Question 3

Regarding Research Question 3, we identified significant and noteworthy correlations between our technology behaviors and social media motives (see Table 1). These results suggest that social media motives are related to other electronic media behavior. First, all four of the motives were significantly (p < .001) related to social media time, TV watching time, addictive smartphone behavior, number of social media platforms, and daily screen time. The strongest correlation was between addictive smartphone behavior and entertainment motive (r = .45, r2 = .20, p < .001) with the convenience motive also demonstrating a robust relationship (r = .32, r2 = .10, p < .001). Not surprisingly, social media time, TV watching time, number of social media platforms used, and daily screen time were positively correlated with every motive. Finally, there was little difference observed in the strength of the correlation coefficients between the number of social media platforms reported and motivations examined.

Research Question 4

Finally, Research Question 4 revealed that almost every health variable had the strongest and significant relation to the entertainment motive compared to the other three motives (see Tables 2­4). Of all the health indicator variables, the strongest correlations with the entertainment motive were identified among sedentary behavior ( r = .16), body appreciation ( r = ­ .21), depressive symptoms (r = .20), negative mood (r = .27), physical symptoms (r = .14), and perceived loneliness (r = .15). Collectively, this suggests that the entertainment motive, generally, is the motivation most applicable to user health indicators.

Discussion

With the burgeoning influence that social media

continues to have on society, it has become imperative to examine how the use of social media influences the user. Most research in the literature has identified relationships that are detrimental to the user, although some have noted benefits between the use of social media, user behavior, and health outcomes (Campisi et al., 2015; Cunningham et al., 2021; Keles et al. 2019; Wright et al., 2020; 2021). Although there are many potential factors, the motivations for using social media have shown some promise in being able to differentiate between these ambiguous results, although the literature lacks an investigation of how these motives are tied to important user behavior and health indicators. In the present study, we examined four prominent social media motives (entertainment, information seeking, personal utility, convenience) and their relationship to user demographic characteristics, social media platforms, as well as user electronic media behavior and user health. Results demonstrated unique relationships within these domains and suggest that social media motivations play a pivotal role in user interface with social media. Although our findings might suggest certain generalizations and effects, our results did indicate a weak effect size on most correlations regardless of a large sample size and significant effects, therefore, caution is suggested in regard to inferences related to this discussion.

First, our results highlight the influence of user demographic characteristics within social media motivations. Through a significant yet small effect size, these findings suggested that those who are younger, earlier in their education, and not married may be more prone to use social media for entertainment purposes, which is particularly important considering the social media platform, user behavior, and user health results. Consistent with prior literature (Wright et al., 2018), we observed social media use declined with age, but our analyses revealed this seems particularly evident within both the entertainment and personal utility motives, suggesting a more nuanced relationship likely depending on user familiarity with the technology. Moreover, within our sample, women reported higher motivations for using social media in every motive than men, which is similar to previous findings regarding use of social media (Twenge & Martin 2020) and specific motivations (Kirkaburun et al., 2020). However, it is interesting to note that the difference in entertainment motive between women and men in the present study was not significant, suggesting that both have similar levels of this motivation. This is particularly important considering the observed differences in the entertainment motivation relative to college year and relationship status. Indeed, this motive decreased for those who are further along in their education (seniors) and for those

Examining Social Media Motives | Ewing, Nienstedt, Wright, and Chambers

who are married, whereas the other motivations did not vary. The entertainment motive was the highest and had the greatest amount of variance among the four social media motivations examined. These findings are similar with that of previous research and may be indicative of attempts at managing or resolving negative mood states, such as anxiety or boredom (Saleh, 2017; Stockdale & Coyne, 2020), consistent with MMT (Sriwilai & Charoensukmongkol, 2016; Wolfers & Utz, 2022). These findings further the work on MMT by suggesting that entertainment motivation may be highest due to its use as a way to manage negative mood states.

Second, our analysis of social media motives by the most prevalent social media platforms yielded interesting consistencies and contrasts. Motivations for social media use may vary by the media platforms the user decides to engage with, suggesting that different platforms will fit different user expectations and needs. Clearly, the entertainment motive was the main highlight across each platform, which is not surprising given the U&G theory (Kircaburun et al., 2020), as many opt to use social media, regardless of platform, to satisfy their desire to be entertained. Interestingly, this was even true for LinkedIn, a professional­based social media platform used primarily for advancing one’s vocational aspirations rather than amusement. TikTok, Twitter, and Snapchat users rated these platforms highest for entertainment motives, which may suggest something unique about these platforms (e.g., new, trendy) or other user characteristics (e.g., undergraduate students) that have been suggested by previous studies (Nienstedt et al., in press; Wright et al., 2020; 2021). Conversely, the other three motives did not display as strong of a presence as the entertainment motive, although they still provide some unique insight into motivations for using particular social media. For instance, it seems a Twitter user is likely to use this platform for personal utility, information, and convenience in addition to entertainment. Combined with the observation that Twitter was only used by 11% of our sample, this suggests that Twitter may be more useful to certain users and in turn, these users may become particularly dedicated to this platform. In contrast, a TikTok user would expect to find TikTok both convenient and entertaining, whereas a Facebook user may not turn to Facebook for personal utility or convenience relative to other platforms. These findings provide support to the U&G theory and add to the literature by inferring that users might become more attached to a certain platform based on the motivations used for that platform.

Third, social media motivations were related in important ways to other electronic media use behaviors, suggesting another avenue whereby social media

motivations can be understood. In line with previous research, we identified an important relationship that the entertainment motive has with each of the other electronic media use variables, such as problematic smartphone use, social media time, and social media use (Meng et al. 2020; Stockdale & Coyne, 2020). These results suggested that motives for social media use may be similar to motives for using other electronic media or that users turn to different sources at different times for similar motivations, particularly in terms of entertainment. Our other three motives (information seeking, personal utility, convenience) involve actions such as messaging friends, reading the news, looking up information, and participating in online chat rooms, which may not exert as strong of an influence as entertainment considering mood management or gratification. For example, entertainment social media platforms such as TikTok, Instagram, Facebook, and YouTube all arguably seek to keep users on social media (Bhargava & Velasquez, 2020; Montag et al., 2019), which could explain the entertainment motivation’s association with problematic smartphone and/or other media use (Purohit et al., 2020). Additionally, because the entertainment motive is aimed at relieving boredom, social media use could become a learned behavior (Donati et al., 2022; Stockdale & Coyne, 2020; Wang, 2018), so that anytime one becomes bored or uncomfortable, they might engage in social media. This interpretation supports MMT, as social media use becomes a form of mood management.

Fourth, our results have important implications for user health and wellness relative to social media motives. Previous findings have identified health and well­being profiles correlated among different social media sites (Masciantonio et al., 2021; Wright et al., 2020, 2021) and our study extended this into unique motivations, connecting with a variety of behavioral, mental, physical, and social health indicators. Indeed, our analysis of entertainment motivation yielded many significant correlations with negative well­being indicators such as poor body image, negative mood, and depressive symptoms. In the case of poor body image, this may be indicative of unrealistic body images commonly found throughout social media (Meier & Gray, 2014) where media created with the purpose of entertainment (e.g., influencers, popular accounts) may perpetuate these unrealistic standards that portray extreme body types (Vandenbosch et al., 2021). In a similar vein, other poor mental health indicators (e.g., negative affect) had a stronger relationship to the entertainment motive than the other motives. Perhaps the attempted mood management or gratification from entertainment social media is poor affective forecasting, which leads to lower levels

of accomplishment and self­esteem (Hall et al., 2019).

Although it may seem that engaging in social media will bring entertainment and pleasure, it might invite negative feelings and behaviors instead (Lin et al., 2016; Primack et al., 2021; Shensa et al., 2018). Thus, using social media for entertainment purposes, more than for other motivations, may be indicative of a void in one’s life that social media is unable to fill. Alternatively, it is plausible that entertainment motivations may lead to these poor health outcomes over a prolonged period of exposure.

Limitations and Future Directions

Although our study has identified several unique associations, these data were collected cross­sectionally, precluding any ability to draw causal conclusions. However, the large sample size and collecting of data multiple times over several semesters bolsters confidence in these associations. Second, all our data are self­report and are susceptible to subjective biases even though motivations are inherently subjective and must be studied via self­report methodologies. Third, measurement issues might have influenced our results. For instance, it is possible that there are more than four social media motives, despite our building on prior research. Moreover, our survey was administered online where data validity may be in question and there is the distinct potential overlapping influence of different social media on the same user behavior and health profiles (a user usually has multiple social media accounts). Fourth, we did not include any other gender options besides men and women in this study. Fifth, outstanding characteristics of our sample might also have influenced the results, including college students, convenience sample, mostly White, mostly women, and all responses were gathered following the COVID­19 pandemic, which might have influenced motivations and behavior regarding social media. Moreover, it may be inappropriate to assume that the findings described in this study would hold true for other communities without further concurrent findings from other studies.

Notwithstanding these potential limitations, our study provided a framework and foundation for future research. First, there is a need for this work to be replicated among other more diverse populations, including younger (adolescents, children), older, ethnically diverse, gender diverse, and less formally educated populations. Whereas this current study found clear evidence of a relationship between social media motivations and subjective behavioral and health­related constructs, further efforts should be made to connect social media motivations to other, more objective outcomes such as grade point average, recorded time spent on a

smartphone or app, or biometric health outcomes (e.g., weight, blood pressure). Going a step further, carefully crafted experimental designs could provide insight into the causal mechanisms behind these relationships, particularly those that are longitudinal or able to isolate the influence of a particular social media motivation. Relatedly, examining how a mismatch between a priori motivations (or expectations) and the outcome of social media use impacts user outcomes may yield important results. For instance, an examination of the common scenario where the original motivation is to gather information, but the user is drawn into spending a great deal of time distracted by content designed to entertain. This type of mismatch may elucidate user outcome differences in the literature. This line of inquiry would also yield itself greatly to the continued study of U&G theory. Although gratification for one motivation was sought in this scenario, it might have been blocked by use of another technology (Pelletier et al., 2020).

In conclusion, social media motives are an important aspect of social media use, influencing user behavior and health. It is upon the findings of this current study that other research efforts can build to identify ways in which the technology interface can be more beneficial for the user. Findings regarding the effects of social media on users are contradicting, and social media motives may explain some of these varying effects. By examining social media motives further, individuals may become more aware of their motivation for using social media and thereby obtain desired outcomes rather than maladaptive ones.

References

Abbasi, I., & Drouin, M. (2019). Neuroticism and Facebook addiction: How social media can affect mood? The American Journal of Family Therapy, 47(4), 199–215. https://doi.org/10.1080/01926187.2019.1624223

Al-Menayes, J. J. (2015). Motivations for using social media: An exploratory factor analysis. International Journal of Psychological Studies, 7(1). https://doi.org/10.5539/ijps.v7n1p43

Arness, D. C., & Ollis, T. (2022). A mixed-methods study of problematic social media use, attention dysregulation, and social media use motives. Current Psychology, 1–20. https://doi.org/10.1007/s12144-022-03472-6

Auxier, B., & Anderson, M. (2022). Social media use in 2021. Pew Research Center: Internet, Science & Tech. Retrieved October 15, 2022, from https://www.pewresearch.org/internet/2021/04/07/social-media-use-in-2021/ Avalos, L., Tylka, T. L., & Wood-Barcalow, N. (2005). The Body Appreciation Scale: Development and psychometric evaluation. Body Image, 2(3), 285–297. https://doi.org/10.1016/j.bodyim.2005.06.002

Berryman, C., Ferguson, C. J., & Negy, C. (2017). Social media use and mental health among young adults. Psychiatric Quarterly, 89(2), 307–314. https://doi.org/10.1007/s11126-017-9535-6

Bhargava, V. R., & Velasquez, M. (2020). Ethics of the attention economy: The problem of social media addiction. Business Ethics Quarterly, 31(3), 321–359. https://doi.org/10.1017/beq.2020.32

Bohannon, R. W., Maljanian, R., & Goethe, J. (2003). Screening for depression in clinical practice: Reliability and validity of a five-item subset of the CESdepression. Perceptual and Motor Skills, 97, 855–861. https://doi:10.2466/pms.2003.97.3.855

Butz, D. A., & Yogeeswaran, K. (2011). A new threat in the air: Macroeconomic threat increases prejudice against Asian Americans. Journal of Experimental

WINTER 2023

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

285

Examining Social Media Motives | Ewing, Nienstedt, Wright, and Chambers

Social Psychology, 47(1), 22–27. https://doi.org/10.1016/j.jesp.2010.07.014

Campisi, J., Folan, D., Diehl, G., Kable, T., & Rademeyer, C. (2015). Social media users have different experiences, motivations, and quality of life. Psychiatry Research, 228(3), 774–780.

https://doi.org/10.1016/j.psychres.2015.04.042

Carpentier Dillman, F. R. (2020). Mood management. The international encyclopedia of media psychology, 1–8.

https://doi.org/10.1002/9781119011071.iemp0255

Chen, L., Zhou, S., & Bryant, J. (2007). Temporal changes in mood repair through music consumption: Effects of mood, mood salience, and individual differences. Media Psychology, 9(3), 695–713.

https://doi.org/10.1080/15213260701283293

Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385.

https://doi.org/10.2307/2136404

Coyne, S. M., Rogers, A. A., Zurcher, J. D., Stockdale, L., & Booth, M. C. (2020). Does time spent using social media impact mental health?: An eight year longitudinal study. Computers in Human Behavior, 104, 106160.

https://doi.org/10.1016/j.chb.2019.106160

Cunningham, S., Hudson, C. C., & Harkness, K. (2021). Social media and depression symptoms: A meta-analysis. Research on Child and Adolescent Psychopathology, 49(2), 241–253. https://doi.org/10.1007/s10802-020-00715-7

Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75.

https://doi.org/10.1207/s15327752jpa4901_13

Dillman Carpentier, F. R., Brown, J. D., Bertocci, M., Silk, J. S., Forbes, E. E., & Dahl, R. E. (2008). Sad kids, sad media? Applying mood management theory to depressed adolescents’ use of Media. Media Psychology, 11(1), 143–166.

https://doi.org/10.1080/15213260701834484

Donati, M. A., Beccari, C., & Primi, C. (2022). Boredom and problematic Facebook use in adolescents: What is the relationship considering trait or state boredom? Addictive Behaviors, 125, 107132.

https://doi.org/10.1016/j.addbeh.2021.107132

Fardouly, J., & Vartanian, L. R. (2016). Social media and body image concerns: Current research and future directions. Current Opinion in Psychology, 9, 1–5.

https://doi.org/10.1016/j.copsyc.2015.09.005

Hall, J. A., Johnson, R. M., & Ross, E. M. (2018). Where does the time go? An experimental test of what social media displaces and displaced activities’ associations with affective well-being and quality of day. New Media & Society, 21(3), 674–692. https://doi.org/10.1177/1461444818804775

Hou, Y., Xiong, D., Jiang, T., Song, L., & Wang, Q. (2019). Social media addiction: Its impact, mediation, and intervention. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 13(1). https://doi.org/10.5817/cp2019-1-4

Huang, C. (2020). A meta-analysis of the problematic social media use and mental health. International Journal of Social Psychiatry, 68(1), 12–33. https://doi.org/10.1177/0020764020978434

Hughes, M. E., Waite, L. J., Hawkley, L. C., & Cacioppo, J. T. (2004). A short scale for measuring loneliness in large surveys. Research on Aging, 26(6), 655–672. https://doi.org/10.1177/0164027504268574

Ivie, E. J., Pettitt, A., Moses, L. J., & Allen, N. B. (2020). A meta-analysis of the association between adolescent social media use and depressive symptoms. Journal of Affective Disorders, 275, 165–174. https://doi.org/10.1016/j.jad.2020.06.014

Jarman, H. K., Marques, M. D., McLean, S. A., Slater, A., & Paxton, S. J. (2021). Motivations for social media use: Associations with social media engagement and body satisfaction and well-being among adolescents. Journal of Youth and Adolescence, 50(12), 2279–2293. https://doi.org/10.1007/s10964-020-01390-z

Keles, B., McCrae, N., & Grealish, A. (2019). A systematic review: The influence of social media on depression, anxiety and psychological distress in adolescents. International Journal of Adolescence and Youth, 25(1), 79–93.

https://doi.org/10.1080/02673843.2019.1590851

Kim, C., & Lee, J. K. (2016). Social media type matters: Investigating the relationship between motivation and online social network heterogeneity. Journal of Broadcasting & Electronic Media, 60(4), 676–693.

https://doi.org/10.1080/08838151.2016.1234481

Kircaburun, K., Alhabash, S., Tosuntaş, Ş. B., & Griffiths, M. D. (2018). Uses and gratifications of problematic social media use among university students: A simultaneous examination of the big five of personality traits, social media platforms, and social media use motives. International Journal of Mental Health and Addiction, 18(3), 525–547.

https://doi.org/10.1007/s11469-018-9940-6

Kross, E., Verduyn, P., Demiralp, E., Park, J., Lee, D. S., Lin, N., Shablack, H., Jonides, J., & Ybarra, O. (2013). Facebook use predicts declines in subjective well-being in young adults. PLoS ONE, 8(8).

https://doi.org/10.1371/journal.pone.0069841

Masciantonio, A., Bourguignon, D., Bouchat, P., Balty, M., & Rimé, B. (2021). Don’t put all social network sites in one basket: Facebook, Instagram, Twitter, TikTok, and their relations with well-being during the COVID-19 pandemic. PLoS ONE, 16(3), 1–14. https://doi.org/10.1371/journal.pone.0248384

Meier, E. P., & Gray, J. (2014). Facebook photo activity associated with body image disturbance in adolescent girls. Cyberpsychology, Behavior, and Social Networking, 17(4), 199–206. https://doi.org/10.1089/cyber.2013.0305

Meng, H., Cao, H., Hao, R., Zhou, N., Liang, Y., Wu, L., Jiang, L., Ma, R., Li, B., Deng, L., Lin, Z., Lin, X., & Zhang, J. (2020). Smartphone use motivation and problematic smartphone use in a national representative sample of Chinese adolescents: The mediating roles of smartphone use time for various activities. Journal of Behavioral Addictions, 9(1), 163–174. https://doi.org/10.1556/2006.2020.00004

Montag, C., Lachmann, B., Herrlich, M., & Zweig, K. (2019). Addictive features of social media/messenger platforms and freemium games against the background of psychological and economic theories. International Journal of Environmental Research and Public Health, 16(14), 2612. https://doi.org/10.3390/ijerph16142612

Naslund, J. A., Bondre, A., Torous, J., & Aschbrenner, K. A. (2020). Social media and mental Health: Benefits, risks, and opportunities for research and Practice. Journal of Technology in Behavioral Science, 5(3), 245–257. https://doi.org/10.1007/s41347-020-00134-x

Nienstedt, C., Smith, N., Braithwaite, H., Gilbert, B. & Wright, R. R. (2023). Swiping away our well-being? Examining well-being indicators among TikTok account holders. Psi Chi Journal of Psychological Research, 28(2), 96–106 https://doi.org/10.24839/2325-7342.JN28.2.96

Pelletier, M. J., Krallman, A., Adams, F. G., & Hancock, T. (2020). One size doesn’t fit all: A uses and gratifications analysis of social media platforms. Journal of Research in Interactive Marketing, 14(2), 269–284. https://doi.org/10.1108/jrim-10-2019-0159

Primack, B. A., Shensa, A., Sidani, J. E., Escobar-Viera, C. G., & Fine, M. J. (2021). Temporal associations between social media use and depression. American Journal of Preventive Medicine, 60(2), 179–188. https://doi.org/10.1016/j.amepre.2020.09.014

Purohit, A. K., Barclay, L., & Holzer, A. (2020). Designing for digital detox: Making social media less addictive with digital nudges. Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3334480.3382810

Reinecke, L. (2016). Mood management theory. The International Encyclopedia of Media Effects, 1–13. https://doi.org/10.1002/9781118783764.wbieme0085

Rodgers, R. F., Slater, A., Gordon, C. S., McLean, S. A., Jarman, H. K., & Paxton, S. J. (2020). A biopsychosocial model of social media use and body image concerns, disordered eating, and muscle-building behaviors among adolescent girls and boys. Journal of Youth and Adolescence, 49(2), 399–409. https://doi.org/10.1007/s10964-019-01190-0

Rosenberg, D. E., Norman, G. J., Wagner, N., Patrick, K., Calfas, K. J., & Sallis, J. F. (2010). Reliability and validity of the Sedentary Behavior Questionnaire (SBQ) for adults. Journal of Physical Activity and Health, 7(6), 697–705. https://doi.org/10.1123/jpah.7.6.697

Saleh, D., Camart, N., & Romo, L. (2017). Predictors of stress in college students. Frontiers in Psychology, 8 https://doi.org/10.3389/fpsyg.2017.00019

Shensa, A., Sidani, J. E., Dew, M. A., Escobar-Viera, C. G., & Primack, B. A. (2018). Social media use and depression and anxiety symptoms: A cluster analysis. American Journal of Health Behavior, 42(2), 116–128. https://doi.org/10.5993/ajhb.42.2.11

Spector, P. E., & Jex, S. M. (1998). Development of four self-report measures of job stressors and strain: Interpersonal Conflict at Work Scale, Organizational Constraints Scale, Quantitative Workload Inventory, and Physical Symptoms Inventory. Journal of Occupational Health Psychology, 3(4), 356–367. https://doi.org/10.1037/1076-8998.3.4.356

Sriwilai, K., & Charoensukmongkol, P. (2015). Face it, don’t Facebook it: Impacts of social media addiction on mindfulness, coping strategies and the consequence on emotional exhaustion. Stress and Health, 32(4), 427–434. https://doi.org/10.1002/smi.2637

Stockdale, L. A., & Coyne, S. M. (2020). Bored and online: Reasons for using social media, problematic social networking site use, and behavioral

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

outcomes across the transition from adolescence to emerging adulthood. Journal of Adolescence, 79(1), 173–183. https://doi.org/10.1016/j.adolescence.2020.01.010

Twenge, J. M., & Martin, G. N. (2020). Gender differences in associations between digital media use and psychological well-being: Evidence from three large datasets. Journal of Adolescence, 79(1), 91–102. https://doi.org/10.1016/j.adolescence.2019.12.018

Van Deursen, A. J. A. M., Bolle, C. L., Hegner, S. M., & Kommers, P. A. M. (2015). Modeling habitual and addictive smartphone behavior. Computers in Human Behavior, 45, 411–420. https://doi.org/10.1016/j.chb.2014.12.039

Vandenbosch, L., Fardouly, J., & Tiggemann, M. (2022). Social media and body image: Recent trends and future directions. Current Opinion in Psychology, 45, 101289. https://doi.org/10.1016/j.copsyc.2021.12.002

Vargo, D., Zhu, L., Benwell, B., & Yan, Z. (2021). Digital technology use during COVID‐19 pandemic: A rapid review. Human Behavior and Emerging Technologies, 3(1), 13–24. https://doi.org/10.1002/hbe2.242

Wang, W. C. (2018). Exploring the relationship among free-time management, leisure boredom, and internet addiction in undergraduates in Taiwan. Psychological Reports, 122(5), 1651–1665. https://doi.org/10.1177/0033294118789034

Whiting, A., & Williams, D. (2013). Why people use social media: A uses and gratifications approach. Qualitative Market Research: An International Journal, 16(4), 362–369. https://doi.org/10.1108/qmr-06-2013-0041

Wolfers, L. N., & Utz, S. (2022). Social media use, stress, and coping. Current Opinion in Psychology, 45, 101305. https://doi.org/10.1016/j.copsyc.2022.101305

Woods, H. C., & Scott, H. (2016). #Sleepyteens: Social media use in adolescence is associated with poor sleep quality, anxiety, depression, and low selfesteem. Journal of Adolescence, 51(1), 41–49. https://doi.org/10.1016/j.adolescence.2016.05.008

Wright, R. R., Broadbent, C., Graves, A., & Gibson, J. (2016). Health behavior change promotion among Latter-day Saint College students. Psi Chi Journal of Psychological Research, 21(3), 200–215. https://doi.org/10.24839/2164-8204.jn21.3.200

Wright, R. R., Evans, A., Schaeffer, C., Mullins, R., & Cast, L. (2021). Social networking site use: Implications for health and wellness. Psi Chi Journal of Psychological Research, 26(2), 165–175.

https://doi.org/10.24839/2325-7342.JN26.2.165

Wright, R. R., Hardy, K., Simpson, S., Egli, M., Mullins, R., & Martin, S. (2018). Loneliness and social media use among religious Latter-day Saint college students: An exploratory study. Journal of Technology in Behavioral Science, 3(1), 12–25. https://doi.org/10.1007/s41347-017-0033-3

Wright, R. R., Larson, J., Richards, S., Hoffmann, S., & Nienstedt, C. (2022). The COVID-19 pandemic: Electronic media use and health among U.S. college students. Journal of American College Health, 1–16. https://doi.org/10.1080/07448481.2022.2155463

Wright, R. R., Nixon, A. E., Peterson, Z. B., Thompson, S. V., Olson, R., Martin, S., & Marrott, D. (2017). The Workplace Interpersonal Conflict Scale: An alternative in conflict assessment. Psi Chi Journal of Psychological Research, 22(3), 163–180. https://doi.org/10.24839/2325-7342.jn22.3.163

Wright, R. R., Schaeffer, C., Mullins, R., Evans, A., & Cast, L. (2020). Comparison of student health and well-being profiles and social media use. Psi Chi Journal of Psychological Research, 25(1), 14–21.

https://doi.org/10.24839/2325-7342.jn25.1.14

Zivnuska, S., Carlson, J. R., Carlson, D. S., Harris, R. B., & Harris, K. J. (2019). Social Media Addiction and social media reactions: The implications for job performance. The Journal of Social Psychology, 159(6), 746–760. https://doi.org/10.1080/00224545.2019.1578725

Author Notes

Carson R. Ewing https://orcid.org/0000­0002­5535­4864

Christian Nienstedt https://orcid.org/0000­0001­9654­6734

Robert R. Wright https://orcid.org/0000­0002­4101­7840

To our knowledge, there are no competing financial interests and or personal relationships that influenced the work, results, or conclusions presented in this paper.

Correspondence concerning this article should be addressed to Carson R. Ewing, Brigham Young University–Idaho, 210 West 4th South, MS2140 Rexburg, ID 83460­2140. Email: carsonreedewing@gmail.com

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN

Where’s the Party? How Clothing and Context Influence Perceptions of Women

Megan Sherman, Regan A. R. Gurung*, Callan Jackman, and Hannah Mather School of Psychological Sciences, Oregon State University

ABSTRACT. The sexual objectification of women is a large societal issue and is related to a host of negative mental and physical health implications for women. Past research has suggested that the objectification of women may be influenced by types of clothing worn and the context in which it is worn. However, it is difficult to discern how much objectification is modified by both clothing and contextual features simultaneously. We manipulated the type of clothing worn by models (high skin exposure or low skin exposure) and the type of context the clothing was allegedly worn for (a party or a job interview). Participants (N = 334) rated models on traits relating to objectification, professionalism, and capabilities. Analyses showed significant main effects for both context and outfit on most variables but no significant interactions. Results suggest that changing the context in which clothing is meant to be worn can significantly change a number of attributions made of the wearer by observers, specifically in terms of objectification. These findings may help build upon the ever ­ growing framework for understanding some potential mechanisms behind the objectification of women.

Keywords: objectification, context, revealing, clothing, women

The sexual objectification of women has a long, negative history that has been woven into many aspects of society (Beaumont et al., 2021). Sexual objectification is prevalent in patriarchal cultural concepts, which play a strong role in much of Western society (McBride & Kwee, 2021). Objectification has received significant research attention with the #Metoo movement, and there has been an increased effort to understand these issues at a psychological level (Rasmussen & Densley, 2017). Unfortunately, the objectification of women not only hurts the specific woman who is being objectified but other women as well. For example, objectification spillover occurs when the increased objectification of one woman leads to an automatic increase in the objectification of others (Guillén & Kakarika, 2020). There are also significant links between objectification, aggression, and assault (Awasthi, 2017; Harsey, 2021). These co ­ occurring factors illustrate just how significant

objectification is; examining these issues may help researchers and advocates better understand and diffuse the objectification of women. We tested if different types of clothing as well as different contexts change the amount of objectification taking place.

Objectification

Objectification is defined as “the experience of being treated as a body (or collection of body parts) valued predominantly for its use to (or consumption by) others” (Fredrickson & Roberts, 1997, p. 174). Female sexual objectification, as the term is most referred to, is when a woman is viewed primarily as an object of male sexual desire (Syzmanski et al., 2011). Objectification is positively related to sexualization and perceptions of promiscuity (Heflick & Goldenberg, 2014; Kellie et al., 2019; Ward, 2016).

Recent theorizing on objectification has explored the link between objectification and dehumanization

Diversity badge earned for conducting research focusing on aspects of diversity.

(Over, 2021). Dehumanization can be commonly defined as the process of depriving a person or group of positive human qualities. There is evidence that women seen in swimsuits or lingerie are considered, at a neural activation level, as akin to animals (Vaes et al., 2011) and even as everyday objects such as shoes instead of as humans (Bernard et al., 2017). Other studies have found that people assign more humanness to themselves in comparison to others. This would indicate that they believe others have fewer human qualities than themselves, which may make it easier to sexually objectify them (Haslam et al., 2005).

What causes the sexual objectification of women specifically? It has been hypothesized that sexualized female bodies, like objects, are recognized for their use rather than seen as human (Philippe, 2015). Men rated a woman as less competent following exposure to sexualized images of other women, implying that exposure to sexualized stimuli is an important factor (Rudman & Borgida, 1995). Given the role of sexualization, most easily operationalized by showing more skin, and past research that shows sexualized bodies, we hypothesized that provocative clothing would be more likely to lead to objectification (Holland & Haslam, 2016).

In addition to addressing how different forms of clothing result in objectification (e.g., tight versus loose), studies have focused on how skin exposure drives sexual objectification. A number of studies have varied the ratio of skin visible to clothing and found that, as more skin is visible, the sexual objectification of a model increases (Anderson et al., 2018). Consequently, there is even a scale to measure sexualized clothing (Smith et al., 2017). Given this past research, a primary focus of our research was on the role of revealing clothing on objectification.

Clothing

If one looks at clothing as a form of nonverbal communication, it is clear that clothing plays a role in one’s identity. Similar to words, the clothes one wears “say” much as well (Awasthi, 2017). Many groups, such as doctors or military personnel, rely on clothing to nonverbally establish authority or profession (Awasthi, 2017). Therefore, garments are part of helping establish one’s identity to the outside world and directly and clearly connecting an individual to a larger group. Because clothing can influence the representation of roles and characteristics, clothing should be examined in the context of social perception (Gurung et al., 2018; Livingston & Gurung, 2021).

Women understand their clothing choices matter when it comes to how men perceive them (Edmonds & Cahoon, 2013). This knowledge may limit women from expressing themselves freely in their choice of clothing.

For example, it is known that clothing is a factor in victim­blaming in cases of sexual assault (Brown et al., 2022). In early studies, participants rating pictures of allegedly sexually assaulted women with varying lengths of skirts attributed greater levels of responsibility for the assault to women in shorter skirts (Workman & Freeburg, 1999). Perceiving a woman in revealing clothes may create the perception that the woman is sexually interested in those around them (Johnson et al., 2016). When dressing to attract, women may be misperceived as trying to seduce. This misperception is more likely in men than in women, thereby increasing the chances of victim blaming by men in cases of unwanted sexual advances (Moor, 2010).

Because clothing is a form of self ­ expression, different types of clothing can lead to the formation of different assumptions, such as ones about sexual morality. After seeing women in outfits varying in promiscuity, participants rated women in more revealing clothes as less moral and having less mental capacity (Kellie et al., 2019). Participants also perceived women in revealing clothes to be more open to casual sex. One may also overestimate a woman’s promiscuousness based on clothing (Guéguen, 2011). Guéguen measured the time it took for men to approach two female confederates sitting in a tavern, one wearing suggestive clothing and one wearing more conservative clothing. The time it took for the men to approach the women after initial eye contact was significantly shorter in the suggestive clothing condition, implying clothing influences first impressions. The location of this study (i.e., in a tavern) also suggests the role of context in sexual objectification based on clothing type.

Context

Clothing and its messages cannot be viewed in isolation. Different situations call for different outfits, and some clothing choices fit some situations better than others. A swimsuit worn on the beach may not draw any attention, whereas a three­piece business suit worn at the beach may attract the eye. Expectations for what is appropriate clothing for different situations (i.e., swimsuits at the beach, suits at a formal event) are reinforced by sociocultural standards that manifest as dress codes. Most schools and workplaces have dress codes for appropriate attire, and when these codes are violated, the perceiver can bear the brunt of prejudice (Gurung et al., 2018). Correspondingly, context is important to consider when one analyzes clothing, and the assumptions people make about what a woman wears.

Contextual priming can affect one’s view of clothing, and clothing could be made to send a different message based on different contextual information or

Sherman, Gurung, Jackman, and Mather | Clothing and Context Perceptions

cues. It is possible that the same outfit may be perceived differently if it is considered to fit the context versus if it does not. In one direct manipulation of this possibility, participants received a message about dating or job interviews presented by a virtual woman dressed either in business attire (i.e., context‐appropriate) or a red short sleeveless dress (i.e., context‐inappropriate; Nowak et al., 2015). The perceived appropriateness of clothing influenced perceptions of the source: women were perceived as less appropriate if their clothing did not match the context given.

Context may also interact with the objectifying properties of provocative clothing, specifically in terms of professions and substance use. Men and women perceived photographs of women dressed provocatively as less competent at their jobs (Glick et al., 2005). Furthermore, much research has examined the role of alcohol in contexts where women are objectified. If a woman is under the influence of alcohol, she is seen as less sexually responsible and perceived as less responsible in comparison to a woman who is sober (Osborn et al., 2018; Radun et al., 2018). Given alcohol is often served at parties, a woman thought to be going to a party may be also be assumed to be open to drinking and less responsible.

Sexism

In exploring the role of clothing and context, it is important to acknowledge and control for the role played by sexism. Sexism is typically seen as a reflection of hostility toward women (Glick et al., 2005), which is pertinent to this study’s focus on negative views of women. Not only is sexism relevant, but Glick et al. also argued that sexism has roots in social and biological conditions such as the patriarchy, which as mentioned before, is a significant tie to objectification. Correspondingly, we measured sexism using the Ambivalent Sexism Inventory (ASI; Glick & Fiske, 1996) and included it as a control variable in all analyses.

The Present Study

The present study differentiates itself from past studies; it examined objectification as it relates not only to clothing but also to context. Much of the previous research has placed emphasis on one factor: clothing or context. Other research has examined the interaction of clothing and context, but only in regard to populations of professional settings (e.g., women in the workforce) and without contrasting different contexts. The present research combined two key variables into one design, analyzing them in conjunction with one another. This research also focused explicitly on perceptions of and by a college population. Past work has built a strong foundation on the definition of objectification, the

ways in which clothing and context are important, the operationalization of objectification, as well as tools to measure these variables. This research built on those discoveries and attempted to bring those different foundational pieces together.

Based on past research on objectification, clothing, and context, we had three main hypotheses. First, we predicted a main effect of outfits, where models in outfits showing more skin (high exposure), would be sexually objectified more, and seen as less professional and capable than models in outfits showing less skin (low exposure), regardless of context. Second, we predicted a main effect of context, where the models when thought to be dressed for a party would be viewed better than if they were thought to be dressed for an interview. They would be more positively viewed and objectified less in the party context, while also being seen as less professional and less capable in the interview context. Third, we anticipated an interaction effect, whereby models in the high exposure outfits showing more skin in a context­incongruent condition (the interview) would be objectified the most. We predicted sexism and gender would be significant covariates.

Method

Participants

Undergraduate introductory psychology students volunteered for this study and received class credit. Participants’ ( N = 343) ages ranged from 18–42 (M = 21.69, SD = 5.39). The sample was mostly women (69.9%), in comparison to men (30.1%) and comprised of European American (63.4%), Asian (12.2%), Hispanic (11.0%), African American (3.7%), Pacific Islander (1.2%), Somalian (1.2%), and American Indian or Alaskan Native (0.6%) individuals. No participant left race/ethnicity blank, but 17 said their race was not listed. We did not have a category for multiracial and unintentionally did not allow participants to select more than one race. One participant reported being transgender, three listed their gender as self­identified, and one said their gender was not listed.

Materials Independent Variable: Outfit

Three White women volunteered from the university for this study. To keep race consistent, we selected a similar race intentionally. Each woman was a friend of a student research assistant and provided mirror selfies, pictures in front of a full­body mirror using a cell phone in a posed (left arm on hip, left leg popped) position (photos provided in Appendix). Research lab members examined the provided photographs and selected the final set, which best satisfied the experimental manipulation. The entire body was visible. In the high­exposure clothing photo,

each model wore an outfit revealing skin (i.e., cleavage, upper thigh, and midriff showing). In the second lowexposure condition, each model wore a tight­fitting outfit that covered undergarments, intimate body parts, and skin around those parts. The models’ hair was pulled back, and their phones covered their faces in the shots to make them less identifiable. All models had similar body shapes and were in the same age range.

Independent Variable: Context

Parties and job interviews are two events that many college students have experience with, if not personally then by proxy. Additionally, these two experiences are very different, representing two sides of the spectrum when it comes to college life. Each context has specific clothing commonly associated with it. For example, one is expected to dress formally for an interview. We used a two­factor design where each factor had two levels. In addition to manipulating outfit for exposure in the visual stimuli, we focused on two contexts: interview and party. We used these contexts because they each represent a specific atmosphere and are on either end of a wide spectrum of situations experienced by college students. A party was chosen because it is a believable situation that a college­aged woman may attend where she may have the freedom to wear more revealing clothes if she so chooses. The job interview context was chosen to represent a situation directly opposite of a party: a situation in which clothing choices may be restricted to professional, modest choices, and where a reserved tone may be expected.

Dependent Measures: Overview

We selected dependent variables pulled from previous research (Chrouser & Gurung, 2007; Glick et al., 2005; Howlett et al., 2015). Our objectification measures and capabilities measures came from Johnson and Gurung (2011). We used descriptive words separated into three categories: capabilities (determined, independent, intelligent, studious, self­respecting), sexual objectification (attractive, desirable, promiscuous, and uses her body to get what she wants), and professional abilities (confident, competent, organized, powerful, and reliable). Participants rated each term on a 5­point Likert scale ranging from 1 (strongly agree) to 5 (strongly disagree).

Reliability of Dependent Variables

We created composite scores for each variable. Two composites had high reliability, professional, Cronbach’s alpha = .87, and capable, Cronbach’s alpha = .90. It should be noted that, although professional and capability­oriented adjectives may seem similar, this distinction was intentional. Variables of the professional nature were selected with the intention of capturing workplace ­ oriented

traits that may relate to one’s professional relationships (i.e., coworkers, managers, and interviewers), whereas capability ­ related adjectives were selected as representations of general personality traits that may relate more to the individual. Reliability for the composite for objectification was low, Cronbach’s alpha = .48, and correspondingly, we analyzed the four items individually.

Main Covariate

We used the Ambivalent Sexism Inventory (ASI, Glick & Fiske, 1996) to measure and control for levels of sexism. The ASI consists of a 22­statement list that participants rated each term on a scale ranging from 1 (strongly agree) to 6 (strongly disagree). Sample items include, “Women are too easily offended” and “Every man ought to have a woman who he adores.” The two subscales of the ASI are Hostile Sexism and Benevolent Sexism. Hostile Sexism is outright prejudice and discrimination toward women. Benevolent Sexism is a combination of attitudes toward women that come off as positive, prosocial, and seeking intimacy, yet are based on stereotypes and restrictive social roles of the female sex. Both benevolent sexism and hostile sexism correlate with stronger tendencies to victim blame (Persson et al., 2018). Cronbach’s alpha for each subscale showed acceptable internal reliability, .89 (Hostile Sexism) and .72 (Benevolent Sexism).

Procedure

The Oregon State University institutional review board approved this study (IRB­2020­0558). All participants accessed this study online through SONA, an online software system where they could choose from a variety of studies. For participating in the study, students received research credit, which is a class requirement in general psychology. We randomly assigned participants to one of four conditions, using a between­group, 2 Outfit (high exposure, low exposure) x 2 Context (party, interview) design. We created the survey using Qualtrics survey software. Participants first read a consent form, then viewed and rated three models (with the image visible during rating). We presented all models in a random order for each participant. Before seeing the models, we told participants in two conditions that they would see three pictures of female college students from [redacted university name] getting ready to go to an interview. The women were said to be sending mirror selfies to their friends to get opinions on their outfits. In the other two conditions, “party” was used in place of “interview.” Women seen in each context were dressed in high or low­exposure clothing. Participants rated the models on all dependent variables, presented on Likert scales. Finally, participants completed the ASI and answered several demographic questions.

Sherman, Gurung, Jackman, and Mather | Clothing and Context Perceptions

Results

As a manipulation check of exposure, participants rated the models’ outfit on the extent to which it was tightfitting, revealing, and modest. Participants rated each term on a 5­point Likert scale rating from 1 (strongly agree) to 5 (strongly disagree). A MANCOVA using outfit as a factor was significant, F(3, 340) = 88.13, p < .001, ηp 2 = .44. As expected, clothing in the high exposure condition was rated as being more exposed (M = 3.64, SD = 0.80), than the clothing in the low exposure condition (M = 2.61, SD = 0.80), F(3, 340) = 58.29, p < .001, ηp 2 = .34. Perhaps not surprising for high exposure clothing styles, this category was also seen as tighter ( M = 4.11, SD = 0.66), and less modest ( M = 2.61, SD = 0.82), than clothing in the low exposure condition M = 3.14, SD = 0.85, F(3, 340) = 53.52, p < .001, ηp 2 = .32, and M = 3.55, SD = 0.67, F(3, 340) = 48.24, p < .001, ηp 2 = .30.

To test the primary hypotheses, we conducted three ANCOVAs using Context (party or interview) and Outfit (high or low exposure) as fixed factors; one each for the two composites with acceptable reliability. We used gender and ASI scores (Hostile Sexism, Benevolent Sexism) as covariates for all analyses. We conducted a multivariate analysis of covariance (MANCOVA) for the objectification variables using the same factors as the ANCOVA.

Given the low numbers of transgender students, analyses did not include participants self­identifying as anything other than men or women. The inclusion of transgender people is important in conducting research, but the low number did not allow us to test this group separately. We list means and standard deviations for all variables by condition in Table 1.

Given the low reliability of the objectification composite, we conducted a MANCOVA with all four individual terms as dependent variables and similar

for Composite Variables

factors and covariates as the ANCOVAs. We found significant multivariate main effects for Context, F (4, 329) = 5.15, p < .001, ηp 2 = .06, and Outfit, F(4, 329) = 3.92, p = .004, ηp 2 = .05, but no significant interaction (p = .23). Only two of the four variables showed significant univariate tests. Participants rated the women in the interview condition as more promiscuous (M = 3.18, SD = 0.85), F(1, 343) = 4.13, p = .04, ηp 2 = .01, and more likely to use her body (M = 2.93, SD = 0.94), F(1, 343) = 10.06, p = .002, ηp 2 = .03, than women in the party condition (promiscuous M = 3.00, SD = 0.93, uses body M = 2.60, SD = 0.94). Participants rated the women in the high exposure outfit condition as more promiscuous (M = 3.24, SD = 0.83), F(1, 343) = 12.04, p < .001, ηp 2 = .03, and more likely to use her body ( M = 2.57, SD = 0.91) F (1, 343) = 12.31, p < .001, ηp 2 = .04, than participants in low exposure condition (promiscuous M = 2.93, SD = 0.94, uses body M = 2.58, SD = 0.91). Neither the interaction of context and outfit nor gender were significant. Only hostile sexism was a significant covariate, F(1, 339) = 9.56. p = .002, ηp 2 = .03.

For the professionalism ANCOVA, we found a significant main effect for context F(1, 339) = 10.63, p = .001, ηp 2 = .03, but not outfit ( p = .10), and no significant interaction ( p = .51). Participants rated models in the interview condition as less professional (M = 3.74, SD = 0.57) than models in the party condition (M = 3.90, SD = 0.57). Only hostile sexism was a significant covariate, F(1, 339) = 19.48, p < .001, ηp 2 = .06. For the capabilities ANCOVA, there were main effects for Context, F(1, 339) = 13.64, p <.001, ηp 2 = .04, and Outfit, F(1, 339) = 6.64, p = .01, ηp 2 = .02. Hostile Sexism was a significant covariate, F(1, 339) = 26.76, p < .001, ηp 2 = .08, as well as Benevolent Sexism, F(1, 339) = 6.12, p = .014, ηp 2 = .02. The interaction was not significant. The perceived “capabilities” of the models were affected by both what outfit they wore as well as the setting in which they were allegedly dressing for. Participants rated models in the interview condition as less capable (M = 3.72, SD = 0.59), in comparison to models in the party condition (M = 3.91, SD = 0.60). Participants rated models in the high­exposure outfits as less capable (M = 3.73, SD = 0.60), in comparison to models in the low ­ exposure clothing condition (M = 3.90, SD = 0.60).

Discussion

In this study, we aimed to replicate past research showing the negative effects of wearing revealing clothing and extended past work to test for the effect of context. Both outfit conditions featured tight­fitting clothing suitable for a party but varied in how much skin was exposed. We also examined the interaction of outfit and

Note. Attractive, desirable, promiscuous, and uses body represent measures of objectification. Professional and capabilities are composite variables.

context and focused on objectification and perceptions of professionalism and capabilities. In strong support of our hypotheses, the context was significant in predicting objectification, professionalism, and capabilities. Women thought to be going to an interview inappropriately dressed were objectified more and seen as being less professional and less capable compared to the same women thought to be going to a party.

We also found a main effect of outfit on objectification and capabilities. Women whose outfits revealed more of their bodies were objectified more and seen as less capable than when the women wore less revealing clothes. As predicted, these results were influenced by sexism where participants higher in hostile sexism rated the women more negatively.

A large body of work has suggested that provocatively dressed women are more sexually objectified (Beaumont et al., 2021; Glick et al., 2005). Many assumptions are made about a woman’s personality based on her clothing, leading to objectification and dehumanization (Guéguen, 2011; Vaes et al., 2011). Beyond objectification, a revealing outfit also affected opinions of a model’s level of professionalism and capabilities. Professional skills are important qualities that many employers look for, and the unfortunate truth that skin exposure may affect the opinions people have of these qualities is an important factor when analyzing workplace environments and hiring practices.

Context

Our results support past findings on clothing and add a new element to this literature with a valuable new focus on context. Similar to research looking at brain activation during perception (Bernard et al., 2019), results showed that revealing clothing by itself may not be the only driver of objectification and negative impressions of a person, but the assumed context the clothing is worn in, or the mismatch of context and outfit, may contribute to negative perceptions of women (Nowak et al., 2015).

Our addition of context to the study of objectification, professionalism, and capabilities paid rich dividends. Context influenced how the generally provocative clothing was perceived. For objectification, the context of the party led to the models being seen more positively than if they were thought to be going to an interview. The context of a party perhaps activates a schema including more provocative clothing. This has large implications for the role that context may play in the objectification of women. Low sexual morality is often associated with revealing clothing (Kellie et al., 2019), but these results show that, when one thinks a woman is going to a less risqué event (for this study’s purposes, an interview), they are objectified more than if they wear the same clothes to a party.

Similar to work on the power of clothing in the workplace (Glick et al., 2005; Howlett et al., 2015), our results show that the context can influence ratings of a woman’s professionalism. Participants had more positive perceptions of the models’ professionalism and capabilities if they thought she was dressed provocatively for a party, but they had more negative judgments if she was dressing in a similar provocative way for an interview. The disjoint between the provocative clothing and the setting of the interview worked against the models. One interesting finding that heavily supported our focus on dressing incongruently to the situation is that women were rated as being less professional when seen in the interview condition. It is likely that the tight clothing, whether revealing or not and which could be worn to a party, was discordant with the idea that they were going to an interview. Wearing party clothes to a party is one thing, wearing them to an interview is another. The fact that the women did not dress appropriately for the context is perhaps the driving force as both outfits were not professional and this was worse when in the interview condition.

The idea that these events in and of themselves have the power to influence how women are perceived by society is important when looking at why women are so objectified. Switching focus from one’s outfit to one’s context is a rather underexplored concept, so there is certainly room to explore this idea further. Further research could analyze why these contexts are so important, pinpoint what is happening cognitively, and eventually, refocus the method of reducing objectification through this newfound understanding of the impact of context.

Interaction of Dependent Variables

We did not find support for our third hypothesis. In contrast to expectations, there was no interaction between the two main variables. One possibility for this finding is that both outfits that could be worn to a party might not have been different enough from each other. Although the outfits were different on exposure level as intended, they also varied in how tight and immodest they were. This suggests the high­exposure outfit was clearly different from the low­exposure one, but the low­exposure outfit might have been provocative as well. Perhaps an interaction effect would have been more likely if we contrasted a conservative outfit with a provocative one or compared a business formal outfit with a party outfit.

We hypothesized that women showing more skin in a context­inappropriate condition (interview) would be objectified more in comparison to the party condition. This lack of significance begs further examination but

Sherman, Gurung, Jackman, and Mather | Clothing and Context Perceptions

could imply that the strength of clothing and context were powerful and acted in different ways and that our use of generally tight clothing in both conditions did not make for a suitable contrast.

Implications and Limitations

In sum, this study showed evidence that clothing and context affect perceptions of the objectification of women as well as perceptions of women’s professionalism and capabilities. This study was a strong first step in exploring ways to diffuse objectification but there are some key limitations to acknowledge. One limitation can be found in the visual stimuli presented. Although there is a significant difference in the high skin exposure vs. low skin exposure in the photos presented to participants, the “low skin exposure” condition still showed some amount of skin on the ankles, wrists, midriff, collar bones, etc. This might have caused the ratings of the participants to be more similar than if the “low skin exposure” condition showed no skin at all. Similarly, we note some stimuli include a bed in the background. This inconsistency could have influenced perceptions in a way the study did not account for.

Another important limitation of this study was that the adjectives in the objectification category may not be as direct measures of the concept as needed, and the low reliability of a composite bears this out. The variables used were used in past research, but it is likely that the words mean something different to today’s students as compared to those from 15 years ago when some of the original studies were conducted. For example, some younger adults may see being promiscuous as a desirable quality whereas past research considers it a sign of objectification.

This study only used White women between the ages of 18–22 as models, which is not generalizable to the larger population. Future research on this topic should expand the types of models included in the study. The models were somewhat homogenous, and varying racial backgrounds for future studies would be imperative. Additional research should examine age and cultural differences in perceptions of objectification. Both clothing conditions featured social wear. It is possible that comparing casual wear with business or business casual wear would have established a stronger and more consistent main effect of outfit.

Lastly, another key limitation is the difference between exposed and revealing clothing. In this study, all models in revealing outfits were vetted by researchers; however, these models’ bodies could have been more exposed. There is a large difference between showing skin versus showing body parts that may be sexual in nature (i.e., cleavage). Further research could replicate

this study with more extreme versions of the outfits presented, with more exposure.

The influence of context in changing perceptions shows promise for the design of interventions to reduce objectification. Whereas much attention has been placed on characteristics of clothing (e.g., if it is revealing or tight), little research has explored the other factors that influence how impressions of certain types of clothing are influenced. This study provided one avenue for exploration. Given the continuing mistreatment faced by women in society and the focus of dress codes on elements of clothing versus contextual factors, it is critical to broaden the focus on factors influencing perceptions of women. Gaining further understanding of the determining factors of the objectification of women may be key in educating people on how to end negative behaviors against women.

References

Anderson, J. R., Holland, E., Heldreth, C., & Johnson, S. P. (2018). Revisiting the Jezebel stereotype. The impact of target race on sexual objectification. Psychology of Women Quarterly, 42(4), 461–476. https://doi.org/10.1177/0361684318791543

Awasthi, B. (2017). From attire to assault: Clothing, objectification, and de-humanization–A possible prelude to sexual violence? Frontiers in Psychology, 8, 338. https://doi.org/10.3389/fpsyg.2017.00338

Beaumont, S., Barton A. M., Jordan, J., & Froideville, S. M. (2021). Look at her, it’s the same old story. Feminist Media Studies, 22(6), 1478–1494. https://doi.org/10.1080/14680777.2021.1893780

Brown, M., Burnett, L., & Boykin, K. (2022). Was she wearing red? The function of victim‐blaming in women’s intrasexual competition. Scandinavian Journal of Psychology. Advance online publication. https://doi-org/10.1111/sjop.12870

Edmonds, E. M., & Cahoon, D. D. (1984). Female clothes preference related to male sexual interest. Bulletin of the Psychonomic Society, 22(3), 171–173. https://doi.org/10.3758/BF03333795

Fredrickson, B. L., & Roberts, T. (1997). Objectification theory: Toward understanding women’s lived experiences and mental health risks. Psychology of Women Quarterly, 21(2), 173–206.  https://doi.org/10.1111/j.1471-6402.1997.tb00108.x

Glick, P., & Fiske, S. T. (1996). Hostile and benevolent sexism: Measuring ambivalent sexist attitudes toward women. Psychology of Women Quarterly, 21(1), 119–135. https://doi.org/10.1111/j.1471-6402.1997.tb00104.x

Glick, P., Larsen, S., Johnson, C., & Branstiter, H. (2005). Evaluations of sexy women in low- and high-status jobs. Psychology of Women Quarterly, 29(4), 389–395. https://doi.org/10.1111/j.1471-6402.2005.00238.x

Guéguen, N. (2011). The effect of women’s suggestive clothing on men’s behavior and judgment: A field study. Psychological Reports, 109(2), 635–638. https://doi.org/10.2466/07.17.PR0.109.5.635-638

Guillén Ramo, L., & Kakarika, M. (2020). The objectification spillover effect: Why and how objectifying a woman affects other women. Academy of Management Proceedings, 2020(1), 20810. https://doi.org/10.5465/AMBPP.2020.20810abstract

Gurung, R. A. R., & Chrouser, C. J. (2007). Predicting objectification: Do provocative clothing and observer characteristics matter? Sex Roles, 57(1–2), 91–99. https://doi.org/10.1007/s11199-007-9219-z

Gurung, R. A. R., Stoa, R., Livingston, N., & Mather, H. (2021). Can success deflect racism? Clothing and perceptions of African American men. The Journal of Social Psychology, 161(1), 119–128. https://doi.org/10.1080/00224545.2020.1787938

Gurung, R. A. R., Punke, E., Brickner, M., & Badalamenti, V. (2018). Power and provocativeness: The effects of subtle changes in clothing on perceptions of working women. The Journal of Social Psychology, 158(2), 252–255. https://doi.org/10.1080/00224545.2017.1331991

Harsey. (2021). Following the script: Pornography’s associations with men’s sexual aggression through objectification and porn sex scripting

Sherman, Gurung, Jackman, and Mather | Clothing and Context Perceptions

eScholarship, University of California.

Haslam, N., Bain, P., Douge, L., Lee, M., & Bastian, B. (2005). More human than you: Attributing humanness to self and others. Journal of Personality and Social Psychology 89(6), 937–950. https://doi.org/10.1037/0022-3514.89.6.937

Heflick, N. A., & Goldenberg, J. L. (2014). Seeing eye to body: The literal objectification of women. Current Directions in Psychological Science, 23(3), 225–229. https://doi.org/10.1177/0963721414531599

Holland, E., & Haslam, N. (2016). Cute little things: The objectification of prepubescent girls. Psychology of Women Quarterly, 40(1), 108–119. https://doi.org/10.1177/0361684315602887

Howlett, N., Pine, K. J., Cahill, N., Orakçıoğlu, İ., & Fletcher, B. (2015). Unbuttoned: The interaction between provocativeness of female work attire and occupational status. Sex Roles, 72(3–4), 105–116. https://doi.org/10.1007/s11199-015-0450-8

Johnson, K. K. P., Ju, H. W., & Wu, J. (2016). Young adults’ inferences surrounding an alleged sexual assault. Clothing and Textiles Research Journal, 34(2), 127–142. https://doi.org/10.1177/0887302X15624550

Johnson, V., & Gurung, R. A. R. (2011). Defusing the objectification of women by other women: The role of competence. Sex Roles, 65(3), 177–188. https://doi.org/10.1007/s11199-011-0006-5

Kellie, D. X., Blake, K. R., & Brooks, R. C. (2019). What drives female objectification? An investigation of appearance-based interpersonal perceptions and the objectification of women. PloS One, 14(8), e0221388–e0221388. https://doi.org/10.1371/journal.pone.0221388

McBride, H. M., & Kwee, J. L. (2021). Inside and out: How western patriarchal cultural contexts shape women’s relationships with their bodies. In D. Stoyanov, B. Fulford, G. Stanghellini, W. Staden, & M. T. H. Wong (Eds.), International perspectives in values-based mental health practice (pp. 103–108). Springer. https://doi.org/10.1007/978-3-030-47852-0_12

Moor, A. (2010). She dresses to attract, he perceives seduction: A gender gap in attribution of intent to women’s revealing style of dress and its relation to blaming the victims of sexual violence. Journal of International Women’s Studies, 11(4), 115–127.

Noll, S. M., & Fredrickson, B. L. (1998). A mediational model linking selfobjectification, body shame, and disordered eating. Psychology of Women Quarterly, 22(4), 623–636. https://doi.org/10.1111/j.1471-6402.1998.tb00181.x

Nowak, K. L., Fox, J., & Ranjit, Y. S. (2015). Inferences about avatars: Sexism, appropriateness, anthropomorphism, and the objectification of female virtual representations. Journal of Computer-Mediated Communication, 20(5), 554–569. https://doi.org/10.1111/jcc4.12130

Osborn, K., Davis, J. P., Button, S., & Foster, J. (2018). Juror decision making in acquaintance and marital rape: The influence of clothing, alcohol, and preexisting stereotypical attitudes. Journal of Interpersonal Violence, 36(5–6), NP2675–NP2696. https://doi.org/10.1177/0886260518768566

Over, H. (2021). Seven challenges for the dehumanization hypothesis. Perspectives on Psychological Science, 16(1), 3–13. https://doi.org/10.1177/1745691620902133

Persson, S., Dhingra, K., & Grogan, S. (2018). Attributions of victim blame in stranger and acquaintance rape: A quantitative study. Journal of Clinical Nursing, 27(13–14), 2640–2649. https://doi.org/10.1111/jocn.14351

Philippe, B., Gervais, S. J., Allen, J., Delmée, A., & Klein, O. (2015). From sex objects to human beings: Masking sexual body parts and humanization as moderators to women’s objectification. Psychology of Women Quarterly,

39(4), 432–446. https://doi.org/10.1177/0361684315580125

Radun, I., Watling, H., Radun, J., Kaistinen, J., & Olivier, J. (2018). Women judging: Is a young male drunk driver perceived as more intelligent than a young female drunk driver? Traffic Injury Prevention, 19(7), 675–679. https://doi.org/10.1080/15389588.2018.1480833

Rasmussen, E. E., & Densley, R. L. (2017). Girl in a country song: Gender roles and objectification of women in popular country music across 1990 to 2014. Sex Roles, 76(3–4), 188–201. https://doi.org/10.1007/s11199-016-0670-6

Rudman, L. A., & Borgida, E. (1995). The afterglow of construct accessibility: The behavioral consequences of priming men to view women as sexual objects. Journal of Experimental Social Psychology, 31(6), 493–517. https://doi.org/10.1006/jesp.1995.1022

Schooler. (2015). The woman next to me: Pairing powerful and objectifying representations of women. Analyses of Social Issues and Public Policy, 15(1), 198–212. https://doi.org/10.1111/asap.12070

Smith, H., Perez, M., Sladek, M. R., Becker, C. B., Ohrt, T. K., & Bruening, A. B. (2017). Development and validation of makeup and sexualized clothing questionnaires. Journal of Eating Disorders, 5(1), 39–39. https://doi.org/10.1186/s40337-017-0171-1

Szymanski, D. M., Moffitt, L. B., & Carr, E. R. (2011). Sexual objectification of women: Advances to theory and research. The Counseling Psychologist, 39(1), 6–38. https://doi.org/10.1177/0011000010378402

Vaes, J., Paladino, P., & Puvia, E. (2011). Are sexualized women complete human beings? Why men and women dehumanize sexually objectified women. European Journal of Social Psychology, 41(6), 774–785. https://doi.org/10.1002/ejsp.824

Ward, L. M. (2016). Media and sexualization: State of empirical research, 1995–2015. The Journal of Sex Research, 53(4–5), 560–577. https://doi.org/10.1080/00224499.2016.1142496

Workman, J. E., & Freeburg, E. W. (1999). An examination of date rape, victim dress, and perceiver variables within the context of attribution theory. Sex Roles, 41(3–4), 261–277. https://doi.org/10.1023/A:1018858313267

Author Note

Megan Sherman https://orcid.org/0000­0003­4758­4384

Regan A. R. Gurung https://orcid.org/0000­0002­3542­4378

Callan Jackman https://orcid.org/0000­0003­3303­4293

Hannah Mather https://orcid.org/0000­0001­6560­4225

Megan Sherman is now in the PsyD program at the University of La Verne in La Verne, CA.

We have no known conflict of interest to disclose. This study was supported by the psychological research lab at Oregon State University and conducted at Oregon State University via Qualtrics software. Special thanks to Regan A. R. Gurung for mentorship during the data collection and writing process.

Positionality Statement: Megan identifies as a heterosexual, cisgender White woman. All authors are nondisabled and acknowledge that their perspectives are influenced by their positions within all of these dimensions of identity.

Correspondence concerning this article should be addressed to Megan Sherman. Email: shermmeg.trinity@gmail.com

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL.

Clothing and Context Perceptions | Sherman, Gurung, Jackman, and Mather APPENDIX

Materials: Photos of Models

295 COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

Predicting Motivation and Learning Strategies in Community College Students

Karen A. Livesey*1, Alison K. Beatty2, Morrison F. Rubin2,

and

Niomi R. Kaiser3

1Department of Psychology, Northern Virginia Community College

2Department of Psychology, George Mason University

3School of Neuroscience, Virginia Polytechnic Institute and State University

ABSTRACT. The present study examined the relationship among student characteristics, types of academic motivation, learning strategies, and grade point average (GPA). Community college students in psychology courses (N =131) completed an online survey, which assessed 3 types of academic motivations (intrinsic, extrinsic, and amotivation), learning strategies (deep and surface), academic self ­ concept, and demographic variables. Results suggested that academic self­concept (β = .30, p < .001) and age (β = .21, p = .02) added to the prediction of intrinsic motivation. Academic self­concept also significantly predicted amotivation (β = –.60, p < .001). The model tested significantly predicted deep learning strategies (Δ R 2 = .37, p < .001) and surface strategies (ΔR2= .13, p < .001) but not consumeristic motivation (ΔR2 = –.02, p = .78) or GPA (ΔR2 = –.01, p = .53). Overall, the results provided mixed support for the proposed model of student learning in which student characteristics predict motivation, which is related to learning strategies and academic performance. Results are discussed in terms of implications for multifactor models of learning in the community college population.

Keywords: motivation, learning strategies, academic self­concept, community college students, women’s health

Several models of student learning have been proposed in the extant literature, including those that focus on the affective, motivational, and cognitive aspects of learning (Ahmed & Bruinsma, 2006; Pintrich, 1994; Valle et al., 2003; Weber et al., 2011). These models have been well­documented in studies of traditional, four­year undergraduate students (also known as university students) or in studies outside the United States (Algharaibeh, 2020; Brotheridge & Lee, 2005; Erdogdu, 2019). However, these models have not been applied to community college students. Community college students are typically older than traditional university students, are from a wider

Open Data, Open Materials, and Preregistration badges earned for transparent research practices. Data and materials are available at https://osf.io/s5b8n. The preregistration can be viewed at https://osf.io/4k7uy/.

range of ethnic backgrounds, and more likely to work than university students (U.S. Census, 2023). Age and academic self ­ concept are positively related to intrinsic motivation in previous research (Ahmed & Bruinsma, 2006; Brotheridge & Lee, 2005). In addition, studying motivation, learning strategies, and academic success in this diverse group of students who represent approximately half of the undergraduate population in the United States (U.S. Census, 2023) can provide guidance for instructors and administrators at these open­access institutions to support student success and contribute to educational equity. These differences in identity may impact student learning. Therefore, it is

Predicting Motivation and Learning Strategies | Livesey, Beatty, Rubin, and Kaiser

important to determine whether these models replicate among community college students.

Recent studies have suggested that consumeristic motivation, motivation based on an extrinsic reward of obtaining a job, may be related to learning behaviors (Bunce & Bennett, 2019; Lashbrook, 2010; Simons et al., 2004). The lack of inclusion of consumeristic motivation in previous student learning models is a clear gap in the literature. As such, our model sought to integrate consumeristic motivation in predicting learning strategies and academic success. The modified model replicates previous examinations of student learning with a new, distinct, population: community college students. For the present study, we aimed to explain how student characteristics are related to academic motivation, with the addition of consumeristic motivation, to learning strategies and academic success.

Academic Motivation

The motivational aspects of learning included in the present model consist of intrinsic motivation, consumeristic, and amotivation motivation. Intrinsic motivation involves performing an activity for enjoyment rather than for an external reward or to attain a separate more instrumental outcome (Ryan & Deci, 2000). The relationship between students’ academic intrinsic motivation, learning behaviors, and academic success or performance has been well ­ documented in university students, but not in community college students (Algharaibeh, 2020; Boyle et al., 2003; Clarke et al., 2014; Donche et al., 2013; Erdogdu, 2019; Kapp et al., 2020; Prat­Sala & Redford, 2010; Vansteeniske et al., 2004). Simons and colleagues (2004) found that higher intrinsic motivation led to the use of more efficient learning strategies and better academic performance. In addition to examining affect, motivation, and cognitive factors, future studies should include more demographic variables that differentiate community college students from university students (e.g., preparation for college, work status, family responsibility) to provide a complete model of factors that could affect this unique population of students. Examining these learning models for community college students in more detail could help educators focus on the concepts that are most relevant to good study skills and academic success. Similarly, normative models of goal orientation linked students’ ability to master academic material with higher levels of performance (Pintrich, 2000). The literature consistently supported the positive relationship between intrinsic motivation, learning strategies, and academic performance in university students. However, studies have not examined whether this relationship holds in the community college population.

Amotivation is the lack of motivation and is assumed to be at the opposite end of the continuum from intrinsic motivation (Ryan & Deci, 2000). Previous literature has suggested that students were prone to self ­ handicapping and lower academic affect when they were less motivated to master material (Pintrich, 2000). Several studies have found that amotivation is negatively correlated with grades (Algharaibeh, 2020; Kapp et al., 2020; Mauer et al., 2013; Vanthournout et al., 2012). Amotivation has been related to dropping out of college and less effective learning strategies in university students (Donche et al., 2013; Vallerand & Bissonette, 1992; Vanthournout et al., 2012). These findings suggest that intrinsic motivation is related to effective learning strategies and academic success, whereas amotivation is related to less effective learning strategies and lower grades. The findings are a bit less clear when examined in the community college population.

Few studies have examined motivation in community college students and these results have been mixed. Liao and colleagues (2012) found that, although motivation was related to academic achievement in international community college students, there was no relationship between the variables in domestic college students. Simon and colleagues (2015) found that motivation was related to achievement and persistence in community college students in the science and technology fields. Given the paucity of studies and inconsistent findings, further examination of the relationship between motivation and academic achievement in community college students is needed.

Extrinsic academic motivation has been the focus of many recent studies; variations of this motivation are referred to as labor­market orientation, academic capitalism, consumerism or consumerist values, and instrumental orientation or degree purchasing orientation (Brotheridge & Lee, 2005; Fairchild & Crage, 2014; Gretsky & Lerner, 2020; Laverghetta, 2018; Nelson & Sandberg, 2017; Quinlan, 2021; Simons et al., 2004). The commonality among all these types of instrumental/ extrinsic motivations is a focus on attending classes in order to obtain a career rather than to learn the course material. Consumeristic motivation has been associated with less elaborative processing, poor study habits, and negatively related to grades (Brotheridge & Lee, 2005). One study, which examined consumeristic motivation in community college students in Hong Kong, found these students to possess high levels of instrumental motivation which was related to poor study skills (Wong, 2022). This orientation is included in the model and is referred to as consumeristic motivation. This type of motivation may be particularly relevant in a community college sample as many students attend these colleges

Livesey, Beatty, Rubin, and Kaiser | Predicting Motivation and Learning Strategies

for financial reasons (Dadgar, 2012; National Center for Public Policy & Higher Education, 2011).

Learning Strategies

Students’ academic motivation is related to the quality of their study skills and types of learning strategies that they use to learn (Boyle et al., 2003; Donche et al., 2013; Lashbrook, 2010; Simons et al., 2004; Vermetten et al., 2001). Biggs (1987) suggested that surface strategies involve less effective processes, such as memorization or minimal processing of the information, whereas deep strategies involve elaborate processing and active pursuit of knowledge beyond the information presented. These deep learning strategies have been found to be related to intrinsic motivation (Boyle et al., 2003; Donche et al., 2013; Prat­Sala & Redford, 2010; Vansteeniske et al., 2004). On the other hand, the less effective, surface processing has been found to be related to consumeristic motivation and amotivation (Brotheridge & Lee, 2005; Bunce & Bennett, 2019; Donche et al., 2013; Vanthournout et al., 2012; Wong, 2022). These studies suggested that different types of motivations are related to different quality learning strategies. Students with intrinsic motivation appear to use deeper, more effective, learning strategies for high quality learning, while students with amotivation or consumeristic motives appear to use surface, or poor quality study processes. These cognitive processes appeared to be the link between the motivational aspects of the model and academic achievement.

The learning strategies that students use are ultimately related to their grades (Ahmed & Bruisma, 2006; Chen et al., 2015; McInerney et al., 2012; Valle et al., 2003; Wu et al., 2021; Zhang et al., 2022). Studies have shown that students who use deep learning strategies tended to have better academic performance than those using surface strategies (Chen et al., 2015; Valle et al., 2003). McInerney and colleagues (2012) found that academic self ­ concept, learning strategies, and academic achievement had reciprocal relationships with each other in a sample of secondary students in Hong Kong. The literature consistently found support for a relationship between learning strategies and academic achievement in university students, but this relationship has not been examined in community college students. Given that community college students tended to be less prepared for college than their university peers, it is critical to extend the examination of these factors to this special group of college students (Bailey et al., 2015; National Center for Education Statistics [NCES], 2016).

Community College Student Characteristics

Most of the research on the factors that predict motivation, learning strategies, and academic success discussed

above is based on traditional university students. Community college students have different characteristics than students who attend four­year institutions. They tend to be older, less prepared for college, have children, and work either full­time or part­time while in college (Bailey et al., 2015; NCES, 2016). These differences could limit the generalizability of previous studies and learning models to this subpopulation of college students.

Since community college students tend to be older than their university counterparts, age was an important variable to include in the proposed model. Several studies found that age was related to motivation and learning strategies (Brotheridge & Lee, 2005; Nelson & Sandberg, 2017; Shillingford & Karlin, 2013; Warden & Myers, 2017; Zhang et al., 2022). Brotheridge and Lee (2005) found that older students were less likely to engage in consumeristic behaviors of academic achievement. In addition, older students embraced a deeper approach to studying, such as testing themselves on important topics, as opposed to their younger peers who used more surface methods, such as only studying course outlines.

Although there do not tend to be gender differences between community college and university students, several constructs included in the model have been related to gender in the literature. Some studies have found that gender is related to motivation (Ajlouni et al., 2022; Brotheridge & Lee, 2005; Brouse et al., 2010; Vallerand & Bissonnette, 1992). Research has shown that men are more likely than women to view their education as a tool for future success, as opposed to, learning for learning’s sake, and that this focus is related to inferior academic performance (Brotheridge & Lee, 2005). Due to the fact that these studies found some gender differences in learning, gender was included as a student characteristic in the present model.

Academic self­concept, one’s affective and cognitive evaluation of their academic capabilities, has also been shown to be related to academic achievement (Ahmed & Bruinsma, 2006; Chen et al., 2015; McInernery et al., 2012; Valle et al., 2003; Wu et al., 2021; Zhang et al., 2022). Several studies have also found that academic self­concept is related to intrinsic motivation which was included in the model (Ahmed & Bruinsma, 2006; Arens et al., 2019). Students who did not perform well academically in high school might not have had high levels of confidence in their academic abilities; community colleges might have been their only option for obtaining a college degree (Bailey et al., 2015). Therefore, it is critical to examine how academic self ­ concept relates to motivation and learning strategies.

The Present Study

There has been a considerable amount of research on academic motivation and learning strategies, but most of the research has focused on university students (e.g., Algharaibeh, 2020; Brotheridge & Lee, 2005; Erdogdu, 2019). The applicability of these models in community college students who have different characteristics than their university student counterparts, has not been examined. These students tend to be older, more likely to be employed, and less prepared for college (Bailey et al., 2015; NCES, 2016; U.S. Census, 2023). It is important to determine whether these models predict learning behaviors in this population. Studying community college students fills a gap in current research and contributes to a more inclusive and informed understanding of motivation in higher education.

The purpose of the present study was to understand how student characteristics impact motivation and to examine the affective, motivational, and cognitive aspects of the learning process in community college students. Our model integrates several learning models and includes the addition of consumeristic motivation (Ahmed & Bruisma, 2006; Valle et al., 2003; Weber et al., 2011; see Figure 1). The predictions based on this model were that:

H1. Student characteristics of age, gender, and academic self ­ concept would predict three types of motivation: (a) intrinsic, (b) consumeristic, and (c) amotivation.

H2. Intrinsic motivation would predict (a) deeper learning strategies when student characteristics and other types of motivation are held constant, (b) whereas amotivation and consumeristic motivation would predict surface learning strategies.

H3. Deep learning strategies would positively predict academic success.

Descriptive Statistics and Correlations for Study Variables

Procedure

Method

The present study was approved by the Northern Virginia Community College institutional review board (IRB) and was conducted in accordance with all ethical guidelines. Information regarding demographics, learning strategies, and motivation was collected in an online survey between March 31, 2022, and May 3, 2022. Limitations on the total number of questions in the survey were set by the IRB. As such, the most reliable questions in each of the three measurement scales: academic self­concept, academic motivation, and learning strategies were used in this study (Biggs et al., 2001; Reynolds et al., 1980; Vallerand et al., 1992). Recruitment announcements were made in classes and through Canvas. Students were offered extra credit to participate in the online survey outside of class time.

Demographics

One hundred thirty­one students from seven psychology classes participated in the study. Participants predominantly identified as women (78% women and 21% men). They ranged in age from 18 to 50 years (M = 23.37, SD = 7.1). The students identified themselves as Hispanic (32.1%), White (21.4%), Black (16%), Asian (16%), mixed ethnicity (6.9%), and Middle Eastern (5.3%). Three participants did not provide ethnic information.

Measures Academic Self-Concept

Participants were asked to answer 11 questions that represented the highest loading factors in each of the five subscales of the Academic Self­Concept Scale according to the original validation study (Reynolds et al., 1980). Items were on a five­point Likert scale that ranged from 1 ( strongly disagree ) to 5 ( strongly agree ). The scale assessed five components of academic self ­ concept including: self ­ confidence, doubt, grade and effort, habits, and external locus of control. Six questions had to be reversed scored; a high score indicated higher levels of academic self­concept. Items selected for the present study can be accessed at https://osf.io/s5b8n. Reynolds et al. (1980) reported a high reliability based on their 40­item scale (α = .91). In the present study, reliability was α = .84 for the 11 items selected from the original scale.

Academic Motivation Scale

Participants answered 14 questions from the Academic Motivation Scale (Vallerand et al., 1992). The modified version of the scale included six intrinsic motivation items, six extrinsic motivation items (four of the extrinsic items were coded as consumeristic motivation items), and two items assessed amotivation. The items from our

modified version can be accessed at https://osf.io/s5b8n.

Participants were instructed to rate the items based upon the extent to which the items corresponded to the reason why they attended college from 1 (does not correspond at all) to 7 (corresponds exactly; e.g., “In order to obtain a more prestigious job later on”). Three participants did not complete all questions, so they were not included in the specific analyses of these variables. Vallerand et al. (1992) reported the reliability of both the intrinsic and amotivation subscales, as α = .85. The reliability of the extrinsic subscales ranged from α = .62 and α = .83. The present study found reliability coefficients of α = .88 for the intrinsic items, α = .80 for the amotivation items, and α = .67 for the new consumeristic subscale.

Learning Strategies

Participants completed six questions from the Study Process Questionnaire (Biggs et al., 2001). We selected items from the original scale that pertained to learning strategies, we did not include items related to motives. Three items assessed deep learning strategies, and three assessed surface learning strategies. Each learning strategy was measured by totaling the three relevant questions in the respective subscales. The list of the items selected for the current measure can be accessed at https://osf.io/ s5b8n. Two participants did not complete all items and were not used in the analyses relating to learning strategies. Biggs et al. (2001) reported reliability coefficients of α = .64 for surface learning strategies and α = .73 for deep learning strategies. The present study found similar reliability coefficients of α = .60 for surface learning strategies and α = .70 for deep learning strategy items.

Analysis

Descriptive statistics were calculated for all predictors and outcome variables. Pearson’s correlation analyses were run for all predictor and outcome variables. These statistics are included in Table 1. We handled missing data by dropping individuals with missing responses from the analyses of those particular measures. Table 1 also contains the number of participants who had complete responses for each variable. Prior to running multiple regression analyses, tests for assumptions of normality, linearity, independence, homoscedasticity, multicollinearity, and outliers were satisfied. Multiple regression analyses were used to predict each of the three types of motivation, the two types of learning strategies, and grade point average (GPA). All analyses were conducted using IBM SPSS 28.0.1.0.

Results

A prediction of the present study was that student characteristics of age, gender, and academic self­concept would be related to the different types of motivation

(H1). Three separate multiple regression analyses were run to predict each motivation subscale from gender, age, and academic self­concept. All assumptions for the use of multiple regression analysis to predict all three types of motivation were met. The multiple regression model significantly predicted intrinsic motivation, F(3, 124) = 9.08, p < .001, R2 = .18, ΔR2 = .16. Academic self­concept (β = .30, p < .001) and age (β = .21, p = .02) were significant positive predictors of intrinsic motivation. The model significantly predicted amotivation, F(3, 124) = 19.73, p < .001, R2 = .32, ΔR2 = .30, with academic self­concept negatively related to amotivation (β = –.62, p < .001). None of the other predictors were significant. The regression model did not predict consumeristic motivation, F(3, 125) = 0.36, p = .78, R2 = .01, ΔR2 = –.02. Regression coefficients and standard errors for all three regression models which predicted motivation types can be found in Table 2.

Learning Strategies and Academic Performance

One prediction was that intrinsic motivation would be

TABLE 2

Multiple Regression Results for Motivation Variables

Note. Mode = “Enter” method in SPSS Statistics; b = unstandardized regression coefficient; CI = confidence interval; LL = lower limit; UL = upper limit, SE b = standard error of the coefficient; β = standardized coefficient; R2 = coefficient of determination; ΔR2 = adjusted

3

Multiple Regression Results for Learning Strategies

related to deep learning strategies beyond the student characteristics of age, gender, and academic self­concept (H2a). A multiple regression was run to predict deep learning strategies from intrinsic motivation, amotivation, and consumeristic motivation as well as gender, age, and academic self­concept. All assumptions for regression analyses predicting both learning strategies were met. The multiple regression model significantly predicted deep learning strategies, F(6, 119) = 13.27, p < .001, R2 = .40, ΔR2 = .37. Of the six predictors, only age (β = .23, p = .005) and intrinsic motivation (β = .48, p < .001) were significant positive predictors of deep learning strategies.

Additionally, another prediction of the present study was that amotivation and consumeristic motivation would be positively related to surface learning strategies (H2b). The model significantly predicted surface learning strategies, F(6, 120) = 4.12, p < .001, R 2 = .17, Δ R 2 = .13. In addition, amotivation was a positive predictor of surface learning strategies (β = .23, p = .03). Unexpectedly, intrinsic motivation was a positive predictor of surface learning strategies (β = .25, p = .01). Regression coefficients and standard errors for all deep learning and surface learning models can be found in Table 3.

Note. Mode = “Enter” method in SPSS Statistics; b = unstandardized regression coefficient; CI = confidence interval; LL = lower limit; UL = upper limit, SE b = standard error of the coefficient; β = standardized coefficient; R2 = coefficient of determination; ΔR2 = adjusted R2

p < .05. ** p < .01. *** p < .001.

4

Multiple Regression Results for Grades

The final prediction was that deep learning strategies would be positively related to GPA (H3). We included student characteristics, types of motivations, and deep learning and surface learning strategies were predictors of GPA and found that the model was not significant, F(8, 107) = 0.88, p = .53, R2 = .06, ΔR2 = –.01. Regression coefficients and standard errors can be found in Table 4.

These results provided a different picture than the proposed model. See Figure 2 for a graphic of relationships among age, academic self­concept, different types of motivation, deep learning and surface learning strategies, and GPA. The results suggested that academic self­concept was a significant predictor of both intrinsic motivation and amotivation. Age and intrinsic motivation were predictors of deep learning strategies, as expected. Gender was not a significant predictor of any affective, motivational, or cognitive variables. Intrinsic and amotivation also predicted surface strategies. None of the variables predicted consumeristic motivation or academic success as measured by GPA.

Discussion

Note. Mode = “Enter” method in SPSS Statistics; b = unstandardized regression coefficient; CI = confidence interval; LL = lower limit; UL = upper limit, SE b = standard error of the coefficient; β = standardized coefficient; R2 = coefficient of determination; ΔR2 = adjusted R2 *** p < .001.

The purpose of the present study was to examine the affective, motivational, and cognitive aspects of the learning process in community college students. Results of the present study supported several of the hypotheses. Some of the student characteristics predicted

the motivational aspects of learning. As predicted (H1a), age and academic self­concept were positively related to intrinsic motivation. This is consistent with previous studies which examined these constructs in a university institution (Shillingford & Karlin, 2013; Warden & Myers, 2017). Academic self­concept also showed a strong positive relationship with intrinsic motivation, which supports the results from other studies of university students and was negatively related to amotivation as we predicted (H1c; Ahmed & Bruinsma, 2006; Arens et al., 2019). Finding similar results in our community college sample extends the external validity of these claims to include nontraditional students who are more prevalent in community colleges.

None of the variables included in our model predicted consumeristic motivation (H1b), nor was this type of motivation related to either of the learning strategies. Although some studies found age differences in consumeristic motivation, the results do not support these findings (Brotheridge & Lee, 2005; Nelson & Sandberg, 2017). Consumeristic motivation was not related to any other variables of interest in the study. This may be because the measure of consumeristic motivation was not sensitive to differences in this construct within this sample. Another possible explanation may be that consumeristic motivation is not relevant to this group of community college students who were predominantly psychology and prenursing majors. Alternatively, the assessment of consumeristic motivation, which consisted of four questions from an existing scale might have contributed to the lack of correlation. Another possible explanation might be that nontraditional students have a better understanding of their academic abilities, therefore, using deep learning strategies due to higher intrinsic motivation. Future studies should test this model with other community college students to support its reliability and validity. Differences in student demographics across geographical locations may result in different levels within the constructs.

Gender was not related to any of the variables in the model, contrary to expectation based on previous studies. Several studies found that female students were more intrinsically motivated, and therefore, less externally regulated than male students (Ajlouni et al., 2022; Brouse et al., 2010; Vallerand & Bissonnette, 1992). It may be that there are fewer gender differences in community college samples, or these null findings might be due to lack of variability in this disproportionately female sample.

The relationship between the different types of motivations and learning strategies was partially supportive of our predictions and consistent with earlier studies. When age was controlled, intrinsic motivation

had a strong positive relationship with deep learning strategies (H2a), which was consistent with previous studies (Fryer et al., 2014; Vansteenkiste et al., 2006). Contrary to our expectations (H2b), both intrinsic and amotivation were positively predictive of surface learning strategies, although these relationships were small. In previous studies, extrinsic motivation and consumeristic motivation were related to surface learning strategies (Donche et al., 2013; Vanthournout et al., 2012). It is possible that community college students, who tend to be less prepared for college, use any types of strategies that they can to learn the required material. Community college students, who often have additional obligations such as maintaining part­time or full­time employment and taking care of their families, may have to alter the strategies for learning new material due to time constraints. Deep strategies are more time­consuming, and many community college students may not have the time

FIGURE 1

Proposed Model of Student Learning

FIGURE 2

Significant Semipartial Correlations Among Major Variables

Predicting Motivation and Learning Strategies | Livesey, Beatty, Rubin, and Kaiser

to utilize them. This result merits further investigation.

Although many studies have found that deep learning strategies and intrinsic motivation are predictive of better performance in academic settings (Algharaibeh, 2020; Clarke et al., 2014; Fryer et al., 2014; Lizzio et al., 2002), this was not the case in the present study . The lack of predictability of GPA in our results may be due to the nature of our sample. The community college student population often includes older students who are returning to college or delayed entrance into a postsecondary institution (Clovis & Chang, 2019). Nelson & Sandberg (2017) found that adults over the age of 25 years old were less predictable when measuring strategy and motivation. The lack of academic preparation in some community college students is a confounding variable that affects all aspects of the learning process, which may influence results based on location.

Other variables related to course format and instructor behaviors should also be taken into consideration in future models as these factors may impact students’ motivation and learning strategies (Weber et al., 2011). Initially, we planned to include classroom variables, such as program of study, whether the course was required or an elective, course type, teaching modality, and teaching behaviors. However, due to confounds, we decided to remove them to focus on a model that concentrated solely on student characteristics.

Several other limitations of the current study reduce the generalizability of these results. Community colleges serve local communities, so this study may not generalize to all community college students (Franklin, 2013). Future research could replicate this model with other samples in a variety of geographical locations to examine if consumeristic motivation impacts learning strategies and academic success among other community college students. Secondly, the fact that the IRB limited the sample to only courses taught by one professor also reduced the generalizability of the results. A longitudinal examination of community college students to determine if their study habits and motivation fluctuate would be informative (Bunce & Bennett, 2019). Additionally, the sample consisted of predominantly female psychology and nursing students from a single professor’s classes. Students in psychology and medical fields might be intrinsically motivated, whereas, other majors such as engineering might be more consumeristically motivated. To enhance the applicability of findings, future researchers should explore these constructs using a more diverse student sample encompassing various disciplines and demographics.

Another factor that might have contributed to our results was the COVID­19 pandemic. It impacted society in many ways including the importance of a

postsecondary education. Older students might have returned to educational institutions to continue their degrees or to pursue a different program of study. This shift could have resulted in individuals who had more intrinsic motivation returning to college as opposed to individuals who never enrolled in higher education or did not complete their initial attempt at a degree. Conversely, students might have been struggling from COVID fatigue and felt less motivated than they would have otherwise. The fact that 17% of these students indicated that they were not sure why they were still attending college suggests this is a possibility. Future studies that examine the relationship between motivational factors and retention would be very useful because community colleges are often criticized for their completion and transfer rates (Bailey et al., 2015).

The present study added to the understanding of the importance of academic motivation and learning strategies in community college students. The variables that were the most predictive of deep learning strategies, in the present study, were intrinsic motivation and academic self­concept. Although they did not predict academic success, when assessed by GPA, research suggested that these higher quality study strategies form the foundation for academic success that could be potentially enhanced by educators in the classroom (Algharaibeh, 2020; Clarke et al., 2014; Fryer et al., 2014; Lizzio et al., 2002). In addition to examining affect, motivation, and cognitive factors, future studies should include more demographic variables that differentiate community college students from university students (e.g., preparation for college, work status, family responsibility) to provide a complete model of factors that could affect this unique population of students. Examining these learning models for community college students in more detail could help educators focus on the concepts that are most relevant to good study skills and academic success.

References

Ahmed, W., & Bruinsma, M. (2006). A structural model of self-concept, autonomous motivation and academic performance in cross-cultural perspective. Electronic Journal of Research in Educational Psychology, 4(3), 551–575. https://eric.ed.gov/?id=EJ804122

Ajlouni, A., Rawadieh, S., Almahaireh, A., & Awwad, F. A. (2022). Gender differences in the motivational profile of undergraduate students in light of self-determination theory: The Case of online learning setting. Journal of Social Studies Education Research, 13(1), 75–103.

https://www.bulenttarman.com/index.php/jsser/article/viewFile/3956/551

Algharaibeh, S. A. S. (2020). Should I ask for help? The role of motivation and help-seeking in students’ academic achievement: A path analysis model. Cypriot Journal of Educational Science, 15(5), 1128–1145. https://doi.org/10.18844/cjes.v15i5.5193

Arens, A. K., Schmidt, I., & Preckel, F. (2019). Longitudinal relations among selfconcept, intrinsic value, and attainment value across secondary school years in three academic domains. Journal of Educational Psychology, 111(4), 663–684. https://doi.org/10.1037/edu0000313

Bailey, T. R., Jaggars, S. S., & Jenkins, D. (2015). Redesigning America’s community colleges: A clearer path to student success. Harvard University Press.

Livesey, Beatty, Rubin, and Kaiser | Predicting Motivation and Learning Strategies

Biggs, J. B. (1987). Student approaches to learning and studying. Hawthorn, Victoria: Australian Council for Educational Research. https://eric.ed.gov/?id=eD308201

Biggs, J. B., Kember, D., & Leung, D. Y. P. (2001). The revised Two Factor Study Process Questionnaire: R-SPQ-2F. British Journal of Educational Psychology, 71(1), 133–149. https://doi.org/10.1348/000709901158433

Boyle, E. A., Duffy, T., & Dunleavy, K. (2003). Learning styles and academic outcome: The validity and utility of Vermunt’s inventory of learning styles in a British higher education setting. British Journal of Educational Psychology, 73(2), 267–290. https://doi.org/10.1348/00070990360626976

Brotheridge, C. M., & Lee, R. T. (2005). Correlates and consequences of degree purchasing among Canadian university students. The Canadian Journal of Higher Education, 35(2), 71–97. https://doi.org/10.47678/cjhe.v35i2.183501

Brouse, C. H., Basch, C. E., LeBlanc, M., McKnight, K. R., & Lei, T. (2010). College student’s academic motivation: Differences by gender, class, and source of payment. College Quarterly, 13(1), 1–10. https://files.eric.ed.gov/fulltext/EJ912093.pdf

Bunce, L., & Bennett, M. (2019). A degree of studying? Approaches to learning and academic performance among student ‘consumers.’ Active Learning in Higher Education, 22(3), 203–214. https://doi.org/10.1177/1469787419860204

Chen, B., Chiu, W., & Wang, C. (2015). The relationship among academic selfconcept, learning strategies, and academic achievement: A case study of national vocational college students in Taiwan via SEM. Asia-Pacific Education Researcher, 24(2), 419–431. https://doi.org/10.1007/s40299-014-0194-1

Clarke, M. H., Middleton, S. C., Ngyuen, D., & Zwick, L. K. (2014). Mediating relationships between academic motivation, academic integration and academic performance. Learning and Individual Differences, 33, 30–38. https://doi.org/10.1016/j.lindif.2014.04.007

Clovis, M. A., & Chang, M. (2019). Effects of academic momentum on degree attainment for students beginning college at 2-year institutions. Journal of College Student Retention: Research, Theory & Practice, 23(2), 322–336. https://doi.org/10.1177/1521025119826245

Dadgar, M. (2012). The academic consequences of employment for students enrolled in community college. CCRC Workshop Paper No. 46. New York, Columbia Teachers College, Community College Research Center.

Donche, V., De Maeyer, S., Coertjens, L., Van Dall, T., & Petegem, P. (2013). Differential use of learning strategies in first-year higher education: The impact of personality, academic motivation, and teaching strategies. British Journal of Educational Psychology, 83(2), 238–251. https://doi.org/10.1111/bjep.12016

Erdogdu, M.Y. (2019). The mediating role of school engagement in the relationship toward learning and academic achievement. International Journal of Education & Literary Studies, 7(2), 75–81. https://doi.org/10.7575/aiac.ijels.v.7n.2p.75

Fairchild, E., & Crage, S. (2014). Beyond the debates: Measuring and specifying student consumerism. Sociological Spectrum, 34(5), 403–420. https://doi.org/10.1080/02732173.2014.937651

Franklin, R. S. (2013). The roles of population, place, and institution in student diversity in American higher education. Growth and Change, 44(1), 30–53. https://doi.org/10.1111/grow.12001

Fryer, L. K., Ginns, P., & Walker, R. (2014). Between students’ instrumental goals and how they learn: Goal content is the gap to mind. British Journal of Educational Psychology, 84(4), 612–630. https://doi.org/10.1111/bjep.12052

Gretsky, M., & Lerner, J. (2020). Students of academic capitalism: Emotional dimensions in the commercialization of higher education. Sociological Research Online, 26(1), 1–17. https://doi.org/10.1177/1360780420968117

Kapp, A., Mostert, K., & de Beer, L. (2020). Investigating the appropriateness and validity of the Academic Motivation Scale-College Version for South African first-year university students. Journal of Student Affairs in Africa, 8(2), 45–58. https://doi.org/10.24085/jsaa.v8i2.4447

Lashbrook, J. (2010). “Into it” or “going through the motions?”: Exploring how college students understand their academic engagement. Children and Youth Speak for Themselves, Sociological Studies of Children and Youth, 13, 351–371. https://doi.org/10.1108/S1537-4661(2010)0000013016

Laverghetta, A. (2018). The relationship between student anti-intellectualism, academic entitlement, student consumerism, and classroom incivility in a sample of college students. College Student Journal, 52(2), 278–282. https://www.ingentaconnect.com/contentone/prin/csj/2018/ 00000052/00000002/art00011

Liao, H. A., Ferdenzi, A. C., & Edlin, M. (2012). Motivation, self-regulated learning efficacy, and academic achievement among international and domestic

students at an urban community college: A comparison. Community College Enterprise, 18(2), 9–38. https://www.proquest.com/docview/1287963109?pqorigsite=gscholar&fromopenview=true

Lizzio, A., Wilson, K., & Simons, R. (2002). University students’ perceptions of the learning environment and academic outcomes: Implications for theory and practice. Studies in Higher Education, 27(1), 27–52 https://doi.org/10.1080/03075070120099359

Mauer, T. W., Allen, D., Gatch, D. B., Shankar, P., & Sturges, D. (2013). A comparison of student academic motivations across three course disciplines. Journal of the Scholarship of Teaching and Learning, 13(5), 77–89. https://digitalcommons.georgiasouthern.edu/cgi/viewcontent.cgi? article=1015&context=ecology-facpubs

McInerney, D. M., Cheng, R. W., & Mok, M. M. C. (2012). Academic self-concept and learning strategies: Direction of effect on student academic achievement. Journal of Advanced Academics, 23(3), 249–269. https://doi.org/10.1177/1932202X12451020

National Center for Education Statistics (NCES). (2016). DATALAB Tables Library. https://nces.ed.gov/datalab/table/library/detail/13903

Nelson, A., & Sandberg, M. (2017). Labour-market orientation and approaches to studying – A study of the first Bologna Students at a Swedish Regional University. Studies in Higher Education, 42(8), 1545–1566. https://doi.org/10.1080/03075079.2015.1007942

Pintrich, P. R. (1994). Continuities and discontinuities: Future directions for research in educational psychology. Educational Psychologist 29(3), 137–148. https://doi.org/10.1207/s15326985ep2903_3

Pintrich, P. R. (2000). Multiple goals, multiple pathways: The role of goal orientation in learning and achievement. Journal of Educational Psychology, 92(3), 544–555. https://doi.org/10.1037/0022-0663.92.3.544

Prat-Sala, M., & Redford, P. (2010). The interplay between motivation, selfefficacy, and approaches to studying. British Journal of Educational Psychology, 80(2), 283–305. https://doi.org/10.1348/000709909X480563

Quinlan, K. M. (2021). Do higher education students really seek ‘value for money’?: Debunking the myth. London Review of Education, 19(1), 3, 1–12. https://doi.org/10.14324/LRE.19.1.03

Reynolds, W. M., Ramirez, M. P., Magrina, A., & Allen, J. A. (1980). Initial development and validation of the Academic Self-Concept Scale. Educational and Psychological Measurement, 40(4), 1013–1016. https://doi.org/10.1177/001316448004000428

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54–67. https://doi.org/10.1006/ceps.1999.1020

Shillingford, S., & Karlin, N. J. (2013). The role of intrinsic motivation in the academic pursuits of nontraditional students. New Horizons in Adult Education & Human Resource Development, 25(3), 91–102. https://doi.org/10.1002/nha3.20033

Simon, R. A., Aulls, M. W., Dedic, H., Hubbard, K., & Hall, N. C. (2015). Exploring student persistence in STEM programs: A motivational model. Canadian Journal of Education, 38(1), 1–27. https://www.jstor.org/stable/canajeducrevucan.38.1.09

Simons, J., Dewitte, S., & Lens, W. (2004). The role of different types of instrumentality in motivation, study strategies, and performance: Know why you learn, so you’ll know what you learn! British Journal of Educational Psychology, 74(3), 343–360. https://doi.org/10.1348/0007099041552314

U.S. Census Bureau (2022). Community College Month: April 2022, Current population survey, school enrollment supplement. https://www.census.gov/newsroom/stories/community-college-month.html

Valle, A., Cabanach, R. G., Nuñez, J. C., González-Pienda, J., Rodriguez, S., & Pineiro, I. (2003). Cognitive, motivational, and volitional dimensions of learning: An empirical test of a hypothetical model. Research in Higher Education, 44(5), 557–580. https://doi.org/10.1023/A:1025443325499

Vallerand, R. J., & Bissonnette, R. (1992). Intrinsic, Extrinsic, and amotivational styles as predictors of behavior: A prospective study. Journal of Personality, 60(3), 599–620. https://doi-org.eznvcc.vccs.edu/10.1111/j.1467-6494.1992.tb00922.x

Vallerand, R. J., Pelletier, L. G., Blais, M. R., Briere, N. M., Senecal, C., & Vallieres, E. F. (1992). The Academic Motivation Scale: A measure of intrinsic, extrinsic, and amotivation in education. Educational and Psychological Measurement, 52(4), 1003–1017. https://doi.org/10.1177/0013164492052004025

Vansteenkiste, M., Lens, W., & Deci, E. L. (2006). Intrinsic versus extrinsic goal contents in self-determination theory: Another look at the quality of academic motivation. Educational Psychologist, 4(1), 19–31 https://doi.org/10.1207/s15326985ep4101_4

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

WINTER 2023

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

305

Predicting Motivation and Learning Strategies | Livesey, Beatty, Rubin, and Kaiser

Vanthournout, G., Gijbels, D., Coertjens, L., Donche, V., & Van Petegem, P. (2012). Students’ persistence and academic success in a first-year professional bachelor program: The influence of students’ learning strategies and academic motivation. Education Research International, 2012 1–10.

https://doi.org/10.1155/2012/152747

Vermetten, Y. J., Lodewijks, H. G., & Vermunt, J. D. (2001). The role of personality traits and goal orientations in strategy use. Contemporary Educational Psychology, 26(2), 149–170. https://doi.org/10.1006/ceps.1999.1042

Warden, D. N., & Myers, C. A. (2017). Nonintellective variables and nontraditional college students: A domain-based investigation of academic achievement. College Student Journal, 51(3), 380–390.

https://www.ingentaconnect.com/content/prin/csj/2017/00000051/ 00000003/art00008

Weber, K., Martin, M. M., & Myers, S. A. (2011). The development and testing of the Instructional Beliefs Model. Communication Education, 60(1), 51–74. https://doi.org/10.1080/03634523.2010.491122

Wong, Y. L. (2022). Student alienation in higher education under neoliberalism and global capitalism: A case of community college students’ instrumentalism in Hong Kong. Community College Review, 50(1), 96–116. https://doi.org/10.1177/00915521211047680

Wu, H., Guo, Y., Yang, Y., Zhao, L., & Guo, C. (2021). A meta-analysis of the longitudinal relationship between academic self-concept and academic achievement. Educational Psychology Review, 33, 1749–1778. https://doi.org/10.1007/s10648-021-09600-1

Zhang, S., Ma, R., Wang, Z., Li, G., & Fa, T. (2022). Academic self-concept mediates the effect of online learning engagement on deep learning in online courses for Chinese nursing students: A cross-sectional study. Nurse Education Today, 117, 105481. https://doi.org/10.1016/j.nedt.2022.105481

Author Note

Karen A. Livesey https://orcid.org/0000­0001­5461­3652

Morrison Rubin graduated with a BA from George Mason University

Niomi Kaiser graduated with a BA from University of Richmond This study was preregistered at https://osf.io/4k7uy/. Materials and data for this study can be accessed at https://osf.io/s5b8n. We have no known conflicts of interest. Special thanks to Kara E. Hokes for insightful comments during the editing process. Correspondence concerning this article should be addressed to Karen A. Livesey. Email: klivesey@nvcc.edu.

COPYRIGHT 2023 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 28, NO. 4/ISSN 2325-7342)

Publish Your Research in Psi Chi Journal

Undergraduate, graduate, and faculty submissions are welcome year round. Only one author (either first author or coauthor) is required to be a Psi Chi member. All submissions are free. Reasons to submit include

• a unique, doctoral­level, peer­review process

• indexing in PsycINFO, EBSCO, and Crossref databases

• free access of all articles at psichi.org

• our efficient online submissions portal View Submission Guidelines and submit your research at www.psichi.org/?page=JN_Submissions

Become a Journal Reviewer

Doctoral­level faculty in psychology and related fields who are passionate about educating others on conducting and reporting quality empirical research are invited become reviewers for Psi Chi Journal. Our editorial team is uniquely dedicated to mentorship and promoting professional development of our authors—Please join us!

To become a reviewer, visit www.psichi.org/page/JN_BecomeAReviewer

Resources for Student Research

Looking for solid examples of student manuscripts and educational editorials about conducting psychological research? Download as many free articles to share in your classrooms as you would like.

Search past issues, or articles by subject area or author at www.psichi.org/journal_past

Learn About Psi Chi

Psi Chi is the International Honor Society in Psychology. Membership is primarily open to undergraduates, graduate students, transfer students, full­time and part­time faculty members, and alumni.

See membership benefits and a link to apply at www.psichi.org/page/member_benefits

Register an account: http://pcj.msubmit.net/cgi-bin/main.plex