Psi Chi Journal of Psychological Research – Summer 2022 by Psi Chi, the International Honor Society in Psychology

SUMMER 2022 | VOLUME 27 | ISSUE 2

ISSN: 2325-7342 Published by Psi Chi, The International Honor Society in Psychology

® ®

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH S U MM E R 2022 | VOLU M E 27, N U M BE R 2

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 TAYLOR BROWN-STONE EDITORIAL ASSISTANTS EMMA SULLIVAN ADVISORY EDITORIAL BOARD GLENA ANDREWS, PhD George Fox University 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

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." Mem bership 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 lo cat ed in Chatta nooga, 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 rec ognition 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 con ven tions, 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 increas ing 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 broad en the dissemination of research across the psychological science community. JOURNAL INFORMATION The Psi Chi Journal of Psychological Research (ISSN 2325-7342) 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) 756-2044. 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. dvertisements that appear in Psi Chi Journal do not represent endorsement by Psi Chi of the A 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.

LISA ROSEN, PhD Texas Women's University CHRISTINA SINISI, PhD Charleston Southern University PAUL SMITH, PhD Alverno College

SUMMER 2022 | VOLUME 27 | ISSUE 2

EDITORIAL: Introducing the Diversity Badge

Steven V. Rouse Social Sciences Division, Pepperdine University Editor, Psi Chi Journal of Psychological Research

Interaction Between Health Self-Efficacy and Race on Self-Reported Health Status

Julianna Berardi, Emily Jobson, and Mora A. Reinka* Department of Psychology, Ursinus College

105

How Perceptions of Obstacles, Stress, and Different Mindsets May Impact a Student’s Self-Beliefs

Kristin P. Rullman, Samantha L. D’Anna, Lauren A. Jacobs, Kristina R. Jacobs, and William A. Jellison* Department of Psychology, Quinnipiac University

113

The Moderating Effect of Mindfulness on the Relationship Between Problematic Smartphone Usage and Depression, Anxiety, and Stress

Christina Stratton, Elizabeth J. Krumrei-Mancuso*, and Cindy Miller-Perrin* Department of Psychology, Pepperdine University

124

Examining the Efficacy of Using a Change Blindness Framework as a Novel Social Media Intervention

Stephanie Misko, Olivia Hays, and Laura Getz* Department of Psychological Sciences, University of San Diego

133

The Classic Stroop Asymmetry in Online Experiments

Mary E. Smith, Micah D. Smith, and Kenith V. Sobel* Department of Psychology and Counseling, University of Central Arkansas

145

Music’s Impact on the Sexualization of Black Bodies: Examining Links Between Hip-Hop and Sexualization of Black Women

Elizabeth A. Otto, Shaina A. Kumar, and David DiLillo* Department of Psychology, University of Nebraska-Lincoln

154

False Memory for Words in Noise: An At-Home Deese-Roediger-McDermott (DRM) Experiment Across Adulthood

Rebecca L. Wagner1 2, Bethany A. Lyon1*, and Angela AuBuchon2* 1 Department of Psychology, University of Nebraska at Omaha 2 Boys Town Medical Research Hospital

166

*Faculty mentor

Predictors of Help-Seeking: Self-Concept Clarity, Stigma, and Psychological Distress Hinza B. Malik and Caroline E. Mann* Department of Psychology, Hollins University

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

https://doi.org/10.24839/2325-7342.JN27.2.96

EDITORIAL: Introducing the Diversity Badge Steven V. Rouse Social Sciences Division, Pepperdine University Editor, Psi Chi Journal of Psychological Research

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

lthough it is possible to learn a lot about an organization from its mission statement, it is sometimes easiest to see how that mission is enacted by looking at how the organization orients its resources. For example, the mission of Psi Chi is “recognizing and promoting excellence in the science and application of psychology” (Psi Chi, 2019). This is a valuable summation of the organization’s goals. However, mission statements are, by design, brief and generalized, and one must often look beyond the mission statement itself to see the ways the mission is used to direct specific programs, procedures, and policies. One might wonder, for example, how exactly the organization would fulfill this mission, and what practical steps are taken toward this goal. A very clear indication of the steps the organi zation has chosen to reach its goals is to examine its strategic initiatives. For example, Psi Chi has a set of four faculty advisory committees, each with its own emphasis to help the honor society fulfill its mission; these are the Diversity Advisory Committee (currently directed by Dr. Gabrielle Smith), the Faculty Support Committee (currently directed by Dr. Seungyeon Lee), the International Advisory Committee (currently directed by Dr. Brien Ashdown), and the Research Advisory Committee (currently directed by Dr. John Edlund). As I look at the organization, I perceive that Psi Chi’s leadership sees the important components of promoting excel lence in the science and application of psychology in order to provide faculty members who teach psychol ogy with the resources they need to do their work, ensure that the organization remains active across national borders, advocate for the highest level of research rigor, and promote diversity and inclusion. As the editor of the Psi Chi Journal of Psychological Research, it is my responsibility to align the journal with the Psi Chi mission, and it helps me to see how to accomplish this by thinking about the journal in the context of those four emphases. How, I ask myself, is the journal supporting faculty members, maintaining an international emphasis, advocat ing for diversity and inclusion, and promoting

high levels of research rigor? I am grateful to follow exemplary editors such as Drs. Melanie M. Domenech Rodríguez and Debi Brannan who accomplished so much in each of these areas. I believe that the journal is already aligned well with each of these four emphases, but it is helpful for me to reflect on what other changes we could enact toward these goals. Recently, I wrote an editorial in which I announced four new initiatives specifically focused on one of these four areas: increasing representa tion of diversity within the pages of the journal (Rouse, 2021). I called for more inclusive and complete reporting of demographics, statements of constraints on generality especially when a sample is highly homogeneous, and optional statements of positionality when a manuscript addresses a historically marginalized community. Moreover, I announced my commitment to increase the diversity of our reviewers. In this second editorial, I am pleased to announce a fifth initiative, the creation of a badge that will designate those articles that focus on aspects of diversity, as seen in Figure 1. Much like our cur rent Open Materials, Open Data, Preregistration, and Replication badges, the new badge will appear on the first page of any article that meets the defin ing criteria.1 Starting in the current issue, we will award Diversity Badges to articles that: (a) examine whether psychological phenomena differ as a func tion of human diversity, (b) highlight psychological characteristics within a historically marginalized group, or (c) identify factors that are related to diversity-based prejudice or discrimination.2 The Open Materials, Open Data, and Preregistration badges were all developed by the Center for Open Science; more information about these three badges can be obtained at https://www.cos.io/initiatives/badges. The Replication badge was created specifically for the Psi Chi Journal of Psychological Research and was introduced by Rouse (2017). 2 These articles should also follow the best practices I outlined previously (Rouse, 2022), such as conscientiously reporting complete demographic data for the participants (including gender, race, ethnicity, and age, as well as any characteristics relevant to the study at hand), and discussing constraints on generality for nondiverse samples. 1

Rouse | Diversity Badge

This journal already publishes many studies with these three emphases of diversity. For instance, Nakai and Gurung (2021) recently exemplified the first emphasis; they showed that a frequently used mea sure of shyness (developed within the United States) does not function in a psychometrically equivalent manner when used in Japan. We look forward to more articles in the future that examine the ways psychological characteristics are stable or variable across different dimensions of identity diversity. In addition, Shepler and colleagues (2021) pro vided an example of the second emphasis by show ing that the relationship satisfaction experienced by gay and bisexual men was systematically related to a variety of constructs in the Minority Stress Model, such as uncertainty and concealment of one’s sexual identity. We encourage the submission of future manuscripts that continue to elucidate the experiences of diverse groups of people. Finally, Almond and colleagues (2021) serve as an example of the third emphasis in their examina tion of the gender-based microaggressions experi enced by graduate-level and early career women in the psychological sciences. We hope that authors continue to submit high-quality manuscripts that examine aspects of identity-based prejudice and discrimination. Important diversity-related research is already being published within the pages of the journal, and the new badge will highlight those articles that use the tools of psychological scientific research to understand and promote diversity.3 I would like to recognize, however, that this new badge differs from the four badges previously awarded by the journal in one important way. The Open Data, Open Materials, Preregistration, and Replication badges all focus on important aspects of open science and transparency in the research process. But while these four existing badges focus on the method of psychological research, this new badge differs in its focus on the content of psycho logical research. One might reasonably wonder

why this topic is different from other important topics in psychological research. Is this, one might ask, the beginning of a proliferation of badges? I recently joked with the editorial board that we would not be able to create too many additional badges; after all, with our red, blue, yellow, green, and (now) purple badges, we are using five of the six primary and secondary colors! Why then, are we choosing to deviate from the previous themes of the methodology badges to highlight the topic of diversity? Ultimately, it ties back to the mission of Psi Chi and its initiatives for reaching its goals. Clearly, an emphasis on diversity is an important means by which Psi Chi seeks to reach its mission, and this emphasis is transparent in many of Psi Chi’s diversity-related initiatives (https://www.psichi.org/ page/Diversity#resources), such as in the diversityrelated grants, informative short videos on YouTube, and an index of diversity-related content. I believe that the decision to create this badge is aligned well with the organization’s mission-based initiatives. Ultimately, this should be the standard that informs and guides changes within the journal. So, as the pages of the Psi Chi Journal of Psychological Research become more colorful, con gratulations to Malik (2022), Otto et al. (2022), and Reinka et al. (2022) for being the first recipients of our new Diversity Badge. I look forward to awarding more of these badges in future issues, and I hope FIGURE 1 Open Data, Open Materials, Preregistration, Replication, and Diversity Badges

DIVERSITY

After specifying these three criteria, we conducted an interrater reliability analysis to determine whether the badges could be assigned consistently. Heather A. Haas (University of Montana Western) and Lisa H. Rosen (Texas Woman's University), both of whom are members of the journal's Editorial Advisory Board, independently read the abstract of each article published in this journal in 2021. Without consulting each other and without being in additional conversations about these criteria, they independently decided whether or not the criteria applied. For the 39 articles in 2021, these raters were in agreement 90% of the time. The product-moment correlation between their ratings was .81 with a Kappa coefficient of .79. This suggests that the badges can be assigned very reliably based on these criteria. I am grateful to Drs. Haas and Rosen for their assistance in this process. 3

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

Diversity Badge | Rouse

that this special program will encourage more diversity-related submissions to our journal, as well as more recognition and attention from across the scientific community for those published on these very important topics.

References Almond, A. L., Ayala, E. E., Moore, M. M., & Mirzoyan, M. D. (2021). Angry, frustrated, and silent: Women’s responses to microaggressions within the discipline that coined microaggressions. Psi Chi Journal of Psychological Research, 26(4), 408–421. https://doi.org/10.24839/2325-7342.JN26.4.408 Berardi, J., Jobson, E., & Reinka, M. A. (2022). Interaction between health self-efficacy and race on self-reported health status. Psi Chi Journal of Psychological Research, 27(2), 99–104. https://doi.org/10.24839/2325-7342.JN27.2.99 Malik, H. B. (2022). Predictors of help-seeking: Self-concept clarity, stigma, and psychological distress. Psi Chi Journal of Psychological Research, 27(2), 166–177. https://doi.org/10.24839/2325-7342.JN27.2.166 Nakai, S. C., & Gurung, R. A. R. (2021). Measurement invariance analysis on the Revised Cheek and Buss Shyness Scale in a U.S. and Japanese college sample. Psi Chi Journal of Psychological Research, 26(4), 383–394. https://doi.org/10.24839/2325-7342.JN26.4.383 Otto, E. A., Kumar, S. A., & DiLillo, D. (2022). Music’s impact on the sexualization

of Black bodies: Examining links between hip-hop and sexualization of Black women. Psi Chi Journal of Psychological Research, 27(2), 145–153. https://doi.org/10.24839/2325-7342.JN27.2.145 Psi Chi. (2019). Mission & purpose. https://www.psichi.org/page/purpose Rouse, S. V. (2017). The red badge of research (and the yellow, blue, and green badges, too). Psi Chi Journal of Psychological Research, 22(1), 2–9. https://doi.org/10.24839/2325-7342.JN22.1.2 Rouse, S. V. (2021). Increasing the representation of diversity in the Psi Chi Journal of Psychological Research. Psi Chi Journal of Psychological Research, 26(4), 360–362. https://doi.org/10.24839/2325-7342.JN26.4.360 Shepler, D. K., Glaros, M. R., & Boot, J. W. (2021). Minority stress and relationship satisfaction among gay and bisexual men. Psi Chi Journal of Psychological Research, 26(3), 296–306. https://doi.org/10.24839/2325-7342.JN26.3.296 Author Note. Steven V. Rouse https://orcid.org/0000-0002-1080-5502 Positionality Statement: Steve Rouse identifies as a cisgender White man. As a bi man, he identifies as part of the LGBTQ+ communities. He is nondisabled. He acknowledges that his perspective is influenced by his position within all of these dimensions of identity. Correspondence concerning this manuscript should be addressed to Steven V. Rouse, Social Sciences Division, Pepperdine University, Malibu, CA 90263-4372, United States. Email: steve.rouse@pepperdine.edu

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

https://doi.org/10.24839/2325-7342.JN27.2.99

Interaction Between Health Self-Efficacy and Race on Self-Reported Health Status Julianna Berardi, Emily Jobson, and Mora A. Reinka* Department of Psychology, Ursinus College

ABSTRACT. The purpose of this study was to analyze the impact of race and ethnicity on the relationship between health self-efficacy (confidence in one’s own health management abilities) and self-reported health status. The relationship between health self-efficacy and self-reported health status was predicted to be weaker in participants of color. A secondary data analysis was conducted on the 2018 National Survey of Health Attitudes with nationally representative participants (N = 7,187). A hierarchical linear regression was performed. Consistent with our hypotheses, our results illustrated a significant main effect for health self-efficacy on health status, where greater health self-efficacy was associated with better health, β = .29, p < .001. Although there was no main effect for race and ethnicity, a significant interaction between the two predictors was also found, suggesting that participants of color demonstrated a weaker relationship between health self-efficacy and self-reported health status compared to White participants, β = –.08, p < .001. Thus, the relationship between health self-efficacy and self-reported health status can be influenced by marginalized racial or ethnic experience.

DIVERSITY

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

Keywords: health self-efficacy, minority health, health status, race, ethnicity

n e ’s h e a l t h s t a t u s i s t h e r e s u l t o f complex, interconnecting systems, including individual beliefs and behavior, interpersonal factors, and institutional structures (Office of Disease Prevention and Health Promotion, 2020). Although these factors may promote positive health outcomes, they may also inhibit wellness, as barriers to health can also be understood using the same connected systems approach (Bauer et al., 2005; Lee et al., 2010). For example, interpersonal and structural forces create health disparities based on race and ethnicity (Lewis et al., 2015). Such disparities occur as both barriers to access (Lee et al., 2010) and as substantial increases in mortality and morbidity of diseases (DiIorio et al., 2011). In addition to macro-level influences on health, *Faculty mentor

there exist important individual influences as well. Health self-efficacy—that is, one’s confidence in their own handling or managing of their individual health—is an established predictor of health status, as higher reported confidence in one’s health man agement abilities has been associated with more optimal health behaviors, and thus better health outcomes (Jerant et al., 2005; Miller & Cronan, 1998). Although prior research has provided a basis for the relationship between health self-efficacy and health status, it has often failed to go beyond the scope of limited populations due to the underrep resentation of minoritized individuals (e.g., Miller & Cronan, 1998). Previous work has also tended to focus on nonrandom, small-scale sampling, which has further contributed to a limited understanding of the breadth of this association (e.g., Jerant et al.,

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

Health Self-Efficacy and Race | Berardi, Jobson, and Reinka

2005). In either scenario, the relationship between health self-efficacy and health status cannot be generalized to the ever-diversifying U.S. popula tion. Thus, larger, more representative samples are needed to corroborate existing research. Among the work that has included racially or ethnically diverse populations, researchers exam ined the reverse relationship, where one’s health status affected their sense of health self-efficacy (Thompson et al., 2017). The purpose of the current research was to examine the impact of minoritized racial and ethnic group membership on the association between health self-efficacy and self-reported health status. Considering the established association between health self-efficacy and health status, and the perva sive role of race and ethnicity on health outcomes, we developed a three-part hypothesis that (a) there would be a positive relationship between health self-efficacy and self-reported health status, where increased levels of health self-efficacy would relate to better rated health; (b) minoritized race or ethnic ity would be inversely associated with self-reported health status, where minoritized participants would report poorer health compared to White par ticipants; and (c) the relationship between health self-efficacy and self-reported health status would be negatively impacted by minoritized race/ethnicity.

Method

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

100

Participants The Robert Wood Johnson Foundation (RWJF) and Research and Development Corporation (RAND) conducted the National Survey of Health Attitudes in 2018 (Carmen et al., 2018) using two internet panels, the American Life Panel (ALP; Santa Monica, CA) and the KnowledgePanel (Paris, France). Through probability-based sampling, address-based samples, and random digit dial ing, the sample was nationally representative of American adults (N = 7,187). Due to participant nonresponse, degrees of freedom varied. Although the ALP was overrepresentative of vulnerable popu lations, such as those with lower per capita income, the panels were pooled into one sample, as no systematic differences arose in RWJF and RAND’s investigation of the panels (Carmen et al., 2018). Participants reported their age in years (M = 53.87, SD = 16.00) as well as their gender (53.8% women; 46.2% men) and were from 30 communities across four major regions of the United States. The data was deidentified and publicly available (see Survey, below); therefore, no IRB approval was necessary. For participant characteristics, see Table 1.

Measures Demographics Racial and ethnic categorization was collected by the internet panels. Due to power concerns, the current study collapsed racial and ethnic categoriza tion into two groups (Non-White 28.9% and White 71.1%). Participants also reported family income (M = $50,000–$59,999) and highest education achieved (M = some college, no degree). Survey The 2018 National Survey of Health Attitudes consisted of 34 questions, with some questions incorporating subsections or multiple parts. The survey and data were obtained through the InterUniversity Consortium for Political and Social Research (ICPSR), a public data forum. Participants were offered the survey online in English or Spanish and were asked questions about their health and beliefs about health equity (Mdncompletion time = 18.5 mins). Internet and technology resources were pro vided to participants who did not have such access by the internet panels ALP and KnowledgePanel. The respondents received a modest, undisclosed, compensation by RWJF and RAND for their participation. To impede potential order effects, some questions were randomized. For the current investigation, we were interested in questions regarding health self-efficacy and health status. Procedure The present study measured the influence of race and ethnicity on the relationship between the participants’ reported health self-efficacy and their reported health status. To quantify health self-efficacy, respondents rated their confidence in their health management abilities: “How confident are you that you can: Manage any health problems you have.” Participants indicated their confidence level on a 4-point scale from 1 (not confident at all) to 4 (very confident; M = 3.20, SD = 0.74). The health status out come measure was self-reported, where respondents rated their health: “Would you say that in general your health is excellent, very good, good, fair, or poor?” (excellent = 1, poor = 5; M = 3.42, SD = 0.94). Analyses The current investigation analyzed the coded responses from the 2018 data using Jamovi (The Jamovi Project, 2020), a free, online, statistical software. The health self-efficacy predictor variable remained on the original 4-point scale but was

Berardi, Jobson, and Reinka | Health Self-Efficacy and Race

transformed into a mean-centered variable for further analysis. The race/ethnicity moderator was dummy coded, such that White participants were coded as 0, and participants who identified as another race or ethnicity were coded as 1. Health status was originally coded on a 5-point scale in the order the responses were presented in the question. Responses were subsequently reverse coded by the current researchers so that higher values equated to better health. Participant income and educa tion were included as covariates to disentangle any effects of race from socioeconomic status. A hierarchical linear regression was conducted using base Jamovi tools and the MedMod package (R Core Team, 2020).

Results To analyze relationships between our three vari ables (i.e., health self-efficacy, race and ethnicity, and health status), we performed a hierarchical linear regression to find main and interaction effects. Health self-efficacy had a significant positive association with self-reported health status (β = .29, B = 0.37, SE = 0.02, 95% CI [0.34, 0.40], p < .001), such that greater health self-efficacy correlated with higher self-reported health status. However, race and ethnicity were not significantly related to health status (β = –.03, B = –0.03, SE = 0.02, 95% CI [–0.07,0.02], p = .22). The main effects and covariates significantly explain almost a tenth of the shared variance between the predictors and the outcome of self-reported health status, adjusted R2 = .09, p < .001. An additional 0.2% of shared variance is explained when including the interaction between health self-efficacy and race into the analysis, p < .001. As predicted, a significant interaction between health self-efficacy and race and ethnicity was found (β = –.08, B = –0.11, SE = 0.03, 95% CI [–0.17, –0.05], p < .001), such that marginalized racial and ethnic demonstrated a weaker relationship between health self-efficacy and self-reported health status than those who identified as White (see Figure 1). It is important to note that both family income (β = .16, B = 0.04, SE = 0.003, 95% CI [0.03, 0.05], p < .001) and education (β = .09, B = 0.04, SE = 0.01, 95% CI [0.03, 0.05], p < .001)—included as covariates in the analyses—were significantly and positively related to self-reported health status. In line with previously established trends (e.g., National Center for Health Statistics, 2012), the greater a participant’s socioeconomic status, the better their reported health. Without education

TABLE 1 Participant Demographic Characteristics Variable

Race/ethnicity Non-Hispanic White

5,111 71.1%

Non-Hispanic Black

671

Hispanic

953 13.3%

Non-Hispanic Asian or Pacific Islander

230

3.2%

0.7%

174

2.4%

Non-Hispanic American Indian or Alaskan Native Non-Hispanic All Other Races Sex Female Male Education Less than 1st grade 1st, 2nd, 3rd, or 4th grade 5th or 6th grade 7th or 8th grade 9th grade 10th grade 11th grade 12th grade, no diploma High school graduate or equivalent Some college, no degree Associate degree Bachelor degree Master degree Professional or doctorate degree

9.3%

3,870 53.8% 3,317 46.2% 14 8 30 52 64 78 77 94 1,602 1,447 778 1,577 969 397

0.2% 0.1% 0.4% 0.7% 0.9% 1.1% 1.1% 1.3% 22.3% 20.1% 10.8% 21.9% 13.5% 5.5%

Less than $5,000

140

1.9%

$5,000–$7,499

0.7%

$7,500–$9,999

105

1.5%

$10,000–$12,499

132

1.8%

$12,500–$14,999

145

2.0%

$15,000–$19,999

213

3.0%

$20,000–$24,999

310

4.3%

$25,000–$29,999

286

4.0%

$30,000–$34,999

332

4.6%

$35,000–$39,999

350

4.9%

$40,000–$49,999

549

7.6%

$50,000–$59,999

593

8.3%

$60,000–$74,999

747 10.4%

$75,000–$99,999

918 12.8%

Family Income

$100,000–$124,999

822 11.4%

$125,000–$199,999

1,039 14.5%

$200,000 or more

446

6.2%

Region Northeast

1,317 18.3%

Midwest

1,487 20.7%

South

2,564 35.7%

West

1,781 24.8%

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

101

Health Self-Efficacy and Race | Berardi, Jobson, and Reinka

and income in the model, race and ethnicity showed a significant main effect (β = –.12, B = –0.11, SE = 0.02, 95% CI [–0.16, 0.06], p < .001), indicating that much of the influence of race on health was driven by socioeconomic factors. Inclusion of the covariates did not substantively alter any other relationships.

Discussion To increase our understanding of the influences on one’s perceived health, we examined the main effects of health self-efficacy and race/ethnicity, along with the subsequent interaction between the two, on self-reported health status. In a three-part hypothesis, we hypothesized that (a) there would be a positive relationship between health self-efficacy and self-rated health status, (b) minoritized race and ethnicity would be inversely associated with self-reported health status, and (c) the relationship between health self-efficacy and self-reported health status would be negatively impacted by minoritized race and ethnicity. A positive association existed between health self-efficacy and self-reported health status, where increased confidence in one’s health management was related to better health, supporting the first part of our hypothesis. However, minoritized race and ethnicity was not significantly related to health status, thus we were not able to support the second part of the hypothesis. Still, the relationship between health self-efficacy and self-reported health status was negatively impacted by the interaction with race and ethnicity, support ing the final part of our hypothesis. The interac tion demonstrates how minoritized participants exhibited a weaker relationship between health FIGURE 1 Interaction of Health Self-Efficacy and Race/Ethnicity on Health Status Health Self-Efficacy

Race/Ethnicity

Interaction between Health Self-Efficacy and Race/Ethnicity

.292 (.015)

−.030 (.023)

Self-Reported Health Status

−.084 (.029)

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

102

Note. Standardized estimates shown, with standard errors in parentheses. Solid lines indicate significance, p < .05. Race/Ethnicity coded as 0 = White; 1 = NonWhite. Regression model includes education and family income as covariates.

self-efficacy and self-reported health status com pared to participants who identified as White. Consistent with the findings of previous litera ture, health self-efficacy continues, in the current study, to be a strong predictor of self-reported health status (Jerant et al., 2005; Miller & Cronan, 1998). Building upon this literature, this relation ship held within a large, nationally representative sample of adults from the United States, spanning across four major regions. Thus, the current study provided support for the previously established association between health self-efficacy and per ceived health. The results in the current work contrast with prior literature on racial or ethnic health dispari ties (e.g., DiIorio et al., 2011) in that we did not see a main effect of race and ethnicity on health status within this nationally representative sample. However, this could be due to the inclusion of edu cation and income as covariates. Indeed, without the included covariates, race and ethnicity did show a significant main effect, underscoring the fact that many social categories are often conflated. Finally, although small, the interaction between health self-efficacy and race/ethnicity significantly described an additional amount of variance, above and beyond the influence of socioeconomic factors. Thus, the interaction may be an important factor in understanding differential health outcomes. Limitations We must note the correlational nature of our data and emphasize that causality cannot be established. Also, because previous literature with minoritized racial and ethnic populations focused on the opposite direction of variables—that is, the impact of health status on health self-efficacy (Thompson et al., 2017)—it is likely that these relationships are bidirectional. For example, our results dem onstrated how lower confidence in one’s health management abilities was associated with lower reported health status, suggesting less favorable perceived health. However, it could also be true that lower reported health status could impact one’s levels of health self-efficacy, as less favorable health could be associated with a lower perceived ability to manage poor health. It is likely that other factors, such as gender or other marginalized social identities, as well as logis tical concerns such as insurance coverage or access to care, would predict health status (e.g., Lasser et al., 2006; Williams et al., 2016; Wright et al., 2018). However, we did not include consideration of these factors here. Future research should investigate how

Berardi, Jobson, and Reinka | Health Self-Efficacy and Race

these, or similar variables, confound the relation ship between health self-efficacy, minoritized race and ethnicity, and self-reported health status. In the current study, we collapsed all minori tized races and ethnicities into a single category of non-White for power concerns, as well as ease of comparison between White Americans and Americans of color. Different racial and ethnic groups likely experience diverse issues regarding health and the healthcare system (e.g., Jaiswal & Halkitis, 2019) that may result in further nuances in the examined relationships. Perhaps most impor tantly, it is imperative to recognize that, although we used racial and ethnic identity as a predictor of health, it is racism that creates pervasive health disparities, not race itself. Indeed, we saw a glimpse of this nuance in our inclusion of education and income as covariates, both of which are not only drivers of health inequity themselves (e.g., National Center for Health Statistics, 2012), but interact with race to affect health (e.g., Brown et al., 2016). Finally, caution is provided on our large sample size. Although it provides national generalizability and greater confidence in our results, it should also be noted that the influences and impacts observed have small effects among such a large sample. The difference between the main effects model and the interaction model, although significant, provided only a fraction of additional explained variance. Because of this minimal effect, some scholars have suggested that such a relationship should be viewed with modesty (Kaplan et al., 2014). Even if viewed with appropriate caution, results from large datasets like those of the current study can provide better generalization; especially regarding nuanced relationships that can be targeted in public health or biomedicine. Conclusion The current study contributed to the understanding of health outcomes, with particular focus on the determinants of perceived health status, specifically among marginalized racial and ethnic populations. Not only did this investigation add to the existing literature on health self-efficacy and health disparities, but we hope it will be foundational for future research. For minoritized individuals, improving health self-efficacy may not be enough to substantively improve health status and reduce health disparities. This research underscored that individual interven tions to improve health and well-being may only go so far, and that more sweeping public policy

addressing the roots of inequality are needed. Although further analysis is necessary to wholly understand the interaction between race and ethnicity and health self-efficacy, this initial study highlights a relationship that can be targeted for intervention among public health officials, social workers, community health organizations, and/or social policy initiatives.

References Bauer, M. S., Williford, W. O., McBride, L., McBride, K., & Shea, N. M. (2005). Perceived barriers to health care access in a treated population. The International Journal of Psychiatry in Medicine, 35(1), 13–26. https://doi.org/10.2190%2FU1D5-8B1D-UW69-U1Y4 Brown, T. H., Richardson, L. J., Hargrove, T. W., & Thomas, C. S. (2016). Using multiple-hierarchy stratification and life course approaches to understand health inequalities: The intersection consequences of race, gender, SES, and age. Journal of Health and Social Behavior, 57(2), 200–222. https://doi.org/10.1177/0022146516645165 Carmen, K. G., Chandra, A., Miller, C., Trujillo, M. D., Yeung, D., Weilant, S., DeMartini, C., Edelen, M. O., Huang, W., & Acosta, J. D. (2018). Development of the Robert Wood Johnson Foundation national survey of health attitudes. RAND Report. https://www.rwjf.org/en/library/research/2019/01/2018national-survey-of-health-attitutes.html DiIorio, C., Steenland, K., Goodman, M., Butler, S., Liff, J., & Roberts, P. (2011). Differences in treatment-based beliefs and coping between African American and White men with prostate cancer. Journal of Community Health, 36(4), 505–512. https://doi.org/10.1007/s10900-010-9334-6 Jaiswal, J., & Halkitis, P. N. (2019). Towards a more inclusive and dynamic understanding of medical mistrust informed by science. Behavioral Medicine, 45(2), 79–85. https://doi.org/10.1080/08964289.2019.1619511 The Jamovi Project. (2020). jamovi. (Version 1.2) [Computer Software]. https://www.jamovi.org Jerant, A. F., von Friederichs-Fitzwater, M. M., & Moore, M. (2005). Patients’ perceived barriers to active self-management of chronic conditions. Patient Education and Counseling, 57(3), 300–307. https://doi.org/10.1016/j.pec.2004.08.004 Kaplan, R. M., Chambers, D. A., & Glasgow, R. E. (2014). Big data and large sample size: A cautionary note on the potential for bias. Clinical and Translational Science, 7(4), 342–346. https://doi.org/10.1111%2Fcts.12178 Lasser, K. E., Himmelstein, D., U., & Woolhandler, S. (2006). Access to care, health status, and health disparities in the United States and Canada: Results of a cross-national population-based survey. American Journal of Public Health, 96(7), 1300–1307. https://doi.org/10.2105/AJPH.2004.059402 Lee, S., Martinez, G., Ma, G. X., Hsu, C. E., Robinson, E. S., Bawa, J., & Juon, H. S. (2010). Barriers to health care access in 13 Asian American communities. American Journal of Health Behavior, 34(1), 21–30. https://doi.org/10.5993/AJHB.34.1.3 Lewis, T. T., Cogburn, C. D., & Williams, D. R. (2015). Self-reported experiences of discrimination and health: Scientific advances, ongoing controversies, and emerging issues. Annual Review of Clinical Psychology, 11, 407–440. https://doi.org/10.1146/annurev-clinpsy-032814-112728 Miller, C., & Cronan, T. (1998). The effects of coping style and self-efficacy on health status and health care costs. Anxiety, Stress and Coping, 11(4), 311–325. https://doi.org/10.1080/10615809808248317 National Center for Health Statistics. (2012). Health, United States, 2011: With special feature on socioeconomic status and health. U.S. Department of Health and Human Services. https://www.cdc.gov/nchs/data/hus/hus11.pdf Office of Disease Prevention and Health Promotion. (2020, October 08). Social determinants of Health. U.S. Department of Health and Human Services. https://www.healthypeople.gov/2020/topics-objectives/topic/socialdeterminants-of-health R Core Team. (2020). R: A language and environment for statistical computing. (Version 3.6) [Computer software]. https://cran.r-project.org/ Thompson, T., Mitchell, J. A., Johnson-Lawrence, V., Watkins, D. C., & Modlin Jr, C. S. (2017). Self-rated health and health care access associated with African American men’s health self-efficacy. American Journal of Men’s Health, 11(5), 1385–1387. https://doi.org/10.1177%2F1557988315598555

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

103

Health Self-Efficacy and Race | Berardi, Jobson, and Reinka

Williams, D. R., Priest, N., & Anderson, N. B. (2016). Understanding associations among race, socioeconomic status, and health: Patterns and prospects. Health Psychology, 35(4), 407–411. https://doi.org/10.1037/hea0000242 Wright, R. R., Perkes, J. L., Schaeffer, C., Woodruff, J. B., Waldrip, K., & Dally, J. L. (2018). Investigating BMI discrepancies in subjective and objective reports among college students. Journal of Human Health Research, 1, 106–115. Author Note. Julianna Berardi https://orcid.org/0000-0002-0852-0399

Mora A. Reinka https://orcid.org/0000-0001-9903-2966 Julianna Berardi is now at the Department of Health Management and Policy, Dornsife School of Public Health, Drexel University. We have no conflicts of interest to disclose. Correspondence concerning this article should be addressed to Mora A. Reinka, Department of Psychology, Ursinus College, 601 E. Main St., P.O. Box 1000, Collegeville, PA, 19426-1000, United States. Email: mreinka@ursinus.edu

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

104

https://doi.org/10.24839/2325-7342.JN27.2.105

How Perceptions of Obstacles, Stress, and Different Mindsets May Impact a Student’s Self-Beliefs Kristin P. Rullman, Samantha L. D’Anna, Lauren A. Jacobs, Kristina R. Jacobs, and William A. Jellison* Department of Psychology, Quinnipiac University

ABSTRACT. The current study explored how the mediational model of perceived task complexity, academic stress, and self-efficacy is moderated by cognitive mindset. College students (N = 140; Mage = 19.2, SD = 1.5) completed self-report measures of task complexity, academic stress, self-efficacy, and cognitive mindset. Results demonstrated that academic stress is a mediator between task difficulty and self-efficacy for students with a fixed mindset, but not for students with a growth mindset. There was a significant indirect effect at the mean of the moderator (indirect effect = –.09, p = .022, 95% CI [–0.20, –0.03]), a significant indirect effect at 1 standard deviation below the mean of the moderator (more fixed mindset; indirect effect = –.17, p = .005, 95% CI [–0.30, –0.06]), but no significant effect at 1 standard deviation above the mean (more growth mindset; indirect effect = –.00, p = .95, 95% CI [–0.09, 0.08]). An implication is that college students may still benefit from growth mindset interventions even at a later state in their academic careers. It can be helpful to make students aware of their cognitive mindset disposition so they can better handle their perceptions of difficulty and in times of stress. Mindset interventions should be explored in future research and implemented in school and university settings. Keywords: college students, mindset, stress, self-efficacy

elf-efficacy is the extent to which one believes in their abilities to accomplish a task (Ye et al., 2018). Albert Bandura’s social cognitive theory (1989) provided a deeper look into self-efficacy and the ways in which it influences individuals’ behavior and outcomes. According to Bandura (1989), selfefficacy is influenced by several factors, such as feedback from teachers, observations of others, and a person’s own performance. Bandura highlighted the importance of the environment because he believed that much of the learning that one does is through the environment (i.e., through observing and interacting with others). To illustrate, research has concluded that a person’s self-efficacy is related to events in the environment that can affect their lives (Dzewaltowski, 1994). Therefore, it is logical to assume that external factors could shape a person’s self-efficacy, such as variables like task complexity and academic stress. *Faculty mentor

The Relationships Between Task Complexity and Self-Efficacy Perceived task complexity can be summarized as how difficult an individual perceives a task; it was important to analyze how this may be related to students’ other subjective qualities (Sides et al., 2017). Past literature has shown perceived task complexity to be negatively correlated with feelings of self-efficacy. Thus, when people think of a task as being difficult, they are also likely to experience a decrease in their belief of success. For example, using tasks involving imaginary scenarios in which participants were asked to rate the perceived com plexity of and their perceived ability to accomplish a task, it was demonstrated that greater levels of perceived task complexity related to a consistent decrease in self-efficacy (Hu et al., 2007). This negative relationship likely exists because, in a more difficult task, people have less information on how

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

105

Academic Perceptions and Cognitive Mindsets | Rullman, D'Anna, Jacobs, Jacobs, and Jellison

to approach it, thus their belief in their ability to be successful decreases. Consequently, when individu als are more familiar with a task and perceive it to be less difficult, they are able to create a plan of action and feel more confident in their abilities. Similar results have been found using hypothetical class schedules; the results concluded that there was a significant negative relationship between perceived task complexity and self-efficacy, thus further emphasizing how the subjectivity of a task’s complexity can be influential in how one perceives their ability to be successful when completing such task (Mangos & Steele-Johnson, 2001). The negative effect of perceived task complexity on self-efficacy has been found with motor tasks as well (e.g., golf task; Sides et al., 2017), which demonstrated how this negative relationship between the two variables extended further than the perception of imaginary scenarios. Therefore, past literature has illustrated a clear negative relationship between perceived task complexity and self-efficacy. The Relationship Between Task Complexity and Academic Stress Past literature has also explored the relationship between task complexity and academic stress. For example, life stress has been examined to see how it relates to errors made on a task (Klein & Barnes, 1994). Having an increased level of stress was shown to increase errors when completing a task, which then suggested that more errors are made when a task is perceived as more difficult. To further expand, it has been demonstrated that individuals who are very aware of their stress responses are more likely to show a decline in performance (Klein & Barnes, 1994). Additionally, it was found that people who perform difficult tasks with high anxiety tend to do worse on the task than people with low anxiety (Brown, 1996). However, the opposite has been found to be true with simple tasks. Therefore, a clear positive relationship between perceived task complexity and academic stress is evident in past literature.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

106

The Relationship Between Academic Stress and Self-Efficacy Oftentimes, academic stress comes about due to a demanding workload, pressure to perform well, and the amount of material that an individual is required to learn (Ye et al., 2018). For example, it has been found that individuals with a greater perception of stress had a decreased sense of self-efficacy; therefore, this indicated a negative relationship between these two variables (Rovira

et al., 2010). Further empirical research focusing on stress and self-efficacy found similar results. A study which utilized middle school teachers found that nearly all teachers were experiencing high stress and also that the majority of them were also characterized as having low self-efficacy (Herman et al., 2020). As shown, research like this supports the claim that individuals with greater stress will report less self-efficacy. The Current Study For the present study, we aimed to investigate whether academic stress mediated the relation ship between perceived task complexity and selfefficacy. Additionally, we aimed to explore how this mediational model is moderated by cognitive mindset. These four variables are thought to be important aspects of a student’s academic career. Thus, we intended to gain a better understanding of how perceived task complexity, academic stress, self-efficacy, and cognitive mindset are related in a student’s life. Cognitive mindset involves the malleability of traits—how much one believes they can improve their abilities (Yeager & Dweck, 2012). There are two different types of cognitive mindsets: fixed and growth. For example, a student with a fixed mindset may be more avoidant of some academic tasks in the fear of appearing unintelligent, as they interpret a lack of success as a reflection of their own, unchangeable ability (Dweck, 2006). These types of students display a fixed mindset because they believe that their intelligence cannot be altered, and their hesitation to participate in difficult tasks grows. On the other hand, a growth mindset is displayed when a student believes that their intelligence can be improved upon and developed. These students may shape their goals around ways to improve their learning and might display more effort in academic tasks (Dweck & Yeager, 2019). Thus, how a student interprets their own intelligence can be paramount for their academic success. In school, students face numerous academic tasks every day with varying difficulty. In referencing cognitive mindset, whether a student displays more of a fixed or growth mindset could be influential in how they handle their everyday stress (Dweck, 2006). When students demonstrate more of a growth mindset, they are less negatively affected by academic challenges and are likely to have a greater work ethic when compared to students of a fixed mindset, leading to a significant improvement in

Rullman, D'Anna, Jacobs, Jacobs, and Jellison | Academic Perceptions and Cognitive Mindsets

classroom motivation and prevention of decreasing grades (Blackwell et al., 2007). This demonstrates how a growth mindset can act as a buffer between task difficulty and academic stress, as those with a stronger growth mindset understand that a chal lenging assignment can be an opportunity to grow and thus feel motivation to perform well rather than feeling stressed and hindered by the potential to fail (Yeager & Dweck, 2020). This could explain why students with a fixed mindset may experience a negative correlation between perceived task difficulty and academic stress. This relationship may be weaker or nonexistent for students with a growth mindset. Additionally, students have a greater potential to be more successful in their academics if they perceive a challenging task as a means to gain aca demic value. It is then possible for students with a growth mindset to report greater self-efficacy when faced with a complex task, consequently influencing behaviors and corresponding environments (Dweck & Leggett, 1988). We therefore proposed that, as perceived task complexity increases, self-efficacy also increases for those with a growth mindset. For example, students with a growth mindset, who perceived an academic challenge as valuable, performed better than those who perceived it as a roadblock (Oyserman et al., 2018). Thus, rather than succumbing to the fear of failure during an academic challenge, students with a growth mindset find themselves rising to the occasion and executing required tasks more confidently than those with a fixed mindset. We hypothesized that cognitive mindset would moderate the mediational model between task complexity, academic stress, and self-efficacy. For fixed mindset, academic stress would mediate the relationship between task complexity and selfefficacy; that is, as perceived task difficulty increases, academic stress will increase, reducing self-efficacy. However, for a growth mindset, we predicted that academic stress would not mediate the relationship between task complexity and self- efficacy. The proposed moderated mediation model is displayed in Figure 1.

Method Participants A total of 140 participants (Mage = 19.2, SD = 1.5) were recruited through the Quinnipiac University Psychology participant pool website. The current study was among other studies or research alterna tives in which the student could choose, and they

received course credit for participation. Of the participants, 106 were women, 33 were men, and one identified as nonbinary. The participants’ class level was also recorded: 95 were first-year students, 18 were sophomores, 14 were juniors, and 13 were seniors. Additionally, the race of participants was recorded: 119 White/European American participants, 10 Hispanic/Latino/Latina participants, 7 Black/ African American participants, 3 Asian or Pacific Islander participants, and 1 Black/White participant. Procedure Before conducting the study, approval was granted by Quinnipiac University’s Institutional Review Board (#11519). Participants completed four measures concerning task complexity, academic stress, self-efficacy, and cognitive mindset online through Google Forms. All participants completed the measures in the same order. Completion of the measures took place during the beginning to middle of the second academic semester of the school year. After signing an informed consent form, participants completed the measures below. After completion of the measures, participants were given a debriefing form, which contained more information about the study, and they were thanked for their participation. Measures Perceived Task Complexity The level of difficulty participants perceived their academic tasks to be was assessed using a modified and shortened version of the Work Design Questionnaire (Morgeson & Humphrey, 2006; α = .74; 15 items; e.g., “My courses require me to analyze a lot of information” and “The job FIGURE 1 Proposed Moderated Mediation Model

Perceptions of Task Difficulty

Academic Stress

Self-Efficacy

Cognitive Mindset: Fixed vs. Growth

107

Academic Perceptions and Cognitive Mindsets | Rullman, D'Anna, Jacobs, Jacobs, and Jellison

required me to keep track of more than one thing at a time”). Participants responded to each item on a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Items were reverse-scored when appropriate and scored such that greater positive values reflected greater perceived task complexity. Academic Stress Anxious-like feelings participants reported because of their academics were assessed using a shortened version of the Perceptions of Academic Stress Scale (Bedewy & Gabriel, 2015; α = .71; 11 items; e.g., “I have enough time to relax after class” and “The size of the curriculum (workload) is excessive”). Participants responded to each item on a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Items were reverse-scored when appropriate and scored such that greater positive values reflected greater academic stress. Self-Efficacy How much the participants believed in their abili ties in relation to academics was assessed using the General Self-Efficacy Scale (Schwarzer & Jerusalem, 1995; α = .82; 10 items; e.g., “I can solve most prob lems if I invest the necessary effort” and “I remain calm when facing difficulties because I can rely on my coping abilities”). Participants responded to each item on a scale ranging from 1 (not true at all) to 4 (exactly true). Items were reverse-scored when appropriate and scored such that greater positive values reflected greater feelings of self-efficacy. Cognitive Mindset How much participants believed in their ability to change their intelligence was assessed using TABLE 1 Descriptive Statistics and Bivariate Correlations for All Participants Measures

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

108

PTC

Perceived Task Complexity (PTC)

3.63

0.42

Academic Stress (AS)

2.48

0.55

.22**

Self-Efficacy (SE)

3.17

0.38

.06

−.43**

Cognitive Mindset (CM)

3.89

0.72

.17

−.12

.12

Note. N = 140. Perceived Task Complexity (PTC) is the level of perceived difficulty of academic tasks. Academic Stress (AS) includes the anxious-like feelings participants felt because of their academics. Self-Efficacy (SE) is how much the participants believed in their abilities in relation to academics. For PTC, AS, and SE, greater positive values reflected greater endorsement of the construct. Cognitive Mindset (CM) is how much participants believed in their ability to change their intelligence. For CM, greater positive values reflected a greater growth mindset, compared to a fixed mindset. ** p < .01. *p < .05.

the Theories of Intelligence Scale (Dweck, 2000; α = .71; 8 items; e.g., “You have a certain amount of intelligence, and you can’t really do much to change it” and “No matter who you are you can significantly change your intelligence level”). Participants responded to each item on a scale ranging from 1 (strongly disagree) to 6 (strongly agree). Items were reverse-scored when appropriate and scored such that greater positive values reflected greater growth mindset.

Results Correlations A series of bivariate correlations were conducted to explore the relationship between perceived task complexity, academic stress, self-efficacy, and cogni tive mindset among all participants. Descriptive statistics and all bivariate correlations are displayed in Table 1. A statistically significant positive correlation was found between perceived task complexity and academic stress, such that as perceived task complex ity increased, academic stress increased. In addition, a statistically significant negative correlation was found between academic stress and self-efficacy, such that, as academic stress increased, self-efficacy decreased. However, for the entire sample, no sta tistically significant relationship was found between perceived task complexity and self-efficacy. Finally, a statistically significant positive correlation was found between cognitive mindset and perceived task complexity demonstrating that participants with more of a growth, compared to a fixed, mindset reported greater perceived task complexity. Moderated Mediation Model To test whether cognitive mindset moderates the proposed mediational model, we used conditional process modeling to test for moderated mediation as outlined by Hayes (2018) using the PROCESS macro. Specifically, we tested to see whether cogni tive mindset moderated the relationships among task difficulty, academic stress, and self-efficacy. We expected that moderated mediation would occur in that path (from task difficulty to academic stress) as it would be statistically significant for those with a fixed mindset, but not for those with a growth mindset. Therefore, only among participants with a fixed mindset would academic stress mediate the relationship between task difficulty and self-efficacy. There was evidence for moderated mediation. We found a significant interaction between the mod erator cognitive mindset and task difficulty on the

Rullman, D'Anna, Jacobs, Jacobs, and Jellison | Academic Perceptions and Cognitive Mindsets

mediator academic stress (b = –0.38, p = .011), thus supporting a moderating (i.e., conditional) effect on path a. The effect from academic stress to the outcome self-efficacy (path b) was significant (b = –.03, p < .001), whereas the direct effect from task difficulty to self-efficacy (path c) was not significant (b = .01, p = .06). The test for mediation at various levels of the moderator resulted in a significant indi rect effect at the mean of the moderator (indirect effect = –.09, p = .022, 95% CI: –.20, –.03), a signifi cant indirect effect at one standard deviation below the mean of the moderator (more fixed mindset; indirect effect = –.17, p = .005, 95% CI: –.30, –.06), but no significant effect at one standard deviation above the mean (more growth mindset; indirect effect = –.00, p = .95, 95% CI: –.09, .08). Hence, the analysis supported the notion that academic stress is a mediator between task difficulty and self-efficacy for students with more of a fixed mindset but not for students with more of a growth mindset.

Discussion For the current study, we intended to investigate whether academic stress mediates the relationship between perceived task complexity and selfefficacy. Additionally, we aimed to examine how this mediational model is moderated by cognitive mindset. Perceived task complexity, academic stress, self-efficacy, and cognitive mindset are crucial components to a student’s academic career. This suggests that each variable is codependent, as negative relationships have shown to impact the connection among one another. Therefore, the goal of our study was to gain a better understanding of how perceived task complexity, academic stress, self-efficacy, and cognitive mindset influence a student’s life. Ideally, the results of our study will be used by educational institutions to improve a student’s academic success. Our results concluded that academic stress is a mediator between task difficulty and self-efficacy for students with a fixed mindset but not for students with a growth mindset. For students who had more of a fixed mindset, our results suggested that academic stress is a mediator between task difficulty and self-efficacy. It is necessary to consider how the relationship between these variables is addressed in previous research and if it aligns with our findings. Therefore, past literature focused on academic stress, task dif ficulty, and self-efficacy had been reviewed, finding similar relationships to our study. For example, the negative relationship between task complexity and self-efficacy has been well documented, focusing

additionally on the cognitive demand of the task (Hu et al., 2007). Meaning that as task complexity increases, the perceived ability to succeed decreases. This also connects to the past research between academic stress and self-efficacy, which additionally demonstrates a negative relationship. Students who experience anxious-like feelings also have reported lower levels of self-efficacy (Ringeisen et al., 2019). Finally, as previously mentioned, individuals who have higher stress levels during a task tend to perceive tasks as more difficult and result in more errors being made, indicating a positive relation ship between these variables (Klein & Barnes, 1994). Thus, this mediational model between task complexity, academic stress, and self-efficacy has been supported by past researchers. With the support of this mediational model, it is also crucial to consider the recognition of cogni tive mindset in previous research. We believe that past literature is aligned with our results demon strating that the mediational model presents itself in those with a fixed, not growth, mindset. This can be supported by the research, which found that a growth mindset in students positively related to classroom motivation and prevented grades from decreasing during times of school transitions. This finding is meaningful because it demonstrates how students with a fixed mindset might not have the luxury of feeling motivated by difficult tasks and are more likely to suffer academically when going through times of stress, such as school transitions. Therefore, it is evident that students with a fixed mindset may suffer academically due to how they perceive their intelligence to be unmalleable, and as our results show, are likely to feel more academic stress, because of perceived task complexity, which will result in lower levels of self- efficacy. Comparing our results to previous literature, we now can show a link between all four variables of perceived task complexity, academic stress, selfefficacy, and cognitive mindset. Prior to this study, we could only make assumptions of how all four variables would be related, but we now know that, for those with a fixed mindset, the mediational model is supported. This also illustrates that, for students who have a growth mindset, their selfefficacy may be affected differently by the variables of perceived task complexity and academic stress or not altered to a significant effect. In sum, we now have evidence of moderated mediation. Implications Because our results are consistent with past studies, it is important to consider how they extend the

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

109

Academic Perceptions and Cognitive Mindsets | Rullman, D'Anna, Jacobs, Jacobs, and Jellison

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

110

findings of the surrounding literature. For example, much of the past literature on cognitive mindset and the other variables of interest predominately utilized samples of students in middle and high school (i.e., Ye et al., 2018; Yeager & Dweck, 2012). Oppositely, our study extended the findings of how cognitive mindset relates to perceived task complexity, aca demic stress, and self-efficacy in a higher education setting as our sample only included college students. By analyzing the makeup of our sample, we can then consider how the disparities between those with a growth and fixed mindset still persist no matter the age. This consideration may lead to some implica tions related to cognitive mindset and interventions to change a student’s mindset. For example, because our results demonstrate that cognitive mindset can be related to academic factors such as stress and self-efficacy in college students, it may be beneficial to have growth mindset interventions even at a later stage in a student’s academic career. Past literature has shown that these interventions can be success ful in middle school and early high school samples because the students who have participated in an intervention have been found to choose more diffi cult academic tasks and improve their performance (Yeager et al., 2016). Therefore, it is possible that a similar intervention could be beneficial even for college students. In addition to considering the potential for growth mindset interventions for college students, our results add important implications for how this research can be applied to a real college setting. For instance, because our results demonstrate that students who have a fixed mindset may experience increased perceived task complexity and academic stress as well as lesser self- efficacy, it is critical that students are aware of their cognitive mindset disposition. Being cognizant of how one interprets the malleability of their traits, such as intelligence, could shape their college experience. This could mean that a student who is self-aware of their own fixed mindset might be more willing to undergo a growth mindset intervention or learn about healthier stress coping mechanisms if they know that their mindset when put under academic stress could lead to less belief in their success. Therefore, schools should be mindful of possible strategies. To make students see the importance of being familiar with their cognitive mindset, schools could utilize various methods of disseminating this information, such as discussions during first-year orientation, so they are better prepared as they enter their college career. The results of the present student can offer some key academic development opportunities

for students, encouraging self-reflection, as well as mindset alteration, for a more positive experience in the pursuit of higher education. Furthermore, pupils with a background that stunted their academic development, such as experiencing traumatic childhood events, external stressors, or specified socioeconomic status, may experience a difficult time focusing when it comes to academics (Yeager & Dweck, 2012). Examples include living below the poverty line, abuse at home, or household dysfunction. The impact of these external variables leads to social development problems, such as aggression toward peers (Fraser, 2018). If not addressed by the school, students then have a higher probability of feeding into the stereotype their life outside of school suggests, and, thus, have a fixed mindset throughout the duration of their education and development. The impact that external variables have on childhood academic development may directly inhibit educational capa bility, making even a growth mindset induction later in their academic career all the more challenging (Yeager & Dweck, 2012). This then highlights the importance of applying a growth mindset into the elementary curriculum earlier and giving greater consideration to how this induction may contribute to their mindset later in life, such as when they attend college. Limitations It is crucial to consider the limitations to external validity and generalizability across a variety of populations. This study focused on a predominately white, female, middle-to upper-class collegiate population, yet cognitive mindset and academic stress exist no matter age, sex, or race (i.e., Yeager & Dweck, 2012; Ye et al., 2018). Past research has demonstrated female students perform better aca demically compared to male students, with a diverse sample, therefore the current study’s results from a majority female sample may not be applicable to all populations (Goldie & O’Connor, 2021). Research also indicates that college populations may not be representative of the general population. This is due to the isolated academic setting students find themselves in, as well as group norms (Sears, 1986). However, it important to note that exploring the experiences of college students does have value (Sears, 2008). There may also be limitations to the inter nal validity of our study. First, the survey mea sures were not counterbalanced, and, thus, our results may have been influenced by order effects. Furthermore, the current study relied on the

Rullman, D'Anna, Jacobs, Jacobs, and Jellison | Academic Perceptions and Cognitive Mindsets

participants’ perception of themselves, coincid ing with their subjected biases, often leading to a possible distorted reality. Notwithstanding, past literature has indicated that a person’s appraisal of their subjective character is not significantly dif ferent from their objective personal estimates. For example, Wright and colleagues (2018) revealed that participants underestimated their body mass index (BMI) by only .51 units. In other words, an individual’s view of themselves may not be substan tially different from reality. The lack of assessing more individuals across a variety of cultural and developmental backgrounds are limitations that restrict generalizability within our study. Our findings fail to address a broader population of differing age groups as well as cultural backgrounds. With this awareness, future research should include a wide variety of partici pants from different demographics. This would add value to our study because it would provide more insight into our participant population. The hope would be that, from surveying the participants, we would have more awareness about personal experiences in crucial stages of their life, potentially making our findings more broadly applicable. Undoubtedly, childhood experiences could impact self-efficacy, one of the primary variables in our study. Consequently, when self-efficacy is impacted, this can have detrimental effects on a student’s academic career. From these results, it can be gathered that academic stress can help explain the relationship between task difficulty and self-efficacy for students who have a fixed mindset, rather than a growth mindset. This research contributed to the extensive research demonstrating the importance of students developing a growth mindset. The hope is that the results and implications of this study are considered and are helpful for students, teachers, mentors, and other strong influences in students’ lives.

References Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44(9), 1175–1184. https://doi.org/10.1037/0003-066X.44.9.1175 Bedewy, D., & Gabriel, A. (2015). Examining perceptions of academic stress and its sources among university students: The perception of academic stress scale. Health Psychology Open, 2(2). https://doi.org/10.1177/2055102915596714 Blackwell, L. S., Trzesniewski, K. H., & Dweck, C. S. (2007). Implicit theories of intelligence predict achievement across an adolescent transition: A longitudinal study and an intervention. Child Development, 78(1), 246–263. https://doi.org/10.1111/j.1467-8624.2007.00995.x Brown, D. W. (1966). The effects of anxiety, stress, and task difficulty on performance on a verbal learning task. The Journal of Psychology: Interdisciplinary and Applied, 64(2), 199–212. https://doi.org/10.1080/00223980.1966.10544844 Dweck, C. S. (2000). Self-theories: Their role in motivation, personality, and development. Taylor & Francis.

Dweck, C. S. (2006). Mindset. Random House Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and personality. Psychological Review, 95(2), 256–273. https://doi.org/10.1037/0033-295X.95.2.256 Dweck, C. S., & Yeager, D. S. (2019). Mindsets: A view from two eras. Perspectives on Psychological Science, 14(3), 481–496. https://doi.org/10.1177/1745691618804166 Dzewaltowski, D. A. (1994). Physical activity determinants: A social cognitive approach. Medicine & Science in Sports & Exercise, 26(11), 1395–1399. https://doi.org/10.1249/00005768-199411000-00015 Fraser, D. M. (2018). An exploration of the application and implementation of growth mindset principles within a primary school. British Journal of Educational Psychology, 88(4), 645–658. https://doi.org/10.1111/bjep.12208 Goldie, P. D., & O’Connor, E. E. (2021). The gender achievement gap: Do teacher–student relationships matter? Psi Chi Journal of Psychological Research, 26(2), 139–149. https://doi.org/10.24839/2325-7342.JN26.2.139 Hayes, A. (2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press. Herman, K. C., Prewett, S. L., Eddy, C. L., Savala, A., & Reinke, W. M. (2020). Profiles of middle school teacher stress and coping: Concurrent and prospective correlates. Journal of School Psychology, 78, 54–68. https://doi.org/10.1016/j.jsp.2019.11.003 Hu, J., Huhmann, B. A., & Hyman, M. R. (2007). The relationship between task complexity and information search: The role of self-efficacy. Psychology & Marketing, 24(3), 253–270. https://doi.org/10.1002/mar.20160 Klein, K., & Barnes, D. (1994). The relationship of life stress to problem solving: Taskcomplexity and individual differences. Social Cognition, 12(3), 187–204. https://doi.org/10.1521/soco.1994.12.3.187 Mangos, P. M., & Steele-Johnson, D. (2001). The role of subjective task complexity in goal orientation, self-efficacy, and performance relations. Human Performance, 14(2), 169– 186. https://doi.org/10.1207/S15327043HUP1402_03 Morgeson, F. P., & Humphrey, S. E. (2006). The Work Design Questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91(6), 1321–1339. https://doi.org/10.1037/0021-9010.91.6.1321 Oyserman, D., Elmore, K., Novin, S., Fisher, O., & Smith, G. C. (2018). Guiding people to interpret their experienced difficulty as importance highlights their academic possibilitiesand improves their academic performance. Frontiers in Psychology, 9. https://doi.org/10.3389/fpsyg.2018.00781 Ringeisen, T., Lichtenfeld, S., Becker, S., & Minkley, N. (2019). Stress experience and performance during an oral exam: The role of self-efficacy, threat appraisals, anxiety, and cortisol. Anxiety, Stress & Coping; An International Journal, 32(1), 50–66. https://doi.org/10.1080/10615806.2018.1528528 Rovira, T., Edo, S., & Fernandez-Castro, J. (2010). How does cognitive appraisal lead to perceived stress in academic examinations? Studia Psychologica, 52(3), 179–192. https://www.studiapsychologica.com/index.php/view-articles/ Schwarzer, R., & Jerusalem, M. (1995). Generalized self-efficacy scale. In J. Weinman, S. Wright, & M. Johnston, Measures in health psychology: A user’s portfolio. Causal andcontrol beliefs (pp. 35–37). NFER-NELSON. Sears, D. O. (1986). College sophomores in the laboratory: Influences of a narrow data base on social psychology’s view of human nature. Journal of Personality and Social Psychology, 51(3), 515–530. https://doi.org/10.1037/0022-3514.51.3.515 Sears, D. O. (2008). College student-itis redux. Psychological Inquiry, 19(2), 72–77. https://doi.org/10.1080/10478400802050181 Sides, R., Chow, G., & Tenenbaum, G. (2017). Shifts in adaptation: The effects of self- efficacy and task difficulty perception. Journal of Clinical Sport Psychology, 11(1), 34–52. https://doi.org/10.1123/jcsp.2016-0020 Wright, R. R., Perkes, J. L., Schaeffer, C., Woodruff, J. B., Waldrip, K., & Dally, J. L. (2018). Investigating BMI discrepancies in subjective and objective reports among college students. Journal of Human Health Research, 1(1), 106–115. Xu, K. M., Koorn, P., de Koning, B., Skuballa, I. T., Lin, L., Henderikx, M., Marsh, H. W., Sweller, J., & Paas, F. (2020). A growth mindset lowers perceived cognitive load andimproves learning: Integrating motivation to cognitive load. Journal of Educational Psychology, 113(6), 1177–1191. https://doi.org/10.1037/edu0000631 Ye, L., Posada, A., & Liu, Y. (2018). The moderating effects of gender on the relationship between academic stress and academic self-efficacy. International Journal of Stress Management, 25(S1), 56–61. https://doi.org/10.1037/str0000089 Yeager, D. S., & Dweck, C. S. (2012). Mindsets that promote resilience: When

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

111

Academic Perceptions and Cognitive Mindsets | Rullman, D'Anna, Jacobs, Jacobs, and Jellison

students believe that personal characteristics can be developed. Educational Psychologist, 47(4), 302–314. https:// doi.org/10.1080/00461520.2012.722805 Yeager, D. S., & Dweck, C. S. (2020). What can be learned from growth mindset controversies? American Psychologist, 75(9), 1269–1284. https://doi.org/10.1037/amp0000794 Yeager, D. S., Romero, C., Paunesku, D., Hulleman, C. S., Schneider, B., Hinojosa, C., Lee, H. Y., O’Brien, J., Flint, K., Roberts, A., Trott, J., Greene, D., Walton, G.M., & Dweck, C. S. (2016). Using design thinking to improve psychological interventions: The case of the growth mindset during the transition to high school. Journal of Educational Psychology, 108(3), 374–391. https://doi.org/10.1037/edu0000098 Author Note. Kristin P. Rullman https://orcid.org/0000-0003-2450-4900

Kristin P. Rullman, Samantha L. D’Anna, Lauren A. Jacobs, Kristina R. Jacobs, and William A. Jellison, Department of Psychology, Quinnipiac University. The four student authors contributed equally to the project and the final manuscript. This research was supported in part by a student research and experiential learning grant from the College of Arts and Sciences at Quinnipiac University. The authors would like to thank Richard Feinn for his assistance with data analysis. Correspondence concerning this article should be addressed to William A. Jellison, Quinnipiac University, Department of Psychology, 275 Mount Carmel Avenue, Hamden, CT 06518, United States. Email: William.jellison@quinnipiac.edu.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

112

https://doi.org/10.24839/2325-7342.JN27.2.113

The Moderating Effect of Mindfulness on the Relationship Between Problematic Smartphone Usage and Depression, Anxiety, and Stress Christina Stratton, Elizabeth J. Krumrei-Mancuso*, and Cindy Miller-Perrin* Department of Psychology, Pepperdine University

ABSTRACT. The purpose of the current study was to investigate mindfulness as a moderator in the relationship between problematic smartphone usage (PSU) and ratings of depression, anxiety, and stress. Participants were 168 undergraduates from a university in southern California, who completed an online survey measuring smartphone addiction, mental health markers, and mindfulness. The current study also investigated the relationship between one’s objective smartphone screen time and ratings of PSU, depression, anxiety, and stress through Pearson’s product-moment correlations and hierarchical regression analyses. Results indicated that, for individuals with high mindfulness, mindfulness significantly moderated the relationship between PSU and anxiety and stress, but not depression. For those high in mindfulness, higher PSU was associated with higher anxiety (B = 0.11, SE = 0.04, 95% CI [0.02, 0.19], p = .02) and stress (B = 0.12, SE = 0.05, 95% CI [–0.08, 0.09], p < .001), an unexpected finding. The relationships between PSU and both anxiety and stress were nonsignificant for those low in mindfulness (p = .80 in both cases). Among the full sample, more objective screen time was associated with more depression (r = .25, p < .001) but was not linked to PSU (r = .13, p = .13). Implications are discussed, as well as limitations and suggestions for future research. Keywords: problematic smartphone usage, mindfulness, depression, anxiety, stress, screen time, smartphone addiction

martphones are widely used across the world and in the United States. Smartphones are most popular among adults aged 18–29, among whom 96% own a smartphone (Pew Research Center, 2021a). Because of the vast prevalence of smartphone use, academics have become increasingly interested in both the positive and negative implications of smartphone use. Smartphones offer the potential to foster relationships through direct communication, to provide efficiency and productivity benefits, and even offer benefits associated with education and entertainment. Research has suggested that the misuse or overuse of smartphones correlates with various aspects of psychological ill-being (Elhai et al., 2018b; Soni et al., 2017). On the other *Faculty mentor

hand, research on mindfulness applications used on smartphones have found positive effects in encouraging autonomous motivation in which an individual engages in behavior consistent with their internal goals and outcomes (Bauer et al., 2017). Problematic Smartphone Usage Smartphone misuse is not a formal condition recognized by psychologists or established in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), and so the operational definition has varied in the literature. Several studies have referred to smartphone addiction (Demirci et al., 2015; Liu et al., 2018; Soni et al., 2017; Yang et al., 2019). Smartphone users display some commonalities to other types of addictions, however, it is imperative

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

113

Mindfulness and Problematic Smartphone Usage | Stratton, Krumrei-Mancuso, and Miller-Perrin

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

114

to note that the DSM-V characterizes these addictive conditions as “substance use disorders” rather than “addictions.” It may be too radical to refer to the misuse of smartphones as an addiction. Other researchers such as Kim et al. (2015) and Merlo (2013) have discussed the problematic use of mobile phones, known as PUMP, which includes people who are depressed playing with mobile phones rather than interacting with people as well as individuals spending increasing amounts of time on their mobile phones thereby avoiding other types of interactions. Although this term cre ates an easy-to-use acronym, much recent research has been primarily concerned with phones that have smartphone capabilities, such as the usage of internet and social media applications, rather than mobile phone usage, in general and has resulted in meta-analyses (Park et al., 2020; Sohn et al., 2019). Smartphones now dominate the mobile phone industry, especially among American adults (Pew Research Center, 2019). Perhaps the best term and definition to char acterize the misuse of smartphones, in particular, is problematic smartphone usage (PSU), a term used by Elhai et al. (2018a), which we adopted in the current study. We likewise define PSU consistent with Elhai and colleagues’ definition of PSU as high scores on the Smartphone Addiction Scale, a measure that is frequently used among smartphone researchers (Demirci et al., 2015; Elhai et al., 2018a; Samaha & Hawi, 2016; Soni et al., 2017). Other researchers have utilized the Mobile Phone Addiction Index (MPAI; Liu et al., 2018; Yang et al., 2019; Zhang et al., 2020). However, studies using the MPAI have only been conducted using Chinese samples. Regardless of the definitions and terms used, PSU has been correlated with several negative psychological health outcomes. Soni et al. (2017) divided participants into three groups based on Smartphone Addiction Scale scores and found that, among a sample of 587 Indian students, those with excessive smartphone usage demonstrated higher depression, anxiety, and stress compared to those with less smartphone usage. In Turkey, Demirci et al. (2015) studied 319 university students and found depression, anxiety, and sleep quality to all be associated with smartphone overuse. These researchers found that depression and anxiety scores were higher in the group of students with high smartphone usage compared to the low smartphone usage group. Volkmer and Lemer (2019) found similar results in a population of

461 German-speaking adults who were dispro portionately female (71%). The study measured well-being, satisfaction with life, and mindfulness, all of which were negatively correlated with mobile phone usage. However, gender differences were also found suggesting that the relationship between mindfulness, well-being, and mobile phone usage was different between men and women in the sample. These findings point to the importance of controlling for demographic variables, such as gender, in future samples. In the United States, Elhai et al. (2018b) also found that increased smartphone use was related to increased levels of specific types of psychopathol ogy, such as depression symptoms and maladaptive emotion regulation skills among an undergradu ate sample of 68 undergraduate students. This study was notable for its use of a smartphone application called “Moment,” which aimed to measure participants’ objective screen time. These researchers asked students to download the app, which measured their screen time during 1 week. Results indicated that depression and expressive suppression of emotions accounted for significant variability in students’ smartphone use, which was objectively measured by their screen time. Similarly, Liu et al. (2019) conducted a study of nearly 12,000 Chinese adolescents and found an association between the total duration of one’s mobile phone use and users’ depressive symptoms. However, this study relied entirely on self-reported screen time, which asked participants how many hours they spend on their phones each day. This is likely an inaccurate measure of one’s actual smartphone use. Despite the fact that most studies on PSU have been cross sectional and/or correlational, the quantity of research and the overlapping, consistent conclu sions of researchers point to a connection between negative psychological effects and mobile phone use that is well-studied and established. Factors Related to PSU To understand the potential for poor outcomes associated with PSU, some researchers have begun to investigate factors that may mediate or moderate the relationship between PSU and negative psycho logical outcomes. However, different researchers have proposed that the relationship between one’s psychological state and PSU works in different directions. Some studies have suggested that one’s psychological state may work to shape their PSU, despite being cross-sectional. For example, Liu et al. (2019) found that self-control partially mediated

Stratton, Krumrei-Mancuso, and Miller-Perrin | Mindfulness and Problematic Smartphone Usage

the association between perceived stress and mobile phone addiction. Kim et al. (2015) pointed out that face-to-face communication may have a moderating role between the use of mobile phones and the development of problematic use of mobile phones. Other studies have suggested that PSU may shape one’s psychological state. In 2017, Bauer et al. investigated the relationship between smartphone usage and autonomous motivations, defined as engaging in a behavior consistent with a person’s intrinsic goals or desired outcomes and emanat ing from the self. The researchers investigated autonomous motivation and mindfulness among 211 participants, aged 16–42. They found that mindfulness while using instant messaging was positively related to users’ well-being. Further, mindfulness was related to more autonomous motivations for using one’s mobile phone. The relationship between participants’ day-to-day mindfulness and positive affect from instant mes saging was fully mediated by a more autonomous motivation to use instant messaging. These studies are important because they point to underlying risk factors that may be related to smartphone use and have implications for methods to protect against the negative outcomes of PSU. Mindfulness as a Relevant Factor Perhaps one of the most promising factors impact ing PSU and psychological outcomes could be mindfulness, as discussed by Yang et al. (2019). Mindfulness is defined as a receptive state of mind in which one’s attention, informed by a sensitive awareness of what is occurring in the present, simply observes what is taking place (Brown & Ryan, 2003). Yang et al. (2019) pointed out that mindfulness could be a moderating factor with regard to mobile phone addiction and depression and anxiety because of a combination of several theories, including the mindfulness stress-buffering hypothesis and the reperceiving model of mindful ness. The mindfulness stress-buffering hypothesis posits that mindfulness can mitigate the detrimental impacts of stress on mental health outcomes. This hypothesis was described in detail by Creswell and Lindsey (2014), who stated that mindfulness miti gates stress appraisals and reduces stress reactivity responses, and that these stress reduction effects explain how mindfulness affects health outcomes. The second hypothesis, the reperceiving model of mindfulness, states that higher levels of trait mind fulness may make it easier to develop the capability of reperceiving stressful experiences, resulting in

less psychological distress in response to negative experiences. Within the lens of smartphone use, more mindfulness may make it easier for a user to reperceive problematic aspects of their smartphone use (e.g., when they feel impatient or anxious when they are not holding their smartphones). Mindfulness has been studied in relation to mobile phone use, but there are varying perspec tives on the direction of the relationship between mindfulness and mobile phone use; different stud ies have suggested that the relationship operates in different directions. Elhai et al. (2018a) found that mindfulness mediated the relationship between depression and anxiety sensitivity with PSU, mea sured after 1 month. This research, which examined 261 college students, found that depression, anxiety sensitivity, and mindfulness were correlated with each other. This study assessed the undergraduate students two times, 1 month apart, using a model where depression and anxiety predicted mindful ness, in turn predicting severity of PSU after the 1 month interval. The researchers suggested that mindfulness can buffer the impact of anxious and depressive psychopathology on behavioral addic tions, such as smartphone use. This mediation suggests that it is mindfulness that buffers one’s depression and anxiety, therefore reducing PSU. Other research has suggested that this relation ship may work in another direction, where mindful ness works to reduce PSU, thus reducing depression and anxiety. Apaolaza et al. (2019) provided a potential process explanation for the beneficial effect of mindfulness on the stress derived from mobile social media use. The researchers found that mindfulness has a beneficial effect on compulsive smartphone social networking site (SNS) use, stat ing that mindfulness decreases the likelihood of developing problematic mobile social media use. The study, which examined 346 undergraduate students in China, found that stress derived from compulsive mobile SNSs was lower in individuals higher in mindfulness. This relationship was medi ated by positive effects on self-esteem, negative effects on social anxiety, and, in turn, reduced compulsive mobile SNS use. The process presented by Apaolaza et al. would suggest that mindfulness decreases compulsive mobile use. This study used the social media site WhatsApp to measure compul sive use. However, the use of WhatsApp may or may not be considered a social media site in the United States, where social media use primarily includes sites such as Facebook, Instagram, and Twitter, where a defining feature includes posts to public

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

115

Mindfulness and Problematic Smartphone Usage | Stratton, Krumrei-Mancuso, and Miller-Perrin

Method

audiences. These applications offer different uses from the limited direct messaging functioning of applications such as WhatsApp. Because different smartphone applications are different in nature, this study may not be generalizable to a United States population where WhatsApp is used by a much smaller percent of adults than other social media sites (Pew Research Center, 2021b), or to the misuse of smartphones in general. Yang et al. (2019) also found that mindfulness moderated the relationship between mobile phone addiction and its negative effects in a similar way. The study examined 1,258 Chinese high school students and found that the associations between mobile phone addiction and anxiety/depression were weaker among adolescents with higher levels of mindfulness compared to those with lower levels of mindfulness. This study further investigated the relationship suggested by Apaolaza et al. (2019) that mindfulness may lead to less PSU, however, it used a Chinese sample. American and Chinese students likely do not use their smartphones in the same ways. Therefore, it would be valuable to extend this research to examine the effects of mindfulness on PSU in a sample of American undergraduate students. Data from Pew Research (2019) sug gested that smartphones are more popular among adults aged 18–29 than any other age bracket, and that college-educated adults displayed a greater percentage of using smartphones than any other education bracket. This points to a need to study moderating factors of PSU, such as mindfulness, among a sample of U.S. undergraduate students.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

116

Purpose of the Current Study The purpose of the current study was to evaluate the moderating effect of mindfulness between PSU and depression, anxiety, and stress ratings among a sample of US college students. We hypothesized that one’s PSU would be positively correlated with one’s ratings of depression, anxiety, and stress, and negatively correlated with mindfulness (Hypothesis 1). Further, we expected that mindfulness would moderate the associations between PSU and depres sion, anxiety, and stress, such that participants with higher mindfulness scores would demonstrate weaker associations between PSU and depression, anxiety, and stress (Hypothesis 2). In other words, we examined mindfulness as a moderating factor (Baron & Kenny, 1986). We also hypothesized that objective daily average smartphone screen time would be positively correlated with ratings of PSU, and depression, anxiety, and stress (Hypothesis 3).

Participants This study included a convenience sample of 168 undergraduate students enrolled in foundational psychology courses at a small, private, Christian, liberal-arts university in Southern California. Participants ranged in age from 18 to 25 years (Mage = 19.18, SD = 1.36). Participants who did not answer at least 95% of items on the survey were excluded from data analysis (n = 7). To minimize the effects of selection bias and maximize power, four participants who skipped only one item on a scale were retained in the analyses, and missing data were replaced with each participant’s mean value for other responses on the scale.1 Among the final sample, 63.7% identified as women, 35.7% men, 0.6% gender nonconform ing. The ethnicity of participants was similar to the population of the college. The most common race reported was White (54.8%), followed by 13.1% Asian, 11.3% Multiracial, 8.9% Latinx, 7.1% Asian American, and 4.8% African American. Approximately 92% of the sample reported using an Apple iPhone, and 100% of the sample reported using a smartphone in their day-to-day life. Procedure Participants were recruited using an online research participation management system, and data were collected through an online survey. Before beginning the survey, each participant was presented with an informed consent form, and the study was approved by the university’s Institutional Research Review Board (Protocol # 20-04-1351). Participants received one research participation credit toward the four total credits required for their psychology course. Participants completed the survey online, which required approximately 10 minutes. Participants completed the following scales: Smartphone Addiction Scale-Short Version (Kwon et al., 2013), Mindful Attention Awareness Scale (Brown & Ryan, 2003), and the Depression Anxiety Stress Scale-21 (DASS-21; Lovibond & Lovibond, 1995). The questionnaire was partially counterbal anced to control for order effects. Half of the participants were randomly assigned to a survey, which presented the DASS-21 first, followed by the Mindful Attention Awareness Scale, while the other half of participants completed the scales in Analysis was also conducted excluding the four participants, and the significance of results did not change, thus we proceeded with our method. 1

Stratton, Krumrei-Mancuso, and Miller-Perrin | Mindfulness and Problematic Smartphone Usage

the reverse order. The Smartphone Addiction Scale was presented third in both versions of the survey in order to prevent completion of this scale from influ encing responses on the other measures. The final items on the questionnaire included demographic information and items that addressed participants’ daily average smartphone screen time. Participants were presented with directions on how to locate their daily average screen time in their iPhone settings. From there, participants were asked to self-report their daily average smartphone screen time. Next, an item asked participants to upload a screenshot of their daily average smartphone screen time. This item assessed the validity of participants’ survey responses by collecting the average daily screen time measured by their smartphone. Measures Problematic Smartphone Usage Participants’ PSU was assessed using the 10-item Smartphone Addiction Scale-Short Version (Kwon et al., 2013). Example items included, “Based on your current situation, to what extent do you agree with the following statements?” “Using my smart phone longer than I had intended,” “Having my smartphone in my mind even when I am not using it,” and “Feeling impatient and fretful when I am not holding my smartphone.” Items were rated on a Likert-type scale ranging from 1 (strongly disagree) to 6 (strongly agree), which measured health and social impairment from smartphone use, as well as symptoms of tolerance and withdrawal. The total score was calculated by summing the responses to all items and scores can range from 10–60. Higher scores represent greater tendencies toward smart phone addiction. Cronbach’s alpha was reported by the test developers to be .91, using adolescent school-aged children in South Korea (Kwon et al., 2013). Internal consistency reliability for the current sample using Cronbach’s alpha was .83. Dispositional Mindfulness Participants’ dispositional mindfulness was assessed using the 15-item Mindful Attention Awareness Scale (Brown & Ryan, 2003). Items included state ments such as “I find myself doing things without paying attention,” “I snack without being aware that I’m eating,” and “I forget a person’s name almost as soon as I’ve been told it for the first time.” Items were rated on a Likert-type scale ranging from 1 (almost always) to 6 (almost never). Total scores were calculated by computing the mean score for all items, which can range from 1–6, with

higher scores representing higher dispositional mindfulness. Cronbach’s alpha was reported by the test developers at .87, using an adult population (Brown & Ryan, 2003). In the current sample, internal consistency reliability using Cronbach’s alpha was .85. Psychological Functioning The psychological functioning of participants was evaluated using the DASS-21 (Lovibond & Lovibond, 1995). The DASS-21 is a set of three scales designed to measure depression, anxiety, and stress. The instructions asked respondents to indicate “how much each statement applied to you over the past week.” Items were rated on a Likerttype scale, ranging from 0 (did not apply to me at all) to 3 (applied to me very much, or most of the time). The DASS-21 included three subscales with 7 items each. The Depression subscale included items such as “I couldn’t seem to experience any positive feeling at all” and measured components such as dysphoria, hopelessness, devaluation of life, self-deprecation, lack of interest/involvement, anhedonia and inertia. The Anxiety subscale included items such as “I felt I was close to panic” and measured autonomic arousal, skeletal muscle effects, situational anxiety, and subjective experi ence of anxious affect. The Stress subscale included items such as “I found it difficult to relax” and measured difficulty relaxing, nervous arousal, and being easily upset/agitated, irritable/over-reactive and impatient. Scores on the DASS-21 included three separate scores for the three subscales: Depression, Anxiety, and Stress, which were calculated by summing the responses for the 7 applicable items. Internal consistency of the DASS-21 was evaluated by the test authors and was strong: (Depression α = .91; Anxiety α = .84; Stress α = .90; Lovibond & Lovibond, 1995). The DASS-21 Anxiety subscale is generally highly correlated with the Beck Anxiety Scale (r = .81), and the Depression subscale is strongly correlated with the Beck Depression Scale (r = .74), suggesting good validity of these subscales (Lovibond & Lovibond, 1995). In the current sample, internal consistency reliability using Cronbach’s alpha was found to be .88 for DASS-21 Depression, .77 for DASS-21 Anxiety, and .79 for DASS-21 Stress subscales. Demographic Questionnaire and Validity Checks Demographic information was collected including age, sex, ethnicity, and socioeconomic status. Two

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

117

Mindfulness and Problematic Smartphone Usage | Stratton, Krumrei-Mancuso, and Miller-Perrin

items on the demographic questionnaire also asked respondents to indicate if they owned and operated a smartphone: “Do you own a personal smartphone that is used in your day-to-day life?” If this question was answered affirmatively, the next question asked was: “Do you own an iPhone that is used in your day-to-day life?” The last section of the demographic questionnaire concerned participants’ smartphone screen-time report, measured by the smartphone company on every iPhone and Android, the self-report of daily average smartphone screen time, and a validity question. An item instructed respondents as follows: “For the next item, you will be asked to utilize data from your phone’s Screen Time Report. Please go to your phone settings and scroll to ‘Screen Time.’” After participants checked a box indicating that this step had been completed, an item asked respondents: “What is your daily average screen time?” This value is calculated by phone manufacturers by averaging the user’s daily screen time over the past week. Respondents reported this number with an open-ended text response. The next item on the questionnaire requested that respondents upload a screenshot of their screen-time report, which we used to verify the self-reported responses. The final item on the questionnaire asked respondents to indicate whether their responses should be used for research purposes. The item read: “Did you answer each item honest and thoughtfully?” with two options: “Yes, my responses are valid” or “No, my responses should be thrown out.” This item was used to ensure the validity of participants’ responses. TABLE 1 Means, Standard Deviations, and Correlations Between Study Variables Scale 1. PSU

30.08

8.70

2. Depression

6.29

4.69

.28**

3. Anxiety

5.16

3.91

.24**

.50**

4. Stress

8.15

4.14

.27**

.58**

5. Mindfulness

3.61

0.75

.32

−.45

6. Self-report ST

295.79

137.21

.20*

.14

.09 −.09

7. Obi ST

311.31

144.41

.13

.25

.13

.08 −.01

8. ST Diff

−5.35

38.50

.58

−.14

−.01

.66** −.47** −.47**

−.08

.04

.96** .06 −.21*

Note. PSU = problematic smartphone usage; Self-report ST = self-report screen time; Obj ST = objective screen time; ST Diff = difference between objective and self-reported screen time. * p < .05. **p < .001.

118

Results Preliminary Analyses Based on previous literature (Salari et al., 2020), we included gender as a control variable in analyses including measures of stress. Indeed, a t test conducted with the current sample indi cated significant gender differences in ratings of stress (t = 2.22, df = 165, p = .04), with women (M = 8.66, SD = 3.85) reporting more stress than men (M = 7.27, SD = 4.52). Men and women did not differ on ratings of PSU (t = 1.58, df = 165, p = .12), depression (t = 1.33, df = 165, p = .18), anxiety (t = 1.69, df = 165, p = .09), or mindfulness (t = –1.47, df = 165, p = .14). We also examined the relationship between age and our variables of interest. Age was not significantly correlated with PSU (r = –.05, p = .50), depression (r = –.05, p = .56), anxiety (r = .00, p = .98), stress (r = .01, p = .93), or mindfulness (r = –.07, p = .37). Means and standard deviations for all dependent measures are displayed in Table 1. We conducted 10 one-way ANOVAs to determine if group differences existed on the variables of interest based on ethnicity or income. Five one-way ANOVAs found no significant group differences based on ethnicity for ratings of PSU, F(5, 162) = 0.98, p = .43, η 2 = .03, depression, F(5, 162) = 1.21, p = .31, η2 = .04, anxiety, F(5, 162) = 0.37, p = .87, η 2 = .01, stress, F(5, 162) = 0.45, p = .82, η2 = .01, or mindfulness, F(5, 162) = 0.66, p = .66, η2 = .01, scores. A second series of five oneway ANOVAs found no significant group differences based on income bracket for ratings of PSU, F(6, 161) = 0.58, p = .75, η2 = .02, depression, F(6, 161) = 1.62, p = .15, η2 = .06, anxiety, F(6, 161) = 0.65, p = .69, η2 = .02, stress, F(6, 161) = 0.64, p = .70, η2 = .02, or mindfulness, F(6, 161) = 0.20, p = .98, η2 = .01, scores. Primary Analyses Hypothesis 1: Factors Related to PSU To examine the relationships between all variables, we conducted Pearson product moment correla tions. PSU was significantly positively correlated with depression (r = .28, p < .001), anxiety (r = .24, p < .001), and stress (r = .27, p < .001). PSU was significantly negatively correlated with mindfulness (r = –.32, p < .001; see Table 1). Hypothesis 2: Mindfulness as a Moderator Between PSU and Depression, Anxiety, and Stress We conducted three hierarchical regression analyses to examine whether mindfulness would

Stratton, Krumrei-Mancuso, and Miller-Perrin | Mindfulness and Problematic Smartphone Usage

moderate links between PSU and depression, anxiety, and stress. The results of these analyses are displayed in Table 2. We conducted a hierarchical regression analysis to examine the moderating effect of mindfulness in the relationship between PSU and depression scores. In Step 1, we entered PSU, which accounted for 7.6% of the variance in ratings of depression. In the second step, mindfulness was added to the model and accounted for 14.3% of the variance in ratings of depression. In the third step, the interaction between mindfulness and PSU was not a significant predictor of depression scores. We conducted a hierarchical regression analysis to examine the moderating effect of mindfulness in the relationship between PSU and anxiety scores. PSU was entered in Step 1, accounting for 5.6% of the variance in ratings of anxiety. In Step 2, we entered mindfulness, which accounted for 17.1% of the variance in anxiety. In the third and final step, mindfulness significantly moderated links between PSU and anxiety, accounting for 2.0% of the variance. Post-hoc probing demonstrated that there was a positive relationship between PSU and anxiety for individuals high in mindfulness (1 SD above the mean; B = 0.11, SE = 0.04, 95% CI [0.02, 0.19], p = .02), but no relationship between PSU and anxiety for those low in mindfulness (1 SD below the mean; –0.01, SE = 0.04, 95% CI [–0.09, 0.07], p = .80). Results of the post-hoc probing are displayed in Figure 1. We conducted a hierarchical regression analysis to examine the moderating effect of mindfulness in the relationship between PSU and stress scores, while controlling for gender. The control variable of gender was entered in the first step, because previous research and our preliminary analyses found significant differences between men and women on ratings of stress. Gender accounted for 2.3% of the variance in ratings of stress. Step 2 included PSU, accounting for 6.4% of the variance in ratings of stress. Mindfulness accounted for 15.9% of the variance in ratings of stress in Step 3. In Step 4, mindfulness significantly moderated links between PSU and stress, accounting for 1.8% of the variance. While controlling for gender, posthoc probing found that the relationship between PSU and stress was significant for individuals high in mindfulness (1 SD above the mean; B = 0.12, SE = 0.05, 95% CI [0.03, 0.21], p < .001), but not for those low in mindfulness (1 SD below the mean; B = .01, SE = 0.04, 95% CI [–0.08, 0.09], p = .80). Results of the post-hoc probing are displayed in Figure 2.

Hypothesis 3: Screen-Time Analyses Screen-time ratings demonstrated few correlations to the other variables. Objective daily average screen time was not significantly correlated with TABLE 2 Hierarchical Regression Analyses Examining Predictions of Depression, Anxiety, and Stress B(SE)

95%CI

ΔR2

Depression Step 1 PSU

.08** 0.15 (0.04)

0.07, 0.23

.28

Step 2 PSU Mindfulness

.14** 0.08 (0.04)

0.00, 0.16

.15

−2.48 (0.45)

−3.38, −1.59

−.40

0.08 (0.04)

0.00, 0.16

.15

−2.48 (0.45)

−3.37, −1.58

−.40

0.03 (0.05)

−0.06, 0.12

.04

Step 3 PSU Mindfulness PSU*Mindfulness

.00

Anxiety Step 1 PSU

.06** 0.11 (0.03)

0.04, 0.17

.24

Step 2 PSU Mindfulness

.17** 0.05 (0.03)

−0.02, 0.11

.10

−2.27 (0.38)

−3.01, −1.53

−.44

Step 3 PSU Mindfulness PSU*Mindfulness

.02* 0.05 (0.03)

−0.02, 0.11

.11

−2.25 (0.37)

−2.99, −1.51

−.43

0.07 (0.04)

0.00, 0.15

.14

Stress Step 1 Gender

.02 −1.12 (0.57)

−2.25, 0.01

−.15

Step 2 Gender PSU

.06** −0.85 (0.56)

−1.96, 0.25

−.12

0.12 (0.04)

0.05, 0.20

.26

Step 3 Gender PSU Mindfulness

.16** −0.60 (0.51)

−1.61, 0.41

−.08

0.06 (0.03)

−0.01, 0.13

.13

−2.31 (0.39)

−3.09, −1.54

−.43

Step 4 Gender PSU Mindfulness PSU*Mindfulness

.02* −0.49 (0.51)

−1.49, 0.52

−.07

0.06 (0.03)

−0.00, 0.13

.14

−2.30 (0.39)

−3.07, −1.53

−.42

0.08 (0.04)

0.00, 0.16

.14

Note. PSU = problematic smartphone usage; PSU*Mindfulness = interaction between problematic smartphone usage and mindfulness. * p < .05. **p < .001.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

119

Mindfulness and Problematic Smartphone Usage | Stratton, Krumrei-Mancuso, and Miller-Perrin

PSU (r = .13, p = .13), anxiety (r = .13, p = .13), or stress (r = .08, p = .35). Objective daily screen time was significantly positively correlated with depres sion (r = .25, p < .001) and self-report screen time (r = .96, p < .001). The difference between participants’ selfreported screen time and objective daily screen time was calculated by subtracting the self-report screen time from the objective daily screen-time measure, in order to calculate a measure of screentime underestimation. Of the 132 participants who reported both measures of screen time, the mean difference was only –5.35 minutes (SD = 8.50) per day. A paired-samples t test indicated that there was no significant difference between the measure of objective daily screen time (M = 311.31, SE = 12.57) and self-reported screen time (M = 305.96, SE = 12.31), t = –1.60, df = 131, p = .11. FIGURE 1 The Relationship Between Problematic Smartphone Usage and Anxiety, Displayed for High Mindfulness and Low Mindfulness Groups 8.00

Mean Anxiety

6.00

4.00

Mindfulness

2.00

0.00

−1 SD +1 SD −1 SD

+1 SD

Problematic Smartphone Usage

FIGURE 2 The Relationship Between Problematic Smartphone Usage and Stress, Displayed for High Mindfulness and Low Mindfulness Groups 12.00

Mean Stress

10.00 8.00 6.00 4.00

Mindfulness −1 SD +1 SD

2.00

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

120

0.00

−1 SD

Problematic Smartphone Usage

+1 SD

A regression analysis was conducted to investi gate the predictive power of objective daily screen time in predicting ratings of PSU. In the analysis, objective daily screen time was not a significant predictor of PSU, F(1, 131) = 2.36, B = 0.01, CI [0.00, 0.02], p = .13.

Discussion The current study examined the moderating effect of mindfulness in the relationship between PSU and the mental health indicators of depression, anxiety, and stress scores. Factors Related to PSU Our first hypothesis that PSU would be positively associated with depression, anxiety, and stress, and negatively associated with mindfulness was supported. PSU was associated with higher levels of depression, anxiety, and stress, accounting for between 5.6 and 7.6% of the variance in these outcomes. This suggests that, on average, as PSU increased, so did participants’ levels of depression, anxiety, and stress. These findings support previous research on the negative psychological outcomes associated with PSU (Apoalaza et al., 2019; Demirci et. al., 2015; Soni et al., 2017; Yang et al., 2019). However, longitudinal research is needed to establish directionality among these variables. If it is found that PSU leads to mental health concerns, this highlights the importance of developing interventions to address PSU, which may in turn decrease negative effects such as anxiety, depres sion, or stress. We observed a significant negative correlation between PSU and mindfulness. In general, greater levels of PSU were related to lower levels of mind fulness. This supports previous research that has begun to investigate the relationship between PSU and mindfulness (Bauer et al., 2017; Elhai et al., 2018a; Volkmer & Lerner, 2019; Yang et al., 2019; Zhang et al., 2020). It is possible that mindfulness might be a mitigating factor that protects indi viduals from developing PSU. On the other hand, PSU may decrease one’s sense of mindfulness. Longitudinal research is needed to determine the direction of this relationship. Mindfulness as a Moderator Between PSU and Depression, Anxiety, and Stress Our second hypothesis that participants who were higher in mindfulness would display weaker associa tions between PSU and depression, anxiety, and stress than those with lower levels of mindfulness

Stratton, Krumrei-Mancuso, and Miller-Perrin | Mindfulness and Problematic Smartphone Usage

was not supported. In fact, we found the opposite. Mindfulness moderated the relationship between PSU and anxiety and stress, with post-hoc probing indicating that, for those high in mindfulness, more PSU was associated with higher anxiety and stress. For those low in mindfulness, there was no association between PSU and anxiety and stress. Mindfulness did not moderate the relationship between PSU and depression. These findings might be due to the fact that individuals higher in mindfulness could be more mindful of the general impacts of their own PSU, which causes the individual greater anxiety and stress. PSU may be more detrimental for those high in mindfulness because it goes against their typical identity as a person who typically displays high levels of attentive awareness. Those low in mindfulness may not demonstrate a significant link between PSU and anxiety and stress because they are not attentive to their excess PSU or because their smartphone use habits may not clash with their values as it does for the high mindfulness group. The moderating effect of mindfulness on the relationship between PSU and anxiety contradicts previous research by Yang et al. (2019), as the moderation relationships worked in opposite direc tions. Yang et al. (2019) found that the relationship between PSU and anxiety was weaker among those high in mindfulness in a Chinese sample. The cur rent study also failed to replicate a finding of Yang et al. (2019) that mindfulness moderated the rela tionship between PSU and depression. Data from the current study did not support the moderating power of mindfulness in analyses examining depres sion. The differing results between the two studies could be due to the fact that students in a Chinese sample and in a U.S. sample use their smartphones differently. A study by Xu and Mocarski (2014) showed that students’ social media usage reflected their cultural values. It could similarly be true that cultural pressure shapes the way that a person per ceives their excessive smartphone use. Perhaps in American culture, a person’s mindful awareness of their PSU causes greater anxiety because of cultural pressure to minimize screen time. Links Between Screen Time and PSU Our research also investigated the relationship of screen time to PSU. We found that objective daily screen time was not a significant predictor of PSU. This supports findings by Elhai et al. (2018b), who measured smartphone screen time over the course of one week.

The paired-samples t test did not indicate significant differences between objective and selfreported screen time, which implies that overall, the sample was fairly accurate in estimating their screen time. This challenges previous assertions by Elhai et al. (2018b) that self-report has consistently been found to be an inaccurate measure of objec tive screen time. Our study suggests that there are specific methods that can be used to achieve fairly accurate self-report data of screen time. Specifically, in the current study, we directed participants on how to find an objective measure of their daily screen time before self-reporting the value, which is a unique approach from other studies (e.g., Liu et al., 2017). The overall findings of the current study regarding screen time suggest that what matters more in developing PSU is not how much time an individual uses their phone, but rather, how the indi vidual uses their phone. There are likely individual differences in how people use their phones, such that some people may have high screen time but may not demonstrate PSU due to the nature of their smartphone usage. Perhaps some people engage in self-improvement or meaningful relational activities through their smartphones, or work or engage in school activities from their smartphones, all of which may increase their screen time without increasing PSU. Limitations The current findings were likely influenced by the study’s timing, which occurred during the COVID-19 lockdown. The university where data were collected had transitioned from in-person classes to online classes prior to data collection, thus all participants were taking online classes. It is possible that at least some participants in the study had greater objective screen time during the study period than during normal circumstances due to circumstances of the lockdown, which required greater reliance on technology. Some participants might have relied on their smartphones for online classes, which would add hours of screen time each day that would not have been the case dur ing normal circumstances. Outside of lockdown, when classes and work responsibilities take place in-person and not online, the time spent on daily responsibilities (such as work and school) would be separate from screen time, as opposed to the cur rent circumstances which may require smartphone use for those same responsibilities. In addition to increasing one’s total screen time, it is also likely that participants were using

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

121

Mindfulness and Problematic Smartphone Usage | Stratton, Krumrei-Mancuso, and Miller-Perrin

their smartphones in generally different ways dur ing these circumstances than before the COVID-19 pandemic. Individuals likely became more reliant on their smartphones for not only work or school, but also connectedness, entertainment, and productivity. This difference would be especially prevalent among a college student population. The primary limitation then is the vast variability in what students are doing on their smartphones that is classified as problematic. Is a student using their smartphone problematically to post on social media? Or is that student online shopping, gaming, or messaging? Smartphone use is both subjective and diverse. Limitations also surround the generalizability of the current sample, all of whom attended the same private, Christian university. It is possible that the university’s generally safe campus, religious affiliations, and temperate climate each play a role in shaping not only students’ general smartphone usage but also mental health correlates. Because of how many factors can shape students’ mental health ratings, we cannot assume the observed relationships between mindfulness, smartphone usage, and depression, anxiety, or stress would generalize to other populations.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

122

Future Research Future research should consider replicating the current study during normal circumstances, outside of the COVID-19 pandemic, to investigate the gen eralizability of the results. Research should aim to evaluate a sample with greater demographic diver sity, in terms of both ethnicity and socioeconomic status. Research may benefit from a more specific measure of objective screen time that would offer insights beyond the total average minutes spent on one’s phone each day. For example, an objective measure of screen time that indicates the time spent within each smartphone application, such as email, video chat, and social media platforms could help provide valuable insight into the relationship between screen-time ratings and PSU, as well as its negative implications. Overall, future research should consider the mechanisms behind PSU in order to further the theoretical and conceptual understanding of the condition. This insight may give way to more precise definitions and measures of PSU. The current study is unique in the finding that PSU might lead to greater anxiety and stress among individuals high in mindfulness, but not among those low in mindfulness. It is likely, though, that

the relationship between PSU and anxiety and stress is bidirectional, as evidenced by existing research (Elhai et al., 2018b). Future research should con sider an experimental approach to provide greater insight into causality.

References Apaolaza, V., Hartmann, P., D’Souza, C., & Gilsanz, A. (2019). Mindfulness, compulsive mobile social media use, and derived stress: The mediating roles of self-esteem and social anxiety. Cyberpsychology, Behavior, and Social Networking, 22(6), 388–396. https://doi.org/10.1089/cyber.2018.0681 Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. https://doi.org/10.1037//0022-3514.51.6.1173 Bauer, A. A., Loy, L. S., Masur, P. K., & Schneider, F. M. (2017). Mindful instant messaging: Mindfulness and autonomous motivation as predictors of well-being in smartphone communication. Journal of Media Psychology: Theories, Methods, and Applications, 29(3), 159–165. https://doi.org/10.1027/1864-1105/a000225 Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84(4), 822–848. https://doi.org/10.1037/0022-3514.84.4.822 Creswell, J. D., & Lindsay, E. K. (2014). How does mindfulness training affect health? A mindfulness stress buffering account. Current Directions in Psychological Science, 23(6), 401–407. https://doi.org/10.1177%2F0963721414547415 Demirci, K., Akgonul, M., & Akpinar, A. (2015). Relationship of smartphone use severity with sleep quality, depression, and anxiety in university students. Journal of Behavioral Addictions, 4(2), 85–92. https://doi.org/10.1556/2006.4.2015.010 Elhai, J. D., Levine, J. C., O’Brien, K. D., & Armour, C. (2018a). Distress tolerance and mindfulness mediate relations between depression and anxiety sensitivity with problematic smartphone use. Computers in Human Behavior, 84, 477–484. https://doi.org/10.1016/j.chb.2018.03.026 Elhai, J. D., Tiamiyu, M. F., Weeks, J. W., Levine, J. C., Picard, K. J., & Hall, B. J. (2018b). Depression and emotion regulation predict objective smartphone use measured over one week. Personality and Individual Differences, 133, 21–28. https://doi.org/10.1016/j.paid.2017.04.051 Kim, J. H., Seo, M., & David, P. (2015). Alleviating depression only to become problematic mobile phone users: Can face-to-face communication be the antidote? Computers in Human Behavior, 51, 440–447. https://doi.org/10.1016/j.chb.2015.05.030 Kwon, M., Kim, D. J., Cho, H., & Yang, S., (2013). The Smartphone Addiction Scale: Development and validation of a short version for adolescents. PLoS ONE, 8(12), e83558. https://doi.org/10.1371/journal.pone.0083558 Liu, Q.-Q., Zhang, D.-J., Yang, X.-J., Zhang, C.-Y., Fan, C.-Y., & Zhou, Z.-K. (2018). Perceived stress and mobile phone addiction in Chinese adolescents: A moderated mediation model. Computers in Human Behavior, 87, 247–253. https://doi.org/10.1016/j.chb.2018.06.006 Liu, J., Liu, C. X., Wu, T., Liu, B.-P., Jia, C.-X., & Liu, X. (2019). Prolonged mobile phone use is associated with depressive symptoms in Chinese adolescents. Journal of Affective Disorders, 259, 128–134. https://doi.org/10.1016/j.jad.2019.08.017 Lovibond, S. H., & Lovibond, P. F. (1995). Manual for the depression anxiety & stress scales. (2nd Ed.) Psychology Foundation. Merlo, L., Stone, A., & Bibbey, A. (2013). Measuring problematic mobile phone use: Development and preliminary psychometric properties of the PUMP scale. Journal of Addiction, 2013, 1–7. https://doi.org/10.1155/2013/912807 Park, C., Zhu, J., Ho Chun Man, R., Rosenblat, J. D., Iacobucci, M., Gill, H., Mansur, R. B., & McIntyre, R. S. (2020). Smartphone applications for the treatment of depressive symptoms: A meta-analysis and qualitative review. Annals of Clinical Psychiatry, 32(1), 48–68. Pew Research Center. (2021a, April 7). Mobile fact sheet. https://www.pewresearch.org/internet/fact-sheet/mobile/ Pew Research Center. (2021b, April 7). Social media fact sheet. https://www.pewresearch.org/internet/fact-sheet/social-media/ Salari, N., Hosseinian-Far, A., Jalali, R., Vaisi-Raygani, A., Rasoulpoor, S., Mohammadi, M., Rasoulpoor, S., & Khaledi-Paveh, B. (2020). Prevalence of stress, anxiety, depression among the general population during the

Stratton, Krumrei-Mancuso, and Miller-Perrin | Mindfulness and Problematic Smartphone Usage

COVID-19 pandemic: A systematic review and meta-analysis. Globalization and Health, 16, 57. https://doi.org/10.1186/s12992-020-00589-w Samaha, M., & Hawi, N. S. (2016). Relationships among smartphone addiction, stress, academic performance, and satisfaction with life. Computers in Human Behavior, 57, 321–325. https://doi.org/10.1016/j.chb.2015.12.045 Sohn, S. Y., Rees, P., Wildridge, B., Kalk, N. J., & Carter, B. (2019). Prevalence of problematic smartphone usage and associated mental health outcomes amongst children and young people: A systematic review, meta-analysis and GRADE of the evidence. BMC Psychiatry, 19(1), 356. https://doi.org/10.1186/s12888-019-2350-x Soni, R., Upadhyay, R., & Jain, M. (2017). Prevalence of smartphone addiction, sleep quality and associated behaviour problems in adolescents. International Journal of Research in Medical Sciences, 5(2), 515–519. https://dx.doi.org/10.18203/2320-6012.ijrms20170142 Volkmer, S. A., & Lermer, E. (2019). Unhappy and addicted to your phone?— Higher mobile phone use is associated with lower well-being. Computers in Human Behavior, 93, 210–218. https://doi.org/10.1016/j.chb.2018.12.015 Xu, Q., & Mocarski, R. (2014). A cross-cultural comparison of domestic American and international Chinese students’ social media usage. Journal of

International Students, 4(4), 374–388. https://doi.org/10.32674/jis.v4i4.456 Yang, X., Zhou, Z., Liu, Q., & Fan., C. (2019). Mobile phone addiction and adolescents’ anxiety and depression: The moderating role of mindfulness. Journal of Child and Family Studies, 28(3), 822–830. https://doi.org/10.1007/s10826-018-01323-2 Zhang, Y., Lv, S., Li, C., Xiong, Y., Zhou, C., Li, X., & Ye, M. (2020). Smartphone use disorder and future time perspective of college students: The mediating role of depression and moderating role of mindfulness. Child and Adolescent Psychiatry and Mental Health, 14(1), 1–11. https://doi.org/10.1186/s13034-020-0309-9 Author Note. Elizabeth J. Krumrei-Mancuso https://orcid.org/0000-0001-6151-7845 Cindy Miller-Perrin https://orcid.org/0000-0002-0093-8037 We have no known conflict of interest to disclose. Correspondence concerning this article should be addressed to Christina Stratton, Social Science Division, Pepperdine University, 24255 Pacific Coast Hwy, Malibu, CA 90263, United States. Email: christina.stratton@pepperdine.edu

123

https://doi.org/10.24839/2325-7342.JN27.2.124

Examining the Efficacy of Using a Change Blindness Framework as a Novel Social Media Intervention Stephanie Misko, Olivia Hays, and Laura Getz* Department of Psychological Sciences, University of San Diego

ABSTRACT. Exposure to idealized, appearance-focused images on social media has been found to be damaging to young women’s body image and self-esteem (Groesz et al., 2001). The goal of the current study was to examine the efficacy of a novel intervention that could serve as a buffer to idealized content, thereby reducing the amount of physical appearance comparisons made by young persons on social media. The intervention consisted of a single disclaimer that informed participants about the difficulty in detecting edited photos from a change blindness framework. Participants (N = 46) were randomly assigned to view either the experimental or control disclaimer before being shown 10 image pairs that depicted a single college-aged woman wearing a bikini. In 5 of the pairs, the second image was edited to reflect the slight changes social media users make to achieve a slimmer look. We found that women who were shown the experimental disclaimer and edited image pairs (M = 3.71, SD = 1.27) more accurately detected changes than those shown the control disclaimer (M = 2.77, SD = 1.11, p < .001). Results suggest that the disclaimer informed women about photo-editing practices, and this change in awareness led to them more accurately detecting changes in edited image pairs. However, we found no effect between disclaimer conditions on physical appearance comparisons. The study’s primary limitation was that the experimental disclaimer functioned as a brief, single-exposure intervention, and thus, more in-depth interventions aimed at informing young persons about their media consumption should be designed and tested. Keywords: change blindness, social media, intervention

M SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

124

any researchers have examined the impact of media exposure on young adults’ mental health. Several studies have examined how particular aspects of traditional media (i.e., magazines, television), such as the inclusion of idealized, appearance-focused photos, impact women’s body image (Silverstein et al., 1986; Tiggemann & Polivy, 2010). Exposure to thin-ideal content in traditional media leads to internalization of the ideal, which in turn greatly increases body dissatisfaction (Groesz et al., 2002; Stice & Shaw, 1994; Stice et al., 1994). In addition to findings from traditional media research, social comparison processes can be elicited from viewing idealized content on these

online social media platforms such as Instagram (Fardouly et al., 2017). Social comparison theory posits that, when people compare themselves, they tend to identify a discrepancy between themselves and the other, which can subsequently alter their self-perceptions (Festinger, 1954). Upward social comparisons are defined as the process in which individuals compare themselves to someone who they deem to be superior on a particular dimension (e.g., intelligence, physical appearance). Those who engage in this type of social comparison may perceive a discrepancy between themselves and the other person and interpret this as an inadequacy on their part. The social comparisons that are specific to one’s physical appearance are known as body

*Faculty mentor

Misko, Hays, and Getz | Change Blindness Social Media Intervention

or appearance comparisons. Engaging in such comparisons has been found to impact body image in young women negatively (Choukas-Bradley et al., 2019; Fardouly et al., 2017). Individuals who report making frequent appearance comparisons also report having intense cognitions related to disordered eating (Fitzsimmons et al., 2016). Although people typically make social compari sons in person, most now have the ability to com pare themselves to thousands of others online with only a few taps on their mobile device (Fardouly et al., 2017). The most popular picture-sharing platform, Instagram, allows social media users to share photos or videos to a large audience and to receive feedback through likes and comments (Clement, 2020). The platform varies widely in content; however, a great deal is focused on physical appearance, and images often depict individuals looking physically flawless. Although Instagram and other social media platforms can be used to cultivate connections with peers and explore interests, young adults have greater accessibility to idealized content and space to engage in appearance comparisons with peers, celebrities, and strangers. As a result of being able to visually compare themselves to others in public and private, there are more opportunities to make more upward social comparisons, which is an underlying mechanism that has been found to facilitate body dissatisfaction (Rodgers et al., 2015; Tiggemann & Polivy, 2010). Brown and Tiggemann (2016) demonstrated how influential these com parison processes can be to both mood and body image, finding that women who viewed pictures of attractive peers were more likely to report expe riencing negative mood and body dissatisfaction than those who viewed travel pictures. Research by Hogue and Mills (2019) additionally supports the impact idealized media has on psychological functioning, and their findings show that compar ing oneself to a known attractive peer, rather than a family member, resulted in increased negative feelings toward one’s body. Moreover, exposure to idealized social media images has been found to decrease self-perceived attractiveness in young women (Sherlock & Wagstaff, 2019). Given these adverse effects, researchers have employed a wide range of intervention methods, including utilizing disclaimers, warning labels, and social media literacy exercises, as ways to mitigate the effects associated with exposure to idealized content (Halliwell et al., 2011; Livingston et al., 2020). The adoption of content disclaimers

(i.e., warnings or labels) in particular is a widely explored, brief intervention method. Borau and Nepomuceno (2019) examined the impact that novel content disclaimers informing individuals about airbrushing practices had on body satisfac tion. In their study, female participants viewed images of female models in beauty magazines who represented the thin ideal, with an experimental group receiving an airbrushing “disclaimer.” Borau and Nepomuceno (2019) found that participant knowledge of airbrush usage did not alleviate personal body dissatisfaction, which did not sup port their hypothesis. Similarly, Livingston et al. (2020) adapted intervention methodology used for traditional media by investigating the effect of having social media influencers convey the reality of idealized Instagram photos through their own cap tions. Their findings also suggest that first-person disclaimer captions are as ineffective at preserving body image as content disclaimers. Although initially promising, content disclaimers have been shown to be inefficacious at protecting young women from experiencing body dissatisfaction (Danthinne et al., 2020; McComb & Mills, 2020;). Media literacy videos have also been a tested intervention method aimed at improving body image and mental health outcomes in young media consumers. Halliwell et al. (2011) found that showing young girls a video of a one-minute media literacy message on the prevalence of Photoshop prior to viewing thin-ideal content, reduced nega tive feelings related to body image. Expanding upon this work, Arendt and colleagues (2017) found a reduction in participant motivation to engage in upward social comparison processes after expos ing their participants to a media awareness video. This intervention engendered more realistic selfperceptions of body image within their participants, which seemed to lessen upward social comparison tendencies in participants. Although researchers are seeing promising findings for extensive media literacy interventions, these interventions can be time consuming to develop and costly to administer. More extensive research is therefore needed to develop effective disclaimers that can briefly inform social media users about idealized content and act as a body image preservation tool. Because social media users may manipulate the images they post on platforms slightly, one brief intervention method that may be effective at inhibiting social comparison processes is informing participants about photo-editing practices and the concept of change blindness. To our knowledge,

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

125

Change Blindness Social Media Intervention | Misko, Hays, and Getz

no other researchers have used content disclaim ers to inform participants about photo-editing practices by explaining the concept of change blindness. Change blindness is a common cognitive phenomenon, which occurs when someone fails to perceive a change in visual stimuli (Pashler, 1988). Interestingly, even when people are aware that they will view changed images, they overestimate their ability to detect those changes (Levin et al., 2000). Participants viewed four different scenarios, each depicted by a series of still images, and imagined them as a movie. Participants were instructed to spot changes between the images and report their confidence in performing the task correctly. Although 83% predicted that they would detect the changes, less than 15% of participants were able to accurately detect changes. The findings from Levin et al. (2000) suggest that social media users may be more confident, rather than accurate, at detecting edited photos posted on platforms. Therefore, those viewing edited appearance-focused images may be unable to detect minute changes that can be made using photo-editing apps, leaving them to perceive a person’s physical appearance to be realistic. Those who do not detect that an appearance-focused photo was edited may think that the pictured physical appearance is attainable and may be more likely to engage in upward social comparison processes (Tiggemann & Polivy, 2010). For the current study, we aimed to design a novel intervention that incorporated the concept of change blindness. We informed participants about the concept of change blindness in an effort to highlight most people’s failure to detect whether a photo has been edited. We predicted that participants informed about photo-editing change blindness before viewing the sequence of images would better detect the changes than those who did not learn of change blindness. Additionally, we predicted that those informed about change blindness would be less likely to engage in physical appearance comparisons.

Method

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

126

Participants This study was approved by the University of San Diego’s (USD) Institutional Review Board. Participants (N = 46) identified as women and recruited from an upper division psychology course using convenience sampling methods. The exclusion criterion included all gender identities except for women; this criterion was set to enhance social comparison processes. Participants were

informed they would receive course credit for participating in this study and given a cover story that the study’s aim was to explore the factors that influence popularity on Instagram. The average age of the participants was 20.57 years (SD = 1.4, range = 19–27). Participants identified as White (n = 34, 70.8%), Asian or Asian American (n = 5, 10.4%), and Black or African American (n = 3, 6.3%), and some participants identified as a race that was not listed (n = 4, 8.3%). Additionally, 68.8% (n = 33) identified their ethnicity as non-Latino or nonHispanic, and most participants (n = 41, 89.1%) reported being active Instagram users. Materials Content Disclaimers Two content disclaimers, an experimental and control, appeared prior to the appearance-focused images (see Appendix A). The purpose of the experimental content disclaimer was to inform participants about the frequent photo-editing practices of social media users and how this relates to the concept of change blindness. The control disclaimer only contained task instructions, which were the same instructions given in the last two sentences of the experimental disclaimer. Images A total of 10 appearance-focused Instagram pictures featuring a woman wearing a bikini were selected from public Instagram accounts, and each photo was seen by the participant twice. The women featured in the images were European American, African American, Asian American, and Latin American. We did not receive permission from the Instagram users to include the publicly posted images in this manuscript, and thus, we have not included them in the appendix. A random number generator was used, while controlling for racial representativeness, to randomly select five images to be edited. We then utilized an application called Bodytune to digitally manipulate those five women’s bodies to appear more like the thin ideal (i.e., reduced waist size, expanded hip width). Change Blindness To account for change detection, or the lack thereof, and the accuracy with which one could detect what was changed about the photo, two items were included in the questionnaire. Items included “Did you detect any changes in the second photo?” and “Where on the body did this change occur?” The second item appeared only if the respondent

Misko, Hays, and Getz | Change Blindness Social Media Intervention

selected “yes” to the first item. These items were presented below the second-appearance focused image in each pair. Popularity Two items related to popularity were used to align with the cover story that participants were presented with at the beginning of the study. Items included, “How many likes do you think this photo received on Instagram?” and “Please indicate your percep tion of the pictured woman’s popularity (0 = not popular at all, 10 = most popular).” These items were not included in the analysis. Participants responded to items regarding popularity after responding to change blindness items. Physical Appearance Comparison Scale-3 (Modified) A modified version of the 20-item Physical Appearance Comparison Scale-3 (PACS-3; Schaefer & Thompson, 2018) was used in this study. We modified the PACS-3 in order to better examine appearance comparisons made on social media rather than traditional media (see Appendix B). This version contained 10 items that asked partici pants how they feel about comparing themselves to peers, influencers, and celebrities who appear on television, social media, and in the public. The following are examples of the items included in the scale: (a) “When I’m out in public, I compare my weight/shape to the weight/shape of others”; (b) “When I see a USD female peer on social media, I compare my weight/shape to her”; and (c) “When I watch television, I compare my weight/shape to the weight/shape of the actors/actresses.” The COVID-19 pandemic has caused many people to quarantine at home and therefore spend less time in public spaces. The original PACS-3 scale contained six items related to comparing one’s physical appearance to public others, so most of these items were omitted from the scale due to irrel evance; the modified version contained only two items related to in-person social comparisons. The following were included in our measure: (a) “When I’m out in public, I compare my weight/shape to the weight/shape of others”; and (b) “When I see a female influencer on social media, I compare my weight/shape to her weight/shape.” Participants responded to all items on a 5-point scale ranging from 1 (always, much better, or extremely good) to 5 (never, much worse, or extremely bad). Schaefer (2017) reported alphas of .85 and higher for each subscale of the PACS-3 when measuring female participants’ responses. In the present study, the

following alpha reliability coefficients were found for the Comparison Amount, Comparison Type, and Comparison Outcome subscales, respectively: .89, .70, and .74. Procedure Participants accessed the online survey via an anonymous link. An online informed consent form was presented first, and participants electronically signed to indicate their consent. Participants then responded to a single item that asked about their gender identity as the inclusion criteria for this study included women. Participants who identi fied as women were shown a brief description that included the study’s cover story (i.e., researchers are investigating factors that influence popularity on Instagram); participants were deceived from the true purpose of the study to reduce the likelihood of participant bias impacting the results. Following this, participants were randomly assigned to be shown either experimental or control disclaimers. The experimental disclaimer explains the concept of change blindness, stating that people typically fail to see a visual change between two images. All participants were presented with a ran domized series of 10 idealized image pairs depicting women wearing bikinis. Half of these image pairs featured an edited second photo, and the other half remained unedited. After viewing each image pair, participants answered four questions, two measuring accuracy of change detection and two assessing perceptions of the pictured woman’s popularity. After viewing the entire paired-image set, participants completed 10 items from the modified version of the PACS-3. Six demographic questionnaire items gathered information about participants’ class level, age, race, ethnicity, and social media usage. At the end of the experiment, participants were debriefed about the study’s true aim. Design A 2 x 2 mixed experimental design was utilized for this study. Participants were randomly assigned to be shown an experimental or control disclaimer; thus, the type of disclaimer was a between-subjects independent variable. Photo condition was a within-participant measure, with five of the images edited, and the rest remained unedited. Accuracy in detecting editing was the main dependent variable and was measured with one forced-choice response item that asked participants to indicate whether the photo had been edited. Participants’ accuracy in

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

127

Change Blindness Social Media Intervention | Misko, Hays, and Getz

detecting editing scores were reported out of five points for both edited and unedited photos. Participant engagement in physical comparison processes was measured using a modified version of the PACS-3 (see Appendix B), which was broken into three categories: Comparison Amount (items 1, 4, 7, 8), Comparison Type (items 2, 5, 9), and Comparison Outcome (items 3, 6, 10). The first category was a frequency measure that assessed how often participants compared their physical appearance to others. The following category, com parison type, measured whether the comparison participants engaged in was an upward or down ward comparison. The final category measured the impact that engaging in a physical appearance comparison had on the participants. Specifically, participants were asked to report their feelings after engaging in a certain comparison.

Results A 2 (experimental vs. control disclaimer) x 2 (edited vs. unedited image-pairs) repeated-measures ANOVA was conducted. A main effect was found for the type of photo, either edited or unedited, F(1, 44) = 14.59, p < .001, Cohen’s d = 1.39, such that people were more accurate in detecting unedited photos (M = 4.11, SD = 0.82) than edited photos (M = 3.26, SD = 1.27). There was no main effect for the disclaimer condition between experimental and control groups, F(1, 44) = 2.49, p = .12, η2p = .05; however, a significant interaction (see Figure 1) was observed between type of photo and disclaimer con dition on change detection scores, F(1, 44) = 7.45, FIGURE 1 Interaction Between Experimental Disclaimer and Ability to Detect Photo Editing Disclaimer Condition

Change Detection Accuracy

Control Disclaimer Experimental Disclaimer

4 3 2 1 0

Edited

Unedited

Type of Paired Image SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

128

Note. A significant interaction, such that the group that viewed the experimental disclaimer was more accurate at identifying when an image had been retouched, was observed between the disclaimer condition and type of paired image on change detection accuracy. Error bars represent 95% confidence intervals.

p = .01, η2p = .15. In other words, there was no difference between the group that received the experimental (M = 3.96, SD = 0.86) and control (M = 4.27, SD = 0.77) disclaimer on change detec tion accuracy for unedited image pairs; however, participants who viewed the experimental dis claimer (M = 3.71, SD = 1.27) were more accurate than those who viewed the control disclaimer (M = 2.77, SD = 1.11) in detecting change when an image pair had been edited. An independent-samples t test demonstrated that participants who viewed the experimental disclaimer (M = 10.67, SD = 4.54) were not sig nificantly different in the amount of physical appearance comparisons than those in the control (M = 12.95, SD = 3.96) as was measured by the PACS-3 Comparison Amount scores, t(44) = 1.82, p = .59, Cohen’s d = 0.05 (Refer back to Design sec tion for PACS-3 subscales). Another independentsamples t test indicated that participants who received the experimental disclaimer (M = 11.46, SD = 1.79) did not make different types of com parisons (i.e., upward or downward) than those who received the control disclaimer (M = 11.73, SD = 1.58) as was measured by the PACS-3 Comparison Type scores, t(44) = 0.54, p = .59, Cohen’s d = 0.16. A third independent samples t test showed that participants receiving the experimental disclaimer (M = 11.46, SD = 1.64) did not have dif ferent fluctuations in their mood (i.e., “When you make these comparisons, how does it usually make you feel?”) after making specific comparisons than participants who received the control disclaimer (M = 11.45, SD = 1.79) as measured by the PACS-3 Comparison Outcome scores, t(44) = –0.01, p = .99, Cohen’s d < 0.01.

Discussion The current experimental study investigated the impact of a brief social media intervention on change detection and physical appearance comparison processes. The data supported our first hypothesis that exposure to an experimental disclaimer would improve the ability to detect changes in images. A significant interaction revealed that, when the images were changed, the experimental disclaimer group scored better on detecting changes in edited photos when compared to the control. From this finding, it appears that the experimental disclaimer successfully informed women about photo-editing practices. This suggests that increasing awareness of distorted media among young people may have led them to question the

Misko, Hays, and Getz | Change Blindness Social Media Intervention

realistic nature of images, and this skepticism contributed to greater accuracy in detecting edits between images. The main effect of photo change condition indicated that participants were more accurate at detecting whether the images had been edited when the images were unedited rather than edited. The results mentioned above supported prior research, which demonstrated that change blindness occurs when viewing visually discrepant images, but that people are largely unaware of this cognitive shortcoming. Contrary to our second hypothesis, the analysis of the modified PACS showed that there was no effect of disclaimer condition on comparison pro cesses (Schaefer and Thompson, 2018). Therefore, participants made more upward comparisons when informed about change blindness and how it relates to photo-editing practices compared to the group that received no intervention. From this, it appears that the experimental disclaimer provided no buffer to viewing idealized, appearance focused images, and therefore did not influence young women to reduce the amount of appearance comparison processes that they engage in. Although this finding did not support our prediction, it contributes to the growing evidence that content disclaimers may be ineffective at reducing upward social comparison processes (Danthinne et al., 2020; Kleemans et al., 2016; McComb & Mills, 2020). Despite these findings, social media platforms have started to employ similar brief interventions that act as a buf fer to media content that is defined as disturbing; however, there are no findings on whether this sensitive content warning has been effective. The transient and low-intensity nature of the intervention might have been the primary reason that there was no reduction in social comparison processes. Although content disclaimers may be cost and time effective, lengthier intervention methods, such as actively watching a short video, have been shown to be more effective at mitigating adverse effects of media exposure (Halliwell et al., 2011). Impressionable adolescents and young adults internalize the idealized appearance-focus content seen in media as realistic, which can lead to skewed perceptions about one’s own body. Interventions that focus on promoting content that features diverse body types may help to change young adults’ perception of “realistic” (Kleemans et al., 2016). A study conducted by Ogden et al. (2020) found that showing women participants images of diverse body types positively impacted their overall self-reported body positivity, whereas

the women who viewed thin-ideal images reported adverse body image outcomes, demonstrating the impacts that the thin-ideal beauty standard has on young adults’ body image. To reduce negative appearance outcomes, researchers have employed intervention videos on media literacy and body positive content, which have reduced negative appearance outcomes and increased overall mood and body image (Cohen et al., 2019; Tiggemann & Polivy, 2010). Experimental evidence from Halliwell et al. (2011) demonstrates that participants who received a warning interven tion reported increased body esteem after viewing images of thin models compared to individuals who did not receive the intervention. In addition to this, Halliwell (2013) found that individuals who report high levels of body appreciation were less negatively impacted by idealized, appearance-focused images. Interventions that foster individual cultivation of body appreciation and positivity may serve longer lasting positive effects on body image and should be designed and tested as a replacement to brief preventative methods, such as content disclaimers, which have consistently shown to be ineffective at preserving body image (McComb & Mills, 2020). Concerning the study’s methodology, the failure to collect data on participants’ visual attention during the change blindness task was an additional weakness in design. Andrighetto et al. (2019) explored how visual attention toward dif ferent body parts may influence change blindness or change detection and found that individuals were more likely to accurately assess changes of body parts that are typically sexualized by the media. Interestingly, the participants in our study who reported that the photo had been changed reported body parts such as the hips, legs, but tocks, and chest being changed, with the waist and stomach region being identified the most by participants. This aligns with Andrighetto et al. (2019) findings that individuals are more attuned to changes made to characteristically sexualized body parts. Eye-tracking software can help gauge participants’ visual attention and shed light on how attention, or lack thereof, impacts change detection or appearance comparison processes. Cho and Lee (2013) used similar software to explore attentional biases in photos and found those experiencing high levels of body dissatisfaction to be more visually drawn toward thin-ideal content, suggesting more frequent opportunities to engage in more upward social comparisons and feel more inadequacy about one’s body. This attentional bias to the thin ideal

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

129

Change Blindness Social Media Intervention | Misko, Hays, and Getz

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

130

may uniquely impact social media users’ feelings and behaviors related to their body image due to the high amount of appearance-focused content on picture sharing platforms. Additionally, some young adults might have read the experimental disclaimer, but chose to disregard the information that was provided due to reactance (Brehm, 1966). The eye-tracking software could elucidate whether an adverse reaction to the disclaimer occurred by providing valuable information regarding partici pant visual engagement. Although participants viewed pictures that were selected from real public Instagram accounts in this study, the overall experimental design had low ecological validity. Participants were instructed to view paired images that were displayed successively on Qualtrics. Each photo was displayed alone on a screen for 10 seconds before its photo pair was presented for an equivalent time interval. This methodology differed from previous studies that used the one shot and flicker paradigm to examine change blindness (Andrighetto et al., 2019; Bracco & Chiorri, 2008; Rensink et al., 1997; Simons, 1996). This novel approach to detecting change blindness is not an inherent design flaw; however, it is important to note the difference in approaches. The reason we conclude that our study has low generalizability across settings is that, on picture sharing platforms such as Instagram, photos are not viewed separately, with one image per screen, but are rather viewed singularly as users scroll through a feed. Although Instagram now enables users to swipe through multiple photos included on the same post, it is uncommon for individuals posting edited images to include the original, unedited photo as their first photo. Therefore, our study requires further replication to better understand the effectiveness of using disclaimers that aim to improve social media users’ experience. Additionally, utilizing scales that measure body dis satisfaction and mood in participants upon viewing idealized images may further aid our understand ing of brief interventions’ impact on young social media users’ mental health. Future studies should also focus on expanding data collection with social media users, who identify differently than White, female undergraduates to gather more robust insights since generalizability of results was limited due to our participants being pooled from a predominantly White, private university. Researchers should also examine body part location as a measure of change detection accuracy. Prior research has supported that images with

sexualized body parts (e.g., waist, hips) promote greater attention to these areas, which thereby leads to greater change detection accuracy for edits made in those areas of the body (Andrighetto et al., 2019). The selected images used in this study contained a woman wearing a two-piece swimsuit, and thus leads us to conclude that the participants’ attentional biases may be important to understand change blindness within specific contexts. Our research contributes to the pre-existing knowledge about interventions aimed at lessening the effects that mass media can have on young adult body image. The primary implication this research has for the general public is that the findings can be taken into consideration when designing similar content disclaimers for social media platforms. The media literacy that is provided from a brief interven tion can aid users in recognizing edited content that might have first been assessed as realistic by the user, and consequently help to optimize social media users’ experience and well-being. Tamplin et al. (2018) found that individuals with low social media literacy skills, or those that do not possess a combination of critical thinking and realism skepti cism when viewing media content, experienced a decrease in body satisfaction after viewing idealized content. This finding suggests that improvement in perceiving, evaluating, and interpreting media content can help protect social media users from experiencing body image disturbance. More research that examines the effectiveness of preexisting interventions that are focused on improv ing media literacy among young adults on social media platforms should also be conducted. Multiple platforms are beginning to regu late their content, including Instagram, Twitter, Facebook, and TikTok. Independent fact-checkers are currently being employed by Instagram to aid users in discriminating fake news from fact, and sensitive content warnings are being implemented on the platform by algorithms that identify and flag posts that users may find disturbing (Instagram, 2020). These design modifications can help to increase media literacy among users and provide an option to not be exposed to particular media content. The abundance of evidence that has shown the harmful effects of consuming idealized, appearance focused media indicates the need for additional platform updates. Therefore, interven tions focusing on mitigating adverse outcomes associated with exposure to such media usage must continue to improve the mental health trajectories of young social media users.

Misko, Hays, and Getz | Change Blindness Social Media Intervention

References Andrighetto, L. Bracco, F. Chiorri, C. Masini, M. Passarelli, M., & Piccinno, T. F. (2019). Nowyou see me, now you don’t: Detecting sexual objectification through a change blindness paradigm. Cognitive Processing, 20(1), 419–429. https://doi.org/10.1007/s10339-019-00927-w Arendt, F., Peter, C., & Beck, J. (2017). Idealized female beauty, social comparisons, and awareness intervention material. Journal of Media Psychology, 29(4), 188–197. https://doi.org/10.1027/1864-1105/a000181 Borau, S., & Nepomuceno, V. (2019). The self-deceived consumer: Women’s emotional and attitudinal reactions to the airbrushed thin ideal in the absence versus presence of disclaimers. Journal of Business Ethics, 154(2), 325–340. https://doi.org/10.1007/s10551-016-3413-2 Bracco, F., & Chiorri, C. (2008). People have the power: Priority of socially relevant stimuli in a change detection task. Cognitive Processing, 10(1), 41–49. https://doi.org/10.1007/s10339-008-0246-7 Brehm, J. W. (1966). A theory of psychological reactance. Academic Press. Brown, Z., & Tiggemann, M. (2016). Attractive celebrity and peer images on Instagram: Effect on women’s mood and body image. Body Image, 19, 37–43. https://doi.org/10.1016/j.bodyim.2016.08.007 Cho, A., & Lee, J. H. (2013). Body dissatisfaction levels and gender differences in attentional biases toward idealized bodies. Body Image, 10(1), 95–102. https://doi.org/10.1016/j.bodyim.2012.09.005 Clement, J. (2020). Instagram: Distribution of global audiences 2020, by age and gender. Statista. https://www.statista.com/statistics/248769/agedistribution-of-worldwide-instagram-user/ Choukas-Bradley, S., Nesi, J., Widman, L., & Higgins, M. K. (2019). Camera-ready: Young women’s appearance-related social media consciousness. Psychology of Popular Media Culture, 8(4), 47–481. https://doi.org/10.1037/ppm0000196 Cohen, R., Fardouly, J., Netwon-John, T., & Slater, A. (2019). BoPo on Instagram: An experimental investigation of the effects of viewing body positive content on young women’s mood and body image. New Media & Society, 21(7), 1546–1564. https://doi.org/10.1177/1461444819826530 Danthinne, E., Giorgianni, F., & Rodgers, R. (2020). Labels to prevent the detrimental effects of media on body image: A systematic review and meta‐analysis. International Journal of Eating Disorders, 53(5), 377–391. https://doi.org/10.1002/eat.23242 Fardouly, J., Pinkus, R. T., & Vartanian, L. R. (2017). The impact of appearance comparisons made through social media, traditional media, and in person in women’s everyday lives. Body Image, 20, 31–39. https://doi.org/10.1016/j.bodyim.2016.11.002 Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117–140. https://doi.org/10.1177/001872675400700202 Fitzsimmons-Craft, E., Bardone-Cone, A. M., Crosby, R. D., Engel, S. G., Wonderlich, S. A., & Bulik, C. M. (2016). Mediators of the relationship between thin-ideal internalization and body dissatisfaction in the natural environment. Body Image, 18, 113–122. http://dx.doi.org/10.1016/j.bodyim.2016.06.006 Groesz, L. M., Levine, M. P., & Murnen, S. K. (2001). The effect of experimental presentation of thin media images on body satisfaction: A meta-analytic review. International Journal of Eating Disorders, 31(1), 1–16. https://doi.org/10.1002/eat.10005 Halliwell, E. (2013). The impact of thin idealized media images on body satisfaction: Does body appreciation protect women from negative effects? Body Image, 10(4), 509–514. https://doi.org/10.1016/j.bodyim.2013.07.004 Halliwell, E., Easun, A., & Harcourt, D. (2011). Body dissatisfaction: Can a short media literacy message reduce negative media exposure effects amongst adolescent girls? British Journal of Health Psychology, 16(2), 396–403. https://doi.org/10.1348/135910710X515714 Hogue, J. V., & Mills, J. S. (2019). The effects of active social media engagement with peers on body image in young women. Body Image, 28, 1–5. https://doi.org/10.1016/j.bodyim.2018.11.002 Instagram. (2020). Why is a post on Instagram marked as false information? https://help.instagram.com/388534952086572 Kleemans, M., Daalmans, S., Carbaat, I., & Anschütz, D. (2018). Picture perfect: The direct effect of manipulated Instagam photos on body image in adolescent girls. Media Psychology, 21(1), 93–110.

https://doi.org/10.1080/15213269/2016.1257392 Livingston, J., Holland, E., & Fardouly, J. (2020). Exposing digital posing: The effect of social media self-disclaimer captions on women’s body dissatisfaction, mood, and impressions of the user. Body Image, 32, 150–154. https://doi.org/10.1016/j.bodyim.2019.12.006 Levin, D. T., Momen, N., Drivdahl, S. B., & Simons, D. J. (2000). Change blindness blindness: The metacognitive error of overestimating change-detection ability. Visual Cognition, 7(1–3), 397–412. https://doi.org/10.1080/135062800394865 McComb, S., & Mills, J. (2020). A systematic review on the effects of media disclaimers on young women’s body image and mood. Body Image, 32, 34–52. https://doi.org/10.1016/j.bodyim.2019.10.010 Ogden, J., Gosling, C., Hazelwood, M., & Atkins, E. (2020). Exposure to body diversity images as a buffer against the thin-ideal: An experimental study. Psychology, Health & Medicine, 25(10), 1–14. https://doi.org/10.1080/13548506.2020.1734219 Pashler, H. (1988). Familiarity and visual change detection. Perception & Psychophysics, 44, 369–378. https://doi.org/10.3758/BF03210419 Rensink, R. A., O’Regan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Sciences, 8(5), 368–373. https://doi.org/10.1111/j.1467-9280.1997.tb00427.x Rodgers, R. F., McLean, S. A., & Paxton, S. J. (2015). Longitudinal relationships among internalization of the media ideal, peer social comparison, and body dissatisfaction: Implications for the tripartite influence model. Developmental Psychology, 51(5), 706–713. https://doi.org/10.1037/dev0000013 Schaefer, L. M. (2017). The Development and Validation of the Physical Appearance Comparison Scale-3 (PACS-3). Graduate Theses and Dissertations. http://scholarcommons.usf.edu/etd/6949 Schaefer, L. M., & Thompson, J. K. (2018). Physical Appearance Comparison Scale–3. PsycTESTS. https://doi.org/10.1037/t67450-000 Sherlock, M., & Wagstaff, D. L. (2019). Exploring the relationship between frequency of Instagram use, exposure to idealized images, and psychological well-being in women. Psychology of Popular Media Culture, 8(4), 482–490. http://dx.doi.org/10.1037/ppm0000182 Silverstein, B., Perdue, L., Peterson, B., & Kelly, E. (1986). The role of the mass media in promoting a thin standard of bodily attractiveness for women. Sex Roles, 14(9), 519–532. http://dx.doi.org/10.1007/BF00287452 Simons, D. J. (1996). In sight, out of mind: when object representations fail. Psychological Sciences, 7(5), 301–305. https://doi.org/10.1111/j.1467-9280.1996.tb00378 Stice, E., Schupak-Neuberg, E., Shaw, H. E., & Stein, R. I. (1994). Relation of media exposure to eating disorder symptomatology: An examination of mediating mechanisms. Journal of Abnormal Psychology, 103(4), 836–840. https://doi.org/10.1037/0021-843X.103.4.836 Stice, E., & Shaw, H. E. (1994). Adverse effects of the media portrayed thin-ideal on women and linkages to bulimic symptomatology. Journal of Social and Clinical Psychology, 13(3), 288–308. https://doi.org/10.1521/jscp.1994.13.3.288 Tamplin, N. C., McLean, S. A., & Paxton, S. J. (2018). Social media literacy protects against the negative impact of exposure to appearance ideal social media images in young adult women but not men. Body Image, 26, 29–37. https://doi.org/10.1016/j.bodyim.2018.05.003 Tiggemann, M. & Polivy, J. (2010). Upward and downward: Social comparison processing of thin idealized media images. Psychology of Women Quarterly, 34(3), 356–364. https://doi.org/10.1111/j.1471-6402.2010.01581.x Author Note. Stephanie Misko https://orcid.org/0000-0003-0622-7866 Olivia Hays https://orcid.org/0000-0002-2108-6697 Laura Getz https://orcid.org/0000-0002-3429-7506 We have no conflicts of interest to disclose. This study was supported by the University of San Diego Psychological Sciences Department. Correspondence concerning this article should be addressed to Laura Getz, Department of Psychological Sciences, University of San Diego, 5998 Alcala Park, San Diego, CA, 92110, United States. Email: lgetz@sandiego.edu SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

131

Change Blindness Social Media Intervention | Misko, Hays, and Getz

APPENDIX A Content Disclaimers Experimental disclaimer: Have you ever done an activity when you had to “spot the changes” between two photos or cartoons? When looking at successive images, people often fail to see changes between them. People might also be unlikely to spot the difference between images with subtle modifications intended to be undetectable. Photoshop is an editing practice among social media users, enabling them to tweak or change their photos in discrete and believable ways. In this study, you will view a sequence of 10 photo pairs displayed in succession. After viewing each set of paired photos, you will answer items related to popularity perceptions of the pictured women. Control disclaimer: In this study, you will view a sequence of 10 photo pairs displayed in succession. After viewing each set of paired photos, you will answer items related to popularity perceptions of the pictured women.

APPENDIX B Physical Appearance Comparison Scale-3 (Modified Version) 1.

When I’m out in public, I compare my weight/shape to the weight/shape of others. Participants selected one of the following options: Always, Most of the time, About half the time, Sometimes, Never.

When I make these comparisons, I typically believe that I look ______ than the person to whom I am comparing myself. Participants selected either: Much better, Somewhat better, About the same, Somewhat worse, Much worse.

When you make these comparisons, how does it usually make you feel? Participants selected either: Extremely good, Somewhat good, Neither good nor bad, Somewhat bad, Extremely bad

When I watch television, I compare my weight/shape to the weight/shape of the actors/actresses.

When I make these comparisons, I typically believe that I look ______ than the person to whom I am comparing myself.

When you make these comparisons, how does it usually make you feel?

When I see a female influencer on social media, I compare my weight/shape to her weight/shape.

When I see an USD female peer on social media, I compare my weight/shape to her.

When I make these comparisons, I typically believe that I look ______ than the person to whom I am comparing myself.

10. After making these comparisons on social media, how does it usually make you feel?

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

132

https://doi.org/10.24839/2325-7342.JN27.2.133

The Classic Stroop Asymmetry in Online Experiments Mary E. Smith, Micah D. Smith, and Kenith V. Sobel* Department of Psychology and Counseling, University of Central Arkansas

ABSTRACT. In a traditional Stroop (1935) task, participants view target words written in colors that are either congruent with their meaning (e.g., “Red” written in red ink) or incongruent (e.g., “Red” written in green). When participants identify the target color, the response time difference between congruent and incongruent targets (i.e., Stroop effect) is typically much larger than when participants identify the target word (i.e., reverse Stroop effect); this is the classic Stroop asymmetry. Recent work has shown that recasting the task so participants localize the target rather than identify it inverts the asymmetry: the Stroop effect is smaller than the reverse Stroop effect. We developed online identification and localization scripts in an attempt to replicate the classic Stroop asymmetry and this recent twist on the classic asymmetry. In Experiment 1, a sample of undergraduate students replicated the classic Stroop asymmetry with an identification task and its inversion with a localization task, p < .001, ηp 2 = .17. In Experiment 2, a sample of participants who were more representative of the general population than in Experiment 1 once again replicated the classic Stroop asymmetry with an identification task and its inversion with a localization task, p < .001, ηp 2 = .20. Our results combine with other recent studies to provide converging evidence that the classic Stroop asymmetry results from the strength of association between task demands (identification vs. localization) and the attended feature (verbal vs. visual). Keywords: Stroop effect, reverse Stroop effect, Stroop asymmetry, visual search, online data collection

he debut of Amazon’s Mechanical Turk website in 2005 inaugurated the era of widespread data collection over the internet. Since then, the number of studies in experimental psychology that include some online data component has steadily increased (Buhrmester et al., 2018) and accelerated in the wake of the COVID-19 pandemic (Greene & NavehBenjamin, 2022). Although online data collection has various advantages, such as the opportunity to access a diverse pool of participants (Newman et al., 2021), an obvious disadvantage is the inability to exercise the same controls as when experiments are conducted in a laboratory setting. Participants who connect remotely to an experiment use their own devices, operating systems, and browsers, in settings of their choosing where they may be concurrently engaged in tasks such as household *Faculty mentor

chores or participating in nearby conversations. These uncertainties threaten the precision of timing for experimental paradigms that rely on response time (RT) as a dependent variable (AnwylIrvine et al., 2021). For much of the history of online data collection, the primary way to gather reliable RT data in an online experiment required expertise in webpage and software design (e.g., Crump et al., 2013), but recently developed tools enable people with a wider range of technical skills to carry out online data collection (Anwyl-Irvine et al., 2020; Henninger et al., 2021). Nevertheless, researchers who are familiar with the point-andclick functionality afforded by Qualtrics (https:// www.Qualtrics.com) and want to run online experiments that gather RT may prefer to ease into the endeavor by developing Qualtrics scripts before jumping into a new and unfamiliar environment.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

133

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

Here we describe experiments based on Qualtrics scripts that attempt to replicate both the classic Stroop asymmetry and a recent study that inverted the classic asymmetry. The venerable Stroop (1935) paradigm is ideal for validating a novel method of capturing RT (Barnhoorn et al., 2015; Crump et al., 2013), because Stroop effects have been so reliably replicated (MacLeod, 2005).

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

134

The Classic Stroop Asymmetry In a typical computer-based Stroop task, par ticipants view a color word written in pixels that either match (e.g., “Red” in red pixels) or do not match (e.g., “Red” in green pixels) its meaning. Participants either report the target color while ignoring its meaning (Stroop condition), or the target word while ignoring its color (reverse Stroop condition). When the unattended feature is incongruent with the attended feature, it can leak through the attentional filter to interfere with participants’ ability to report the attended feature (Linzarini et al., 2017). The difference in RT between congruent and incongruent targets represents the magnitude of the interference (Whitehead et al., 2018). Interference from the unattended feature is typically greater in the Stroop condition, in which participants report the target color, than in the reverse Stroop condition, in which they report the target word; this is the classic Stroop asymmetry (Melara & Algom, 2003). One prominent explanation for the classic Stroop asymmetry asserts that visual and verbal information are encoded and processed in two different systems (Song & Hakoda, 2015; Virzi & Egeth, 1985), and interference can occur when the stimulus is encoded in one format while the response is encoded in the other. For example, in the traditional Stroop condition, participants vocally report the target color, so the visually encoded stimulus (target color) must be translated into a verbal code in order to vocalize it (Durgin, 2000). Even if the response is manual rather than vocal, as when participants press a key to report the target color, the manual response is verbally mediated (Bearden et al., 2021; Blais & Besner, 2006; Parris et al., 2019; Sugg & McDonald, 1994). Whereas translation is required when participants report the target color in the Stroop condition, when participants report the target word in the reverse Stroop condition, no translation is required between the (verbally encoded) target word stimu lus and the (verbally mediated) response. Thus, the translation account explains why the Stroop

effect is typically larger than the reverse Stroop effect, but it also predicts that if the response were changed to require a visual rather than verbal code, translation would be required in the reverse Stroop condition between the (verbally encoded) target word stimulus and the (visually encoded) response. As a result, the asymmetry would be inverted: the Stroop effect should be smaller than the reverse Stroop effect. Durgin (2000; replicated by Miller et al., 2016) tested this prediction of the translation hypothesis. Each display contained a target word surrounded by four rectangular patches of color. The target word was either a color word with an incongru ent pixel color, or neutral (noncolor words with colored pixels in the Stroop condition or color words written with gray pixels in the reverse Stroop condition). Participants moved the cursor from the target toward the color patch that was consistent with either the target color in the Stroop condition or the target word in the reverse Stroop condition. Translation was not required in the Stroop condi tion between the (visually encoded) target color stimulus and the (visually encoded) color patch response, but translation was required in the reverse Stroop condition between the (verbally encoded) target word stimulus and the (visually encoded) color patch response. Consistent with the transla tion account, the Stroop effect was smaller than the reverse Stroop effect: an inversion of the classic Stroop asymmetry. Although the results from Durgin’s (2000) color matching experiment were consistent with the translation account, Blais and Besner (2007) identi fied a potential confound in Durgin’s method. While it is true that Durgin encouraged visual pro cessing in a Stroop task by using visually encoded color patch responses, the task he designed also relied on visual processing. Whereas the traditional Stroop paradigm requires participants to identify the target color or word, Durgin required participants to localize the color patch matching the target color or word. Blais and Besner proposed that traditional identification tasks are more strongly associated with verbal than visual processing, but Durgin’s localization task is more strongly associated with visual than verbal processing. Thus, the strengthof-association account implies that an identification task should confer an advantage on the target word, whereas a localization task should confer an advan tage on the target color, so manipulating the task may be sufficient to invert the Stroop asymmetry even without any need for translation.

Smith, Smith, and Sobel | Classic Stroop Asymmetry in Online Experiments

Sobel et al. (2020) developed a variation on Durgin (2000) and manipulated the task to verify this prediction of the strength-of-association account. Participants identified the target in the first experiment and localized it in the second. As expected, the identification task in the first experi ment replicated the classic Stroop asymmetry. Then for the localization task in the second experiment, the center of each display was occupied by a cue indicating what participants should search for on that trial: a rectangular color patch in the Stroop condition, or a color word written in a neutral color in the reverse Stroop condition. Surrounding the cue were four color words that were all either congruent or incongruent with their pixel color. Participants reported the location of the item that had the same pixel color as the color patch cue in the Stroop condition or the same meaning as the word cue in the reverse Stroop condition. Because the cued feature and the relevant target feature relied on the same type of code (i.e., color patch cue and target color in Stroop, word cue and target word in reverse Stroop), there was no need for translation. Nevertheless, as predicted by the strength-of-association account, the localization task itself was sufficient to invert the classic Stroop asymmetry: The Stroop effect was smaller than the reverse Stroop effect. Hypotheses For this project, we aimed to develop a Qualtrics script that replicates the classic Stroop asymmetry for an identification task, primarily to demonstrate that the script is sufficiently powerful to detect the effect even with the uncertainties inherent in a remotely delivered task. Also, we aimed to extend on this validity check by attemtpting to replicate the inversion of the classic Stroop asymmetry with a localization task as in Sobel et al. (2020). Experimental psychologists commonly rely on samples comprising undergraduate students because this population is readily accessible, but the results from studying such convenience samples may not generalize to the overall population (Hanel & Vione, 2016; Peterson & Merunka, 2014). For that reason, after releasing the script to a convenience sample of undergraduate students, we wanted to release it to a sample of participants who are more representative of the general population. We had two hypotheses. First, we hypothesized that, for a convenience sample of university students in Experiment 1, the identification task would replicate the classic Stroop asymmetry, and

the localization task would replicate Sobel et al. (2020) by inverting the asymmetry. And second, we hypothesized that the results from Experiment 1 would generalize to a more representative sample in Experiment 2.

Experiment 1a: Identification Task With University Students Method Participants We submitted an application for review entitled Color and Meaning that was approved by our uni versity’s Institutional Review Board, and treated all participants in accordance with the ethical stan dards established by the American Psychological Association (2017). In Experiments 1a and 1b, all participants were undergraduate students at a mid-sized school in the southern United States and received credit in a variety of psychology courses in exchange for their participation. We used the results from a pilot experiment to determine an appropriate sample size that would reliably detect a difference between a Stroop effect and reverse Stroop effect. The pilot experiment yielded a Cohen’s d of 0.21, which would require a sample size of 79 participants to achieve 80% power at an alpha of .05 (Bausell & Li, 2002). Our sample for Experiment 1a included 83 participants, 13 of whom identified as male, 69 as female, and one as nonbinary. Participants’ ages ranged from 18 to 61 with a mean of 22.07 and standard deviation of 6.85. There were 62 participants who described them selves as White, 15 as Black or African American, three as Hispanic, one as Asian, one as American Indian or Alaska Native, and one preferred not to report a race. The Qualtrics script randomly assigned each participant to carry out either the identification task (Experiment 1a) or the localiza tion task (Experiment 1b). All participants were exposed to a Stroop condition in one block and a reverse Stroop condition in the other block. Block order was counterbalanced across participants. Apparatus Participants logged in to their accounts on Sona (https://UCA.sona-systems.com) to sign up for the experiment. The Sona website then redirected them to the script on Qualtrics, which they could complete by using their own personal devices. Stimuli Qualtrics scripts typically require participants to answer one or more questions on each page, then move on to the next page by clicking the Advance

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

135

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

button located in the lower right corner of each page. However, in our study, if the participant’s cursor (for anyone using a personal computer) or finger (for anyone using a tablet device) were situated in the lower right section of the display at the beginning of each trial, we thought they would respond more quickly to response buttons near the Advance button on the right side of the display. To mitigate these biasing effects, we wanted to force participants to position their cursor or finger in the middle of the display just before the response buttons appeared. To do so, each trial began with a page containing a prompt (i.e., “Click here to start the next trial”) above a button situated in the middle of the display. If participants tried to click the Advance button on the lower right side of the display rather than the cursor-centering button, the Qualtrics script insisted that they select the cursor-centering button before they could advance. To make the script automatically advance to the next page without participants having to click the Advance button, we found a JavaScript routine posted on the Qualtrics support forum. After par ticipants clicked the cursor-centering button, the JavaScript routine automatically advanced to the next page before participants had the opportunity to click the Advance button. After the JavaScript routine in the cursorcentering button triggered an auto-advance, a page containing the experimental stimulus and responses appeared. We generated the stimulus displays in PowerPoint then exported them into JPEG files that we uploaded to our Qualtrics image library. Screenshot examples appear in Figure 1. At the top of the page was a target item that was one of four capitalized color words (“Red,” “Green,” “Blue,” or “Yellow”) that matched the pixel color of the respective word against a black background. Below the target were four radio buttons, each of which was labelled with either a red, green, blue, FIGURE 1 Screenshots Depicting Stimuli and Response Buttons in Experiments 1a and 2a

Note. The screenshot on the left was selected from a Stroop task with a congruent target, and the screenshot on the right was selected from a reverse Stroop task with an incongruent target.

136

or yellow patch of color (Stroop condition), or the word “Red,” “Green,” “Blue,” or “Yellow” written in black pixels (reverse Stroop condition). In the Stroop condition, participants were instructed to select the color patch that matched the target color, and in the reverse Stroop condition, participants were instructed to select the word that matched the target word. When participants selected the correct response, the JavaScript routine automatically advanced to the next trial’s cursor centering page. If participants clicked the Advance button instead of one of the four response buttons, Qualtrics registered that response as incorrect. In that case, and when participants actually did select an incorrect response, the script presented a message saying “The answer you provided was incorrect. Please try to select the right answer.” To continue the experiment after receiving this admonition, participants needed to click the Advance button, which then advanced to the next trial’s cursor centering page. We intended this extra step for incorrect responses to provide an incentive to select the correct responses. Procedure The first three pages in the Qualtrics script con tained an informed consent letter, a set of demo graphic questions, and instructions, in that order. The instructions page informed participants that they would view a series of target items, one at a time, and that they should report the target color (for participants assigned to the Stroop condition in the first block) or word (for participants assigned to the reverse Stroop condition in the first block). After participants finished the instructions, the script proceeded through a series of four practice trials that were excluded from analysis, followed by 32 experimental trials. The 32 experimental trials in each block contained 16 displays in which the target color and word were congruent (e.g., the word “Red” written in red pixels) and 16 in which they were incongru ent (e.g., the word “Red” written in green pixels). The congruent displays included four repetitions of each of the four words “Red,” “Green,” “Blue,” and “Yellow,” written in pixels that matched the word’s meaning. The incongruent displays included two repetitions of each of the following eight word-color combinations: the word “Red” written in green pixels, “Red” in blue pixels, “Green” in red pixels, “Green” in yellow pixels, “Blue” in red pixels, “Blue” in yellow pixels, “Yellow” in green pixels, and “Yellow” in blue pixels. The target displays were the

Smith, Smith, and Sobel | Classic Stroop Asymmetry in Online Experiments

same for both the Stroop and the reverse Stroop conditions, with the primary difference between the two conditions being that the response buttons were labelled with color patches in the Stroop condition or words written in black in the reverse Stroop con dition. Thus, for example, when presented with an incongruent target such as the word “Red” in green pixels, the correct response in the Stroop condition would have been the green color patch, and the correct response in the reverse Stroop condition would have been the word “Red.” Given these 16 congruent displays and 16 incongruent displays, each of the four responses was correct eight times in each block. The 32 displays were presented in random order. Each stimulus display page included a timer that was invisible to the participants. Timers in Qualtrics scripts record four pieces of data pertain ing to the page where they are placed. For the parameters First click, Last click, and Page submit, the clock starts ticking when the page appears, and ends with the participant’s first mouse click on the page, the participant’s last click on the page, or when the page advances, respectively. The param eter Click count represents the number of times the participant clicked anywhere on the page before it advanced to the next page. In most Experiment 1a trials, participants clicked just once on the stimulus display page, which was when they clicked one of the response buttons. In trials with just one click, First click and Last click had an identical value, and Page submit was a few milliseconds longer, with the delay representing how long the JavaScript routine took to advance the page after a response button was selected. Because each page with a timer only had one question (asking participants to report the target color or word), in trials with just one click, we interpreted First click as the RT for that trial. Trials with more than one click were not as readily interpretable as those with just one click, as we will describe in the Results section. After participants completed the four practice trials and 32 experimental trials, the script informed participants that they had just completed the first half of the experiment, and that for the remainder of the experiment they should attend to the other target feature. After reading the instructions, participants proceeded through four practice trials followed by 32 experimental trials presented in random order, while attending to the target feature that they had not attended in the first block. Once both blocks were completed, each participant was presented with a debriefing page.

Results The data from three participants were excluded from analysis. One participant did not finish the experiment, and the other two provided incorrect responses for more than half of the trials in at least one of the four conditions. In addition to the three excluded participants, we also removed RTs from individual trials for three reasons. First, we removed trials with incorrect responses, such as when a participant clicked the button for a blue color patch when the target’s pixels were red. The second reason for removal concerns the trials in which the Click count was greater than one. As mentioned previously in the Procedure section, for trials with a Click count of one, we interpreted the timer’s First click parameter as the RT for that trial. We did not initially understand why any trial would have more than one click, but eventually we discovered that either a finger swipe on a phone or tablet device to reposition the display, or a mouse click anywhere on the page not on a response button would be counted as a click. Because there is no way to determine whether an extra click was attributable to a finger swipe event on a tablet device or an actual extra click on a computer, we decided to exclude from analysis any trials in which Click count was greater than one. We should mention here that, on some trials, the Click count was zero, which occurred when a participant clicked the Advance button rather than one of the response buttons. Such instances were rare, and as mentioned previously in the Method section, they were recorded as incorrect responses and thus were removed along with the other incorrect responses. After removing any trials with incorrect responses or more than one click, we calculated mean RTs for each participant in the Stroop condition and the reverse Stroop condition. Any RT that was more than three standard deviations above the mean for its participant and condition was removed as an out lier. The percentage of trials that were removed for each of these three reasons is presented in Table 1. TABLE 1 Percentages of Trials Excluded From Analysis in Experiment 1 Congruent Errors Identification Stroop Localization

Incongruent

Clicks Outliers

Errors

Clicks Outliers

0.23% 3.75% 1.33%

2.19%

3.20% 2.66%

Reverse 0.47% 2.03% 0.78%

1.80%

2.19% 2.27%

SUMMER 2022

Stroop

1.50% 4.11% 1.50%

2.53%

2.93% 2.14%

Reverse 0.87% 4.11% 1.26%

5.78%

5.85% 2.14%

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

137

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

Participants’ mean RT after errors, extra clicks, and outliers had been removed (depicted in Figure 2) were submitted to a two-way ANOVA with congru ity and attended feature as within-subjects factors. Responses were faster for congruent targets than incongruent, F(1, 79) = 16.96, p < .001, ηp2 = .18, but RTs were not significantly different when the participants attended to the target color (Stroop) than when they attended to the target word (reverse Stroop), F(1, 79) = 3.22, p = .08, ηp2 = .04. The inter action between congruity and attended feature, F(1, 79) = 11.03, p = .001, ηp2 = .12, shows that the Stroop effect was different than the reverse Stroop effect. Simple effects analysis confirmed that the Stroop effect was significant, F(1, 79) = 57.36, p < .001, ηp2 = .42, but the reverse Stroop effect was not, F(1, 79) = 0.02, p = .88, ηp2 = .003, and the Stroop effect size FIGURE 2 Mean Response Times for an Identification Task in Experiment 1a and a Localization Task in Experiment 1b 3500

Experiment 1: University Student Sample Stroop/Congruent

Stroop/Incongruent

Reverse/Congruent

Reverse/Incongruent

Response time (ms)

3000

2500

2000

1500

1000

Identification

Localization

Note. Error bars represent 95% confidence intervals (calculation based on Loftus & Masson, 1994).

FIGURE 3 Screenshots Depicting Stimuli and Response Buttons in Experiments 1b and 2b

Note. The screenshot on the left was selected from a Stroop task with a congruent target positioned at 1 o’clock, and the screenshot on the right was selected from a reverse Stroop task with an incongruent target positioned at 8 o’clock.

138

(ηp2 = .42) was larger than the reverse Stroop effect size (ηp2 = .003). Discussion The results from the identification task replicated the classic Stroop asymmetry; the Stroop effect was larger than the reverse Stroop effect. Because the classic Stroop asymmetry is sufficiently reliable to be called a classic, it may seem unsurprising that we succeeded. However, an online version of a Stroop task introduces much variability in the apparatuses and experimental contexts that could have prevented a successful replication. By obtaining the classic effect, the identification task lays the groundwork to suggest that, if the inversion of the classic Stroop effect in Sobel et al. (2020) is replicable, our remote version of their localization task should be sufficiently powerful to replicate it.

Experiment 1b: Localization Task With University Students Method Participants Our sample included 90 participants, 13 of whom identified as male, 75 as female, and two as nonbi nary. Participants’ ages ranged from 18 to 55 with a mean of 21.98 and standard deviation of 6.07. There were 64 participants who described them selves as White, 17 as Black or African American, three as Hispanic, two as Asian, two as American Indian or Alaska native, and one preferred not to state a race. As in the identification task, all participants were exposed to a Stroop condition in one block and a reverse Stroop condition in the other block, and block order was counterbalanced across participants. Apparatus As in the identification task, participants accessed the Qualtrics script using their own personal devices through their Sona account. Stimuli The basic structure of each trial was the same as in the identification task: each trial began with a cursor centering page, followed by a page contain ing the stimulus display and response buttons, and finally an admonishment page when participants selected an incorrect response. Also, the same JavaScript routines triggered automatic advance ment in the cursor centering and stimulus display pages. However, the stimulus displays and response labels were different than in the identification task.

Smith, Smith, and Sobel | Classic Stroop Asymmetry in Online Experiments

Each display contained not just one word as in the identification task, but all four of the color words (“Red,” “Green,” “Blue,” and “Yellow”), arranged on an imaginary circle centered on a fixation mark, as can be seen in the example screenshots in Figure 3. All four words were either congruent or incongruent with their pixel color. The items were placed at 90-degree intervals from each other, in one of two arrangements; in one arrangement, the four items appeared at clockface positions 1 o’clock, 4 o’clock, 7 o’clock, and 10 o’clock, and in the other arrangement the items appeared at 2 o’clock, 5 o’clock, 8 o’clock, and 11 o’clock. Below the target display was a cue indicat ing what the participant should search for in that trial. In the Stroop condition, the cue was a color patch containing pixels that were either red, green, blue, or yellow, and in the reverse Stroop condi tion, the cue was one of the words “Red,” “Green,” “Blue,” or “Yellow” written in black pixels. Below the target display and cue were two response buttons; the button on the left was labelled “Left side” and the button on the right was labelled “Right side.” Procedure As in the identification task, the first three pages in the Qualtrics script contained an informed consent letter, demographic questions, and instructions, in that order. The instructions page informed partici pants that they would view a series of displays and cues, and that they should report whether the target was positioned on the left side or the right side of the fixation cross. After reading the instructions, participants proceeded through a series of four practice trials that were excluded from analysis, followed by 32 experimental trials. The 32 experimental trials in each block con tained 16 congruent displays and 16 incongruent displays; congruent targets were written with pixels that were congruent with their meaning, and incon gruent displays included the same eight word-color target combinations as in the identification task: “Red” in green pixels, “Red” in blue pixels, “Green” in red pixels, “Green” in yellow pixels, “Blue” in red pixels, “Blue” in yellow pixels, “Yellow” in green pixels, and “Yellow” in blue pixels. In the 16 congruent displays and the 16 incongruent displays, each word and each color appeared twice in each of the eight clockface positions. The target appeared on the right side of the fixation cross in half of the trials, and on the left side in the other half of the trials. The 16 congruent trials and 16 incongruent trials were presented in random order. At the end

of the first block, participants were informed that they had completed half of the experiment, and they should attend to the other target feature for the remainder of the experiment. After completing both blocks, participants were presented with a debriefing statement. Results The data from 11 participants were excluded from analysis because in at least one of the four condi tions more than half of their trials had more than one click. This is in sharp contrast to the results in Experiment 1a, in which none of the participants had so many trials with extra clicks. In our attempt to figure out why so many participants had extra clicks, we tried, among other things, running the script on a phone positioned horizontally. Because the stimulus displays were larger in Experiment 1b than the identification displays in Experiment 1a, as we ran the simulation we needed to swipe each display upward to access the response buttons below. Thus we believe that 11 participants carried out the script on a phone they held horizontally, forcing them to swipe the display just to access the response buttons before they could even select one. In addition to the 11 excluded participants, we also removed RTs from individual trials for the same three reasons as in Experiment 1a. The percentages of removed trials are presented in Table 1. Participants’ mean RTs (depicted in Figure 2) were submitted to a two-way ANOVA with congru ity and attended feature as within-subjects factors. Responses were faster for congruent targets than incongruent, F(1, 78) = 66.39, p < .001, ηp2 = .46, and were faster when participants attended to the target color (Stroop) than when they attended to the target word (reverse Stroop), F(1, 78) = 39.78, p < .001, ηp2 = .34. The significant interaction between congruity and attended feature, F(1, 78) = 23.31, p < .001, ηp2 = .23, shows that the Stroop effect was different than the reverse Stroop effect. Simple effects analysis confirmed that both the Stroop effect, F(1, 78) = 7.06, p = .010, ηp2 = .08, and the reverse Stroop effect, F(1, 78) = 54.80, p < .001, ηp2 = .41, were significant, and the Stroop effect size (ηp2 = .08) was smaller than the reverse Stroop effect size (ηp2 = .41). In Experiment 1a, the two-way interaction between congruity and attended feature indicated that the Stroop effect was larger than the reverse Stroop effect, but in Experiment 1b, the two-way interaction indicated that the Stroop effect was smaller than the reverse Stroop effect. This suggests

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

139

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

140

that the three-way interaction between congruity, attended feature, and task (identification versus localization) should be significant. With that in mind, we submitted mean correct RTs from both Experiments 1a and 1b to a three-way ANOVA with congruity and attended feature as within-subjects factors, and task as a between-subjects factor. The results confirmed that the three-way interaction was significant, F(1, 157) = 32.83, p < .001, ηp2 = .17.

any money changing hands. Alas, members of the general public cannot be compensated with class credit. To gather a more representative sample for Experiment 2, we acquired external funding through a Psi Chi Undergraduate Research Grant to provide monetary incentives to members of the general public.

Discussion The main effect of congruity in the localization task in Experiment 1b echoed the same effect from the identification task Experiment 1a, but unlike the identification task, the main effect of attended feature was significant in Experiment 1b but not Experiment 1a. The Experiment 1b localization task required participants to search through several items before they could select one item for further processing, whereas no selection was required for the Experiment 1a identification task because there was only one display item. The main effect of attended feature in Experiment 1b indicates that responses were faster when participants attended to the target color in the Stroop condition than when they attended to the target word in the reverse Stroop condition. This effect was also present in Sobel et al. (2020), who argued that selecting one of the search items is more efficient for a visual feature such as color than for a semantic feature such as a word’s meaning (Wolfe & Horowitz, 2004). The effect that is more relevant to our first hypothesis is the three-way interaction between congruity, attended feature, and task, which con firms our first hypothesis that an identification task would yield the classic Stroop asymmetry, and that a localization task would invert the asymmetry. This outcome represents more than a simple replication of Sobel et al. (2020), because our online version of identification and localization tasks introduced variability into the apparatuses and experimental contexts that would be absent from an experiment run on a single computer in a laboratory setting. Nevertheless, our sample of undergraduate students is more homogeneous along such dimen sions as age and educational level than the general population (Hanel & Vione, 2016; Peterson & Merunka, 2014). Of course, a major reason for the common practice in psychology departments (including our own) of allowing undergraduate students to gain course credit for their participation in experiments is that researchers can compensate participants for their time and attention without

Method Participants With funding from a Psi Chi Undergraduate Research Grant, we were able to pay Qualtrics to recruit participants for Experiments 2a and 2b. Although the quote from Qualtrics did not specify how much they paid each participant, a simple cal culation can estimate an upper boundary. We paid Qualtrics $1,050 and received data from 305 partici pants, which amounts to $3.44 per person, minus any brokering fees retained by Qualtrics. Because Qualtrics reports the latitude and longitude of the device used to connect to the online script, we could determine each participant’s general loca tion. There was at least one participant from the four corners of the contiguous United States: from San Diego, CA, in the southwest; Seattle, WA, in the northwest; Miami, FL, in the southeast; Boston, MA, in the northeast; and many places in between. Of the 161 participants in Experiment 2a, 53 identified as male, 107 as female, and one as nonbinary. There were 110 participants who described themselves as White, 22 as Black or African American, 14 as Hispanic, five as Asian, five as American Indian or Alaska Native, two as Native Hawaiian or Pacific Islander, one as Creole, and two preferred not to report a race. Participants’ ages ranged from 18 to 75 years with a mean of 38.43 and standard deviation of 14.88.

Experiment 2a: Identification Task With a Nationwide Sample

Apparatus, Stimuli, and Procedure As in Experiment 1, participants used their own per sonal devices to access the script on Qualtrics.com, which was identical to the one used in Experiment 1. Results The data from 40 participants were excluded from analysis because for each of these participants more than half of the trials from at least one of the four conditions included an incorrect response (17 participants), extra clicks (19 participants), or a combination of errors and extra clicks (four participants). In addition, the data from another

Smith, Smith, and Sobel | Classic Stroop Asymmetry in Online Experiments

than when they attended to the target word (reverse Stroop), F(1, 119) = 0.22, p = .64, ηp2 = .002. The interaction between congruity and attended feature, F(1, 119) = 37.42, p < .001, ηp2 = .24, shows that the Stroop effect was different than the reverse Stroop effect. Simple effects analysis confirmed that the Stroop effect was significant, F(1, 119) = 106.96, p < .001, ηp2 = .47, but the reverse Stroop effect was not, F(1, 119) = 3.25, p = .07, ηp2 = .03, and the Stroop effect size (ηp2 = .47) was larger than the reverse Stroop effect size (ηp2 = .03). Discussion The results from Experiment 2a echoed those from Experiment 1a, with a significant main effect of congruity but not attended feature, and a significant interaction between congruity and attended feature. Further, the presence of a classic Stroop asymmetry, as indicated by the interaction effect, lends preliminary support to our second hypothesis that the results from a convenience sample of undergraduate students would general ize to a more representative sample. To see if the TABLE 2 Percentages of Trials Excluded From Analysis in Experiment 2 Congruent Errors

Localization

Incongruent

Clicks Outliers

Errors

Clicks Outliers

0.68% 5.10% 1.15%

4.11%

6.61% 2.71%

Reverse 0.89% 7.81% 1.93%

2.40%

7.92% 2.24%

1.92% 2.91% 1.63%

5.18%

2.77% 1.92%

Reverse 1.78% 2.70% 3.20%

10.08%

2.56% 4.47%

Identification Stroop Stroop

FIGURE 4 Mean Response Times for an Identification Task in Experiment 2a and a Localization Task in Experiment 2b 3500

Experiment 2: Nationwide Sample Stroop/Congruent Stroop/Incongruent

Reverse/Congruent Reverse/Incongruent

3000

Response time (ms)

participant was excluded because their mean RT was more than three standard deviations greater than the mean RT of the other participants. We have little insight into why there was a much higher proportion of participants excluded from the general population sample than from the sample of university students in Experiment 1, so any explanation we offer would necessarily be speculative. One possiblity might be that university students can more readily imagine themselves as the researcher carrying out the project in which they participate than members of the general popula tion, and thus the sample of university students included a higher proportion of individuals who conscientiously attended to the stimuli. To be clear, we do not mean to suggest that every member of the sample from the general population paid less attention than every member of the sample of university students. Instead, we are arguing that the sample from the general population might have included more individuals who were less invested in the experiment than the sample of university students. Researchers who carry out projects online are commonly urged to include attention checks (e.g., “If you are paying attention select strongly disagree”) as a way to remove the data from inat tentive participants (Newman et al., 2021), but our task has built-in attention checks in the form of errors and extra clicks. With that in mind, after excluding inattentive participants as described above, we consider the remaining data to represent participants who were paying sufficient attention. In any event, we sought grant funding to obtain a more heterogeneous sample than could be obtained from students at our university, and heterogeneity is just what we got: demographic, geographic, and attentional. Perhaps the need to remove a relatively high proportion of participants is simply the cost of doing business with heterogeneous samples. As in Experiment 1, we also removed RTs from individual trials that were incorrect, had more than one click, or were more than three standard deviations greater than the mean RTs for that participant and task. The percentage of trials that were removed for each of these three reasons is presented in Table 2. Participants’ mean RTs (depicted in Figure 4) were submitted to a two-way ANOVA with congru ity and attended feature as within-subjects factors. Responses were faster for congruent targets than incongruent, F(1, 119) = 50.04, p < .001,ηp2 = .30, but RTs were not significantly different when the participants attended to the target color (Stroop)

2500 2000 1500 1000

SUMMER 2022 Identification

Localization

Note. Error bars represent 95% confidence intervals (calculation based on Loftus & Masson, 1994).

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

141

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

results from a localization task would also support our second hypothesis, in Experiment 2b, a sample from the general population completed the same online script as the participants in Experiment 1b.

Experiment 2b: Localization Task With a Nationwide Sample Method Participants With money from our Psi Chi grant, we paid Qualtrics to recruit 144 participants, who were even a bit more geographically widespread than in Experiment 2a; one participant accessed the script from a device in Fairbanks, AK. Our sample included 47 participants who identified as male, 96 as female, and one as nonbinary. There were 92 participants who described themselves as White, 28 as Black or African American, 14 as Hispanic, seven as Asian, two as Native Hawaiian or Pacific Islander, and one as American Indian or Alaska Native. Participants’ ages ranged from 19 to 79 years with a mean of 39.74 and a standard deviation of 14.91. Apparatus, Stimuli, and Procedure The apparatus was the same as in Experiment 2a. The stimuli and procedure were the same as in Experiment 1b.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

142

Results The data from 56 participants were excluded from analysis for having errors (36 participants), extra clicks (15 participants), or a combination of errors and extra clicks (five participants) in more than half of the trials from at least one of the four conditions. As mentioned in the Results section for Experiment 2a, a higher proportion of partici pants were excluded from the general population sample than the sample of university students for reasons which elude us. Nevertheless, we can identify a possible reason why a higher proportion of participants were excluded from Experiment 2b than Experiment 2a. In contrast to the identifica tion task, the localization task in Experiment 2b required two mental processing stages: a selection stage followed by a decision stage (Smith & Sewell, 2013). In each localization trial, participants were presented with not just one item, but four, and they needed to select one of the items before they could decide whether it was the target. Apparently, the extra layer of processing in localization led to more errors than just the single decision stage in identification. The percentage of individual trials excluded from analysis due to error, extra clicks, or

being an outlier, is presented in Table 2. Participants’ mean RTs (depicted in Figure 4) were submitted to a two-way ANOVA with congru ity and attended feature as within-subjects factors. Responses were faster for congruent targets than incongruent, F(1, 87) = 57.01, p < .001,ηp2 = .40, and responses were faster when the participants attended to the target color (Stroop) than when they attended to the target word (reverse Stroop), F(1, 87) = 23.86, p < .001,ηp2 = .22. The interaction between congruity and attended feature, F(1, 87) = 21.44, p < .001, ηp2 = .20, shows that the Stroop effect was different than the reverse Stroop effect. Simple effects analysis confirmed that both the Stroop effect, F(1, 87) = 18.86, p < .001, ηp2 = .18, and the reverse Stroop effect, F(1, 87) = 48.76, p < .001, ηp2 = .36, were significant, but the Stroop effect size (ηp2 = .18) was smaller than the reverse Stroop effect size (ηp2 = .36). As in Experiment 1, the Stroop effect was larger than the reverse Stroop effect for an identification task, but smaller than the reverse for localization, which suggests that the three-way interaction between congruity, attended feature, and task (iden tification versus localization) should be significant. And also as in Experiment 1, a three-way ANOVA with congruity and attended feature as withinsubjects factors, and task as a between-subjects factor confirmed that the three-way interaction was significant, F(1, 206) = 51.83, p < .001, ηp2 = .20. Discussion The results from Experiment 2b echoed those from Experiment 1b, with significant main effects of both congruity and attended feature; as in Experiment 1b, the main effect of attended feature suggests that target selection is more efficient on the basis of color than word. Also, the inversion of the classic Stroop asymmetry resulting from manipulating the task between Experiments 2a and 2b, as indicated by the three-way interaction of congruity, attended feature, and task, fully supports our second hypoth esis that the results from a convenience sample of undergraduate students would generalize to a more representative sample.

General Discussion Replications of the Stroop effect are vastly more common in the Stroop literature than replications of the reverse Stroop effect (Blais & Besner, 2006; Diaz-Piedra et al., 2022). Of course, it is impos sible to determine whether the relative scarcity of reported replications indicates that noone

Smith, Smith, and Sobel | Classic Stroop Asymmetry in Online Experiments

is looking for the effect, or that everyone who looks for an effect fails to find one. Even so, the strength-of-association account (Blais & Besner, 2007) suggests that the latter explanation is more likely, primarily because the typical task in Stroop experiments relies on identification of the target features, which is more strongly associated with the verbal feature. In contrast, many studies that report either a reverse Stroop effect by itself (Blais & Besner, 2007; Diaz-Piedra et al., 2022; Yamamoto et al., 2016) or larger reverse Stroop effect than Stroop (Durgin, 2000; Miller et al., 2016; Sobel et al., 2020; Song & Hakoda, 2015; Uleman & Reeves, 1971) diverge from the traditional identification by using a dif ferent task (e.g., scanning, matching, searching, or pointing a gun) that is more strongly associated with visual than verbal processing. Sobel et al. (2020), and more recently, Diaz-Piedra et al. (2022) argued that the strength of association between the task demands and the attended feature is the key for eliciting a robust reverse Stroop effect. In fact, if visual tasks were as common as verbal tasks throughout the long history of Stroop research, perhaps the classic Stroop asymmetry would never have become a classic, because inverted Stroop asymmetries would have been observed just as often as the asymmetry that has become known as the classic. In this project, we aimed to develop online ver sions of Stroop identification and localization tasks, in the hope that a classic Stroop asymmetry for the identification task would validate our script, and an inversion of the classic Stroop asymmetry would support the strength-of-association account. The three-way interaction between congruity, attended feature, and task in Experiment 1 confirmed our first hypothesis, and the three-way interaction in Experiment 2 confirmed our second hypothesis. It would be tempting to conclude from these results that experiments conducted online are just as reli able as those in a laboratory, and researchers can readily generalize their findings obtained from a convenience sample of undergraduate students to the overall population. Nevertheless, we must acknowledge limitations of our project that prevent us from drawing such bold conclusions. Limitations As we acknowledged in the Introduction, online experiments entail sources of variability in com puter hardware, software, and ambient distractions that cannot be eliminated (Anwyl-Irvine et al.,

2021). In addition, students often report that they can more easily pay attention to lectures when they attend class in person than when they connect remotely (Becker et al., 2020), and the same may be true of participation in experiments. These sources of variability, as well as others that we are failing to imagine, seem likely to bias online experiments toward type II error, insofar as the variability might conceal an actual difference between conditions. While this is a limitation of our experiment as well as any other online experiment, in the Conclusion we will emphasize the upside. A second limitation concerns whether the results obtained from a convenence sample general ize to a sample that is more representative of the overall population. We found that the results from a convenience sample in Experiment 1 did indeed generalize to a more representative sample in Experiment 2, but this is inconsistent with previous research showing that the results obtained from a convenience sample diverge unpredictably from more representative samples (Hanel & Vione, 2016; Peterson & Merunka, 2014). Of course, one relevant difference between our study and theirs is that our tasks assessed basic perceptual and attentional processing, whereas their studies exam ined participants’ explicit attitudes. This suggests that certain kinds of findings, such as those that describe basic cognitive mechanisms of perception and attention, may more readily generalize from convenience samples of undergraduate students to the overall population than other kinds of findings, such as those that describe explicit attitudes. The second limitation of our study is that we studied just one kind of cognitive task; before researchers who study cognitive processing in convenience samples can be confident that their results generalize to the overall population, many other cognitive tasks need to be compared across convenience samples and representative samples. Conclusion In our description of the first limitation, we acknowledged that known and unknown sources of variability entailed by online experiments bias them toward type II errors. This suggests that online experiments may be too insensitive to detect subtle effects, but could uniquely suit them for replication studies. If an effect can overcome a bias for a type II error, that would provide persuasive evidence in support of the hypothesis. Furthermore, our results combine with results from other recent studies (Diaz-Piedra et al., 2022; Sobel et al., 2020)

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

143

Classic Stroop Asymmetry in Online Experiments | Smith, Smith, and Sobel

to provide converging evidence supporting the strength-of-association account of the Stroop asymmetry. Our second limitation recognized that our examination of the generalizability from conve nience samples only looked at one kind of cognitive task, but this is nevertheless an important first step. In summary, we believe that our study makes a valu able contribution to the current understanding of the classic Stroop asymmetry.

References American Psychological Association. (2017). Ethical principles of psychologists and code of conduct. http://www.apa.org/ethics/code Anwyl-Irvine, A., Dalmaijer, E. S., Hodges, N., & Evershed, J. K. (2021). Realistic precision and accuracy of online experiment platforms, web browsers, and devices. Behavior Research Methods, 53(4), 1407 1425. https://doi.org/10.3758/s13428-020-01501-5 Anwyl-Irvine, A. L., Massonnié, J., Flitton, A., Kirkham, N., & Evershed, J. K. (2020). Gorilla in our midst: An online behavioral experiment builder. Behavior Research Methods, 52(1), 388–407. https://doi.org/10.3758/s13428-019-01237-x Barnhoorn, J. S., Haasnoot, E., Bocanegra, B. R., & van Steenbergen, H. (2015). QRTEngine: An easy solution for running online reaction time experiments using Qualtrics. Behavior Research Methods, 47(4), 918–929. https://doi.org/10.3758/s13428-014-0530-7 Bausell, R. B., & Li, Y. F. (2002). Power analysis for experimental research: A practical guide for the biological, medical, and social sciences. Cambridge University Press. Bearden, R. L., Asgari, S., Sobel, K. V., & Scoles, M. T. (2021). Manual responses are verbally mediated in Stroop identification tasks. Psi Chi Journal of Psychological Research, 26(2), 121–129. https://doi.org/10.24839/2325-7342.JN26.2.121 Becker, S. P., Breaux, R., Cusick, C. N., Dvorsky, M. R., Marsh, N. P., Sciberras, E., & Langberg, J. M. (2020). Remote learning during COVID-19: Examining school practices, service continuation, and difficulties for adolescents with and without attention-deficit/hyperactivity disorder. Journal of Adolescent Health, 67(6), 769–777. https://doi.org/10.1016/j.jadohealth.2020.09.002 Blais, C., & Besner, D. (2006). Reverse Stroop effects with untranslated responses. Journal of Experimental Psychology: Human Perception and Performance, 32(6), 1345–1353. https://doi.org/10.1037/0096-1523.32.6.1345 Blais, C., & Besner, D. (2007). A reverse Stroop effect without translation or reading difficulty. Psychonomic Bulletin & Review, 14(3), 466–469. https://doi.org/10.3758/BF03194090 Buhrmester, M. D., Talaifar, S., & Gosling, S. D. (2018). An evaluation of Amazon’s Mechanical Turk, its rapid rise, and its effective use. Perspectives on Psychological Science, 13(2), 149–154. https://doi.org/10.1177/1745691617706516 Crump, M. J., McDonnell, J. V., & Gureckis, T. M. (2013). Evaluating Amazon’s Mechanical Turk as a tool for experimental behavioral research. PloS One, 8(3), e57410. https://doi.org/10.1371/journal.pone.0057410 Diaz-Piedra, C., Gianfranchi, E., Catena, A., & Di Stasi, L. L. (2022). Electrophysiological correlates of the reverse Stroop effect: Results from a simulated handgun task. International Journal of Psychophysiology, 175, 32–42. https://doi.org/10.1016/j.ijpsycho.2022.02.006 Durgin, F. H. (2000). The reverse Stroop effect. Psychonomic Bulletin & Review, 7(1), 121–125. https://doi.org/10.3758/BF03210730 Greene, N. R., & Naveh-Benjamin, M. (2022). Online experimentation and sampling in cognitive aging research. Psychology and Aging, 37(1), 72–83. https://doi.org/10.1037/pag0000655 Hanel P. H. P., & Vione K. C. (2016). Do student samples provide an accurate estimate of the general public? PLoS One, 11(12), e0168354. https://doi.org/10.1371/journal.pone.0168354 Henninger, F., Shevchenko, Y., Mertens, U. K., Kieslich, P. J., & Hilbig, B. E. (2021). lab. js: A free, open, online study builder. Behavior Research Methods, 1–18. https://doi.org/10.3758/s13428-019-01283-5 Linzarini, A., Houdé, O., & Borst, G. (2017). Cognitive control outside of conscious

–

SUMMER 2022

awareness. Consciousness and Cognition: An International Journal, 53, 185–193. https://doi.org/10.1016/j.concog.2017.06.014 Loftus, G. R., & Masson, M. E. J. (1994). Using confidence intervals in withinsubject designs. Psychonomic Bulletin & Review, 1(4), 476–490. https://doi.org/10.3758/BF03210951 MacLeod, C. M. (2005). The Stroop Task in Cognitive Research. In A. Wenzel & D. C. Rubin (Eds.), Cognitive methods and their application to clinical research. (pp. 17–40). American Psychological Association. https://doi.org/10.1037/10870-002 Melara, R. D., & Algom, D. (2003). Driven by information: A tectonic theory of Stroop effects. Psychological Review, 110(3), 422–471. https://doi.org/10.1037/0033-295X.110.3.422 Miller, H. C., Kubicki, S., Caffier, D., Kolski, C., & Naveteur, J. (2016). The Stroop and reverse Stroop effects as measured by an interactive tabletop. International Journal of Human-Computer Interaction, 32(5), 363–372. https://doi.org/10.1080/10447318.2016.1150642 Newman, A., Bavik, Y. L., Mount, M., & Shao, B. (2021). Data collection via online platforms: Challenges and recommendations for future research. Applied Psychology: An International Review, 70(3), 1380–1402. https://doi.org/10.1111/apps.12302 Parris, B. A., Sharma, D., Weekes, B. S. H., Momenian, M., Augustinova, M., & Ferrand, L. (2019). Response modality and the Stroop task: Are there phonological Stroop effects with manual responses? Experimental Psychology, 66(5), 361–367. https://doi.org/10.1027/1618-3169/a000459 Peterson, R. A., & Marunka, D. R. (2014). Convenience samples of college students and research reproducibility. Journal of Business Research, 67(5), 1035–1041. https://doi.org/10.1016/j.jbusres.2013.08.010 Smith, P. L., & Sewell, D. K. (2013). A competitive interaction theory of attentional selection and decision making in brief, multielement displays. Psychological Review, 120(3), 589–627. https://doi.org/10.1037/a0033140.supp Sobel, K. V., Puri, A. M., & York, A. K. (2020). Visual search inverts the classic Stroop asymmetry. Acta Psychologica, 205. https://doi.org/10.1016/j.actpsy.2020.103054 Song, Y., & Hakoda, Y. (2015). An fMRI study of the functional mechanisms of Stroop/reverse-Stroop effects. Behavioural Brain Research, 290, 187–196. https://doi.org/10.1016/j.bbr.2015.04.047 Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18(6), 643–662. https://doi.org/10.1037/h0054651 Sugg, M. J., & McDonald, J. E. (1994). Time course of inhibition in color-response and word-response versions of the Stroop task. Journal of Experimental Psychology: Human Perception and Performance, 20(3), 647–675. https://doi.org/10.1037/0096-1523.20.3.647 Uleman, J. S., & Reeves, J. (1971). A reversal of the Stroop interference effect, through scanning. Perception & Psychophysics, 9(3-A), 293–295. https://doi.org/10.3758/BF03212651 Virzi, R. A., & Egeth, H. E. (1985). Toward a translational model of Stroop interference. Memory & Cognition, 13(4), 304–319. https://doi.org/10.3758/BF03202499 Whitehead, P. S., Brewer, G. A., Patwary, N., & Blais, C. (2018). Contingency learning is reduced for high conflict stimuli. Acta Psychologica, 189, 12–18. https://doi.org/10.1016/j.actpsy.2016.09.002 Wolfe, J. M., & Horowitz, T. S. (2004). What attributes guide the deployment of visual attention and how do they do it? Nature Reviews Neuroscience, 5, 1–7. https://doi.org/10.1038/nrn1411 Yamamoto, N., Incera, S., & McLennan, C. T. (2016). A reverse Stroop task with mouse tracking. Frontiers in Psychology, 7, 670. https://doi.org/10.3389/fpsyg.2016.00670 Author Note. Mary E. Smith is now at the Department of Psychology, University of Memphis. This research was supported by an Undergraduate Research Grant from Psi Chi at the University of Central Arkansas. We have no known conflict of interest to disclose. Correspondence concerning this article should be addressed to Kenith V. Sobel, Department of Psychology and Counseling, University of Central Arkansas, Conway, AR 72035, United States. Email: ksobel@uca.edu

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

144

https://doi.org/10.24839/2325-7342.JN27.2.145

Music’s Impact on the Sexualization of Black Bodies: Examining Links Between Hip-Hop and Sexualization of Black Women Elizabeth A. Otto, Shaina A. Kumar, and David DiLillo* Department of Psychology, University of Nebraska-Lincoln

ABSTRACT. The pervasiveness of sexualization in Western societies is harmful to women, regardless of racial or ethnic identity. However, predictors of sexualization among Black women are understudied. To address this gap, we examined whether listening to and liking hip-hop music would each independently relate to the sexualization of Black women in everyday life, and if this relation unfolded through greater exposure to objectification of Black women in music. A sample of 215 college students completed self-report questionnaires that assessed preferences for liking and listening to hip-hop music, exposure to objectification of Black women in music, and biases toward sexualizing Black women in everyday life. Results revealed that more exposure to objectification of Black women in music mediated the relation between increased listening to hip-hop music and greater sexualization of Black women in everyday life, B = 0.08, SE = 0.03, 95% CI [0.03, 0.13]. Similarly, the link between liking hip-hop music and sexualization of Black women in everyday life was mediated by exposure to objectification of Black women in music, B = 0.09, SE = 0.03, 95% CI [0.04, 0.15]. Results provide an initial step in understanding how preferences toward hip-hop music and exposure to objectification of Black women in music contributes to sexualization of Black women. Future research should continue contributing to conversations that challenge the hypersexualization of Black women.

DIVERSITY

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

Keywords: Black women, hip-hop, music, objectification, sexualization

he widespread occurrence of sexualization in Western societies creates an environment in which the female body is under constant scrutiny, resulting in negative outcomes for women, such as anxiety, internalization of objectified views, and body shame (American Psychological Association [APA] Task Force, 2010; Fredrickson & Roberts, 1997; Szymanski et al., 2011). Although sexualization impacts the lives of all women regardless of racial or ethnic identity, most research *Faculty mentor

related to sexualization has focused on White women (APA, 2010; Heimerdinger-Edwards et al., 2011). This is concerning given that Black women are sexualized as much or more as their White counterparts (Heimerdinger-Edwards et al., 2011), yet factors that influence their sexualization are largely understudied (Heimerdinger-Edwards et al., 2011; Moradi, 2011). The present study addressed this gap by examining two potential predictors of the sexualization of Black women, in addition to a

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

145

Hip-Hop and Sexualization of Black Women | Otto, Kumar, and DiLillo

mediator of this relation. Specifically, we examined whether listening to and liking hip-hop music would each independently relate to the sexualization of Black women in everyday life through exposure to the objectification of Black women in music. Sexualization of Women The APA (2010) defines sexualization as occurring when: (a) a person’s value is reduced to their sexual appeal or behavior, (b) a person’s attractiveness is held on a high pedestal, (c) a person is reduced to an object meant for personal pleasure, and/or (d) sexuality is imposed on a person. Women are constantly exposed to sexualizing images through virtually every media form (e.g., product advertis ing, movies, music) and interpersonal interactions, such as parental pressures to strive toward beauty (APA, 2010). The pervasiveness of women’s sexual ization makes it difficult for them to live their lives without hearing or viewing sexualizing materials; for instance, hip-hop music consistently sexualizes women by reducing them to sexualized body parts (Conrad et al., 2009; Zillman et al., 1995). An empirical study by Kilster and Lee (2009) also found that men who were exposed to hip-hop music with high sexual content expressed greater likelihood of sexually objectifying women. This indicates that, when combined with various societal factors, such as the media’s negative portrayal of women, sexualiza tion becomes a problematic system that can lead to a multitude of negative outcomes for women. Being exposed to instances of sexual objec tification may lead women to experience more anxiety than men (Fredrickson & Roberts, 1997; Swim et al., 2001). Not knowing how or when they will experience objectification creates an added layer of physical safety concerns for many women (Calogero, 2004; Watson et al., 2012). Moreover, sexual objectification is linked to body shame, which results in part from Western society’s idealization of thin bodies (Buchanan et al., 2008; Fredrickson & Roberts, 1997; Kozee et al., 2007; Watson et al., 2012). These ideals leave many women excluded, resulting in perceived unattractiveness and inter nalization of the objectified gaze (Calogero, 2004; Heimerdinger-Edwards et al., 2011; Kozee et al., 2007; Watson et al., 2012).

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

146

Sexualization of Black Women Black women experience as much sexualization as their White counterparts (Heimerdinger-Edwards et al., 2011), yet factors that influence their sexualization are largely understudied (Anderson

et al., 2018; Heimerdinger-Edwards et al., 2011; Moradi, 2011; Watson et al., 2012; Watson et al., 2015). Researchers have predominantly stud ied societal sexualization related to women of European descent, with Black women excluded entirely or included only as a means of comparison (Heimerdinger-Edwards et al., 2011; Watson et al., 2012). This lack of inclusion is problematic because Black women may be sexualized in ways that are different than women of other ethnic groups (Watson et al., 2012). For instance, unlike their White counterparts, Black women are expected to conform to two seemingly incompatible body types simultaneously. They are expected to be curvy and shaped like an hourglass, while still conforming to societal expectation of thinness (Watson et al., 2012). This emphasis on the body shape of Black women indicates that their sexualization may be influenced by unique factors. This, as well as other potential differences in Black women’s sexualiza tion, may contribute to negative outcomes that are unique to their racial group. For example, Black women experience more fear of crimes, such as rape, than women of other racial groups (Callanan, 2012; Watson et al., 2015), which may result in their being more hyperaware of experiences of objectification (Watson et al., 2012). Internalized sexual objectification also leads some women of color to view their sexuality as one of their only assets (Szymanski et al., 2011). History of Sexualization of Black Women Sexualization and objectification of Black women predominately has roots in slavery, a time in which Black women were reduced to objects by slave own ers who controlled their reproductive rights in an effort to further their mistreatment (Wallace et al., 2011). As a result, generations of commodification and auctioning of Black women based on their ability to bear children reduced them to sexual ized body parts (Dagbovie-Mullins, 2013). During the period of U.S. colonialism and slavery, several stereotypical images of Black women also emerged: the promiscuous, light-skinned Jezebel; the asexual, dark-skinned Mammy, and the domineering and aggressive Sapphire, among others (Thomas et al., 2004; Wallace et al., 2011; Watson et al., 2012). Together, the Jezebel, Mammy, and Sapphire con tinue to diminish Black women to stereotypic roles. The Jezebel stereotype has particularly been used to sexually objectify and exploit Black women by marking them as hypersexualized (DagbovieMullins, 2013). By unfairly viewing them as

Otto, Kumar, and DiLillo | Hip-Hop and Sexualization of Black Women

responsible for others’ sexual behavior toward them, this stereotype results in beliefs that Black women cannot be raped (Dagbovie-Mullins, 2013). Today, the Jezebel has further transformed into various stereotypes, such as the Diva, who chooses men based on their social status; the Gold Digger, who barters her sexuality for material wealth; and the Freak, who seeks to satisfy her own physical desires (Stephens & Phillips, 2005; Watson et al., 2012). The prevalence of these damaging images can result in numerous consequences for Black women, including self-sexualization and selfobjectification (Wallace et al., 2011). One way that Black women attempt to protect themselves from negative sexualized stereotypes is through their perception of and engagement in sexual activities. Although some Black women who internalize the hypersexual representation of the Jezebel stereotype aim to take ownership of their sexuality through sexual exploration, most exhibit increased hypervigilance about embracing their sexual desires or engaging in sexual behaviors (Leath et al., 2021). Additionally, many Black women alter the way they dress and consequently feel the need to present themselves more conservatively (Wilson & Russel, 1996). These are just two examples among many of how the Jezebel stereotype severely affects the way Black women present themselves. Hip-Hop Music and the Objectification of Black Women The origins of hip-hop can be found in slave spirituals, blues, jazz, and soul music, which were all meant to articulate emotional and physical hard ships (Conrad et al., 2009; Zillmann et al., 1995). Later, artists continued to extend this tradition by promoting social justice through its lyrics (Zillmann et al., 1995). However, since its appearance in mainstream media during the 1980s, hip-hop artists began expressing a greater number of controversial messages with regard to sex, violence, drugs, mate rial wealth, and the sexualization and maltreatment of Black women (Conrad et al., 2009; Peterson, et al., 2007; Zillmann et al., 1995). The pervasiveness of these controversial themes in hip-hop music, in particular the sexualization of Black women, is problematic, given that hip-hop music and its subgenres have grown to extreme popularity and are enjoyed by fans across gender and racial boundaries. Indeed, on the 2019 Billboard Year-End Chart, 46 hip-hop songs made the Billboard Hot 100 Songs and 108 hip-hop albums made the Billboard 200 Albums, six of which earned a spot in the top

10 (Billboard, 2020a; Billboard, 2020b). Hip-hop music has portrayed Black women in sexually stereotyped ways for decades (Conrad et al., 2009; Peterson et al., 2007) and, in more recent years, artists have been accused of spreading greater misogynistic and degrading messages to listeners (Pough, 2015). Due to hip-hop music’s tendency to be associated with Black male identity and masculin ity, Black women are often depicted as overly sub missive (Conrad et al., 2009) and perpetuating the Jezebel stereotype (Conrad et al., 2009; Peterson et al., 2007; Zillmann et al., 1995). Consequently, Black women fall victim to stereotypes present in hip-hop music that deem them to be hypersexual, lacking in morals, and/or extremely materialistic (Peterson et al., 2007). For example, lyrics to popu larized songs such as “Birthday Song” by 2 Chainz, “Candy Shop” by 50 Cent, and “Tip Drill” by Nelly include conversations about sexualizing women’s body parts at length, while intentionally describing women’s actions in a manner that depicts them as provocative and hypersexual. In addition, the music videos that accompany hip-hop songs also play an important role in perpetuating negative sexual stereotypes toward Black women. Indeed, Black women are commonly portrayed as less than human and objectified, such that they are seen dancing and posing provocatively while wearing swimsuits and lingerie. These videos exemplify how hip-hop music likely plays a part in reinforcing the belief that Black women should be viewed as hypersexual commodities. Exposure to hip-hop music with high sexual content has been found to increase sexual objec tification of women, sexual permissiveness, and acceptance of rape across racial groups (Kilster & Lee, 2009). For example, Stephens and Phillips (2005) found that White college students were more likely to perceive Black women in a negative manner following exposure to hip-hop music. Similarly, in a qualitative study that examined how White, Hispanic, and Asian American individuals perceived Black women following exposure to hip-hop music, researchers found a salient theme related to Black women being perceived as sexual objects for men’s desires (Jacobson, 2015). Such findings indicate that hip-hop music’s impact is pervasive enough to create a negative image of Black women throughout society, and thus all racial groups should be considered when examining biases toward the sexualization of Black women. As such, the present study’s sample included a variety of racial groups in order to test proposed

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

147

Hip-Hop and Sexualization of Black Women | Otto, Kumar, and DiLillo

hypotheses more broadly. Because hip-hop is a diverse genre, listening to and liking this type of music does not inherently mean that all consumers are exposed to the prob lematic themes listed above. Rather, because the prevalence of sexualization of Black women in hiphop music is high (Conrad et al., 2009; Peterson et al., 2007), those who listen to and like hip-hop music might engage with content that objectifies Black women more often, and this may in turn encour age increased sexualization of Black women more generally. Prior empirical work supports this belief, as researchers have found that exposure to Black media, including hip-hop, is directly linked to stron ger endorsement, internalization, and identification with sexualized images of Black women in everyday life (Conrad et al., 2009; Gordon, 2008; Peterson et al., 2007; Szymanski et al., 2011; Wingood et al., 2003). Furthermore, due to the widespread popular ity of hip-hop music, listening to and liking hip-hop music might be distinct constructs. As with all forms of media, individuals can consume hip-hop content in daily life without necessarily liking the genre (e.g., on the radio, in television advertisements). Thus, for the current study, we aimed to disentangle individual preferences related to listening to and liking hip-hop music and their relation with the objectification and sexualization of Black women. The Present Study In sum, prior literature has suggested that listening to and liking hip-hop music, as well as exposure to objectification of Black women through music, are likely to lead to sexualization of Black women in everyday life. The primary aim of the present study was to test these possibilities by examining exposure to objectification of Black women through music as a mediator of the association between listening to and liking hip-hop music and sexualization of Black women. First, we expected that individuals who listen to hip-hop music would be more likely to endorse sexualization of Black women in everyday life through greater exposure to objectification of Black women in music (H1). Second, we similarly predicted that individuals who like hip-hop music would be more likely to endorse sexualization of Black women in everyday life through greater exposure to objectification of Black women in music (H2). SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

148

Method Participants Participants were 168 women and 46 men

undergraduate students, in addition to one under graduate student who identified as “other” with regard to their gender. Participants ranged in age from 17 to 31 years old (Mage = 20.56, SD = 2.50). Among participants, 70.7% identified as European American/White (n = 152), 11.2% as Asian/Pacific Islander (n = 24), 8.4% as Hispanic/Latinx (n = 18), 2.8% as African American/Black (n = 6), and 7.0% as “other” (n = 15). Measures Demographics Participants completed a demographics measure including questions regarding their gender, age, and race/ethnicity. Music Preferences Participants were asked to respond to questions about (a) how often they listened to and (b) how much they liked the following genres: country, pop, hip-hop, rap, R&B, reggae, rock, alternative rock, indie, EDM, metal/heavy metal, and folk. Response options were rated using a 5-point scale (listened to question: 1 = I don’t listen to it, 5 = I listen all the time; liked question: 1 = I hate it, 5 = I love it). For the purpose of the current study, responses regarding genres related to hip-hop, rap, and R&B were averaged to create our predictor variables (i.e., how often participants listened to hip-hop and how much participants liked hip-hop). To note, we collapsed across the hip-hop, rap, and R&B music genres given their conceptual overlap. Specifically, hip-hop is considered a broad movement from which rap has originated, such that rap is commonly viewed as part of hip-hop music and culture (Next Level, 2018). Further, in recent years, contemporary R&B has also been known to borrow from hip-hop and rap inspired instrumen tals and themes, such as the albums Trapsoul by Bryson Tiller and The Miseducation of Lauryn Hill by Lauryn Hill (New World Encyclopedia, 2015). Thus, the hip-hop, rap, and R&B music genres were consolidated based on these similarities. Objectification of Women in Music Participants were asked to report how often the music they listen to is objectifying toward Black women using the 9-item Perception of Portrayals of Sexual Stereotypes in Rap Music Questionnaire (Peterson et al., 2007). Example items include, “Are Black women treated disrespectfully by men?” and “Are Black women portrayed as sex objects?” Response options were coded on a 6-point scale (0 = never, 5 = always). Items were averaged to create

Otto, Kumar, and DiLillo | Hip-Hop and Sexualization of Black Women

a total score of exposure to the objectification of Black women in music as our mediator variable. This measure appears to be face valid (Peterson et al., 2007), and Cronbach’s alpha for the Perception of Portrayals of Sexual Stereotypes in Rap Music Questionnaire in the current sample suggested internal consistency was excellent (α = .92). Sexualization of Black Women The 7-item modern Jezebel subscale of the Stereotypic Role for Black Women Scale (SRBWS) was utilized to measure sexualization of Black women in everyday life (Thomas et al., 2004; Townsend et al., 2010). Example items include, “Black girls always want to have sex,” and “Black girls use sex to get what they want.” All items were rated on a 5-point scale (1 = strongly disagree, 5 = strongly agree) and then averaged to create a total score of sexualization of Black women in everyday life as our outcome variable. Past work has sup ported the validity of this measure (see Thomas et al., 2004, and Townsend et al., 2010). Cronbach’s alpha for the modern Jezebel subscale in the cur rent sample suggested internal consistency was excellent (α = .93). Procedures Participant recruitment took place at the University of Nebraska-Lincoln. Participants were recruited using a mixture of flyers across campus, classroom recruitment, and the Psychology Department’s undergraduate research pool system. After provid ing informed consent, participants were prompted to complete self-report questionnaires via Qualtrics using a personal computer at a location of their choosing. Participants who completed this survey through the Psychology Department’s research pool system were rewarded with research credit for their courses, and all other participants were entered into a raffle for a $45 Amazon Gift Card upon completion of the survey. All procedures were approved by the UNL Institutional Review Board.

Results

women in music (M = 1.59; SD = 1.09) and minimal sexualization of Black women in everyday life (M = 1.91; SD = 0.85). Further, bivariate correlations estimated using Pearson’s r formula indicated that listening to hip-hop music (r = –.01, p = .86) and liking hip-hop music (r = .07, p = .29) were not sig nificantly related to sexualization of Black women in everyday life. However, listening to hip-hop music (r = .45, p < .001) and liking hip-hop music (r = .41, p < .001) were positively related to exposure to music that objectified Black women. Finally, expo sure to music that objectified Black women was positively related to sexualization of Black women in everyday life (r = .22, p = .001). Given the gender breakdown in the current study, we also examined whether gender was cor related with our variables of interest to determine its inclusion as a covariate. Gender was not associ ated with listening to hip-hop music, liking hip-hop music, exposure to music that objectified Black women, or sexualization of Black women (ps > .05). Thus, gender was not included as a covariate in any further analyses. Path Analyses: Objectification of Black Women in Music as a Mediator Hip-Hop Listening Using Hayes’ PROCESS macro for SPSS (Hayes, 2017), we first conducted a path analysis to exam ine relations among hip-hop listening, exposure to objectification of Black women in music, and sexualization of Black women in everyday life. Here, we employed Hayes’ (2017) mediation criteria such that the inference of indirect effects is based on the product of paths a and b (i.e., ab) as opposed to hypothesis tests of a and b separately. Through this approach, a bias-corrected bootstrap technique with 5,000 resamples was used to derive the 95% CIs TABLE 1 Descriptive Statistics and Correlations Hip-Hop Listening

Hip-Hop Liking

Hip-Hop Listening

Descriptives and Bivariate Correlations Descriptive information and bivariate correlations among study variables are displayed in Table 1. All variables were within acceptable limits of skew (< |3|) and kurtosis (< |10|; Kline, 2015). On aver age, participants reported moderately listening to hip-hop music (M = 3.28; SD = 1.16) and liking hip-hop music (M = 3.49; SD = 1.01). Participants reported some exposure to objectification of Black

Objectification of Sexualization of Black Women in Music Black Women

–

Hip-Hop Liking

.88***

–

Objectification of Black Women in Music

.45

.41

–

Sexualization of Black Women

.07

–.01

.22*

–

3.28

3.50

1.59

1.92

1.17

1.01

1.09

0.85

Observed Score Range

1.00–5.00

0.00–3.89

1.00–4.71

***

Note. N = 215. p < .01. p < .001. *

***

149

Hip-Hop and Sexualization of Black Women | Otto, Kumar, and DiLillo

for direct and indirect effects in the model. This non-parametric resampling method accounts for nonnormal distribution of data and performs well in small samples by maximizing power and minimiz ing Type I error rate (Shrout & Bolger, 2002). As shown in Figure 1, results indicated a non significant direct path between listening to hip-hop music and sexualization of Black women, B = –0.02, SE = 0.05, 95% CI [–0.13, 0.08]. However, results indicated a significant indirect effect between listen ing to hip-hop music and sexualization of Black women through exposure to objectification of Black women in music, B = 0.08, SE = 0.03, 95% CI [0.03, 0.13]. These results support our first hypothesis and suggest that exposure to objectification of Black FIGURE 1 Relations Among Listening to Hip-Hop Music, Exposure to Objectification of Black Women in Music, and Sexualization of Black Women in Everyday Life Indirect Effect: B = 0.08, 95% CI [0.03, 0.13]

Objectification of Black women in music

b-path: B = 0.18, 95% CI [0.07, 0.29]

Hip-hop listening

Sexualization of Black women

Note. Nonsignificant effects are dashed and significant effects are solid. All estimates are unstandardized.

FIGURE 2 Relations Among Liking Hip-Hop Music, Exposure to Objectification of Black Women in Music, and Sexualization of Black Women in Everyday Life Indirect Effect: B = 0.09, 95% CI [0.04, 0.15]

Objectification of Black women in music

a-path: B = 0.45, 95% CI [0.32, 0.56]

Hip-hop liking

b-path: B = 0.21, 95% CI [0.10, 0.32]

c' Direct Effect: B = −0.10, 95% CI [−0.23, 0.02]

Sexualization of Black women

Note. Nonsignificant effects are dashed and significant effects are solid. All estimates are unstandardized.

150

Hip-Hop Liking Next, we conducted a second path analysis to examine relations among hip-hop liking, exposure to objectification of Black women in music, and sexualization of Black women in everyday life using the same procedures outlined above. As shown in Figure 2, results indicated a nonsignificant direct path from liking hip-hop music to sexualizing Black women, B = –0.10, SE = 0.06, 95% CI [–0.23, 0.02]. Further, results indicated a significant indirect effect between liking hip-hop music and sexualization of Black women through exposure to objectification of Black women in music, B = 0.09, SE = 0.03, 95% CI [0.04, 0.15]. These results support our second hypothesis and suggest that exposure to objectification of Black women in music is a potential mechanism of the relation between liking hip-hop music and explicit sexualization of Black women in everyday life (H2).

Discussion

a-path: B = 0.42, 95% CI [0.32, 0.52]

c' Direct Effect: B = −0.02, 95% CI [−0.13, 0.08]

women in music is a potential mechanism of the relation between listening to hip-hop music and explicit sexualization of Black women in everyday life (H1).

The goal of this study was to examine relations among individual preferences for hip-hop music and sexualization of Black women in everyday life. Our primary hypotheses that listening to and liking hip-hop music would each be related to greater self-reported sexualization of Black women through exposure to objectification of Black women in music were supported. We discuss these findings below. Consistent with prior work (Conrad et al., 2009; Peterson et al., 2007; Stephens & Phillips, 2005), our results show that hip-hop music preferences are indirectly related to greater sexualization of Black women. However, we extend this work by illustrating that exposure to the objectification of Black women in music might be one mechanism through which this relation unfolds. Hip-hop music has a history of perpetuating negative sexual stereo types specific to Black women, such as the Jezebel (Conrad et al., 2009; Peterson et al., 2007; Zillmann et al., 1995), and thus it is possible that individuals who both listen to and like hip-hop music, and are then exposed to objectification of Black women in music, become influenced to view Black women through a sexualized lens. The lack of a significant direct link among listening to and liking hip-hop music and sexualiza tion of Black women may be due to the diversity

Otto, Kumar, and DiLillo | Hip-Hop and Sexualization of Black Women

of music covered by the hip-hop genre. That is, while hip-hop has been identified as a genre that commonly sexualizes Black women, the degree to which these women are sexualized varies greatly across artists, songs, and subgenre of hip-hop (e.g., 90s hip-hop, mumble rap, trap R&B). Therefore, reports of listening to or liking hip-hop more will not necessarily relate to sexualization of Black women in everyday life directly; rather, this link becomes particularly salient as individuals are more exposed specifically to music that objectifies Black women, which supports our initial hypotheses. Our finding of an indirect path to sexualization of Black women from listening to and liking hip-hop music suggests that greater exposure to music that objectifies Black women may reinforce traditional gender attitudes and negative sexual stereotypes of Black women (Gordon, 2008; Peterson et al., 2007). Limitations and Future Directions Although the current study contributes to prior knowledge about hip-hop music preferences and sexualization of Black women, its findings should be viewed in light of several limitations. First, the cross-sectional design precludes conclusions about causation and the temporal order of variables in our model. Specifically, it is possible that listening to and liking hip-hop could also mediate relations between exposure to objectification of Black women in music and sexualization of Black women in everyday life, although the proposed ordering of our variables is supported by prior empirical work (Conrad et al., 2009; Gordon, 2008; Peterson et al., 2007; Szymanski et al., 2011; Wingood et al., 2003). Further exami nation of hip-hop music preferences, exposure to objectification of Black women in music, and biases toward the sexualization of Black women in the context of a longitudinal study is needed. Second, given the focus of the current study and the possibility that agreement with some items on our measures might appear problematic (e.g., “Black girls always want to have sex”), self-report bias could have impacted our findings. That is, par ticipants (especially White women and men) might have wanted to avoid appearing biased or racist and adhered to social desirability norms. Thus, it might be beneficial to explore other ways of measuring sexual objectification in combination with validated self-report measures. Using eye-tracking technol ogy, for instance, would allow researchers to map the visual behaviors of individuals who engage in hip-hop culture and listen to its music as they view images of Black women versus women from other

racial groups. Indeed, it is possible that utilizing additional behavioral measurement methods might promote less biased responding and better represent everyday actions. Third, even though hip-hop music and its subgenres appeal to a wide variety of individuals, our sample of minority participants, especially Black individuals, was small. Although significant effects were found in our primary models of interest, and our results align with findings that White college students are more likely to perceive Black women in a negative manner after exposure to hip-hop music (Stephens & Phillips, 2005), the lack of diversity in our sample creates an inability to fully generalize these results to other populations. Future researchers might consider examining a sample recruited from other regions of the United States to attain greater diversity, and thus provide a better understanding of hip-hop music’s impact on Black women’s sexualization across racial groups. This need to examine group differences is underscored by research that shows exposure to hip-hop music with high sexual content increases sexual objectification of women, sexual permissiveness, and acceptance of rape among a diverse sample of individuals (Kilster & Lee, 2009). Hip-hop music’s impact seems to be pervasive enough to create a particularly negative image of Black women throughout society, and thus all racial groups should be considered when examining biases toward the sexualization of Black women. Finally, to further address gaps in the literature, future researchers should also focus on collecting data from Black participants, including Black women, in order to better determine how the proposed relations unfold among Black individuals and whether unique internalization of sexualized stereotypes impact findings. Indeed, it would be valu able to explore Black women’s perceptions as a result of their objectification and, given that many hip-hop artists and performers are Black women, whether the perpetuation of negative sexualizing stereotypes by other Black women influence their views. Finally, the current study only examined prefer ences toward hip-hop music with regard to listening or liking the music. One element not examined is how much of the music is being consumed in a video format, which offers an additional mode in which to objectify women visually. It is possible that individuals who watch, not only listen, to music that objectifies Black women will be even more likely to sexualize Black women in everyday life. Future researchers should aim to examine how much objectifying content is being portrayed (and

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

151

Hip-Hop and Sexualization of Black Women | Otto, Kumar, and DiLillo

consumed) through music videos versus only listen ing to lyrics in order to identify how consumption of different forms of sexualizing media may influence the sexualization of Black women in society. Conclusion In sum, our findings indicate that hip-hop music as a genre does not necessarily objectify Black women as a whole. Rather, those who listen to hip-hop music that includes greater objectifying themes may come to hold greater biases toward sexualizing Black women in daily life. To counteract these biases, there needs to be greater availability of popular media that presents Black women in a nonsexualized manner should our findings hold true. Particularly, greater effort needs to be made to create and disseminate music and other media that do not portray Black women as primarily objects of sexual attraction. In addition, results suggest there might be a need to make hip-hop artists aware of their active contribu tion to the hypersexualization of Black women through the content of the music they produce. Although monetary gain likely drives much of the sexualization of Black women in hip-hop and related music, we nonetheless recommend informing artists about their impact to raise awareness and begin to decrease the focus on sexual objectification in this group. It is critical to continue contributing to a conversation that will challenge the objectification and sexualization of Black women in everyday life by raising awareness of these phenomena.

References

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

152

American Psychological Association. (2010). Report of the American Psychological Association Task Force on the sexualization of girls. https://www.apa.org/pi/women/programs/girls/report 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 Billboard. (2020a). Year-end charts: Billboard 200 albums. https://www.billboard.com/charts/year-end/2019/top-billboard-200-albums Billboard. (2020b). Year-end charts: Hot 100 songs. https://www.billboard.com/charts/year-end/2019/hot-100-songs Buchanan, T. S., Fischer, A. R., Tokar, D. M., & Yoder, J. D. (2008). Testing a culturespecific extension of objectification theory regarding African American women’s body image. The Counseling Psychologist, 36(5), 697–718. https://doi.org/10.1177/0011000008316322 Callanan, V. J. (2012). Media consumption, perceptions of crime risk and fear of crime: Examining race/ethnic differences. Sociological Perspectives, 55(1), 93–115. https://doi.org/10.1525/sop.2012.55.1.93 Calogero, R. M. (2004). A test of objectification theory: The effect of the male gaze on appearance concerns in college women. Psychology of Women Quarterly, 28(1), 16–21. https://doi.org/10.1111/j.1471-6402.2004.00118.x Conrad, K., Dixon, T. L., & Zhang, Y. (2009). Controversial rap themes, gender portrayals and skin tone distortion: A content analysis of rap music videos. Journal of Broadcasting & Electronic Media, 53(1), 134–156. https://doi.org/10.1080/08838150802643795 Dagbovie-Mullins, S. A. (2013). Pigtails, ponytails, and getting tail: The infantilization and hyper-sexualization of African American females in

popular culture. The Journal of Popular Culture, 46(4), 745–771. https://doi.org/10.1111/jpcu.12047 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 Gordon, M. K. (2008). Media contributions to African American girls’ focus on beauty and appearance: Exploring the consequences of sexual objectification. Psychology of Women Quarterly, 32(3), 245–256. https://doi.org/10.1111/j.1471-6402.2008.00433.x Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). Guilford. Heimerdinger-Edwards, S. R., Vogel, D. L., & Hammer, J. H. (2011). Extending sexual objectification theory and research to minority populations, couples, and men. The Counseling Psychologist, 39(1), 140–152. https://doi.org/10.1177/0011000010383894 Jacobson, G. (2015). Racial formation theory and systemic racism in hip-hop fans’ perceptions. Sociological Forum, 30(3), 832–851. https://doi.org/10.1111/socf.12186 Kilster, M. E., & Lee, M. J. (2009). Does exposure to sexual hip-hop music videos influence the sexual attitudes of college students? Mass Communication and Society, 13(1), 67–66. https://doi.org/10.1080/15205430902865336 Kline, R. B. (2015). Principles and practice of structural equation modeling (4th ed.). Guilford. Kozee, H. B., Tylka, T. L., Augustus-Horvath, C. L., & Denchik, A. (2007). Development and psychometric evaluation of the interpersonal sexual objectification scale. Psychology of Women Quarterly, 31(2), 176–189. https://doi.org/10.1111/j.1471-6402.2007.00351.x Leath, S., Jerald, M. C., Perkins, T., & Jones, M. K. (2021). A qualitative exploration of Jezebel stereotype endorsement and sexual behaviors among Black college women. Journal of Black Psychology, 47(4–5), 244–283. https://doi.org/10.1177/0095798421997215 Moradi, B. (2011). Objectification theory: Areas of promise and refinement. The Counseling Psychologist, 39(1), 153–163. https://doi.org/10.1177/0011000010384279 New World Encyclopedia. (2015). Rhythm and blues. https://www.newworldencyclopedia.org/entry/rhythm_and_blues Next Level. (2018). Rap vs. hip hop. https://www.nextlevel-usa.org/blog/rap-vs-hip-hop Peterson, S. H., Wingood, G. M., DiClemente, R. J., Harrington, K., & Davies, S. (2007). Images of sexual stereotypes in rap videos and the health of African American female adolescents. Journal of Women’s Health, 16(8), 1157–1164. https://doi.org/10.1089/jwh.2007.0429 Pough, G. D. (2015). Check it while I wreck it: Black womanhood, hip-hop culture, and the public sphere. Northeastern University Press. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-experimental studies: New procedures and recommendations. Psychological Methods, 7(4), 422–445. https://doi.org/10.1037/1082-989X.7.4.422 Stephens, D. P., & Phillips, L. (2005). Integrating black feminist thought into conceptual frameworks of African American adolescent women’s sexual scripting processes. Sexualities, Evolution & Gender, 7(1), 37–55. https://doi.org/10.1080/14616660500112725 Swim, J. K., Hyers, L. L., Cohen, L. L., & Ferguson, M. J. (2001). Everyday sexism: Evidence for its incidence, nature, and psychological impact from three everyday diary studies. Journal of Social Issues, 57(1), 31–53. https://doi.org/10.1111/0022-4537.00200 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 Thomas, A. J., Witherspoon, K. M., & Speight, S. L. (2004). Toward the development of the Stereotypic Role for Black Women Scale. Journal of Black Psychology, 30(3), 426–442. https://doi.org/10.1177/0095798404266061 Townsend, T. G., Neilands, T. B., Thomas, A. J., & Jackson, T. R. (2010). I’m no Jezebel; I am young, gifted, and black: Identity, sexuality, and black girls. Psychology of Women Quarterly, 34(3), 273–285. https://doi.org/10.1111/j.1471-6402.2010.01574.x Watson, L., Marszalek, J., Dispenza, F., & Davids, C. (2015). Understanding the relationships among White and African American women’s sexual objectification experiences, physical safety anxiety, and psychological distress. Sex Roles, 72(3–4), 91–104.https://doi.org/10.1007/s11199-014-0444-y Watson, L. B., Robinson, D., Dispenza, F., & Nazari, N. (2012). African American

Otto, Kumar, and DiLillo | Hip-Hop and Sexualization of Black Women

women’s sexual objectification experiences. Psychology of Women Quarterly, 36(4), 458–475. https://doi.org/10.1177/0361684312454724 Wilson, M., & Russell, K. (1996). Divided sisters: Bridging the gap between Black women and White women. Anchor Books. Wingood, G. M., DiClemente, R. J., Bernhardt, J. M., Harrington, K., Davies, S. L., Robillard, A., & Hook III, E. W. (2003). A prospective study of exposure to rap music videos and African American female adolescents’ health. American Journal of Public Health, 93(3), 437–439. https://doi.org/10.2105/AJPH.93.3.437 Zillmann, D., Aust, C. F., Hoffman, K. D., Love, C. C., Ordman, V. L., Pope, J. Seigler, P. D., Gibson, R. J. (1995). Radical rap: Does it further ethnic division? Basic & Applied Social Psychology, 16(1–2), 1–25. https://doi.org/10.1080/01973533.1995.9646098 Author Note. Elizabeth A. Otto https://orcid.org/0000-0002-6363-6562 Shaina A. Kumar https://orcid.org/0000-0003-1928-646X This study was supported by the University of Nebraska-

Lincoln McNair Scholars Program and the Undergraduate Creative Activities and Research Experience (UCARE) program. None of the authors have any conflicts of interest to disclose. Positionality Statement: Elizabeth Otto identifies as a cisgender Black woman. She is also an immigrant from South Sudan, Africa. Shaina Kumar identifies as a cisgender biracial (White, South Asian) woman. She also identifies as part of the Jewish community. David DiLillo identifies as a cisgender White man. All authors acknowledge that their perspectives are influenced by their positions within all of these dimensions of identity. Correspondence concerning this article should be addressed to Elizabeth A. Otto, Department of Psychology, University of Nebraska-Lincoln, 238 Burnett Hall, Lincoln, NE 68588-0308, United States. Email: ottoelizabeth41@gmail.com.

153

https://doi.org/10.24839/2325-7342.JN27.2.154

False Memory for Words in Noise: An At-Home DeeseRoediger-McDermott (DRM) Experiment Across Adulthood Rebecca L. Wagner1 2, Bethany A. Lyon1*, and Angela AuBuchon2* 1 Department of Psychology, University of Nebraska at Omaha 2 Boys Town Medical Research Hospital

ABSTRACT. The Deese-Roediger-McDermott paradigm is an experimental manipulation that induces false memory creation (Deese, 1959; Roediger & McDermott, 1995). This study presented auditory DRM lists either in silence or in background noise to both young and older adult samples. Background noise significantly reduced participants' recall accuracy, F(1, 93) = 14.14, p < .001, ηp2 = .13, and false memory production, F(1, 93) = 4.02, p = .05, ηp2 = .04, but there were no significant differences between age groups. Moreover, older adults’ self-reported measures of cognitive load correlated with their accuracy, r(47) = –.39. No such relationship was observed in the younger adult group suggesting that, although both groups reached similar levels of performance, they likely approached the task differently. These findings also highlight the need for further examination into how individual differences mediate memory performance in the DRM, especially in older adults. Keywords: cognitive, memory, older adults, free recall, background noise

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

154

emory is fallible. Knowing the conditions under which memory fails highlights points for intervention and contributes to theory of underlying processes that support memory. Typically, memory failures are associated with information that is presented but later forgotten. Another type of memory failure is when information is seemingly “remembered” but was never presented. These occurrences, known as false memories, are difficult to confirm in daily life because they depend on reliable documentation of the presented information. However, false memories are easily observed in laboratory experiments where the presented information is known. One method commonly used to elicit false memories is the Deese-Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger & McDermott, 1995). The DRM uses lists of 12 to 15 highly related words that fall within a semantic network (e.g., nap, rest, lay, dream, awake). The

Open Data and Open Materials badges earned for transparent research practices. Data and materials are available at https://osf.io/gu7aw/

words are typically presented individually with a free recall or recognition test completed immediately afterward. Most importantly, each presented list excludes a “critical lure” word (e.g., sleep), which is closely related to the listed words (e.g., nap, rest, lay). Recall of the critical lure serves as a marker for a false memory provoked by the DRM paradigm (Deese, 1959; Roediger & McDermott, 1995). Many different factors influence the rate of false memory production in the DRM, such as aging, including associated changes in processing approaches, list structure, and more. Aging is associated with decreased recall of presented words and an increase in inclusion of critical lures and other list intrusions (Watson et al., 2005). Changes to both cognition and sensory systems appear to contribute to older adults’ increased susceptibility to false memories. Declines in cognitive functioning, particularly in attention allocation abilities, increase the task difficulty for

*Faculty mentor

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

older adults (Kahneman, 1973). When the DRM list is presented aurally, task difficulty is further increased for individuals with reduced hearing acuity, common among older adult populations. The Role of Suppression Mechanisms in Preventing False Memories Production of false memories in the DRM arises from the inability to suppress inappropriately activated items, the critical lures, during recall of presented lists. Attention and source monitoring, a set of processes that aid in differentiating between original sources of information (Hashtroudi et al., 1990; Johnson et al., 1993), are integral for suppressing these errors in activation elicited by the DRM’s design. Older adults’ reduced access to attentional resources can cause source monitor ing failures during recall, typically resulting in an overall higher recall of words than their younger counterparts, but more of the words are critical lures, outside list intrusions, and repeated words (Dywan & Jacoby, 1990; Watson et al., 2005). In memory, related words, phrases, and con cepts are linked, allowing an entire network of related concepts to be activated by a single word or idea. Typically, this spreading activation is beneficial during memory recollection, because singular memory cues can activate entire networks of information, making that information more readily accessible, rather than needing specific cues for each individual memory (Anderson, 1983). However, the DRM relies on this spreading activation bias in memory to activate critical lures during list recall, creating a false memory which leads to inaccurate recollections (Pierce et al., 2005). Adequate attentional control and source monitoring processes are then critical for inhibition of spreading activation caused by the DRM list pre sentation (Meade et al., 2006; Watson et al., 2005). The effects of spreading activation and need for accurate source monitoring on DRM listmemory are amplified by the changes in processing approaches that occur with aging. Younger adults tend to rely on item-based, or verbatim, processing. In other words, each list item is processed and encoded on its own (Tun et al., 1998). This method is known to require more cognitive resources and, in general, be more effortful to complete (Brainerd & Reyna, 2005). Older adults, on the other hand, tend to rely more on gist-based processing, focusing on commonalities among items rather than the items themselves (Tun et al., 1998). Differences in gist processing and item-based processing align with underpinnings of the fuzzy trace theory

(Reyna & Brainerd, 1998). This theory posits that individuals can create two representations of the same information, one that is exact to the informa tion presented, and one that is the gist, or general idea of the information presented. Reliance on gist processing in a DRM paradigm leads participants to generate false memories, as the critical lure matches the gist of the entire list. Older adults’ preference for gist-based processing, coupled with the noted attentional demands of a DRM task, increase the likelihood that a critical lure will be recalled. List Structure Influences False Memory Production Accompanying age-related changes that facilitate false memory production, the structure of the DRM lists can further increase the likelihood that a critical lure will be recalled. Each DRM list consists of words that are highly related to the unpresented critical lure; however, some lists have greater associative connections to this word compared to others. When this association, called the backward associative strength (BAS), is stronger, the rate of false memory production tends to be higher (Knott et al., 2012). Background Noise Imposes Further Cognitive Demand Presenting auditory memoranda in background noise further interferes with processing and impairs recall by imposing additional cognitive demands on adult listeners (Hintzman, 1988; Peelle, 2018). Background noise reduces perceptual information available to a listener, increasing the chances that presented information is misidentified and stored inaccurately in memory, which would manifest as a memory error (Peelle, 2018). To decrease the likeli hood of an error, attentional resources are diverted away from semantically processing the words to perceptually processing the individual sounds of the words. This results in better identification of list words but decreases the total number of words recalled. Additionally, perceptual processing limits the amount of spreading activation that occurs, which decreases the number of critical lures and list intrusions present in recall (Kjellberg et al., 2008; Marsh et al., 2015; McCoy et al., 2005). Individual Differences in Cognitive Load Cognitive load refers to the combination of cognitive abilities and mental effort that indi viduals allocate toward a task, as well as one’s metacognitive awareness of the demands of that task (Tomporowski, 2002, 2003). Cognitive abilities, including attention allocation, perceptual

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

155

DRM Across Adulthood | Wagner, Lyon, and AuBuchon

processing, memory, and perceptions of task demands, all vary with individuals. Mental effort, or the effort an individual puts toward a task, also varies between individuals, as it has been shown to be affected by past experience, conceptions sur rounding skill level and ability, motivation, and age (Van Gerven et al., 2000). The National Aeronautics and Space Administration Raw Task Load Index (NASA RTLX) has been a validated tool to assess the subdivisions of cognitive load in both younger adults (Longo, 2018), older adults (Devos et al., 2020), and listening-in-noise tasks (Mackersie & Cones, 2011). Closer examination of individual differences in cognitive load using the DRM, with and without background noise, is useful in teasing apart the complex nature of false memories, their formation, and why some are more susceptible to their formation than others. Current Study For the current study, we aimed to assess recall accuracy and false memory production in adults on an at-home, auditory DRM task. Based on previous literature, we expected that adding noise to an auditory DRM task would increase the cognitive load of the task, decreasing both recall accuracy and the number of false memories produced. We anticipated this change in performance to be more pronounced in older adults due to changes in hear ing ability, noted shifts in processing approaches, and additional cognitive loads that are introduced by background noise. TABLE 1 Study Demographics Young Adult Age

Older Adult

M = 23.25, SD = 3.05

M = 68.78, SD = 2.87 Sex

Male

Female

Other

0 Ethnicity

Hispanic or Latino

Not Hispanic of Latino

Asian

White/Caucasian

Race SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

156

Black or African American

More than one

Method Participants were anonymously recruited and remotely tested using web-based platforms prolific. co and gorilla.sc. Prolific.co was utilized for com pletely anonymous, online participant recruitment in which participants were prescreened to meet study inclusion criteria by the platform. Following their completion of the study, participants were automatically paid by the platform. In the event that participants needed to contact a researcher, they could use the platform’s messaging system, which maintained their anonymity. Gorilla.sc, on the other hand, was used to administer all experiment procedures completely online via a web browser on participants’ personal computers. The link, generated on gorilla.sc, was provided to prolific. co, allowing participants to be seamlessly directed between the two platforms. In the current study, participants received DRM lists presented auditorily over headphones either in silence or embedded in multitalker background noise. Prior to hearing the noisy lists, participants completed a 15-word speech perception task that assessed their ability to hear words in noise. After completing both blocks, participants were asked to self-report their conceptions about the task’s demands and their ability to meet those demands using the NASA RTLX. Treatment of all participants met the guidelines set forth by the American Psychiatric Association and a university-based institutional review board. Participants One hundred participants were screened and recruited anonymously using prolific.co. This platform uses participant self-report questionnaires to allow researchers to screen their participants based on a variety of criteria. Our young adult participants all reported (a) being between 19 and 29 years of age (M = 23.25; SD = 3.05), (b) being from, and currently living in, the United States (U.S.), (c) learning and speaking English as their first language, (d) and having perceived normal or corrected to normal vision, no known hearing difficulties, and no known long-term health condi tions or disabilities. Our older adult participants all reported being between ages 65 and 75 (M = 68.78; SD = 2.87) and met Criteria 2 through 4 mentioned above. Five participants, four young adults and one older adult, reported writing words down or receiv ing help from someone else on the task and were excluded. The final analyses included 46 younger adults and 49 older adults (see Table 1 for reported

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

participant demographics). All participants passed the speech perception task (M = 0.91, SD = 0.08), the attention checks within the study, and were compensated $4 through prolific.co. Materials DRM Lists Twenty-four DRM lists were recorded from Roediger and McDermott (1995). These were recorded in a female voice using an anechoic chamber. Using Audacity for Windows (Version 2.4.2, Audacity), the word lists were created to present a word every 2 sec onds in the same order as Roediger and McDermott (1995). Lists were leveled to present the words between 60 and 70 dB with the computer set at 50% volume. Of the 24 lists recorded, we utilized 23 of them throughout the task. King and river lists were used as practice lists because they included proper nouns (i.e., England, George, and Mississippi). The girl list was used in the speech perception task. The spider list was excluded from presentation, as it included words such as “tarantula” and “arachnid,” which are commonly misspelled. The remaining 20 lists were used during the memory conditions of the task. Background Noise In order to mimic a natural environment, cafeteria ambience noise recorded using a TASCAM DR-40 portable recorder at 48kHz 24bit (Stereo) was selected from freesound.org (https://freesound. org/people/TaXMaNFoReVeR/sounds/325438/). For presentation during the lists, the first 40 sec onds of the sound clip was extracted and dampened to 45 to 55 dB, roughly 10 to 15 dB quieter than the word presentation decibel level. The sound presentation was then combined with the word lists, maintaining the decibel difference during presen tation, even across different participant-selected volumes. All sound files reported are available on OSF (https://osf.io/gu7aw/). NASA Raw Task Load Index Our NASA RTLX procedure was adopted from the NASA TLX pen-and-paper version available online (https://humansystems.arc.nasa.gov/groups/ TLX/tlxpaperpencil.php). Participants were asked to use a slider to rate the following categories: men tal demand, physical demand, temporal demand, performance, effort, and frustration. Mental demand asks how mentally demanding participants interpreted the task to be. Physical demand asks how physically demanding participants perceived the task to be. Temporal demand asks how hurried

or rushed the pace of the task seemed. Performance asks participants to rate how successful they felt they were at accomplishing what was asked of them during the task. Effort asks participants how hard they had to work in order to accomplish the level of performance they achieved. Lastly, frustration asks participants the level to which they felt “insecure, discouraged, irritated, stressed, or annoyed” during the task. Each response, aside from performance, was ranked from very low to very high. Performance was rated from perfect to failure. The original NASA TLX procedure proposed by Hart and Staveland (1988) included both the aforementioned rating scale task and a pair-wise rating task where participants are presented two of the six item titles (i.e., mental demand, physical demand, temporal demand, performance, effort, frustration) and are asked to choose which was more important to them when completing the task (Hart & Staveland, 1988). A weighted score would then be calculated from both the rating item responses and the pair-wise rating task. The NASA RTLX procedure only collects participants’ item raw scores, which can then be averaged or analyzed individually (Hart, 2006). Longo (2018) reported an alpha reliability coefficient of .65 when administer ing the NASA RTLX with young adults. With older adults, Devos and colleagues (2020) reported an interclass correlation coefficients (ICCs) between .71 and .81, indicating good reliability for the NASA RTLX with older adults. The present study reported an alpha reliability coefficient of .54 for the NASA RTLX across both age groups. Procedures All experimental procedures were conducted via gorilla.sc (Anwyl-Irvine et al., 2019). Prior to starting the experiment, participants completed a headphone-check task adopted from Parker and colleagues (2021) in which participants had to determine which of three tones was the quietest. They had two attempts to pass four of six trials. If they were unable to pass the headphone check, they were disenrolled from the study. After passing the headphone check, participants were randomly assigned to start with either the silent list block or the noisy list block to minimize practice effects. All participants received both block types. At the conclusion of the experiment, partici pants completed the NASA RTLX and were asked if they noticed anything about the words and the list structures. No participants reported any prior knowledge regarding the DRM.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

157

DRM Across Adulthood | Wagner, Lyon, and AuBuchon

Experimental Tasks Participants completed two blocks of trials each consisting of 12 DRM lists: two practice and 10 trial lists. One block presented the lists in silence while the other presented lists embedded in the back ground noise, counterbalanced across participants. Trial lists were randomly selected for each block by the experimental program and only appeared once for each participant. After listening to one DRM list, participants were prompted to type the words they recalled from the list. After finishing their recall task, participants advanced to the next list until all 10 trial lists were completed for the block. When using these measures, we reported an alpha reliability coefficient of .70 for the silent list blocks and .71 for the noise list blocks. Prior to beginning the noisy block, participants completed the speech perception task. All 15 words from the girl list were presented individually within background noise. Following each word, the free response box would appear, prompting participants to enter the word they heard into the box. Analysis Plan Data cleaning to correct misspellings of typed responses was completed prior to calculating accuracy and false memory rates. Three raters who were blind to the experiment procedures and expectations analyzed the spelling of response. If all three raters agreed on what word the participant was attempting to spell, the spelling was changed and the response was included in the analyses. Many of the corrected responses were common spelling mistakes or homophones (e.g., switching FIGURE 1 Group Accuracy Between Block Conditions 15

List Item Recall

Group Older Adult Young Adult 5

0 Noise

Block Condition

Silent

Note. Accuracy scores were calculated based on whether or not a presented word was present in participants’ recall. Because of this, participants could have up to 15 items correctly recalled.

158

“ie” and “ei” or “course” and “coarse”). All spelling corrections were held constant across both groups. Accuracy and false memory were calculated based on whether the word or critical lure, respectively, were present in participants’ recalled responses. Averaged accuracy was calculated from the number of presented list words recalled per list. Average rates of false memory were calculated from the number of critical lures present in the recall responses. BAS values for the 20 DRM lists were taken from Roediger and colleagues (2001). NASA RTLX item scores were left in their raw form. RTLX total load scores were calculated by averaging participants’ responses, a process that produces more or equally sensitive results as the original weighting procedure (Bittner et al., 1989; Hendy et al., 1993). All mixed ANOVA analyses were conducted using the rstatix R package (version 0.7.0; Kassambara, 2021) in RStudio (version 1.4.1717; 2021). The anova_test function was used to cal culate mixed methods ANOVAs for accuracy and false memories across both age groups (young adult, older adult) and task type (silent, noise). Pearson correlation analyses were conducted using the stats R package. The cor.test function was used to perform correlations between the NASA RTLX self-report scores, accuracy, and false memory rates between the age groups.

Results We conducted a 2 (age group) x 2 (block type) mixed methods ANOVA for recall accuracy. The main effect of age group for recall accuracy, F(1, 93) = 0.84, p = .36, ηp2 = .01, was not significant. A main effect of block type was found, F(1, 93) = 14.14, p < .001, ηp2 = .13, with better accuracy in the silent block, as anticipated. Younger adults performed slightly better in silence (M = 7.47, SD = 2.45) than in noise (M = 7.16, SD = 2.55). The same was found for older adults, who performed better in silence (M = 7.92, SD = 2.56) than in noise (M = 7.41, SD = 2.44). The interaction was not significant, F(1, 93) = 0.87, p = .35, ηp2 = .01 (see Figure 1. We also conducted a 2 (age group) x 2 (block type) mixed methods ANOVA for false memories measured as the number of critical lures recalled. The findings for accuracy were mirrored in the false memory ANOVA pattern: the main effect for age group, F(1, 93) = 0.11, p = .74, ηp2 = .00, was not significant, but the main effect for block type was significant, F(1, 93) = 4.02 p = .05, ηp2 =

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

.04. Within groups, between the silent and noise blocks, younger adults recalled more critical lures during the silence (M = 0.30, SD = 0.46) than noise (M = 0.22, SD = 0.42). Older adults had similar degrees of false memories across silence (M = 0.28, SD = 0.45) and noise (M = 0.27, SD = 0.45). The interaction was not significant, F(1, 93) = 2.61, p = .11, ηp2 = .03 (see Figure 2). Similar to previous research, positive correla tions between each list’s BAS and the recall of the list’s critical lure were significant. Young adult critical lure recall was positively correlated with the list BAS, r(44) = .42, p = .003. In addition, older adult recollection of the critical lure was positively correlated with the list BAS, r(47) = .53, p < .001 (see Figure 3).

Total Cognitive Load Averaged total load scores were negatively cor related with older adult accuracy, r(47) = –.39, p = .01 (see Figure 4A), indicating that, as older adults reported higher levels of cognitive load through their reports, their accuracy for the presented words decreased. Older adult rates of false memory were not significantly correlated with their perceived cognitive load, although this was nearing significance, r(47) = .27, p = .06 (see Figure 4B). Young adult averaged cognitive load accuracy, r(44) = .02, p = .91, and false memory, r(44) = .06, p = .69, correlations were not significant (see Figures 4A & 4B).

Group False Memory (False Alarm) Rates Between Block Conditions 1.00

False Alarms

0.75

Group 0.50

Older Adult Young Adult

0.25

0.00 Noise

Silent

Block Condition

Note. False memories were measured based on whether the critical lure word was present in participant recall.

FIGURE 3 Group False Memory Production by Size of the List Backward Associative Strength (BAS) 1.00

Average False Memory Production

Exploratory Analysis It was important for us to tease apart the individual differences in cognitive load for the DRM task, as this differs across individuals, and possibly across age groups. To examine participants’ cognitive load on the DRM task more closely, we conducted a cor relational analysis between the NASA RTLX total load score, the item scores (i.e., mental demand, physical demand, temporal demand, performance, effort, frustration), and group-level accuracy and false memory rates. Because we only administered the NASA RTLX once at the end of the task, we collapsed across the sound condition blocks to perform the correlational analyses. Accuracy and false memory relations with the total load score and physical demand, performance, frustration scores reached significance, primarily with older adults. Further elaboration of the total load score, physical demand, performance, and frustration correlations are outlined below, but all correlations and their significance are reported in Table 2.

FIGURE 2

Group Older Adult Young Adult

0.75

0.50

0.25

0.00 0.1

0.2

0.3

0.4

DRM List Backward Associative Strength

TABLE 2 Correlations Between NASA RTLX Subscores, Group Accuracy, and Group False Alarms Young Adult Accuracy False Alarms

Older Adult Accuracy

False Alarms

Total Cognitive Load

.02

−.06

−.39*

.27*

Mental Demand

.13

−.07

.33

−.01

Physical Demand

.04

.03

−.42*

.19

.17

−.24

−.17

−.01

−.32*

.20

−.49**

Temporal Demand Performance Effort Frustration

.38*

.16

.05

.18

−.17

−.02

−.16

−.30*

.32*

Note. Removal of one outlier removed the significance and changed the correlation to r(46) = .18, p = .22. * significance < .05. **significance < .001. 1

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

159

DRM Across Adulthood | Wagner, Lyon, and AuBuchon

also not significant (see Figures 5A & 5B).

Physical Demand Older adult physical demand subscores were nega tively correlated with their performance, r(47) = –.43, p = .002 (see Figure 5A), implying that older adults who reported the task requiring less physical effort to complete had greater accuracy scores than older adults who reported that the task required more physical effort. Older adult false memory rates, r(47) = .19, p = .19 (see Figure 5B), along with young adult accuracy, r(44) = .04, p = .80, and false memory rates, r(44) = .03, p = .82, correlations were

Performance Performance subscores were negatively correlated with older adult accuracy, r(47) = –.49, p = < .001, and younger adult accuracy, r(44) = –.32, p = .03 (see Figure 6A), indicating that, regardless of age, participants who reported feeling more successful at completing what the task asked of them (i.e., rating scores closer to 0, or perfect) were more accurate in their list recall. However, performance subscores

FIGURE 4 NASA RTLX Averaged Total Cognitive Load Scores Correlated With Group Accuracy (A) and Group False Memory Rates (B) A

B Group

Group

1.00

Older Adult Young Adult

Older Adult Young Adult 0.75

Accuracy

False Alarms

10 0.50

5 0.25

0.00 0

100

NASA Total Cognitive Load Score

100

NASA Total Cognitive Load Score

FIGURE 5 NASA RTLX Physical Demand Subscores Correlated With Group Accuracy (A) and Group False Memory Rates (B) A

Group

1.00 Older Adult Young Adult

Older Adult Young Adult 0.75

Accuracy

False Alarms

10 0.50

5 0.25

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

160

0.00 0

100

NASA Physical Demand

100

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

only correlated with false memory rates for older adults, r(47) = .38, p = .01 (see Figure 6B). Therefore, older adults who thought poorly about their per formance (i.e., reported scores closer to 100, or failure) on the task had an increased number of critical lures present in their recall of the presented lists. The correlation between false memory rates and performance subscores was not significant for younger adults, r(44) = .20, p = .17 (see Figure 6B).

was a significant negative correlation between frustration subscores and accuracy, r(47) = –.30, p = .04, meaning that, as older adults reported greater levels of these feelings due to the task’s demands, their recall accuracy for the presented list words decreased (see Figure 7A). There was also a significant positive correlation between frustration scores and false memory rates in older adults, r(47) = .32, p = .03, indicating that, when older adult participants reported greater levels of the feelings covered under the frustration item, the appearance of critical lures in the participants’ list recall increased (see Figure 7B). In the younger adult group, both accuracy, r(44) = –.02, p = .88,

Frustration The frustration item assessed participants’ feelings of insecurity, discouragement, irritation, stress, or annoyance during the task. For older adults, there

FIGURE 6 NASA RTLX Performance Subscores Correlated With Group Accuracy (A) and Group False Memory Rates (B) A

B Group

Group

1.00

Older Adult Young Adult

Older Adult Young Adult 0.75

Accuracy

False Alarms

10 0.50

5 0.25

0.00 100

100

NASA Performance

FIGURE 7 NASA RTLX Frustration Scores Correlated With Group Accuracy (A) and False Memory Rates (B) A

Group

1.00

Older Adult Young Adult

Older Adult Young Adult 0.75

Accuracy

False Alarms

0.50

5 0.25

SUMMER 2022

0.00

0 0

NASA Frustration

100

NASA Frustration

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

161

DRM Across Adulthood | Wagner, Lyon, and AuBuchon

and false memory, r(44) = –.16, p = .30, correlations were not significant (see Figures 7A & 7B).

Discussion

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

162

For the current study, we aimed to examine agerelated differences in accuracy and false memory rates between younger and older adults on a remote, auditory DRM task. We expected to amplify these differences with the addition of background noise, which would increase the cognitive load of the task. This was expected to affect older adults more severely because of changes in hearing acuity, default processing type, and the ability of older adults to meet the cognitive load of the task. These manipulations were designed to help us better understand adult susceptibility to false memory production in everyday listening environments. We did observe a significant main effect between noise and silent conditions on both participants’ accuracy as well as the number of false memories they produced. However, this effect was similar for both younger and older adults. Background noise, even when it is lower than conversational speech levels, appeared to decrease the number of false memories created through the DRM’s associative, semantic spreading activation. However, how the DRM phenomenon arises in noise conditions, whether through affecting per ceptual access to the presented words, individual differences, or some interaction of the two, still remains unclear. In addition, we found a significant positive cor relation between the BAS of each DRM list and the rate of false memory production on that list across both age groups, further supporting the importance of strong inner list word association to produce false memories in the DRM. However, it is important to highlight that these positive correlations were similar for both our younger and older adult age groups, indicating that, in our sample, the strength of the BAS affected false memory production recall in both groups in a similar way. Given previous research and understanding of memory changes with age, it is intriguing that we did not find significant differences between age groups. When comparing younger and older adults, equivalent performance on any memory task is atypical. A post-hoc power analysis determined a weak power level for the current study (1-β = .67), however, Norman and Schacter (1997), which provided the basis of the current study, detected age-group differences in a free recall DRM task with only 24 participants in each group. Given that

the current study doubled that of Norman and Schacter, it is unlikely that we were underpowered to find age-group differences. Comparable performance in both accuracy and false memory rates between our younger and older adult participants may be a driving force behind our lower power level and may indicate differences in our adult samples compared to those of previous literature. Older adult participants in this study were active on prolific.co, which may imply they use the internet more frequently, engage with studies more often, and thus may differ from older adults who generally participate in research in-person or from the general older adult population in the United States. Further research is needed to better understand whether, or what, age-related differ ences may affect how people approach listening to and remembering information presented noise that more closely resembles our everyday environment. Although our older adult population might have differed from the typical older adult popula tion of the United States, our exploratory analyses attempted to tease apart perceived differences in task demands that might have affected task performance. Through NASA RTLX self-reports, we found that self-conceptions surrounding our task, its demands, and participants’ own cognitive abilities (i.e., the cognitive load) related to their accuracy and susceptibility to false memory produc tion on the DRM task, particularly for our older adults. These results may indicate differences in older adults’ perceptions of the cognitive load of the task at hand, causing differences in the cogni tive resources allocated to the task and the mental effort they put toward accomplishing the task. The exploratory results may also highlight differences in older adults’ level of cognitive self-efficacy, or their beliefs surrounding their own ability to reach certain levels of performance on cognitively demanding tasks (Seeman et al., 1996). Those who reported that the task required more physical demand from them, that their per formance was inadequate, or that they were more frustrated with the task may have weaker cognitive self-efficacy. Weaker self-efficacy surrounding cognitive ability could also lead to older adult individuals putting less effort toward the task or prevent them from adequately meeting the task’s demands (Seeman et al., 1996). Looking forward, these perceptions of task demands, and conceptions of participants’ own cognitive abilities can provide direction for further understanding when there is a lack of age differences found in performance for

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

memory-dependent tasks, similar to the findings of the current study. Limitations The at-home, independent testing removed some experimental and environmental control in our experiment. For example, because participants could set their own volume levels, we had to rely on relative signal-to-noise ratios without being able to set absolute listening levels. It is also possible that participants did not adhere to our direc tions (writing down words without disclosing it, etc.). Nevertheless, precautionary checks, such as requiring headphones to limit nonexperimental background noise, were implemented to catch these potential limitations brought about by remote testing procedures. Aside from limitations in procedural control within the present study, our participant pool may also present an additional, unexpected limitation. Unlike previous, in-person administrations of the DRM with older adult participants, our study did not find age-related differences in memory accuracy or false memory production. This finding brings about additional questions in how our older adult sample was different than that of previous studies. For example, self-selecting into Prolific’s participant pool likely requires a certain level of comfort learning new technologies, which may be related to overall cognition. Or older adults on prolific may be practiced at memory tests if they have been exposed to other similar experiments on the platform. Unfortunately, due to the anonymity of our participant pool, it is unknown if there were specific demographic factors that prevented us from finding age-related differences in memory ability. Future studies that utilize remote testing of older adults on a DRM paradigm should include more in-depth demographic questionnaires to account for these factors when interpreting their results. In addition to the aforementioned limitations, the low reliability coefficient for the NASA RTLX exploratory analyses in the present study does bring a sense of hesitancy when interpreting the correlational results. Although our reliability for this measure was considered low, other studies with both younger and older adults have reported higher reliability coefficients (see Longo, 2018, and Devos et al., 2020, respectively), potentially highlighting the need for more standardized administration of the RTLX across studies, particularly those in which the participant completes the measure without the guidance of a present researcher.

Application and Future Directions The present study highlights the benefits of including measures of individual difference in studies of memory and aging, particularly in stud ies on false memory. As stated above, older adults tested using asynchronous and completely remote procedures performed comparably to younger adults—a departure from previous findings in the false memory literature (see Norman & Schacter, 1997). Uncovering the individual characteristics that support accurate free recall without increasing false memory in older adult populations will pro vide guidance to promote healthy cognitive aging. Future studies, especially those that implement remote procedures, should take additional steps to account for potential individual differences, especially that are brought about by online recruit ment platforms like prolific.co. In addition to including more in-depth demo graphic questionnaires to address potential sample differences, the present study also reinforced the usefulness of the NASA RTLX in understanding older adults’ perceptions of cognitive load during memory tasks. Although the reliability of this mea sure has fluctuated across studies and administra tion procedures, assessing cognitive load during memory tasks like a DRM paradigm is essential for understanding memory accuracy, particularly when testing older adults. NASA RTLX items, particularly performance and frustration, were sensitive to older adult free recall accuracy (see Table 2). Despite also completing the NASA RTLX, younger adults’ self-reports, aside from performance, were not correlated to their accuracy (see Table 2), suggest ing that younger and older adults approached the tasks differently, despite reaching similar levels of memory performance. With these findings, it is apparent that measures of cognitive load should be an addition to any testing procedure, whether remote or in-person, where task demands can increase perceived cognitive load by a participant. Conclusion This study expands our understanding of how task demands intersect with individual differences to promote false memory formation. Although previous research has examined DRM list perfor mance under broadband noise presentation in young adults (Marsh et al., 2015), the multitalker babble used here reflects a common source of background noise faced by both older and younger adults in every-day listening situations. This type of background noise decreases list recall accuracy,

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

163

DRM Across Adulthood | Wagner, Lyon, and AuBuchon

which also decreases the number of false memories produced during recall, similar to what has been noted in previous literature (see Kjellberg et al., 2008; Marsh et al., 2015; McCoy et al., 2005). This study also makes a notable contribution to the DRM literature by introducing assessments of participants’ self-conceptions regarding the task and their performance on the task in relation to both accuracy and susceptibility to form false memories. These self-conceptions provide further insight into individual differences in recall accuracy and false memory production within a DRM task, particularly when participants are older adults. Future studies utilizing the DRM paradigm should consider including similar measures, in addition to measures of hearing thresholds, especially when testing older adults.

References

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

164

Anderson, J. R. (1983). A spreading activation theory of memory. Journal of Verbal Learning and Verbal Behavior, 22(3), 261–295. https://doi.org/10.1016/S0022-5371(83)90201-3 Anwyl-Irvine, A. L., Massonnié, J., Flitton, A., Kirkham, N., & Evershed, J. K. (2019). Gorilla in our midst: An online behavioral experiment builder. Behavior Research Methods, 52(1), 388–407. https://doi.org/10.3758/s13428-019-01237-x Bittner, A. C., Byers, J. C., Hill, S. G., Zaklad, A. L., & Christ, R. E. (1989). Generic workload ratings of a mobile air defense system (LOS-F-H). Proceedings of the Human Factors Society Annual Meeting, 33(20), 1476–1480. https://doi.org/10.1177/154193128903302026 Brainerd, C. J., & Reyna, V. F. (2005). The science of false memory. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195154054.001.0001 Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58(1), 17–22. https://doi.org/10.1037/h0046671 Dywan, J., & Jacoby, L. (1990). Effect of aging on source monitoring: Differences in susceptibility to false fame. Psychology and Aging, 5(3), 379–387. https://doi.org/10.1037/0882-7974.5.3.379 Hart, S. G. (2006). Nasa-Task Load Index (NASA-TLX); 20 years later. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 50(9), 904–908. https://doi.org/10.1177/154193120605000909 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 Hashtroudi, S., Johnson, M. K., & Chrosniak, L. D. (1990). Aging and source monitoring. Psychology and Aging, 4(1), 106–112. https://doi.org/10.1037/0882-7974.4.1.106 Hendy, K. C., Hamilton, K. M., & Landry, L. N. (1993). Measuring subjective workload: When is one scale better than many? Human Factors, 35(4), 579–601. https://doi.org/10.1177/001872089303500401 Hintzman, D. L. (1988). Judgements of frequency and recognition memory in a multiple-trace memory model. Psychological Review, 95(4), 528–551. https://doi.org/10.1037/0033-295X.95.4.528 Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 144(1), 3–28. https://doi.org/10.1037/0033-2909.114.1.3 Longo, L. (2018). Experienced mental workload, perception of usability, their interaction and impact on task performance. PloS one, 13(8), e0199661. https://doi.org/10.1371/journal.pone.0199661 Kahneman, D. (1973). Attention and effort. Prentice-Hall. Kassambara, A. (2021). rstatix: Pipe-friendly framework for basic statistical tests. R package version 0.7.0. https://CRAN.R-project.org/package=rstatix Kjellberg, A., Ljung, R., & Hallman, D. (2008). Recall of words heard in noise. Applied Cognitive Psychology, 22(8), 1088–1098. https://doi.org/10.1002/acp.1422 Knott, L. M., Dewhurst, S. A., & Howe, M. L. (2012). What factors underlie associative and categorical memory illusions? The roles of backward

associative strength and interitem connectivity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38(1), 229–239. https://doi.org/10.1037/a0025201 Mackersie, C. L., & Cones, H. (2011). Subjective and psychophysiological indices of listening effort in a competing-talker task. Journal of the American Academy of Audiology, 22(2), 113–122. https://doi.org/10.3766/jaaa.22.2.6 Marsh, J. E., Ljung, R., Nostl, A., Threadgold, E., & Campbell, T. A. (2015). Failing to get the gist of what’s being said: Background noise impairs higher-order cognitive processing. Frontiers in Psychology, 6(548). https://doi.org/10.3389/fpsyg.2015.00548 McCoy, S. L., Tun, P. A., Cox, L. C., Colangelo, M., Stewart, R. A., & Wingfield, A. (2005). Hearing loss and perceptual effort: Downstream effects on older adults’ memory for speech. The Quarterly Journal of Experimental Psychology, 58(1), 22–33. https://doi.org/10.1080/02724980443000151 Meade, M. L., Watson, J. M., Balota, D. A., & Roediger, H. L. (2006). The roles of spreading activation and retrieval mode in producing false recognition in the DRM paradigm. Journal of Memory and Language, 56(3), 305–320. https://doi.org/10.1016/j.jml.2006.07.007 Norman, K. A., & Schacter, D. L. (1997). False recognition in younger and older adults: Exploring the characteristics of illusory memories. Memory & Cognition, 25(6), 838–848. https://doi.org/10.3758/bf03211328 Parker, A. J., Woodhead, Z. V. J., Thompson, P. A., & Bishop, D. V. M. (2021). Assessing the reliability of an online behavioural laterality battery: A preregistered study. Laterality, 26(4), 359–397. https://doi.org/10.1080/1357650X.2020.1859526 Peelle, J. E. (2018). Listening effort: How the cognitive consequences of acoustic challenge are reflected in brain and behavior. Ear and Hearing, 39(2), 204–214. https://doi.org/10.1097/AUD.0000000000000494 Pierce, B. H., Gallo, D. A., Weiss, J. A., & Schacter, D. L. (2005). The modality effect in false recognition: Evidence for test-based monitoring. Memory & Cognition, 33(8), 1407–1413. https://doi.org/10.3758/BF03193373 Reyna, V. F., & Brainerd, C. J. (1998). Fuzzy-trace theory and false memory: New frontiers. Journal of Experimental Child Psychology, 71(2), 194–209. https://doi.org/10.1006/jecp.1998.2472 Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21(4), 803–814. https://psycnet.apa.org/buy/1995-42833-001 Roediger, H. L, Watson, J. M., McDermott, K. B., & Gallo, D. A. (2001). Factors that determine false recall: A multiple regression analysis. Psychonomic Bulletin & Review, 8(3), 385–407. https://doi.org/10.3758/BF03196177 RStudio Team. (2021). RStudio: Integrated Development Environment for R. RStudio, PBC, Boston, MA. http://www.rstudio.com/ Seeman, T., McAvay, G., Merrill, S., Albert, M., & Rodin, J. (1996). Self-efficacy beliefs and change in cognitive performance: MacArthur studies on successful aging. Psychology and Aging, 11(3), 538–551. https://doi.org/10.1037/0882-7974.11.3.538 Tomporowski, P. D. (2002). The psychology of skill: A life-span approach. Praeger Pub Text. Tomporowski, P. D. (2003). Performance and perceptions of workload among young and older adults: Effects of practice during cognitively demanding tasks. Educational Gerontology, 29(5), 447–466. https://doi.org/10.1080/713844359 Tun, P. A., Wingfield, A., Rosen, M. J., & Blanchard, L. (1998). Response latencies for false memories: Gist-based processes in normal aging. Psychology and Aging, 13(2), 230–241. https://doi.org/10.1037//0882-7974.13.2.230 Van Gerven, P. W. M., Paas, F. G. W. C., Van Merriënboer, J. J. G., & Schmidt, H. G. (2002). Cognitive load theory and aging: Effects of workload examples on training efficiency. Learning and Instruction, 12(1), 87–105. https://doi.org/10.1016/S0959-4752(01)00017-2 Watson, J. M., Bunting, M. F., Poole, B. J., & Conway, A. R. A. (2005). Individual differences in susceptibility to false memory in the Deese-RoedigerMcDermott paradigm. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(1), 76–85. https://doi.org/10.1037/0278-7393.31.1.76 Author Note. Rebecca L. Wagner https://orcid.org/0000-0003-1659-6016 Bethany A. Lyon https://orcid.org/0000-0002-8165-3075 Angela AuBuchon https://orcid.org/0000-0002-6746-4462 This project was funded by the Fall 2020 Psi Chi Undergraduate Research Grant. Creation of the stimuli

Wagner, Lyon, and AuBuchon | DRM Across Adulthood

was supported, in part, by the National Institute of General Medical Sciences, U54 GM115458, which funds the Great Plains IDeA-CTR Network. We would like to thank Dr. Chris Stecker assistance with recording and processing stimuli. We would also like to thank Dr. Aryn Kamerer for guidance on administering and

analyzing the NASA Task Load Index. We have no conflicts of interest to disclose. Materials and data for this study can be found at https://osf.io/gu7aw/. Correspondence concerning this article should be addressed to at Rebecca L. Wagner at 402-269-0729. United States. Email: rebeccawagner.rlw@gmail.com

165

https://doi.org/10.24839/2325-7342.JN27.2.166

Predictors of Help-Seeking: Self-Concept Clarity, Stigma, and Psychological Distress Hinza B. Malik and Caroline E. Mann* Department of Psychology, Hollins University

ABSTRACT. Multiple studies have shown that individuals with low self-concept clarity (SCC) are highly susceptible to mental health problems (depression and anxiety). However, despite the increased vulnerability to psychopathology, prior research has not examined the relationship between SCC and help-seeking. Hence, to develop a comprehensive understanding of the aforementioned relationship, well-established predictors of help-seeking (psychological distress and stigma) were included in this study. A total of 111 students completed an online survey. Results indicated that lower SCC was associated with higher psychological distress, lower help-seeking propensity, and higher stigma. However, SCC was not found to be a unique predictor of help-seeking above and beyond the established predictors in the multiple regression analysis. Stigma was further divided into perceived public, personal, and perceived peer stigma. The past literature showed no association between perceived public stigma and help-seeking. In addition, perceived public stigma has been found to be higher than personal stigma. Thus, the current study altered the perceived stigma reference group (from “public” to “peer”) to investigate if this change would influence the association with help-seeking. Consistent with prior research, a significant mean difference was found such that perceived public stigma remained significantly higher than personal stigma (95% CI [1.45, 2.34]) and was not correlated with help-seeking or personal stigma. However, both personal and perceived peer stigma were negatively correlated with help-seeking and positively correlated with each other, such that high personal and peer stigma were associated with lower help-seeking. The results can provide insight for future help-seeking intervention programs and mental health stigma reduction campaigns.

DIVERSITY

Open Data and Materials badges earned for transparent research practices. Data and materials can be viewed at https://osf.io/edfhg/files/. Diversity badge earned for conducting research focusing on aspects of diversity.

Keywords: self-concept clarity, help-seeking, mental illness stigma, psychological distress, perceived peer stigma

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

166

pidemiological data showed that about 46.4% of the general population has a lifetime prevalence of being affected by a mental illness (Kessler et al., 2005). According to

the 2019 National Survey on Drug Use and Health (NSDUH) by the Substance Abuse and Mental Health Services Administration (SAMHSA, 2019), one in every five U.S. adults (51.5 million) lives with

*Faculty mentor

Malik and Mann | Predictors of Help-Seeking

a mental illness. Despite the widespread prevalence of mental illnesses, only about 23.0 million adults out of 51.5 million received treatment. As early as 1999, the discrepancy was attributed to stigma, cited in the mental health report by the U.S. Surgeon General David Satcher (U.S. Department of Health & Human Services, 1999). Moreover, mental health stigma has been shown to be a foremost barrier in help-seeking, not only in the United States, but also globally (Thornicroft, 2007). Approximately 30% of the worldwide population is affected by a mental illness every year; however, two-thirds of people do not seek help (Kessler et al., 2005). Hence, the topic of help-seeking and its predictors in addition to stigma remains essential in an effort toward improving mental health campaigns. One such predictor that has not been exam ined before is self-concept clarity (SCC). A lower SCC has been consistently associated with psy chological maladjustment that contributed to the development of internalizing and externalizing disorders (Bigler et al., 2001; Parise et al., 2019). However, despite the strong association between SCC and psychological maladjustment, there was a lack of literature exploring the influence of SCC on help-seeking. Hence, for the current study, we attempted to bridge the gap by examining the association and the strength of SCC as a predictor of help-seeking behavior. Self-Concept The broad definition of self-concept is the per ception of oneself, influenced by the interaction between the environment and subsequent experi ences. There are two key aspects of self-concept: SCC and self-concept differentiation (SCD). SCC is defined as the extent to which self-beliefs are clearly and confidently defined, internally consistent, and stable over time (Campbell et al., 1996). For example, if the description of one’s personality fluctuated from one day to the other, that would indicate a lower SCC. The results from Campbell et al. (1996) suggested that a lower SCC is associated with psychological maladjustment, which can be operationally defined as the individual’s inability to meet the demands of life resulting in psychological distress. Moreover, respondents with a lower SCC had lower self-esteem, conscientiousness, agreeable ness, and higher neuroticism. On the other hand, self-concept differentiation is the extent to which self-conceptions vary across different roles (e.g., general, student, friend,

romantic partner, offspring, and worker; Donahue et al., 1993). Donahue et al. (1993) suggested that differentiation across roles results in fragmentation of the self-system, which correlates with psycho logical maladjustment. Hence, respondents with a higher self-concept differentiation were more vulnerable to depression, neuroticism, and lower self-esteem (Donahue et al., 1993). Bigler et al. (2001) is among the few studies that examined SCC and SCD together. Researchers found that both SCC and SCD were independent constructs. Both SCC and SCD predicted psy chological maladjustment, but SCC was a better predictor in both samples (college students and inpatient schizophrenia patients). A similar trend was supported by Diehl and Hay (2011). Hence, for the current study, we decided to focus on SCC as a predictor of help-seeking instead of SCD. Self-Concept Clarity An important developmental task during ado lescence is establishing more clarity in one’s self-concept. A longitudinal study showed small increases in SCC during adolescence supporting that SCC is subject to change (Schwartz et al., 2012). One of the factors that have been shown to mediate the association between SCC and psy chological maladjustment is emotional regulation (Parise et al., 2019). Emotional regulation refers to how effectively an individual regulates their affect (Caprara et al., 2008). If adolescents regulated their negative and positive affect relatively well, they were more likely to have a higher SCC (Parise et al., 2019). As a result, the likelihood of develop ing internalizing and externalizing problems or expericing psychological distress decreased. Hence, efficient emotional regulation was directly associ ated with higher SCC and inversely associated with psychological maladjustment. Furthermore, another factor that has been shown to influence SCC is parental bonding. In Perry et al. (2008), parental care and warmth were positively associated with SCC. The increase in SCC was attributed to the confidence that parents gave to their children during their self-concept exploration phase. In addition, the likelihood of developing more clarity in self-concept was higher if the communication between parents and children was clear and open compared to if the child kept secrets (Frijns & Finkenauer, 2009; Van Dijk et al., 2014). Moreover, a longitudinal study showed that a higher SCC in adolescents was associated with a better relationship with their parents that was

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

167

Predictors of Help-Seeking | Malik and Mann

indicated by a lower frequency of negative interac tion (Becht et al., 2017). SCC has also been shown to predict healthy identity development (Campbell et al., 1996; Schwartz et al., 2012). A higher SCC was associated with higher self-esteem, affect balance, and lower depression and anxiety (Bigler et al., 2001; Butzer & Kuiper, 2006; Cicei, 2012). On the contrary, people with lower SCC were found to be more susceptible to internalizing disorders, such as anxiety and depression indicative of psychological distress. Psychological distress can be operationally defined as a state of emotional suffering that involves symptoms of depression and anxiety (Mirowsky & Ross, 2002). This is also supported by several longitudinal studies through adolescence that showed the relationship between lower SCC and internalizing problems (Schwartz et al., 2012; Van Dijk et al., 2014). Elements of internalizing problems (e.g., loneliness, chronic self-analysis, low internal state awareness, and rumination) have been shown to be significantly associated with lower SCC (Campbell et al., 1996; Richman et al., 2016). Intolerance of uncertainty model and SCC are good predictors of generalized anxiety disorder (GAD; Kusec et al., 2016). The intolerance of uncertainty model posited that the development of GAD could be explained through excessive worry, which is a natu ral response to modulate feelings of uncertainty (Dugas et al., 1998). As prior research showed that lower SCC was indicative of uncertainty about the self, it was found to be a salient predictor of GAD (Kusec et al., 2016). In addition, Stopa et al. (2010) found that a lower SCC was associated with higher social anxiety. Moreover, previous literature showed that lower SCC was not only associated with internal izing and externalizing disorders, but also with eat ing disturbances (Perry et al., 2008). Hence, a lower SCC was indicative of higher psychological distress and increased susceptibility to psychopathology.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

168

Help-Seeking The theory of planned behavior can be useful in understanding human behavior, such as helpseeking. The theory outlines three types of beliefs: behavioral beliefs (experiences or consequences associated with help-seeking), normative beliefs (expectations or behaviors of significant others regarding help-seeking), and control beliefs (factors that increase/decrease help-seeking). According to this theory, an attitude toward a behav ior such as help-seeking is formed by behavioral

beliefs associated with mental health help-seeking (e.g., prior experience with seeking help for mental health concerns). Normative beliefs tap into perceived social norms and control beliefs tap into how much power the person believe they have in seeking help. A positive attitude and perceived social norm together with higher perceived control are associated with an increase in help-seeking (Ajzen, 1991). The two salient factors that have been found to consistently predict help-seeking behavior are psychological distress and mental health stigma (Boerema et al., 2016; Wadman et al., 2017). Mental health stigma can be further classified into personal and perceived public stigma. Personal stigma is defined as an individual’s stereotypes, prejudices, and behavior, whereas perceived public stigma is defined as an individual’s perception of the extent to which the public holds negative attitudes, stereotypes, and prejudice towards those who seek mental health treatment (Corrigan, 2004). Several studies have found that perceived public stigma is higher than personal stigma and both are positively correlated (Eisenberg et al., 2009; Lally et al., 2013). In addition, personal stigma has been shown to be negatively associated with help-seeking, but per ceived public stigma was not found to be associated with help-seeking (Boerema et al., 2016; Eisenberg et al., 2009). A longitudinal study conducted with college students by Golberstein et al. (2009) con tinued to support that perceived public stigma was not associated with help-seeking behavior. Hence, personal stigma independently influenced helpseeking. However, several studies showed otherwise that higher public stigma is indeed a barrier to help-seeking (Kulesza et al., 2015; Nearchou et al., 2018; Phelan et al., 2000). As personal attitudes are shaped by public attitudes (Link, 1987), it is logical to think that both personal stigma and perceived public stigma should be related to help-seeking. Current Study The relationship between low SCC and psychologi cal maladjustment and what factors may influence the association are well established. However, previous research has not studied whether the indi viduals who have a lower SCC and are subsequently highly susceptible to psychological maladjustment would be more or less likely to seek help. Hence, the purpose of the current study was to explore the influence of SCC on the likelihood of help-seeking. In addition to SCC, psychological distress and stigma were included in this study because they

Malik and Mann | Predictors of Help-Seeking

have been established as reliable predictors of helpseeking. The inclusion of these predictors allowed us to answer an additional research question of whether SCC predicted help-seeking above and beyond the established predictors. How an individual perceives a behavior or an attitude of their peers toward something is a key predictor of health behavior. This is because the perception creates a perceived social pressure that pushes the individual either to engage or not in a particular behavior (Ajzen, 2001). Hence, in the current study in addition to perceived public stigma and personal stigma, we measured perceived peer group stigma. Perceived peer group stigma gauges to what extent one thinks their peers hold negative attitudes toward those who seek mental health treat ment. If the peer group held negative attitudes or perceptions toward those who sought help for their mental health, that would demonstrate a higher perceived peer group stigma. The current study addressed the research question of whether the respondent’s personal stigma would also be higher to be concordant with their peers given their high perceived peer group stigma or vice versa. In the literature, there are mixed findings regarding the relationship of public and personal stigma with help-seeking. Some studies showed that public stigma was related to help-seeking (Kulesza et al., 2015; Nearchou et al., 2018; Phelan et al., 2000), whereas other studies showed that only per sonal stigma was related to help-seeking (Boerema et al., 2016; Eisenberg et al., 2009). In addition, public stigma was not found to be associated with help-seeking (Golberstein et al., 2009). Hence, the current study disentangled the relationship between not only the different types of stigmas (personal, perceived public, and perceived peer group stigma), but also their association with help-seeking. The following hypotheses were formulated in the light of literature: (a) SCC would positively cor relate with help-seeking; (b) SCC would negatively correlate with psychological distress; (c) a positive correlation would be found between personal and peer-group stigma; and (d) no correlation would be found between perceived public stigma and help-seeking, but a negative correlation would be found between help-seeking behavior and personal stigma as well as perceived peer stigma.

Method Participants A total of 111 undergraduate students from Hollins University, a small, women’s liberal

arts institution, between the ages of 18 and 46 (M = 20.69, SD = 3.83) participated in the study. Most participants identified as women (n = 101). Nine participants identified as nonbinary, and one participant preferred not to say. The sample was predominantly White/non-Hispanic/European (n = 61), with Black/African American (n =14), Asian (n =13), multiracial (n = 8), Asian American (n = 6), Hispanic/Latino (n = 4), American Indian/Alaska Native (n = 2), Native Hawaiian or Pacific Islander (n = 1), and Middle Eastern (n = 1) participants. Measures Self-Concept Clarity Scale The scale was developed by Campbell et al. (1996) and consists of 12 items (α = .86). All items are reverse scored except 6 and 11. The scale uses a 5-point Likert scale with anchors (strongly disagree to strongly agree) and gauges the beliefs held by the individual about themselves (e.g., “Sometimes I think I know other people better than I know myself”). Higher scores indicate higher SCC and vice versa. The internal reliability for this scale in this sample was good (α = .86). Help-Seeking To measure help-seeking, the Inventory of Attitudes Toward Seeking Mental Health Services (IASMHS) and Mental Help Seeking Intention Scale (MHSIS) were combined. The questions of MHSIS were interspersed in the IASMHS as per the guidelines provided by the scoring key of MHSIS (Hammer & Spiker, 2018). Inventory of Attitudes Toward Seeking Mental Health Services. IASMHS was developed by Mackenzie et al. (2004), which consists of 24 items (α = .87) and has three factors. Each fac tor has eight items whose internal consistency is good. The first factor is Psychological Openness, which is an individual’s openness to acknowledge their psychological problems and seek help for those problems (α = .82; e.g., “There are certain problems which should not be discussed outside of one’s immediate family”). Help-Seeking Propensity is the second factor, which is an individual’s belief about their willingness and ability to seek help for psychological problems (α = .76; e.g., “I would have a very good idea of what to do and who to talk to if I decided to seek professional help for psychological problems”). Lastly, Indifference to Stigma refers to an individual’s concern about how the significant others will think or feel when they find out that they are seeking help for their

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

169

Predictors of Help-Seeking | Malik and Mann

psychological problems considering social norms (α = .79; e.g., “I would not want my significant other [spouse, partner, etc] to know if I were suffering from psychological problems”). Participants indi cate their level of agreement with each statement using a 5-point Likert scale with anchors (disagree to agree). A higher cumulative score indicates hav ing a more positive attitude toward help-seeking behavior. The internal reliability for this scale in the cur rent sample was good (α = .84). The Psychological Openness subscale had a questionable internal consistency (α = .66); Help-Seeking Propensity (α = .78) and Indifference to Stigma (α = .79) had acceptable internal consistency. Mental Health Seeking Intention Scale. MHSIS was developed by Hammer and Spiker (2018) and consists of three items (e.g., “If I had a mental health concern, I would intend to seek help from a mental health professional”). Participants indicate their intention to seek help for their mental health concerns using a 7-point Likert scale with anchors (extremely unlikely to extremely likely). The internal reliability for this scale in this sample was excellent (α = .96).

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

170

Perceived Need for Help and Mental Health Services Utilization Perceived need for help was measured by an item, “In the past 12 months, did you think you needed help for emotional or mental health problems, such as feeling sad, blue, anxious or nervous?” To measure mental health services utilization, a single item asked participants whether, in the past year, they received treatment for mental health problems. If the participant answered yes, they were asked what type of treatment they received (psychotropic medication or therapy/counseling). These items were taken from the Healthcare for Communities Study, a national study of mental health care use questionnaire (Wells et al., 2006), and were also utilized by Eisenberg et al. (2009). An additional item asking if mental health problems were affecting the respondent’s aca demic performance was taken from Eisenberg et al. (2009). If so, the respondents were asked if they would talk to no one or write whom they would talk to as a text entry. The responses were categorized into four categories: mental health professional (therapist/counselor) or primary care physician, family member (parents, siblings, and grandparents), friend or partner, and college faculty (professor, advisor, and dean).

Kessler Psychological Distress Scale The scale was developed by Kessler et al. (2003) and consists of 10 questions (α = .93) e.g., “During the last 30 days, about how often did you feel nervous?”). Participants indicate how they have been feeling over the last 30 days (i.e., 4 weeks) using a 5-point Likert-type scale with anchors (none of the time to all of the time). The cumulative score range is 0–40 where higher scores are indicative of higher psychological distress and vice versa. The cumulative score can be categorized according to the following cutoffs mentioned in the Victorian Population Health Survey (Department of Human Services, Victoria, 2001): 10–19 (likely to be well), 20–24 (likely to have a mild disorder), 25–29 (likely to have a moderate disorder), and 30–50 (likely to have a severe disorder). The reliability for this scale in this sample was excellent (α = .93). Adapted Devaluation-Discrimination Scale for Stigma Perceived Public/Peer Stigma Scale. The current study used the devaluation-discrimination scale which was adapted by Lally et al. (2013) from the original version (Link, 1987). One of the main changes in the adapted version is altering the word ing from “mental patients” to “people who have received mental health treatment.” The Perceived Public Stigma Scale consists of 12 statements (α = .86; e.g., “Most people think less of a person who has received mental health treatment”). The inter nal reliability for this scale in the current sample was good (α = .85). For this study, the Perceived Public Stigma Scale by Lally et al. (2013) was adapted to measure perceived peer stigma by changing the wording from “most people” to “my peers” (e.g., “My peers would think less of a person who has received mental health treatment”). The perceived peer stigma scale also consists of 12 statements and the internal reliability in this sample was found to be good (α = .89). The perceived public and peer stigma scales were equally randomized between participants and indicated their level of agreement using a 5-point Likert scale with anchors (strongly agree to strongly disagree). Question 5, 6, 7, 9, 11, and 12 are reverse coded, and the higher the average score, the higher is the perceived public or peer-group stigma held by the respondent. Personal Stigma Scale. Lally et al. (2013) adapted four items (α = .78) from the original version by changing the wording from “most people” to “I.” The first two statements measure

Malik and Mann | Predictors of Help-Seeking

a respondent’s behavior, (e.g., “I would willingly accept a person who has received mental health treatment as a close friend”) and the last two state ments measure an attitude, (e.g., “I believe that a person who has received mental health treatment is just as trustworthy as the average citizen”). The scale uses a 5-point Likert scale with anchors (strongly disagree to strongly agree), and items 3 and 4 are reverse coded. The internal reliability for this scale in this sample was poor (α = .52). Procedure After receiving approval from the Human Research Review Committee at Hollins University, an online Qualtrics survey was open for approximately two months for the undergraduate students (age 18 and or above). The survey link was distributed through class announcement emails by psychology profes sors, international departmental emailing list, and Facebook posts on various university pages. The completion time was between 20–25 minutes, and the survey began with an informed consent. If the participants chose to participate, they completed a demographics questionnaire (age, major, gender, race/ethnicity) followed by a series of measures. The first measure was the SCC questionnaire that focused on the beliefs held by the individual about themselves. The second questionnaire was the IASMHS combined with the MHSIS, both of which have the same instructions (definitions of the terms that will be used in IASMHS and MHSIS) to measure help-seeking behavior. The third question naire was the Kessler Psychological Distress Scale followed by the Personal Stigma Scale. The fifth questionnaire presented gauged the perceived need for help and mental health utilization. The sixth questionnaire was equally randomized between participants, so they either received the Perceived Public Stigma or Perceived Peer-Group Stigma Scale. In between all the measures, each participant answered two instructed response items that were randomly embedded to check participant atten tion. After the completion of the questionnaires, participants read a debriefing statement and were redirected to the extra credit information portion if the participant wished to receive extra credit in one psychology undergraduate-level course.

Results Out of the total 113 participants, two participants were removed because they answered all instructed response items incorrectly, resulting in 111 partici pants. All results are reported at an alpha level of .05. Internal reliability of all scales was found to be

close to the original internal reliability values except for the IASMHS Psychological Openness subscale, which had questionable internal consistency (α = .66), and the Personal Stigma Scale, which had poor internal consistency (α = .52). All analyses were conducted using SPSS statistical software. Correlations Between SCC, Help-Seeking Behavior, and Psychological Distress To test the relationship between SCC and helpseeking, a Pearson correlation coefficient was calculated between SCC and help-seeking scales (IASMHS and MHSIS). A weak, positive correlation between SCC and IASMHS was found, r(103) = .25, p = .005, indicating a significant relationship between the variables. Hence, a higher SCC was associated with a more positive attitude toward seeking mental health services. A significant correlation was not found between SCC and MHSIS, r(108) = .14, p = .08. A ceiling effect was found as respondents clustered around higher scores on the MHSIS as evident by a skew of –0.67. After recoding the data as a dichotomous variable (1 and 0), no significant changes in the results were found. Next, the correlations between SCC and IASMHS inventory subscales were calculated. There was not a significant correlation between SCC and Psychological Openness. However, a weak, posi tive correlation between SCC and Indifference to Stigma was found, r(104) = .25, p = .005, indicating a significant relationship between the variables in the expected direction. Having a high SCC was associated with lower mental health stigma. A weak, positive correlation between SCC and Help-Seeking Propensity was found, r(107) = .21, p = .02, indicat ing a significant relationship between the variables. Participants with higher SCC responded as having a greater tendency towards help-seeking. To test the relationship between SCC and psychological distress, a Pearson correlation coef ficient was calculated between SCC and Kessler Psychological Distress Scale. A moderate, negative correlation was found, r(108) = –.58, p < .001, indicating a significant relationship between the variables such that higher SCC was associated with lower psychological distress. All means, standard deviations, and correlations between self-concept, help-seeking measures, and psychological distress can be found in Table 1. Correlations Between Stigma (Personal and Perceived Public/Peer) Perceived public stigma was found to be higher

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

171

Predictors of Help-Seeking | Malik and Mann

(M = 2.89, SD = 0.60) than perceived peer stigma (M = 2.27, SD = 0.64) and personal stigma (M = 1.34, SD = 0.52) as shown in Table 2. A paired-samples t test was calculated to find if the mean difference in the perceived public stigma and personal stigma was significant. Results show that the difference in means was significant such that perceived public stigma was higher than personal stigma, t(54) = 14.11, p < .001, 95% CI [1.46, 2.35]. To test the relationship of personal stigma with perceived public stigma and perceived peer stigma, a Pearson correlation coefficient was calculated. A moderate, positive correlation between personal stigma and perceived peer stigma was found, r(52) = .34, p = .006, indicating that a lower personal stigma was associated with a lower perceived peer stigma and vice versa. A significant correlation was not found between personal and perceived public stigma, r(53) = .06, p = .32. TABLE 1 Descriptive Statistics and Correlations Between Self–Concept Clarity Scale, Help-Seeking Measures, and Psychological Distress Measure

1 –

1. Self-Concept Clarity

31.91

9.42

2. IASMHS

86.20

12.64

.25**

3. Psychological Openness

27.90

5.28

.08

4. Indifference to Stigma

27.90

6.69

.25**

5. Help-Seeking Propensity

28.58

6.19

.21*

4.84

1.78

.14

29.85

9.41

−.58

6. MHSIS 7. Psychological Distress

.66** −.28

.23** −.03

.38** .81**

–

−.32 −.24 **

−.21*

Note. IASMHS = Inventory of Attitudes Toward Seeking Mental Health Services. MHSIS = Mental Health Seeking Intention Scale. * p < .05. **p < .01.

TABLE 2 Descriptive Statistics and Correlations Between Help-Seeking and Stigma Measures Measure

28.58

6.19

–

2. MHSIS

4.84

1.79

.81** –

3. Personal Stigma

1.34

4. Peer Stigma 5. Public Stigma

1. Help-Seeking Propensity

6. Indifference to Stigma SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

172

0.52 −.19* −.14

–

2.27

0.64 −.27 −.15

.34** –

–

2.89

0.60

.03 −.12

.06

–

29.70

6.69

.36

–

.38

–

Note. Peer and Public stigma measures were equally randomized between participants thus, their correlation was not computed. MHSIS = Mental Health Seeking Intention Scale. * p < .05. **p < .01.

Correlations Between Help-Seeking Propensity, MHSIS, and Stigma To test the relationship between help-seeking measures and stigma scales, a Pearson correlation coefficient was calculated. A weak, negative correla tion between the Help-Seeking Propensity subscale and Personal Stigma Scale was found, r(108) = –.19, p = .02, indicating that a high help-seeking pro pensity was associated with having a lower personal stigma. A significant correlation was not found between the Help-Seeking Propensity subscale and Perceived Public Stigma Scale, r(53) = .03, p = .41 A weak, negative correlation between the Help-Seeking Propensity subscale and Perceived Peer Stigma Scale was found, r(51) = –.27, p = .03, indicating that high help-seeking propensity was associated with having a lower perceived peer stigma. There were no significant correlations between MHSIS and stigma: personal stigma: r(108) = –.14, p = .13; perceived public: r(53) = –.12, p = .18; and perceived peer: r(52) = –.15, p = .14). A ceiling effect for MHSIS was observed. However, a moderate, positive correlation between Indifference to Stigma subscale and MHSIS was found, r(105) = .38, p = .001, indicating that partici pants who were indifferent to stigma had a higher mental help seeking intention. Multiple Linear Regression A multiple linear regression was conducted to predict the Help-Seeking Propensity subscale based on SCC, personal stigma, and psychological distress. The data was screened for assumptions and outli ers, and no outliers were found. All assumptions of linearity, normality, and homoscedasticity were found to be met, and multicollinearity was not found; however, the correlation between SCC and the Kessler Psychological Distress Scale was reach ing the cut off of .6–.8 as seen in Table 3. A significant regression equation was found, F(3, 105) = 5.29, p = .002, R2adj. = .11. Accounting for all other variables in the model, a 1 unit increase in personal stigma was associated with a –3.02 decrease in help-seeking, SE = 1.11, 95% CI for b [–5.21, –0.83], β = –.26, p = .007. Accounting for all other variables in the model, a 1 unit increase in psycho logical distress was associated with a –.16 decrease in help-seeking, SE = 0.07, 95% CI for b [–.30, –.01], β = –.24, p = .04. Accounting for all other variables in the model, a 1 unit increase in SCC was associated with a .07 increase in help-seeking, but this was not significant, SE = 0.07, 95% CI for b [–.07, .22], β = .11, p = .31.

Malik and Mann | Predictors of Help-Seeking

Need for Help and Mental Health Services Utilization Only 18% (n = 18) of the total participants (N = 111) responded that they would talk to no one if mental health problems were affecting their academic performance. Many other participants (n = 91) indicated they would speak with someone. Most respondents reported they would talk with a mental health professional or primary care physi cian (n = 44), college faculty (n = 42), friend or partner (n = 36), and family member (n = 32). If the participant’s answer belonged to more than one category, the answer was counted in each category. In the past year, 76 respondents reported that they needed help with their mental health prob lems, 25 selected maybe, and 10 selected the option no. In the past year, 50 of the respondents reported having received treatment for mental health prob lems, whereas only three refused treatment, and 58 reported that they did not receive any treatment. From those 50 respondents who received mental health treatment, 13.5% received psychotropic medication (n = 15) and 31.5% received therapy/ counseling (n = 35).

Discussion The purpose of the study was to develop a compre hensive understanding of the relationship between SCC, stigma, and help-seeking behavior. Despite the numerous studies including longitudinal that have established a strong association between lower SCC and susceptibility to psychopathology, there was a lack of literature on SCC and help-seeking behavior (Bigler et al., 2001; Kusec et al., 2016; Perry et al., 2008; Stopa et al., 2010; Schwartz et al., 2012; Van Dijk et al., 2014). To our knowledge, this was the first study to explore the concept of SCC and helpseeking together. The first hypothesis (a) predicted a positive correlation between SCC and help-seeking. We correlated SCC Scale with help-seeking measures (IASMHS and MHSIS). Results from the IASMHS and Help-Seeking Propensity subscale show that individuals who have a higher SCC have a more positive attitude toward seeking mental health services and are more likely to seek help. Hence, the results show that lower SCC is associated with lower help-seeking. The Indifference to Stigma subscale was found to be positively correlated with SCC, indicating that lower SCC is associated with higher stigma. We did not find a significant correlation between SCC and the MHSIS. Although the

internal validity of the 3-item MHSIS in our sample was excellent, there was a ceiling effect (i.e., all respondents clustered around higher scores on the scale). This ceiling effect further echoed in the results as only 18% of the respondents were not willing to talk to anyone if their mental health was affecting their academic performance. Hence, the inability to find a significant correlation between SCC and MHSIS is unique to the population used by the study. Moreover, a dissertation showed that the internal validity for MHSIS in their sample was unacceptable (Miller, 2020). They recoded their data as a dichotomous variable (0 and 1a), but when the same was done for this study, it made no significant changes in the results. In addition, the MHSIS was developed using volunteers from ResearchMatch, limiting the generalizability of the scale because it has not been validated outside of the original study (Hammer & Spiker, 2018). Hence, the Help-Seeking Propensity subscale of the IASMHS was used as the dependent variable for the multiple linear regression analysis instead of MHSIS. The second hypothesis (b) predicted a nega tive correlation between SCC and psychological distress. This is because previous research has supported that lower SCC predicts psychological maladjustment, which increases the susceptibility to developing internalizing and externalizing disorders (Bigler et al., 2001; Campbell et al., 1996; Stopa et al., 2010; Van Dijk et al., 2014). We found that as SCC increased, the psychological distress gauged by the Kessler Psychological Distress Scale decreased supporting the previous literature and current hypothesis. In conjunction with the results of the previous hypothesis, this means that having a lower SCC is associated with higher psychological distress but despite that, there is a negative asso ciation between SCC and help-seeking such that a lower SCC was associated with lower help-seeking. To i n v e s t i g a t e w h e t h e r S C C p r e d i c t s TABLE 3 Multiple Regression Results for Predictors of IASMHS Help-Seeking Propensity Subscale Collinearity Statistics SCC

0.07

Tolerance

VIF

1.01

.314

.65

1.53

−3.02

1.11 −2.73

.007

.95

1.05

Psychological Distress −0.16

0.07 −2.21

.036

.65

1.54

Personal Stigma

Note. SCC = self-concept clarity. IASMHS = Inventory of Attitudes Toward Seeking Mental Health Services.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

173

Predictors of Help-Seeking | Malik and Mann

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

174

Help-Seeking Propensity above and beyond the established predictors (mental health stigma and psychological distress), a multiple regression analysis was conducted. Personal stigma and psycho logical distress continued to be unique predictors of help-seeking propensity as suggested by the previous literature (Boerema et al., 2016; Wadman et al., 2017). Results show that personal stigma was the strongest predictor followed by psychological distress. SCC was not found to be a unique predictor of help-seeking in the presence of personal stigma and psychological distress. This means that there is a possibility that SCC does predict help-seeking to a certain degree, but the strength of its predictability does not surpass that of the established predictors. Moreover, the inability to find SCC as a significant predictor could be due to the high correlation of –5.8 between SCC and psychological distress that was reaching the cutoff of .6–.8 for multicollinearity. The third hypothesis (c) predicted a positive correlation between personal and peer-group stigma. This was the first study to investigate peer-group stigma and the results supported the hypothesis such that, when personal stigma increased, peer-group stigma also increased. This is consistent with the conceptualization presented by Ajzen (2001) that how people perceive their peers attitude toward a behavior creates a perceived social pressure that pushes the individual to engage or not in that behavior. In the context of the results, this means that a higher peer group stigma is indicative of a perception that peers hold a negative attitude toward mental health help-seeking. As perceptions help shape personal beliefs, the direction of respon dents’ personal stigma matched their perceived peer group stigma. Hence, the positive correlation between personal and peer group stigma. Consistent with previous literature, a significant mean difference was found such that perceived public stigma was higher than personal stigma (Eisenberg et al., 2007; Lally et al., 2013). However, no association between personal and perceived public stigma was found contrary to the findings reported in Eisenburg et al. (2009). This implies that the perception of the level of stigma held by the public is not associated with personal stigma and vice versa, which indicates that both public and personal stigma might be independent constructs. The fourth hypothesis (d) predicted no cor relation between perceived public stigma and help-seeking, but a negative correlation between both personal and peer stigma with help-seeking behavior. Consistent with previous literature,

personal stigma continued to be associated with help-seeking, whereas perceived public stigma was not associated with help-seeking (Golberstein et al., 2009). In addition, a negative correlation was found between peer stigma and help-seeking propensity. Hence, changing the reference group from perceived “public” stigma to perceived “peer” stigma did help change the association with helpseeking. To summarize, personal and perceived peer stigma were inversely associated with helpseeking propensity, and perceived public stigma had no relationship with help-seeking. In terms of the need for help, most participants were willing to speak with someone if mental health services were affecting their academic performance. About half of the participants received mental health services, which is reassuring especially because the psychological distress mean score fell in the likely to have a moderate disorder range. This is a college student sample that cannot be directly compared to the adult sample in the SAMHSA (2019) report. Nonetheless, the results align well as the SAMHSA (2019) also showed that only half of the U.S. adults sought help. This shows that barriers to help-seeking are still persistent, thus research exploring this critical juncture should continue to inform the mental health campaigns. Limitations One of the major limitations that warrant attention is that the data was collected during a worldwide pandemic, COVID-19. This is evident by the mean score of the Kessler Psychological Distress Scale, which fell in the likely to have a moderate disorder range. The ceiling effect in the Mental Health Services Intention Scale (MHSIS) could also be influenced by the increase in mental health awareness and utilization amid the pandemic. In addition, the research participants consisted of college students who had access to mental health services via university counseling services. As a self-report survey, social desirability bias is a key consideration while interpreting results, especially for variables such as personal stigma. Moreover, we utilized a convenience sample (under graduate students from Hollins University, mostly women and White/non-Hispanic/European) which greatly limits the generalizability of our results. If compared to a co-ed institution or different ethnicities, there is a chance for the results to vary dramatically. Hence, results should be interpreted with caution, and the special demographics should be taken into account.

Malik and Mann | Predictors of Help-Seeking

Implications and Future Research One of the major strengths of the current study is that we investigated the relationship between variables that have not been studied before to help fill the gaps in the literature. Being the first study that we know of to explore SCC with help-seeking, replication is highly recommended. According to the current study, lower SCC is associated with a decrease in help-seeking, despite the association trending toward higher psychological distress. Moreover, lower SCC was found to be associated with higher stigma. Hence, this is a population that needs to be targeted by not only mental health helpseeking intervention programs, but also antistigma campaigns. The results also call for direct intervention at improving SCC, especially when research shows that self-concept is changeable during the adolescent period (Schwartz et al., 2012). Other studies can focus on what are the factors that help shape selfconcept. For example, Parise et al. (2019) showed that emotional regulation is a mediator between SCC and psychological maladjustment. Knowing that self-concept is shaped during the adolescence, schools should focus on developing emotional regulation skills. In addition, mental health care providers should also place a greater emphasis on emotional regulation skills, especially when seeing an adolescent. Moreover, a longitudinal study over the course of early developmental years until late adolescence could bolster knowledge regarding the developmental trajectory of SCC. In addition to emotional regulation, parental factors such as warmth and clear communication (Perry et al., 2018; Van Dijk et al., 2014) influence the development of SCC. These factors should be clearly communicated to the public while explain ing the importance of a healthy identity via SCC that influences psychological distress, stigma, and subsequent help-seeking. Caregivers should also be encouraged to speak to their child regarding mental health in order to work toward reducing personal stigma and improving help-seeking behavior. If one piece of the equation, such as emotional regulation, is improved, there might be a good chance to increase SCC that is associated with lower psychological distress, lower stigma, and higher help-seeking propensity as per our results. However, as this is not an experimental or a mediation-model study, causal statements cannot be made and warrant further research. In terms of predicting help-seeking, SCC was not found to be a valid predictor, and future

studies should aim to replicate the finding to gauge whether this is true across different settings. Moreover, it is imperative to research additional fac tors that might influence help-seeking in order to inform and improve the mental health help-seeking interventions. Stigma and psychological distress continued to be salient predictors of help-seeking. This shows that stigma continues to be a barrier to seeking mental health services even after several decades and personal stigma stands to be the stron gest predictor of help-seeking in the current study. Hence, there is a need to streamline mental health and help-seeking interventions to target personal stigma. Some of the ways include contact and psy choeducation (Corrigan et al., 2012). Contact can be increased by providing an opportunity to meet people with lived experience such as done in the School Space intervention (Chisholm et al., 2016). In terms of psychoeducation, the importance and ways to seek mental health should be highlighted while also focusing on improving mental health literacy to aid familiarity. This is because familiarity with mental health has been shown to be associated with lower stigma (Corrigan et al., 2012). This is the first study to our knowledge that investigated perceived peer stigma in addition to personal stigma and perceived public stigma. Hence, this study should act as a catalyst to explore this new domain. This is especially important in the context of college students as the lifetime age of onset of 75% of mental health disorders is before the age of 24 (Kessler et al., 2005). In addition, mental health problems early in life have been shown to be associated with adverse outcomes in academics (Breslau et al., 2008); for example, depression is a predictor of lower GPA and drop ping out (Eisenberg et al., 2009). Therefore, antistigma interventions should target ways to reduce perceived peer group stigma. Our results show that perceived peer group stigma is associated with personal stigma. Given the theory that peer perceptions (i.e., perceived peer group stigma) shape personal beliefs (i.e., personal stigma), future research should focus on causal experiments. Generally, increasing the sample size and random sampling would be beneficial for future studies. Overall, these findings hold immense significance for mental health educational programs that aim at reducing stigma and increasing awareness regarding the different types of stigma that impede help-seeking behavior.

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

175

Predictors of Help-Seeking | Malik and Mann

References

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

176

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-t Ajzen, I. (2001). Nature and operation of attitudes. Annual Review of Psychology, 52(1), 27–58. https://doi.org/10.1146/annurev.psych.52.1.27 Becht, A. I., Nelemans, S. A., Van Dijk, M. P., Branje, S. J., Van Lier, P. A., Denissen, J. J., & Meeus, W. H. (2017). Clear self, better relationships: Adolescents’ selfconcept clarity and relationship quality with parents and peers across 5 years. Child Development, 88(6), 1823–1833. https://doi.org/10.1111/cdev.12921 Bigler, M., Neimeyer, G. J., & Brown, E. (2001). The divided self revisited: Effects of self-concept clarity and self-concept differentiation on psychological Adjustment. Journal of Social and Clinical Psychology, 20(3), 396–415. https://doi.org/10.1521/jscp.20.3.396.22302 Boerema, A. M., Kleiboer, A., Beekman, A. T., van Zoonen, K., Dijkshoorn, H., & Cuijpers, P. (2016). Determinants of help-seeking behavior in depression: A cross-sectional study. BMC Psychiatry, 16(1). https://doi.org/10.1186/s12888-016-0790-0 Breslau, J., Lane, M., Sampson, N., & Kessler, R. C. (2008). Mental disorders and subsequent educational attainment in a US national sample. Journal of Psychiatric Research, 42(9), 708–716. https://doi.org/10.1016/j.jpsychires.2008.01.016 Butzer, B., & Kuiper, N. A. (2006). Relationships between the frequency of social comparisons and self-concept clarity, intolerance of uncertainty, anxiety, and depression. Personality and Individual Differences, 41(1), 167–176. https://doi.org/10.1016/j.paid.2005.12.017 Campbell, J. D., Trapnell, P. D., Heine, S. J., Katz, I. M., Lavallee, L. F., & Lehman, D. R. (1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70(1), 141–156. https://doi.org/10.1037/0022-3514.70.1.141 Caprara, G. V., Di Giunta, L., Eisenberg, N., Gerbino, M., Pastorelli, C., & Tramontano, C. (2008). Assessing regulatory emotional self-efficacy in three countries. Psychological Assessment, 20(3), 227–237. https://doi.org/10.1037/1040-3590.20.3.227 Chisholm, K., Patterson, P., Torgerson, C., Turner, E., Jenkinson, D., & Birchwood, M. (2016). Impact of contact on adolescents’ mental health literacy and stigma: The schoolspace cluster randomised controlled trial. BMJ Open, 6(2). https://doi.org/10.1136/bmjopen-2015-009435 Cicei, C. C. (2012). Examining the association between self-concept clarity and self-esteem on a sample of Romanian students. Procedia - Social and Behavioral Sciences, 46, 4345–4348. https://doi.org/10.1016/j.sbspro.2012.06.252 Corrigan, P. (2004). How stigma interferes with mental health care. American Psychologist, 59(7), 614–625. https://doi.org/10.1037/0003-066x.59.7.614 Corrigan, P. W., Morris, S. B., Michaels, P. J., Rafacz, J. D., & Rüsch, N. (2012). Challenging the public stigma of mental illness: A meta-analysis of outcome studies. Psychiatric Services, 63(10), 963–973. https://doi.org/10.1176/appi.ps.201100529 Department of Human Services, Victoria (2001). Victorian Population Health Survey. Victorian Government Department of Human Services. https://www.health.vic.gov.au/population-health-systems/earlier-reports Diehl, M., & Hay, E. L. (2011). Self-concept differentiation and self-concept clarity across adulthood: Associations with age and psychological well-being. The International Journal of Aging and Human Development, 73(2), 125–152. https://doi.org/10.2190/ag.73.2.b Donahue, E. M., Robins, R. W., Roberts, B. W., & John, O. P. (1993). The divided self: Concurrent and longitudinal effects of psychological adjustment and social roles on self-concept differentiation. Journal of Personality and Social Psychology, 64(5), 834–846. https://doi.org/10.1037/0022-3514.64.5.834 Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, M. H. (1998). Generalized anxiety disorder: A preliminary test of a conceptual model. Behaviour Research and Therapy, 36(2), 215–226. https://doi.org/10.1016/s0005-7967(97)00070-3 Eisenberg, D., Downs, M. F., Golberstein, E., & Zivin, K. (2009). Stigma and help seeking for mental health among college students. Medical Care Research and Review, 66(5), 522–541. https://doi.org/10.1177/1077558709335173 Frijns, T., & Finkenauer, C. (2009). Longitudinal associations between keeping a secret and psychosocial adjustment in adolescence. International Journal of Behavioral Development, 33(2), 145–154. https://doi.org/10.1177/016025408098020

Golberstein, E., Eisenberg, D., & Gollust, S. E. (2009). Perceived stigma and helpseeking behavior: Longitudinal evidence from the healthy minds study. Psychiatric Services, 60(9), 1254–1256. https://doi.org/10.1176/ps.2009.60.9.1254 Hammer, J. H., & Spiker, D. A. (2018). Dimensionality, reliability, and predictive evidence of validity for three help-seeking intention instruments: ISCI, GHSQ, and MHSIS. Journal of Counseling Psychology, 65(3), 394–401. https://doi.org/10.1037/cou0000256 Kessler, R. C., Barker, P. R., Colpe, L. J., Epstein, J. F., Gfroerer, J. C., Hiripi, E., Howes, M. J., Normand, S. T., Manderscheid, R. W., Walters, E. E., & Zaslavsky, A. M. (2003). Screening for serious mental illness in the general population. Archives of General Psychiatry, 60(2), 184–189. https://doi.org/10.1001/archpsyc.60.2.184 Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the national comorbidity survey replication. Archives of General Psychiatry, 62(6), 593–602. https://doi.org/10.1001/archpsyc.62.6.593 Kulesza, M., Pedersen, E. R., Corrigan, P. W., & Marshall, G. N. (2015). Helpseeking stigma and mental health treatment seeking among young adult veterans. Military Behavioral Health, 3(4), 230–239. https://doi.org/10.1080/21635781.2015.1055866 Kusec, A., Tallon, K., & Koerner, N. (2016). Intolerance of uncertainty, causal uncertainty, causal importance, self-concept clarity and their relations to generalized anxiety disorder. Cognitive Behaviour Therapy, 45(4), 307–323. https://doi.org/10.1080/16506073.2016.1171391 Lally, J., ó Conghaile, A., Quigley, S., Bainbridge, E., & McDonald, C. (2013). Stigma of mental illness and help-seeking intention in university students. The Psychiatrist, 37(8), 253–260. https://doi.org/10.1192/pb.bp.112.041483 Link, B. G. (1987). Understanding labeling effects in the area of mental disorders: An assessment of the effects of expectations of rejection. American Sociological Review, 52(1), 96–112. https://doi.org/10.2307/2095395 Mackenzie, C. S., Knox, V. J., Gekoski, W. L., & Macaulay, H. L. (2004). An adaptation and extension of the Attitudes Toward Seeking Professional Psychological Help Scale. Journal of Applied Social Psychology, 34(11), 2410–2433. https://doi.org/10.1111/j.1559-1816.2004.tb01984.x Miller, C. A. (2020). Help seeking for mental health problems in mental health counselors: Prevalence, barriers, and predictors [Doctoral dissertation]. https://irl.umsl.edu/dissertation/1005 Mirowsky, J., & C. E. Ross. (2002). Selecting outcomes for the sociology of mental health: Issues of measurement and dimensionality. Journal of Health and Social Behaviour, 43(2), 152–170. Nearchou, F. A., Bird, N., Costello, A., Duggan, S., Gilroy, J., Long, R., McHugh, L., & Hennessy, E. (2018). Personal and perceived public mental-health stigma as predictors of help-seeking intentions in adolescents. Journal of Adolescence, 66(1), 83–90. https://doi.org/10.1016/j.adolescence.2018.05.003 Parise, M., Canzi, E., Olivari, M. G., & Ferrari, L. (2019). Self-concept clarity and psychological adjustment in adolescence: The mediating role of emotion regulation. Personality and Individual Differences, 138, 363–365. https://doi.org/10.1016/j.paid.2018.10.023 Perry, J. A., Silvera, D. H., Neilands, T. B., Rosenvinge, J. H., & Hanssen, T. (2008). A study of the relationship between parental bonding, self-concept and eating disturbances in Norwegian and American college populations. Eating Behaviors, 9(1), 13–24. https://doi.org/10.1016/j.eatbeh.2007.01.007 Phelan, J. C., Link, B. G., Stueve, A., & Pescosolido, B. A. (2000). Public conceptions of mental illness in 1950 and 1996: What is mental illness and is it to be feared? Journal of Health and Social Behavior, 41(2), 188–207. https://doi.org/10.2307/2676305 Richman, S. B., Pond, R. S., Dewall, C. N., Kumashiro, M., Slotter, E. B., & Luchies, L. B. (2016). An unclear self leads to poor mental health: Self-concept confusion mediates the association of loneliness with depression. Journal of Social and Clinical Psychology, 35(7), 525–550. https://doi.org/10.1521/jscp.2016.35.7.525 Schwartz, S. J., Klimstra, T. A., Luyckx, K., Hale, W. W., & Meeus, W. H. (2012). Characterizing the self-system over time in adolescence: Internal structure and associations with internalizing symptoms. Journal of Youth and Adolescence, 41(9), 1208–1225. https://doi.org/10.1007/s10964-012-9751-1 Stopa, L., Brown, M. A., Luke, M. A., & Hirsch, C. R. (2010). Constructing a self: The role of self-structure and self-certainty in social anxiety. Behaviour Research and Therapy, 48(10), 955–965. https://doi.org/10.1016/j.brat.2010.05.028 Substance Abuse and Mental Health Services Administration (2019). National Survey of Drug Use and Health (NSDUH) Releases. U.S. Department of Health and Human Services. https://www.samhsa.gov/data/release/2019-

Malik and Mann | Predictors of Help-Seeking

national-survey-drug-use-and-health-nsduh-releases Thornicroft, G. (2007). Most people with mental illness are not treated. The Lancet, 370 (9590), 807–808. https://doi.org/10.1016/s0140-6736(07)61392-0 U.S. Department of Health and Human Services (1999). Mental health: A report of the surgeon general. Rockville, MD: U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health. Van Dijk, M. P., Branje, S., Keijsers, L., Hawk, S. T., Hale, W. W., & Meeus, W. (2014). Self-concept clarity across adolescence: Longitudinal associations with open communication with parents and internalizing symptoms. Journal of Youth and Adolescence, 43(11), 1861–1876. https://doi.org/10.1007/s10964-013-0055-x Wadman, R., Webster, L. A. D., & Stain, H. (2017). Help-seeking and psychological distress in university students: Findings from a crosssectional online survey in the UK. European Psychiatry, 41(S1). https://doi.org/10.1016/j.eurpsy.2017.01.2101 Wells, K. B., Sturm, R., & Burnam, A. (2006). National survey of alcohol, drug, and mental health problems [Healthcare for Communities], 2000-2001. ICPSR Data Holdings. https://doi.org/10.3886/icpsr04165.v1

Author Note. Hinza B. Malik https://orcid.org/0000-0001-5088-1561 Caroline E. Mann https://orcid.org/0000-0001-9568-0552 Materials and data for this study can be accessed at https://osf.io/edfhg/. We have no known conflicts of interest to disclose. The Faculty Development and Student Research Funds Committee at Hollins University awarded the Janet L. MacDonald & Beatrice E. Gushee to assist in covering the costs of conferences in 2021 for the dissemination of this study. The lead author of this manuscript was awarded the F.J. McGuigan Psychology Award of Excellence for her honors theses. Correspondence concerning this article should be addressed to Hinza B. Malik who is now at Department of Psychology, University of North Carolina Wilmington, 601 South College Road UNCW Station 22734, Wilmington, NC, 28407, United States. Email: MalikHB@hollins.edu

177

Turn Your Compassion Into a Career in Mental Health Our online degree programs feature relationship-focused curriculum, innovative faculty, and lead to licensure.

alliant.edu/psychology

SUMMER 2022 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

178

179

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, doctoral-level, peer-review 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 Doctoral-level 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, full-time and part-time faculty members, and alumni. See membership benefits and a link to apply at www.psichi.org/page/member_benefits