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

Psi Chi Journal of

Psychological Research WINTER 2019 | VOLUME 24 | ISSUE 4

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

Â®

® ®

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH W I NTE R 2019 | VOLU ME 24, N U MBE R 4

EDITOR DEBI BRANNAN, PhD Western Oregon University Telephone: (503) 751-4200 E-mail: debi.brannan@psichi.org ASSOCIATE EDITORS ERIN AYALA, PhD St. Mary's University of Minnesota JENNIFER L. HUGHES, PhD Agnes Scott College TAMMY LOWERY ZACCHILLI, PhD Saint Leo University STEVEN V. ROUSE, PhD Pepperdine University ROBERT R. WRIGHT, PhD Brigham Young University-Idaho EDITOR EMERITUS MELANIE M. DOMENECH RODRIGUEZ, PhD Utah State University MANAGING EDITOR BRADLEY CANNON DESIGNER TAYLOR STONE EDITORIAL ASSISTANT REBECCA STEMPEL ADVISORY EDITORIAL BOARD GLENA ANDREWS, PhD George Fox University RUTH L. AULT, PhD DePaul University AZENETT A. GARZA CABALLERO, PhD Weber State University MARTIN DOWNING, PhD NDRI ALLEN H. KENISTON, PhD University of Wisconsin–Eau Claire MARIANNE E. LLOYD, PhD Seton Hall University DONELLE C. POSEY, PhD Washington State University PAUL SMITH, PhD Alverno College

ABOUT PSI CHI Psi Chi is the International Honor Society in Psychology, founded in 1929. Its mission: "recognizing and promoting excellence in the science and application of psychology." 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 Central Office 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 Central Office, 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 third-party 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.

Psi Chi Journal of

WINTER 2019 | VOLUME 24 | ISSUE 4

210

Examining the Effects of Exercise on Frustration-Induced Anxiety-Like Behavior in Rats

J. Eason Taylor , Bailee Ficzere, Jonathan St. Louis, and Timothy J. Schoenfeld* Belmont University

222

Eating Pathology in International Vietnamese and White American Undergraduate Women in the United States

Ngoc Nguyen and Champika K. Soysa* Worcester State University

235

An Experimental Analysis of Emotion Induction Prior to Reading a Health Narrative on Personal Risk Perception, Health Intentions, and Behavior

Aisha L. Udochi, University of Central Arkansas; Lindsay A. Kennedy*

Marc A Sestir*

246

, Hendrix College; and

, University of Central Arkansas

The Effects of Conflicting Dietary Information on Dieting Self-Efficacy and Motivation

Vivian Wei Lin Leung , Elizabeth J. Krumrei-Mancuso* Pepperdine University

255

Young Women's Sexist Beliefs and Internalized Misogyny: Links With Psychosocial and Relational Functioning and Political Behavior

Adrian J. Dehlin and Renee V. Galliher* Utah State University

Social and Cognitive Effects of Smartphone Use in Face-to-Face Verbal Interactions

265

, and Janet Trammell*

Timothy J. Johnson, Marygrace Y. Kaiser*. Alexander B. Swan* Eureka College

Helicopter Parenting and Emotion Regulation in U.S. College Students

274

Susan J. Wenze , Anna B. Pohoryles, and Jennifer M. DeCicco* Lafayette College

284

Invited Editorial: Psychological Science Plays a Critical Role in Addressing the Environmental Crisis

Ethan A. McMahan Western Oregon University

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH *Faculty mentor

209

https://doi.org/10.24839/2325-7342.JN24.4.210

Examining the Effects of Exercise on Frustration-Induced Anxiety-Like Behavior in Rats J. Eason Taylor , Bailee Ficzere, Jonathan St. Louis, and Timothy J. Schoenfeld* Belmont University

ABSTRACT. Frustration is an emotional event arising from decreases in expected reward following motivated behavior and is associated with stress and anxiety in humans, albeit rarely studied in rodents. Rodent studies have shown that anxiety-like behavior is a potential side effect of frustration, although the mechanisms and potential preventative actions for frustration are unknown. To study anxiety interventions in rodents, running wheels are used to consistently decrease anxiety-like behavior. However, wheel running has not been used to study its effects on frustration-induced anxious behavior. Thus, we modeled frustration in both control and running rats, and predicted that running would buffer anxiogenic effects of frustration. Long-Evans rats (N = 16) were randomly assigned to either control or exercise conditions. All rats were trained on a progressive variable ratio (VR) lever pressing schedule up to VR20. After reaching criterion, rats went through a frustration trial, during which no reward was given. After both VR20 and frustration trials, corticosterone was measured from tail blood, and anxiety-like behavior was analyzed in an open field. Last, hippocampal tissue was analyzed for dendritic spine density. Control rats had increased anxiety-like behavior, t(7) = 4.84, p = .002, and corticosterone levels, t(8) = 3.31, p = .011, following induced-frustration. However, running rats showed no such increases, t(7) = -0.24, p = .82, and had higher spine density throughout the hippocampus, t(4) = -8.21, p = .001. The present findings suggest exercise as a preventative intervention against the maladaptive effects of frustration on physiology and anxious behavior. Keywords: frustration, anxiety-like behavior, running, corticosterone, hippocampus

A WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

210

nxiety disorders are the most prevalent mental illnesses among adults in the United States, with a lifetime prevalence rate upward of 33% in the general population (Kessler, Petukhova, Samplson, Zaslavsky, & Wittchen, 2012), and this number is unlikely to have changed over the past two decades (Bandelow & Michaelis, 2015). Additionally, anxiety disorders create a burden, with an annual cost between 50â&#x20AC;&#x201C;100 billion dollars in the United States (Kessler & Greenberg, 2002; Shirneshan, 2013). This demonstrates a critical need to develop more effective and accessible treatment options.

Anxiety-Like Behavior in Rodents Rodent studies allow the ability to model anxiety disorders in an ethologically appropriate manner to discover their etiology and mechanisms for potential treatments (Lister, 1990). For example, the open field test, a common procedure used to measure general locomotor behavior, can be used to assay anxiety in rats (Prut & Belzung, 2003). Reduced exploration of the more threatening parts of the open field (e.g., the center regions) is a behavior that models novelty avoidance and caution, which are observed in anxiety disorders in humans (Frenkel et al., 2015). Although most

*Faculty mentor

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

rodent studies have investigated baseline anxiety levels, it is important that studies also understand the environmental impact on developing anxietylike behavior in rodents (reviewed in Schoenfeld & Cameron, 2015). Two different environmental experiences, frustration and running, exert opposite effects on anxious behavior in both humans and rodents (Anderson & Shivakumar, 2013; Cuenya, Fosacheca, Mustaca, & Kamenetzky, 2012; Fulk et al., 2004; Keenan & Newton, 1984), so we sought to investigate the interplay of both on anxiety-like behavior in rats. Frustration as an Anxiogenic Tool Frustration is the emotional sensation experienced when progress toward seemingly achievable goals is hindered by obstacles such that an individual perceives a lack of control (Meindl et al., 2018). Frustration is related to human anxiety in the workplace (Keenan & Newton, 1984) and school (Brotman, Kircanski, Stringaris, Pine, & Leibenluft, 2017; Wigfield & Meece, 1988). In rodents, frustra tion has been induced through the removal of expected reward in operant tasks such as lever press ing (Burokas, Gutiérrez-Cuesta, Martín-García, & Maldonado, 2012). Scull, Davies, and Amsel (1970) operationally defined this frustration effect origi nally as an increase in the intensity of response rate following a sudden period of nonreward. Although more rarely studied, there are more generalized emotional side effects to frustrative-nonreward in addition to those standard effects on lever pressing behavior. Behavioral analysis during frustration trials suggested that rats displayed consummatory behavior that resembles the state of anxiety in other tests (Cuenya et al., 2012). Specifically, Cuenya and associates (2012) demonstrated that social isola tion produced both anxiety-like behavior in the elevated plus-maze—another common apparatus used to measure anxiety in rodents—and atypical consummatory behavior during a sucrose test (ambulatory behavior and rearing) theorized to resemble anxiety. In addition, anxiolytic medica tions reduced these anxiety-like consummatory behaviors during frustration trials (Flaherty, 1990; Mustaca, Bentosela, & Papini, 2000), suggesting that frustration elicited emotional side effects that resemble anxiety in rodents. However, no studies have directly investigated if frustration produces anxious behavior outside of the frustrating sce nario, like the open field, nor if exercise prevents such anxiety-like behavior. Therefore, frustrativenonreward is an understudied mechanism we seek

to utilize to measure potential effects on anxiety-like behavior of a more emotional kind. Exercise Effects on Anxiety and Stress Hormones Conversely, aerobic exercise reduces anxiety in patients with anxiety disorders (Anderson & Shivakumar, 2013), and wheel running is con sidered a rodent model of aerobic exercise that routinely decreases anxiety-like behavior (Fulk et al., 2004; Schoenfeld et al., 2014). Wheel running, then, is commonly used with rodents to study effects of exercise on anxiety-like behavior and possible biological mechanisms that underlie its anxiolytic effects. Chronic wheel running buffered the release of corticosterone, the main stress hormone in rodents, which was increased by stressful experiences such as electric shock and physical restraint (BenaroyaMilshtein et al., 2004; Hare, Beierle, Toufexis, Hammack, & Falls, 2014), and further buffered the anxiogenic effects of stress (Lapmanee, Charoenphandhu, & Charoenphandhu, 2013). Despite these findings using physical stressors, little is known about the effects of wheel running to buf fer more emotionally driven sources of anxiety-like behavior. Therefore, the primary purpose of the present study was to investigate the role of wheel running in rodents as a potential intervention on frustration-induced increases in anxiety-like behav ior. In addition, corticosterone was measured to investigate physiological measures of stress follow ing frustration in addition to behavioral measures of anxiety-like activity. Structural Change in the Hippocampus The hippocampus is a brain area involved in the perception of stressful contexts and the production of anxious behavior (Adhikari, Topiwala, & Gordon, 2010) and is implicated as a key brain area for devel oping new interventions to treat anxiety disorders (Gorman, 2003). All three major subregions of the hippocampus (i.e., the dentate gyrus (DG) and areas CA3 and CA1) form the trisynaptic circuit and each have been functionally implicated in moderating anxious behavior (Jimenez et al., 2018; Kheirbek et al., 2013; Schumacher et al., 2018). Interestingly, these areas are all highly plastic and undergo many structural changes in response to environmental stimuli. One of these alterations— increases in dendritic spines—reflects neuron growth and is positively associated with stress recov ery and resilience (Sousa, Lukoyanov, Madeira, Almeida, & Paula-Barbosa, 2000; Yang, Shirayama, Zhang, Ren, & Hashimoto, 2015). Chronic running

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

211

Exercise Effects on Frustration-Induced Anxiety | Taylor, Ficzere, St. Louis, and Schoenfeld

(at least ten days) produced increases in dendritic spine density of all subregions of the hippocampus (Eadie, Redila, & Christie; 2005; Lin et al., 2012; Stranahan, Khalil, & Gould, 2007), suggesting that structural plasticity throughout the hippocampus is both sensitive to environmental manipulations and a mechanism for stress resilience, although these changes take time. Changes in dendritic spines within the hippocampus likely reflect key adapta tions in the brain produced by exercise to prevent future anxiety-like behavior (Schoenfeld, Rada, Pieruzzini, Hsueh, & Gould, 2013). Therefore, we investigated dendritic structure as a potential mechanism involved in frustration-induced anxietylike behavior. Although anxiolytic medications reversed learning deficits following frustration (Morales, Torres, Megias, Candido, & Maldonado, 1992), and exercise prevented effects of physical stressors on anxiety-like behavior (Greenwood et al., 2013; Lapmanee, Charoenphandhu, Teerapornpuntakit, Krishnamra, & Charoenphandhu, 2017), it is unknown how an experience like exercise impacts the effects of a less physical form of stress (i.e., frus tration) on anxiety-like behavior. Currently, there is a gap in knowledge of the emotional side effects and physiological mechanisms of frustration. There fore, we utilized frustration to explore the effects of wheel running on anxiety-like behavior and the physiological and neuronal circuits that underlie these behavioral effects. We hypothesized that running rats would display less anxiety-like behavior in the open field compared to nonrunning control rats, and that this difference would be heightened following reward-based frustration. In addition, we predicted that running rats would have decreased corticosterone expression and increased spine density within hippocampal neurons compared to control rats that would provide information about possible mechanisms for exercise effects on frustration-induced anxious behavior.

Method

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

212

Participants Eight-week-old male Long-Evans rats (Envigo) were used and randomly assigned to running and control conditions. Rats were given ad libitum access to food and water, then kept on a restricted diet (~16g chow/rat/day) schedule starting the week prior to operant training. All rats were group-housed (2–3 rats/cage) and maintained on a 12-hour light-dark cycle (lights on at 6 a.m.). All animal protocols conformed to the Institute of Laboratory Animal

Research and approved by Belmont University IACUC. Apparatuses Rats were housed in two different environments, based on treatment group. Control rats (n = 8) were housed in standard laboratory cages (25 cm x 45 cm), whereas running rats (n = 8) were housed in an otherwise standard cage, albeit larger (40 cm x 50 cm), because it contained a large running wheel directly in the cage (Lafayette Instruments) and provided them free access to running all day. Although running cages were bigger, the nonwheel areas were similar in size (23 cm x 50 cm) to stan dard nonrunning cages. For operant training, rats learned to lever press for 45 mg sucrose pellets (Bioserv) in Stu dent Learning Chambers (Lafayette Instruments) equipped with two levers and lights to act as discriminative stimuli. For anxiety-like behavior, rats were tested in an open field (1 m x 1 m). To measure corticosterone levels following anxiety testing, rats were briefly restrained (Harvard Apparatus) and tail blood serum was collected using microvette capillary tubes (KentScientific) and a centrifuge (Eppendorf). Serum was analyzed using a corticosterone ELISA kit (Enzo Life Sciences) and 1420 multilabel counter (PerkinElmer). To analyze brain tissue, brains were cut on a sliding microtome (American Optical), reacted with a Golgi stain (FD Neurotechnologies), and analyzed using an Olympus BX50 brightfield microscope aided by an Infinity 3S-1 camera (Lumenera) and ImageJ software (NIH). Procedure Experimental design. As depicted in Figure 1A, both control and running rats (n = 8 each) were housed in their respective cage environments for two weeks before operant training began at 10 weeks old. Rats first went through magazine train ing and then performed progressively higher VR schedules until completing the VR20 schedule (see below). After the first VR20 trial, all rats were tested in the open field for anxiety-like behavior, and blood was analyzed from a subset of rats (n = 5 from each group) for corticosterone levels. Next, all rats went through a frustration trial, immediately after which blood was collected again for corticosterone, and anxiety-like behavior was tested in the same open field with new context cues. At the end of testing, brains were extracted, and dendritic spines

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

were analyzed from a subset of rats (n = 3). Operant learning and frustration. Progressive ratio operant training was adapted from Rossi and Yin (2012). The same 45 mg sucrose pellets used for magazine training were given to rats in their home cage to develop motivation for reward. Rats were first trained on a fixed ratio of one (FR1), with each lever press yielding one pellet. Rats were then trained on an increasing VR schedule, receiving one pellet after an average of n lever presses. Starting at VR2, rats progressed through VR3, VR5, VR10, VR15, ending in VR20. Criterion of a successful trial was operationally defined as achieving 50 rewards within an hour. Three consecutive successful trials at a given VR schedule were required to move to the next level. Time to reach criterion (with a maximum of 60 minutes) was recorded each trial. For each VR schedule, time to reach criterion was averaged across all attempted trials (successful and unsuccessful) and used for data analysis. Twentyfour hours after the final VR20 trial, frustration was induced by placing rats back into the operant chamber for a duration of 30 minutes but with the lever deactivated from delivering reward. The num ber of lever presses every 5 minutes was recorded. All lever press and reward data were tabulated by automatic counters (Lafayette Instruments). Corticosterone measurements and anxiety-like behavior. After the first VR20 trial, rats were tested for baseline exploratory behavior during a 10-min ute interval in the open field. The open field had a smooth plastic floor finish with walls covered in laminated green construction paper and scented with lavender (LorAnn Oils) for context. Onto the floor of the open field, a 5 x 5 grid was created using colored tape to produce 25 equal squares (20 cm x 20 cm each) over which rats could freely explore. The total number of grid intersections crossed, center intersections crossed (operationally defined as crossings within the middle 3x3 grid of 9 squares), and total time in the center was collected by hand during the entire 10-minute window. Center crossings and time are inversely related with anxiety-like behavior, while total crossings reflect general locomotion (Prut & Belzung, 2003). Directly following the frustration trial, rats were placed back into the open field to measure frustration-induced anxiety-like behavior. To prevent habituation effects, context was altered during the second open field test by covering the floor with purple fine grade sandpaper, decorating the wall with laminated black and white diagonal stripes, and rosemary scent (LorAnn Oils) was used.

For both trials, blood was collected from the tail vein 20 minutes after removal from the open field to assess anxiety-related corticosterone levels. To do so, rats were briefly restrained in a clear Plexiglas tube (Harvard Apparatus) and the tail vein near the end was nicked with a razor blade. Blood was collected into microvette capillary tubes, and after 1–2 hours, blood was centrifuged at 14,000 RPM, plasma was extracted, and samples were frozen until the assay was performed. For the corticosterone assay, procedures were followed directly from the ELISA kit manual. Dendritic spine density in the DG, CA3, and CA1 of the hippocampus. To measure how wheel running induces changes in dendritic complexity in the hippocampus, rats were rapidly decapitated and whole brains extracted and processed using Golgi impregnation to fully label neurons in the hippocampus. Briefly, brains were rinsed, cut into one-inch chunks, and incubated in Golgi-Cox solu tion (Rapid GolgiStain kit; FD Neurotechnologies) for 3 weeks in the dark. Chunks were transferred to RapidGolgi Stain solution for 48 hours, after which brains were lightly frozen with dry ice, and 100 µm sections throughout the hippocampus were cut using a Microtome and mounted onto SuperFrost slides (ThermoFisher Scientific). Neurons were visualized by using a RapidGolgi Stain developing solution, graded ethanol, and cleared with xylene before being coverslipped with Permount. All slides were coded before analysis and analyzed blind to treatment condition. Neurons that were fully impregnated (with full dendritic trees for granule neurons in DG and both apical and basal dendritic trees for pyramidal neurons in CA3 and CA1, without trees being clearly cut off due to sectioning or incomplete staining) and in isolation (without confound of overlapping neigh boring neurons) were located using brightfield microscopy and analyzed for dendritic spine density (analyzed at 100x objective). Within each subregion of the hippocampus, five separate neurons with a pronounced nucleus and clearly defined dendrites were selected for analysis. One secondary or tertiary dendrite was isolated from each granule cell in the DG and apical dendrites of pyramidal neurons in CA3 and CA1 and captured with an Infinity 3S-1 monochrome camera (Lumenera). Photographed sections were then analyzed using ImageJ. For each dendritic branch, a 30 µm section was isolated and manually counted for dendritic spines and used for density measurements. The average of all five neurons per subregion was calculated and used for

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

213

Exercise Effects on Frustration-Induced Anxiety | Taylor, Ficzere, St. Louis, and Schoenfeld

all analyses. For all analyses, an alpha level of .05 was used to determine significant effects.

FIGURE 1

Operant training

Running wk1

wk2

wk3

wk4 1st VR20 trial - Open field - Blood sample

Control Runner 400 350 300

Running distance (meters/rat/week)

Weight (grams)

Results wk5 Frustration trial - Open field - Blood sample

25000 20000 15000 10000 5000 0

1 2 3 4 5 ek ek ek ek ek We We We We We

Figure 1A. After two weeks of access to running wheels, both control and running rats were trained on a progressive variable ratio (VR) schedule up to VR20. Anxiety-like behavior was measured in the open field test and blood was collected for corticosterone analysis after both the first VR20 trial and the frustration trial to measure the effect of frustration and wheel running on stress and anxious behavior. Figure 1B. Both control and running rats had similar weight throughout the experiment. Figue 1C. Rats consistently run throughout the experiment. All graphs represent means+SEM.

statistical analysis.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

214

Statistical Analysis For all analyses, running was a between-subjects variable; however, all other variables were repeatedmeasures variables and analyzed as such. To measure weight gain over five weeks of the experi ment, a 2 x 5 (Running x Week) mixed-factorial Analysis of Variance (ANOVA) was conducted. To measure running distances over the four weeks of running, group cage distances were measured and analyzed on a per-rat basis, and then a oneway repeated-measures ANOVA was conducted. To measure the effect of running on lever press behavior throughout the six VR schedules learned, a 2 x 6 (Running x VR Schedule) mixed-factorial ANOVA was conducted. To measure the effect of running on extinction during the frustration trial, a 2 x 6 (Running x Time) mixed-factorial ANOVA was conducted to analyze lever pressing behavior at 5-minute intervals. To measure the effect of running and frustration on open field behavior and corti costerone release, 2 x 2 (Running x Frustration) mixed-factorial ANOVAs were conducted compar ing post-VR20 and postfrustration trials. To measure the effect of running on dendritic spine density in the hippocampus, independent-samples t tests were conducted for each subregion separately. For all ANOVAs, Bonferroni pairwise comparisons were used to analyze main effects of repeated-measures variables, and effect sizes (h2) were determined for

Basic Characteristics of Control and Running Rats To compare weight gain over the course of the experiment between control and running rats, a 2 x 5 (Running x Week) mixed-factorial ANOVA was performed on weight of all rats (Figure 1B). A main effect of week followed by Bonferonni pairwise comparisons showed that both control and running rats gained weight while on an ad libitum diet, but had no change in weight after Week 2, when the restricted diet began, F(4, 56) = 24.71, p < .001, h2 = .64. There were no effects of running on weight overall or at any given week of the experiment: main effect running, F(1, 14) < .01, p = .96, h2 < .01, interaction, F(4, 56) = 1.86, p = .13, h2 = .12. Because rats were group-housed, individual rodent running distances could not be calculated. However, to verify group activity in the running cages, distance was measured weekly in each run ning cage (Figure 1C). A one-way within-subjects ANOVA followed by Bonferroni pairwise compari sons showed that running distances increased from Weeks 1 and 2 to Weeks 3 and 4, F(3, 21) = 25.63, p < .001, h2 = .79. Running Hastens Learning and Buffers Frustration-Induced Anxiety-Like Behavior To determine if running impacts the progression of rats through increasing-VR ratios during operant training, the average time to reach criterion (across all attempted trials) among control and running rats were compared at each schedule of VR train ing (Figure 2A). A 2 x 6 (Running x VR Schedule) mixed-factorial ANOVA showed a main effect of running, with running rats reaching criterion faster than control rats, F(1, 14) = 4.85, p = .045, h2 = .26. A main effect of VR schedule, F(5, 70) = 11.80, p < .001, h2 = .46, followed by Bonferroni pairwise comparisons showed that all rats spent similar amounts of time reaching criteria for the first three VR schedule (all p > .1). However, the last three VR schedule all took more time to reach criterion than VR5 (all p < .001). There was no interaction effect of running on learning VR schedule (p > .1). To validate frustration during the last trial, lever press rate was measured during the 30-minute frustration trial in 5-minute intervals for all rats (Figure 2B). A 2 x 6 (Running x Time) mixedfactorial ANOVA showed a main effect of time, F(5, 70) = 12.90, p < .001, h2 = .48, with Bonferroni

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

Discussion In the present study, we investigated the effects of exercise on frustration-induced anxiety-like FIGURE 2

Control Runner

40 20 0

Control

Control Runner

40 20

5 10 15 20 25 30 Minutes of Trial

400

Total crossings

% Time in center

VR20 Frustration

5 0

2 3 5 0 5 0 VR VR VR VR1 VR1 VR2

15 10

B Lever press/min

Time to reach criteria (min)

VR20 Frustration

300 200 100 0

Runner

Control

Runner

Figure 2A. Running rats progressed through variable ratio schedules more quickly, especially during later schedule with higher motivation. Figure 2B. All rats, independent of exercise, extinguished lever pressing behavior during the frustration trial. Figure 2C. Sedentary rats displayed more anxiety-like behavior by exploring the center of the open field less, however runners were buffered from this response. Figure 2D. All rats explored the open field as a whole similarly. * p < .05 compared to control or beginning of frustration trial. All graphs represent means+SEM.

FIGURE 3

200

150 100 50 0

Control Runner

VR20 Frustration

B Dendritic Spines / 30μm

Running Decreases Frustration-Elicited Corticosterone Release and Increases Dendritic Spines Throughout the Hippocampus To identify a possible biological mechanism for dif ferences in anxiety-like behavior, blood was drawn from a subset of rats after both open field tests and examined for corticosterone levels (Figure 3A). A 2 x 2 (Running x Frustration) mixed-factorial ANOVA did not yield a main effect of frustration nor an interaction effect (p > .1), but running rats had significantly less corticosterone overall than control rats, F(1, 8) = 7.29, p = .027, h2 = .48. Simple effects analyses showed that this main effect was entirely driven by corticosterone levels after the frustration trial, when running rats had significantly less corticosterone than control rats, t(8) = 3.31, p = .011, whereas running and control rats had similar corticosterone levels following VR20 trial, t(8) = 0.82, p = .44. Because the hippocampus provides negative feedback onto the hypothalamic pituitary adrenal (HPA) axis stress response (Herman & Cullinan, 1997), we measured dendritic spines throughout the hippocampus in the brains of a subset of both

running and control rats (Figure 3B). Independentsamples t tests were conducted and in all three regions of the hippocampus measured (DG, CA3, and CA1), running rats had increased spine density compared to control rats: DG, t(4) = -8.21, p = .001, h2 = .94; CA3, t(4) = -3.39, p = .027, h2 = .74; CA1, t(4) = -3.42, p = .027, h2 = .75.

Corticosterone (ng/ml)

pairwise comparisons detailing that lever press rate was slower at the end of the frustration trial than the beginning for all rats (p = .003). There was no main effect of running nor a Running x Time interaction on lever press rate (p > .1), suggesting that all rats extinguished similarly during the frustration trial. To determine if frustrative-nonreward dif ferentially affects anxiety-like behavior in running and control rats, time exploring the center of an open field test was compared between running and control rats after both the first VR20 and frustration trials (Figure 2C). A 2 x 2 (Running x Frustration) mixed-factorial ANOVA showed no main effects of either running or anxiety trial on exploration through the open field. However, a marginal interaction effect, F(1, 14) = 3.43, p = .085, h2 = .20, followed by simple effects analyses showed that control rats had an increase in anxiety-like behavior from VR20 to frustration trial by explor ing the center of the open field less, t(7) = 4.84, p = .002, whereas running rats had no increase in anxiety-like behavior, t(7) = -0.24, p = .82. This effect on exploration was not due to general locomotion because there were no main effects of running and anxiety trial, nor an interaction on total explora tion through the open field, depicted by the total number of grid crossings during the test (all ps > .1, see Figure 2D).

CA3

CA1

Control Runner

20 10 0

* Figure 3A. Running rats had lower corticosterone release during open field exploration, especially following the frustration trial. Figure 3B. Running increased dendritic spine density within the dentate gyrus (DG), CA3, and CA1 regions of the hippocampus. * p < .05 compared to control. All graphs represent means+SEM.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

215

Exercise Effects on Frustration-Induced Anxiety | Taylor, Ficzere, St. Louis, and Schoenfeld

behavior in rats and potential mechanisms mediat ing these effects. All rats were trained to lever press at high rates for reward following a progressively increasing VR schedule. Although runners performed moderately faster than control rats, all rats reached criterion for learning the VR20 schedule. To induce frustra tion, all rats were put into a nonrewarded trial for 30 minutes, during which all rats, regardless of experimental group, extinguished lever pressing behavior, suggesting that all rats experienced nonreward similarly. Importantly, a few days before the frustration trial, runner and control rats showed similar anxiety-like behavior in the open field test and corticosterone levels in blood serum follow ing a rewarded operant trial. However, after the frustration trial, control rats displayed more anxious behavior and had increased blood corticosterone levels, compared to runners. In a subset of these rats, dendritic morphology was analyzed in the hippocampus, and runners were found to have widespread increases in dendritic spine density in each major subregion of the hippocampus, providing a possible biological mechanism for the exercise effects on stress and anxiety-like behavior. Together, these effects suggest that wheel running rats have a blunted stress response to frustration, which may act toward a prevention of anxiety-like behavior following frustrating events.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

216

Effects of Running on Operant and Anxiety-Like Behavior Wheel running in rodents is commonly utilized as a rodent model of human exercise. Similar to anxiolytic actions of physical exercise in human patients (Anderson & Shivakumar, 2013; Carek, Laibstain, & Carek, 2011; Salmon, 2001), wheel run ning decreased basal anxiety-like behavior in both unstressed and chronically stressed experimental rodents across many studies (Fulk et al., 2004; Lapmanee et al., 2013; Schoenfeld et al., 2014). In addition, running prevented increases in anxietylike behavior following both physical and social stressors (Greenwood et al., 2013; Lapmanee et al., 2017; Patki et al., 2014), concluding that exercise can buffer the effects of environment stress. Our findings extend this literature by suggesting that wheel running prevented frustration-induced anx ious behavior in rats. Frustration-induced emotional changes have been compared similarly to anxious behaviors (Gray, 1978), suggesting that similar mechanisms work to reduce emotional behaviors in running rats following stressful or frustrating events

One additional explanation is that exercised rats were less frustrated following removal of a reward because of enhanced motivated behavior, and therefore their emotional behavior was less affected. Running rats in our study demonstrated enhanced motivation because they were faster to reach criterion across different VR schedules, but also progressed from one VR level to the next more quickly than sedentary controls. Long-Evans rats given access to running wheels overate following a period of caloric restriction (Evans, Messina, Knight, Parsons, & Overton; 2005), which sug gests that our running rats, on food restriction for operant training, may rebound by being more motivated to lever press for food when given the chance. Although the internal state of rats is impossible to determine, both sedentary and running rats behaved in the same manner during the frustration trial, which serves as an important manipulation check. Both rats lever pressed at a high rate at the beginning of the frustration trial and extinguished lever pressing behavior as the trial continued. Therefore, even if running rats were more motivated to lever press for reward, they were not hypermotivated to the point of ignoring the nonreward case. Assuming both running and sed entary rats notice the lack of reward, only running rats were prevented from this frustrating situation impacting their emotional behavior. Interestingly, exercised rats were no less anxious than controls when tested in the open field test after the VR20 trial, despite having run for four weeks by that point. Other studies have shown that four weeks of running was sufficient to be anxiolytic in rodents (Binder, Droste, Ohl, & Reul, 2004; Salam et al., 2009), although we only showed differences due to exercise following the frustration trial. One difference in our study is that all rats were trained to lever press for reward in addition to just running or being sedentary. Although reported research is unclear whether instrumental reward learning produces changes in anxiety-like behavior in rodents, one well-regarded theory of the anxiolytic effects of exercise is that it is dependent on the rewarding nature of running in rodents (Brené et al., 2007). Many forms of exer cise (forced and voluntary, wheel and treadmill) produced anxiolytic responses, activated reward circuitry in the brain, and produced conditioned place preference to running arenas, indicative of reward value (Greenwood et al., 2011; Herrera et al., 2016). Therefore, because of the reward ing nature of operant training, it is possible that

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

sedentary rats have less anxiety-like behavior than expected. Importantly, despite no difference in baseline anxiety-like behavior, sedentary rats displayed increased anxiety-like behavior following frustration, which was prevented by chronic exer足 cise. Although we measured anxiety-like behavior in the open field test, this test measures avoidance behavior in only one domain, exploration of a novel arena. Exploration through the center of the field is accepted to reflect anxious behavior in the open field; however, some other factors are worth considering as potential confounds. First, increased exploration in the open field may not reflect anxiety but may just reflect increased activity levels because of wheel running behavior. Second, one potential caveat in repeated open field testing is the use of fine-grade sandpaper as a tactile component on the second test that may add a stressful component to that specific test compared to the first test with a smooth floor. To address this second factor, although home cage sandpaper flooring (instead of normal sawdust bedding) is slightly aversive to rodents (Tokunaga et al., 2007), sandpaper flooring is commonly used in behavioral tests to manipulate tactile textures without affect足 ing behavior by sandpaper alone (Brydges & Hall, 2017). In addition, the smooth floor of the open field for the first test has a glossy finish that reflects light, so any anxiogenic qualities of sandpaper would be outweighed by increased reflection of aversive light in the first test. Importantly for both potential confounds, total exploration, defined by the number of grid crossings throughout the arena, did not change as a result of running or arena floor, suggesting that neither of these variables influenced general locomotion throughout the arena. The only exploration that was changed was into the more threatening parts of the field, the center, so we are confident that behavioral differ足 ences between running and control rats following frustration reflect anxiety-like behavior. In future studies, counterbalancing of open field arena configuration would be helpful to fully eliminate any contextual element of the second open field from potentially increasing anxiety-like behavior on its own. Last, other tests of anxiety-like behavior, the elevated plus-maze and light-dark box, utilize more salient threats, namely height and intense white light (Lezak, Missig, & Carlezon, 2017). Comparatively, it is possible that the open field is an anxiety test that is less sensitive to environmental stressors like frustration. In addition, other anxiety tests measure more defensive behaviors such as the

marble burying test and acoustic startle. Therefore, future studies should utilize various anxiety tests measuring both avoidance and defensive behaviors to capture the whole effects of frustration and exercise on anxiety-like behaviors in rats. Impact of Stress Processing and the Hippocampus on Behavior Frustration induced an increase in corticosterone in sedentary rats, but not running rats, mirroring changes in anxiety-like behavior. Previous research showed that a frustration effect elicited activation of the HPA axis stress response through increased release of corticosterone (Goldman, Coover, & Levine, 1973; Romero, Levine, & Sapolsky, 1995). Although open field exploration released corticosterone on its own (Campbell, Morrison, Walker, & Merchant, 2004), prior stress further increased corticosterone, negatively correlating with exploratory activity (Marin, Cruz, & Planeta, 2007). This suggests that increased corticosterone released through frustration acts as a stressor with potential causative effects on anxiety-like behavior. Although the previous study (Marin et al., 2007) described a correlational relationship between glucocorticoid release and anxiety-like behavior, pharmacological activation of the HPA axis through injections of corticotropin-releasing hormone (CRH) is anxiogenic. Global injections of CRH decreased exploratory and anxiety-like behavior in the open field when it is novel and under bright light conditions (Koob & Thatcher-Britton, 1985; Kumar & Karanth, 1996; Valdez et al., 2002). In addition, through a host of studies using agonist and antagonist drugs (see review by Lalonde & Strazielle, 2017), it is a central tenet that activation of the HPA axis produces anxiety-like behavior in the open field. Similar to our findings, previous research suggested that rodent exercise not only reduced stress-elicited anxiety-like behavior, but prevented corticosterone increases measured 30 minutes later (Benaroya-Milshtein et al., 2004). One potential mechanism for this corticosterone difference is that exercise quickened the reduction of corticosterone levels following activation of the HPA axis during stress (Hare et al., 2014). Following this, it is known that the hippocampus functions as a negative feed足 back loop onto the HPA axis, inhibiting HPA axis activity following its activation (Herman & Cullinan, 1997). Therefore, alterations in the hippocampus through long-term exercise provide one mechanism to affect responses to environmental stressors.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

217

Exercise Effects on Frustration-Induced Anxiety | Taylor, Ficzere, St. Louis, and Schoenfeld

To directly examine this, we measured den dritic spine density of granule and pyramidal neurons within the DG and regions CA3 and CA1 of the hippocampus as an indicator of hippocampal growth. Like previous studies, we found robust increases in dendritic spines within all subregions of the hippocampus (Eadie et al., 2005; Lin et al., 2012; Stranahan et al., 2007). Although increases in dendritic spines in the hippocampus were typically correlated with memory enhancements (Sorra & Harris, 2000), they should enhance other functions of the hippocampus, like negative feedback of the HPA axis as well. Indeed, an inverse relationship between dendritic spines in the hippocampus and stress-induced corticosterone is typically reported in the literature. Stress resulted in spine loss and elevated corticosterone (Magariños, McEwen, Flügge, & Fuchs, 1996), whereas enrichment increased spine density and decreased corticoste rone response (Hutchinson et al., 2012). Therefore, induced growth of hippocampal dendritic spines through exercise may be sufficient to prevent negative effects of frustration by inhibiting HPA axis activation during frustrating situations.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

218

Limitations Only a subset of rats was used to measure corti costerone response and dendritic spine density in running and control rats due to budgetary and time constraints, potentially limiting the power and reliability of these specific findings in relation to behavior. However, moderate running effect sizes for corticosterone data, notably a highly vari able measure (Segar, Kasckow, Welge, & Herman, 2009), and robust effect sizes for hippocampal spine density in running rats, a reliable finding in the literature, suggest that these effects are strong, despite the limited sample size. The big gest limitation with these subsets, then, was the inability to run correlational analyses to determine if dendritic spine density was negatively correlated with corticosterone release and positively correlated with exploratory behavior in the open field within each experimental condition separately. However, group differences suggested this correlation to be a strong possibility and future experiments should investigate this in greater depth. Additionally, it would be informative to ana lyze the correlation between individual rat wheel running distances and our dependent measures to determine if running is an all-or-none effect or graded, depending on the amount of exercise exerted. However, because rats were group housed,

it is impossible to quantitate actual individual rat running distances. Research suggested that the positive neurological effects of wheel running was prevented by social isolation housing (Stranahan, Khalil, & Gould, 2006), so group housing was purposefully chosen to maintain positive effects of wheel running. Although this makes it difficult to determine if all rats ran the same amount, quali tatively we observed all rats taking turns running, and even having multiple rats running at a time, so it is possible the wheel running distances actually underestimate the average individual distance for given rats. Future research may be able to utilize activity-tracking technology (Freund et al., 2013) to keep group housing of rats but track individual run ning distances to perform these sorts of analyses. Similar to this, we were only able to test the relationships between running, corticosterone, hippocampus structure, and anxiety-like behavior following frustration. Although running rats were less anxious, had lower corticosterone in their blood, and increased dendritic spines throughout the hippocampus, this does not mean that changes in hippocampal structure and/or corticosterone release caused changes in anxiety-like behavior. Manipulating dendritic spines within the hip pocampus without affecting other processing is currently unattainable in neuroscience methods. However, experiments can be performed that directly manipulate HPA axis function. To address this issue, adrenalectomized (ADX) rats given low-dose corticosterone replacement treatment in drinking water allows for baseline corticosterone levels to be present without stress-induced increases in corticosterone following activation of the HPA axis. Comparing ADX rats to sham controls would allow us to test whether increases in corticosterone are what causes the increase in anxiety-like behavior following a frustrating event. Although beyond the scope of the present study, future studies should look to investigate this. Due to limitations in rodent colony room avail ability, we were only able to measure our effects in male rats, so we are unable to generalize our results to female rats. Wheel running reduced anxiety-like behavior and corticosterone release in female rats (Jones, Gupton, & Curtis, 2016; Robinson, Christ, Cahill, Aldrich, & Taylor-Yeremeeva, 2019). How ever, corticosterone responses in male and female rats differed following environmental manipula tions (Kent et al., 2017). To our knowledge, no studies have been conducted on frustration-induced anxious behavior in female rats, so it would be

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

interesting to see if there are sex differences in anxiety-like responses to emotional experiences, given reliable sex differences in stress responses to emotional stressors in humans (Kirschbaum, Wüst, & Hellhammer, 1992). Last, another potential limitation is the dif ficulty in dissociating anxiolytic effects of running from potential anxiolytic effects of instrumental learning using reward. Because both running and sedentary rats displayed similar anxiety-like behavior following VR20, it is possible that lever pressing for reward has similar anxiolytic properties as running, at least at baseline and when rats are sated. Although future experiments should specifi cally test whether the rewarding nature of operant conditioning produces anxiolytic responses, we are confident that our results showed that running does cause rats to be more resilient toward environ mental stressors, allowing for greater exploration despite frustrating circumstances. The impacts of exercise on stress resilience are well-known (Kochi et al., 2017; Sciolino et al., 2015), and this resilience allows rats to remain active and adaptive to chang ing environments. Conclusion This study utilized wheel running as a rodent model of exercise to investigate the effects of exercise in preventing anxiety-like behavior in rats following an emotionally stressful situation. We found that running caused a decrease in anxious behavior following frustration, coinciding with lower stress hormone release and increased dendritic spine density in the hippocampus. Overall, the results sug gested that exercise buffers an anxiety-like response following a frustrating situation in rats and that this effect may be mediated by inhibition of stress hormones and dendritic spine changes within the hippocampus. These findings extend the literature on the anxiolytic effects of exercise and broaden the types of environmental experiences that reliably elicit anxious behavior in rodents. From an etho logical viewpoint, rats demonstrating less anxious behavior in the open field following frustration may correspond to increased exploratory behavior in the wild despite previous exposure to unrewarding situations. For a sedentary rat, acting cautiously in a new environment after failing to find a reward is not adaptive nor conducive to proactive foraging behavior needed to survive in the wild, so exercise reflects adaptive behavior attributed to strong hippocampal functioning (Glasper, Schoenfeld, & Gould, 2012). These findings also implicate aerobic

activity as a potential stress buffer for frustrating and unrewarding experiences for people worldwide. Frustration has been linked with aggression, anxiety disorders, and clinical depression (Harrington, 2006; Hokanson, 1961), so these findings extend exercise as a potential therapeutic intervention to decrease effects of frustration.

References Adhikari, A., Topiwala, M. A., & Gordon, J. A. (2010). Synchronized activity between the ventral hippocampus and the medial prefrontal cortex during anxiety. Neuron, 65(2), 257–269. https://doi.org/10.1016/j.neuron.2009.12.002 Anderson, E. H., & Shivakumar, G. (2013). Effects of exercise and physical activity on anxiety. Frontiers in Psychiatry, 4, 27. https://doi.org/10.3389/fpsyt.2013.00027 Bandelow, B., & Michaelis, S. (2015). Epidemiology of anxiety disorders in the 21st century. Dialogues in Clinical Neuroscience, 17, 327–335. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4610617/pdf/ DialoguesClinNeurosci-17-327.pdf Benaroya-Milshtein, N., Hollander, N., Apter, A., Kukulansky, T., Raz, N., Wilf, A., . . . Pick, C. G. (2004). Environmental enrichment in mice decreases anxiety, attenuates stress responses and enhances natural killer cell activity. European Journal of Neuroscience, 20, 1341–1347. https://doi.org/10.1111/j.1460-9568.2004.03587.x Binder, E., Droste, S. K., Ohl, F., & Reul, J. M. (2004). Regular voluntary exercise reduces anxiety-related behaviour and impulsiveness in mice. Behavioural Brain Research, 155, 197–206. https://doi.org/10.1016/j.bbr.2004.04.017 Brené, S., Bjørnebekk, A., Åberg, E., Mathé, A. A., Olson, L., & Werme, M. (2007). Running is rewarding and antidepressive. Physiology and Behavior, 92, 136–140. https://doi.org/10.1016/j.physbeh.2007.05.015 Brotman, M. A., Kircanski, K., Stringaris, A., Pine, D. S., & Leibenluft, E. (2017). Irritability in youths: A translational model. American Journal of Psychiatry, 174, 520–532. https://doi.org/10.1176/appi.ajp.2016.16070839 Brydges, N. M., & Hall, L. (2017). A shortened protocol for assessing cognitive bias in rats. Journal of Neuroscience Methods, 286, 1–5. https://doi.org/10.1016/j.jneumeth.2017.05.015 Burokas, A., Gutiérrez-Cuesta, J., Martín-García, E., & Maldonado, R. (2012). Operant model of frustrated expected reward in mice. Addiction Biology, 17, 770–782. https://doi.org/10.1111/j.1369-1600.2011.00423.x Campbell, B. M., Morrison, J. L., Walker, E. L., & Merchant, K. M. (2004). Differential regulation of behavioral, genomic, and neuroendocrine responses by CRF infusions in rats. Pharmacology Biochemistry and Behavior, 77, 447–455. https://doi.org/10.1016/j.pbb.2003.12.010 Carek, P. J., Laibstain, S. E., & Carek, S. M. (2011). Exercise for the treatment of depression and anxiety. The International Journal of Psychiatry in Medicine, 41, 15–28. https://doi.org/10.2190/pm.41.1.c Cuenya, L., Fosacheca, S., Mustaca, A., & Kamenetzky, G. (2012). Effects of isolation in adulthood on frustration and anxiety. Behavioural Processes, 90, 155–160. https://doi.org/10.1016/j.beproc.2012.01.003 Eadie, B. D., Redila, V. A., & Christie, B. R. (2005). Voluntary exercise alters the cytoarchitecture of the adult dentate gyrus by increasing cellular proliferation, dendritic complexity, and spine density. Journal of Comparative Neurology, 486, 39–47. https://doi.org/10.1002/cne.20493 Evans, S. A., Messina, M. M., Knight, W. D., Parsons, A. D., & Overton, J. M. (2005). Long-Evans and Sprague-Dawley rats exhibit divergent responses to refeeding after caloric restriction. American Journal of Physiology: Regulatory, Integrative and Comparative Physiology, 288, R1468–R1476. https://doi.org/10.1152/ajpregu.00602.2004 Flaherty, C. F. (1990). Effect of anxiolytics and antidepressants on extinction and negative contrast. Pharmacology and Therapeutics, 46, 309–320. https://doi.org/10.1016/0163-7258(90)90097-l Frenkel, T. I., Fox, N. A., Pine, D. S., Walker, O. L., Degnan, K. A., & Chronis-Tuscano, A. (2015). Early childhood behavioral inhibition, adult psychopathology and the buffering effects of adolescent social networks: A twenty-year prospective study. Journal of Child Psychology and Psychiatry, 56, 1065– 1073. https://doi.org/10.1111/jcpp.12390 Freund, J., Brandmaier, A. M., Lewejohann, L., Kirste, I., Kritzler, M., Krüger, A., . . . Kempermann, G. (2013). Emergence of individuality in genetically identical

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

219

Exercise Effects on Frustration-Induced Anxiety | Taylor, Ficzere, St. Louis, and Schoenfeld

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

220

mice. Science, 340(6133), 756–759. https://doi.org/10.1126/science.1235294 Fulk, L. J., Stock, H. S., Lynn, A., Marshall, J., Wilson, M. A., & Hand, G. A. (2004). Chronic physical exercise reduces anxiety-like behavior in rats. International Journal of Sports Medicine, 25, 78–82. https://doi.org/10.1055/s-2003-45235 Glasper, E. R., Schoenfeld, T. J., & Gould, E. (2012). Adult neurogenesis: Optimizing hippocampal function to suit the environment. Behavioural Brain Research, 227, 380–383. https://doi.org/10.1016/j.bbr.2011.05.013 Goldman, L., Coover, G. D., & Levine, S. (1973). Bidirectional effects of reinforcement shifts on pituitary adrenal activity. Physiology and Behavior, 10, 209–214. https://doi.org/10.1016/0031-9384(73)90299-0 Gorman, J. M. (2003). New molecular targets for antianxiety interventions. The Journal of clinical psychiatry, 64, 28–35. Retrieved from https://psycnet.apa.org/record/2003-03224-007 Gray, J. A. (1978). The neuropsychology of anxiety. British Journal of Psychology, 69, 417–434. https://doi.org/10.1111/j.2044-8295.1978.tb02118.x Greenwood, B. N., Foley, T. E., Le, T. V., Strong, P. V., Loughridge, A. B., Day, H. E., & Fleshner, M. (2011). Long-term voluntary wheel running is rewarding and produces plasticity in the mesolimbic reward pathway. Behavioural Brain Research, 217, 354–362. https://doi.org/10.1016/j.bbr.2010.11.005 Greenwood, B. N., Spence, K. G., Crevling, D. M., Clark, P. J., Craig, W. C., & Fleshner, M. (2013). Exercise-induced stress resistance is independent of exercise controllability and the medial prefrontal cortex. European Journal of Neuroscience, 37, 469–478. https://doi.org/10.1111/ejn.12044 Hare, B. D., Beierle, J. A., Toufexis, D. J., Hammack, S. E., & Falls, W. A. (2014). Exercise-associated changes in the corticosterone response to acute restraint stress: Evidence for increased adrenal sensitivity and reduced corticosterone response duration. Neuropsychopharmacology, 39, 1262– 1269. https://doi.org/10.1038/npp.2013.329 Harrington, N. (2006). Frustration intolerance beliefs: Their relationship with depression, anxiety, and anger, in a clinical population. Cognitive Therapy and Research, 30, 699–709. https://doi.org/10.1007/s10608-006-9061-6 Herman, J. P., & Cullinan, W. E. (1997). Neurocircuitry of stress: Central control of the hypothalamo–pituitary–adrenocortical axis. Trends in Neurosciences, 20, 78–84. https://doi.org/10.1016/s0166-2236(96)10069-2 Herrera, J. J., Fedynska, S., Ghasem, P. R., Wieman, T., Clark, P. J., Gray, N., . . . Greenwood, B. N. (2016). Neurochemical and behavioural indices of exercise reward are independent of exercise controllability. European Journal of Neuroscience, 43, 1190–1202. https://doi.org/10.1111/ejn.13193 Hokanson, J. E. (1961). The effects of frustration and anxiety on overt aggression. The Journal of Abnormal and Social Psychology, 62, 346–351. https://doi.org/10.1037/h0047937 Hutchinson, K. M., McLaughlin, K. J., Wright, R. L., Ortiz, J. B., Anouti, D. P., Mika, A., . . . Conrad, C. D. (2012). Environmental enrichment protects against the effects of chronic stress on cognitive and morphological measures of hippocampal integrity. Neurobiology of Learning and Memory, 97, 250–260. https://doi.org/10.1016/j.nlm.2012.01.003 Jimenez, J. C., Su, K., Goldberg, A. R., Luna, V. M., Biane, J. S., Ordek, G., . . . Paninski, L. (2018). Anxiety cells in a hippocampal-hypothalamic circuit. Neuron, 97, 670–683. https://doi.org/10.1016/j.neuron.2018.01.016 Jones, A. B., Gupton, R., & Curtis, K. S. (2016). Estrogen and voluntary exercise interact to attenuate stress-induced corticosterone release but not anxiety-like behaviors in female rats. Behavioural Brain Research, 311, 279–286. https://doi.org/10.1016/j.bbr.2016.05.058 Keenan, A., & Newton, T. J. (1984). Frustration in organizations: Relationships to role stress, climate, and psychological strain. Journal of Occupational and Organizational Psychology, 57, 57–65. https://doi.org/10.1111/j.2044-8325.1984.tb00148.x Kent, M., Bardi, M., Hazelgrove, A., Sewell, K., Kirk, E., Thompson, B., . . . Lambert, K. (2017). Profiling coping strategies in male and female rats: Potential neurobehavioral markers of increased resilience to depressive symptoms. Hormones and Behavior, 95, 33–43. https://doi.org/10.1016/j.yhbeh.2017.07.011 Kessler, R. C., & Greenberg, P. E. (2002). The economic burden of anxiety and stress disorders. Neuropsychopharmacology: The Fifth Generation of Progress, 67, 982–992. Retrieved from https://acnp.org/wp-content/uploads/2017/11/ch67_981-992.pdf Kessler, R. C., Petukhova, M., Sampson, N. A., Zaslavsky, A. M., & Wittchen, H. U. (2012). Twelve-month and lifetime prevalence and lifetime morbid risk of anxiety and mood disorders in the United States. International Journal of Methods in Psychiatric Research, 21, 169–184.

https://doi.org/10.1002/mpr.1359 Kheirbek, M. A., Drew, L. J., Burghardt, N. S., Costantini, D. O., Tannenholz, L., Ahmari, S. E., . . . Hen, R. (2013). Differential control of learning and anxiety along the dorsoventral axis of the dentate gyrus. Neuron, 77, 955–968. https://doi.org/10.1016/j.neuron.2012.12.038 Kirschbaum, C., Wüst, S., & Hellhammer, D. (1992). Consistent sex differences in cortisol responses to psychological stress. Psychosomatic Medicine, 54, 648–657. https://doi.org/10.1097/00006842-199211000-00004 Kochi, C., Liu, H., Zaidi, S., Atrooz, F., Dantoin, P., & Salim, S. (2017). Prior treadmill exercise promotes resilience to vicarious trauma in rats. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 77, 216–221. https://doi.org/10.1016/j.pnpbp.2017.04.018 Koob, G. F., & Thatcher-Britton, K. (1985). Stimulant and anxiogenic effects of corticotropin releasing factor. Progress in Clinical and Biological Research, 192, 499–506. Retrieved from https://europepmc.org/abstract/med/3878524 Kumar, K. B., & Karanth, K. S. (1996). Alpha-helical CRF blocks differential influence of corticotropin releasing factor (CRF) on appetitive and aversive memory retrieval in rats. Journal of Neural Transmission, 103(8–9), 1117–1126. https://doi.org/10.1007/bf01291796 Lalonde, R., & Strazielle, C. (2017). Neuroanatomical pathways underlying the effects of hypothalamo-hypophysial-adrenal hormones on exploratory activity. Reviews in the Neurosciences, 28, 617–648. https://doi.org/10.1515/revneuro-2016-0075 Lapmanee, S., Charoenphandhu, J., & Charoenphandhu, N. (2013). Beneficial effects of fluoxetine, reboxetine, venlafaxine, and voluntary running exercise in stressed male rats with anxiety-and depression-like behaviors. Behavioural Brain Research, 250, 316–325. https://doi.org/10.1016/j.bbr.2013.05.018 Lapmanee, S., Charoenphandhu, J., Teerapornpuntakit, J., Krishnamra, N., & Charoenphandhu, N. (2017). Agomelatine, venlafaxine, and running exercise effectively prevent anxiety-and depression-like behaviors and memory impairment in restraint stressed rats. PloS One, 12(11), e0187671. https://doi.org/10.1371/journal.pone.0187671 Lezak, K. R., Missig, G., & Carlezon Jr, W. A. (2017). Behavioral methods to study anxiety in rodents. Dialogues in Clinical Neuroscience, 19, 181. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5573562/ Lin, T. W., Chen, S. J., Huang, T. Y., Chang, C. Y., Chuang, J. I., Wu, F. S., . . . Jen, C. J. (2012). Different types of exercise induce differential effects on neuronal adaptations and memory performance. Neurobiology of Learning and Memory, 97, 140–147. https://doi.org/10.1016/j.nlm.2011.10.006 Lister, R. G. (1990). Ethologically-based animal models of anxiety disorders. Pharmacology and Therapeutics, 46, 321–340. https://doi.org/10.1016/0163-7258(90)90021-s Magariños, A. M., McEwen, B. S., Flügge, G., & Fuchs, E. (1996). Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. Journal of Neuroscience, 16, 3534–3540. https://doi.org/10.1523/jneurosci.16-10-03534.1996 Marin, M. T., Cruz, F. C., & Planeta, C. S. (2007). Chronic restraint or variable stresses differently affect the behavior, corticosterone secretion and body weight in rats. Physiology and Behavior, 90, 29–35. https://doi.org/10.1016/j.physbeh.2006.08.021 Meindl, P., Yu, A., Galla, B. M., Quirk, A., Haeck, C., Goyer, J. P., . . . Duckworth, A. L. (2018). A brief behavioral measure of frustration tolerance predicts academic achievement immediately and two years later. Emotion (Washington, DC). https://doi.org/10.1037/emo0000492 Morales, A., Torres, M. D. C., Megias, J. L., Candido, A., & Maldonado, A. (1992). Effect of diazepam on successive negative contrast in one-way avoidance learning. Pharmacology Biochemistry and Behavior, 43, 153–157. https://doi.org/10.1016/0091-3057(92)90651-u Mustaca, A. E., Bentosela, M., & Papini, M. R. (2000). Consummatory successive negative contrast in mice. Learning and Motivation, 31, 272–282. https://doi.org/10.1006/lmot.2000.1055 Patki, G., Solanki, N., Atrooz, F., Ansari, A., Allam, F., Jannise, B., . . . Salim, S. (2014). Novel mechanistic insights into treadmill exercise based rescue of social defeat-induced anxiety-like behavior and memory impairment in rats. Physiology & Behavior, 130, 135–144. https://doi.org/10.1016/j.physbeh.2014.04.011 Prut, L., & Belzung, C. (2003). The open field as a paradigm to measure the effects of drugs on anxiety-like behaviors: A review. European Journal of Pharmacology, 463(1–3), 3–33.

Taylor, Ficzere, St. Louis, and Schoenfeld | Exercise Effects on Frustration-Induced Anxiety

https://doi.org/10.1016/s0014-2999(03)01272-x Robinson, S., Christ, C. C., Cahill, M. M., Aldrich, S. J., & Taylor-Yeremeeva, E. (2019). Voluntary exercise or systemic propranolol ameliorates stressrelated maladaptive behaviors in female rats. Physiology and Behavior, 198, 120–133. https://doi.org/10.1016/j.physbeh.2018.10.012 Romero, L. M., Levine, S., & Sapolsky, R. M. (1995). Adrenocorticotropin secretagog release: stimulation by frustration and paradoxically by reward presentation. Brain Research, 676, 151–156. https://doi.org/10.1016/0006-8993(95)00111-3 Rossi, M. A., & Yin, H. H. (2012). Methods for studying habitual behavior in mice. Current Protocols in Neuroscience, 8–29. https://doi.org/10.1002/0471142301.ns0829s60 Salam, J. N., Fox, J. H., DeTroy, E. M., Guignon, M. H., Wohl, D. F., & Falls, W. A. (2009). Voluntary exercise in C57 mice is anxiolytic across several measures of anxiety. Behavioural Brain Research, 197, 31–40. https://doi.org/10.1016/j.bbr.2008.07.036 Salmon, P. (2001). Effects of physical exercise on anxiety, depression, and sensitivity to stress: A unifying theory. Clinical Psychology Review, 21, 33–61. https://doi.org/10.1016/s0272-7358(99)00032-x Schoenfeld, T. J., & Cameron, H. A. (2015). Adult neurogenesis and mental illness. Neuropsychopharmacology, 40, 113–128. https://doi.org/10.1038/npp.2014.230 Schoenfeld, T. J., Kloth, A. D., Hsueh, B., Runkle, M. B., Kane, G. A., Wang, S. S. H., & Gould, E. (2014). Gap junctions in the ventral hippocampalmedial prefrontal pathway are involved in anxiety regulation. Journal of Neuroscience, 34, 15679–15688. https://doi.org/10.1523/jneurosci.3234-13.2014 Schoenfeld, T. J., Rada, P., Pieruzzini, P. R., Hsueh, B., & Gould, E. (2013). Physical exercise prevents stress-induced activation of granule neurons and enhances local inhibitory mechanisms in the dentate gyrus. Journal of Neuroscience, 33, 7770–7777. https://doi.org/10.1523/jneurosci.5352-12.2013 Schumacher, A., Villaruel, F. R., Ussling, A., Riaz, S., Lee, A. C., & Ito, R. (2018). Ventral hippocampal CA1 and CA3 differentially mediate learned approachavoidance conflict processing. Current Biology, 28, 1318–1324. https://doi.org/10.1016/j.cub.2018.03.012 Sciolino, N. R., Smith, J. M., Stranahan, A. M., Freeman, K. G., Edwards, G. L., Weinshenker, D., & Holmes, P. V. (2015). Galanin mediates features of neural and behavioral stress resilience afforded by exercise. Neuropharmacology, 89, 255–264. https://doi.org/10.1016/j.neuropharm.2014.09.029 Scull, J., Davies, K., & Amsel, A. (1970). Behavioral contrast and frustration effect in multiple and mixed fixed-interval schedules in the rat. Journal of Comparative and Physiological Psychology, 71, 478–483. https://doi.org/10.1037/h0029160 Segar, T. M., Kasckow, J. W., Welge, J. A., & Herman, J. P. (2009). Heterogeneity of neuroendocrine stress responses in aging rat strains. Physiology & Behavior, 96, 6–11. https://doi.org/10.1016/j.physbeh.2008.07.024 Shirneshan, E. (2013). Cost of illness study of anxiety disorders for the ambulatory adult population of the United States. (Unpublished doctoral dissertation). University of Tennessee Health Science Center, Memphis, TN.

https://doi.org/10.21007/etd.cghs.2013.0289 Sorra, K. E., & Harris, K. M. (2000). Overview on the structure, composition, function, development, and plasticity of hippocampal dendritic spines. Hippocampus, 10, 501–511. https://doi.org/10.1002/10981063(2000)10:5%3C501::aid-hipo1%3E3.3.co;2-k Sousa, N., Lukoyanov, N. V., Madeira, M. D., Almeida, O. F. X., & Paula-Barbosa, M. M. (2000). Reorganization of the morphology of hippocampal neurites and synapses after stress-induced damage correlates with behavioral improvement. Neuroscience, 97, 253–266. https://doi.org/10.1016/s0306-4522(00)00050-6 Stranahan, A. M., Khalil, D., & Gould, E. (2006). Social isolation delays the positive effects of running on adult neurogenesis. Nature Neuroscience, 9, 526. https://doi.org/10.1038/nn1668 Stranahan, A. M., Khalil, D., & Gould, E. (2007). Running induces widespread structural alterations in the hippocampus and entorhinal cortex. Hippocampus, 17, 1017–1022. https://doi.org/10.1002/hipo.20348 Tokunaga, S., Takeda, Y., Shinomiya, K., Yamamoto, W., Utsu, Y., Toide, K., & Kamei, C. (2007). Changes of sleep patterns in rats with chronic constriction injury under aversive conditions. Biological and Pharmaceutical Bulletin, 30, 2088–2090. https://doi.org/10.1248/bpb.30.2088 Valdez, G. R., Inoue, K., Koob, G. F., Rivier, J., Vale, W., & Zorrilla, E. P. (2002). Human urocortin II: Mild locomotor suppressive and delayed anxiolyticlike effects of a novel corticotropin-releasing factor related peptide. Brain Research, 943, 142–150. https://doi.org/10.1016/s0006-8993(02)02707-5 Wigfield, A., & Meece, J. L. (1988). Math anxiety in elementary and secondary school students. Journal of Educational Psychology, 80, 210–216. https://doi.org/10.1037//0022-0663.80.2.210 Yang, C., Shirayama, Y., Zhang, J. C., Ren, Q., & Hashimoto, K. (2015). Regional differences in brain-derived neurotrophic factor levels and dendritic spine density confer resilience to inescapable stress. International Journal of Neuropsychopharmacology, 18. https://doi.org/10.1093/ijnp/pyu121 Author Note. J. Eason Taylor, https://orcid.org/0000-0003-0067-7540, Department of Psychology, Belmont University; Bailee Ficzere, Department of Psychology, Belmont University; Jonathan St. Louis, Department of Psychology and Neuroscience Program, Belmont University; Timothy J. Schoenfeld, https://orcid.org/0000-0001-7134-7296, Department of Psychology and Neuroscience Program, Belmont University. This study was supported by a Winter Undergraduate Research Grant provided by Psi Chi, the International Honor Society in Psychology, and internal research support from the College of Science and Mathematics at Belmont University. Special thanks to Psi Chi Journal reviewers for their support. Correspondence concerning this article should be addressed to Timothy J. Schoenfeld, 1900 Belmont Blvd, McWhorter Rm 413, Psychology, Belmont University, Nashville, TN, 37212. E-mail: timothy.schoenfeld@belmont.edu

221

https://doi.org/10.24839/2325-7342.JN24.4.222

Eating Pathology in International Vietnamese and White American Undergraduate Women in the United States Ngoc Nguyen and Champika K. Soysa* Worcester State University

ABSTRACT. Vietnamese women are understudied in the eating pathology literature (Ko et al., 2015). Addressing this gap, we used a social-cognitive perspective (Fiske & Taylor, 2017) to investigate aspects and predictors of eating pathology in 44 international Vietnamese and 40 White American undergraduate women, both studying in the United States. Our hypotheses were partially supported. Contributing to the literature, we found that there were significant differences across aspects of eating pathology between the two ethnic/cultural groups (p = .003, ηp2 = .25), where international Vietnamese undergraduate women reported significantly greater pathological eating than their White American counterparts. We confirmed that body dissatisfaction was the predominant aspect of eating pathology in both ethnic/cultural groups, and that culturally shaped self-schemata might have influenced compensatory pathological eating behaviors. Finally, we found that friend influence positively predicted pathological eating in international Vietnamese (β = .40, p = .014), and partner influence positively predicted pathological eating in White Americans (β = .46, p = .003). Our results may reflect cognitive schemata shaped by sociocultural norms in the relatively collectivistic culture in Vietnam compared to the more individualistic culture in the United States (Parks & Vu, 1994). Our findings could inform ethnic variations in interventions that target specific aspects of eating pathology in college students. Keywords: eating pathology, body dissatisfaction, culture, ethnic differences, college students

E WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

222

arly research on eating pathology focused primarily on middle-class, White women in the United States and other Western countries (Pike & Walsh, 1996). Over the past two decades, however, many researchers have shifted their focus to eating pathology in other ethnic groups, especially Asians (Cummins, Simmons, & Zane, 2005). Despite a significant increase in the number of studies that examined eating pathology among Asians, findings from these investigations were inconsistent. For example, early cross-cultural research on eating pathology reported a lower (Tsai, Hoerr, & Song, 1998) or comparable (Cachelin, Veisel, Barzegarnazari, & Striegel-Moore, 2000)

prevalence in Asians compared to Whites. Other investigations, however, found that women from Asian countries had the highest perception of being overweight and the greatest number of attempts to lose weight, when compared to their White counterparts in several Western countries (Wardle, Haase, & Steptoe, 2006). Given that earlier studies often referred to Asians as a homogeneous ethnic group, their failure to account for regional/ national heterogeneity might have contributed to the inconsistency of their results (Cummins et al., 2005). As a result, later cross-cultural research on eating pathology started to focus on one specific Asian population rather than including individuals

*Faculty mentor

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

with diverse Asian ethnic backgrounds in a single sample (Jackson & Chen, 2007; Roh Ryu, Lyle, & Mccabe, 2003). Among Asian populations, eating pathology in Vietnamese women remains understudied (Ko et al., 2015). In the present study, we examined aspects and predictors of eating pathology in ethnically homogeneous samples of women who were Vietnamese international students and White Americans, both studying in the United States as undergraduates. Despite the consideration of Asian ethnic heterogeneity, evidence from the literature sug gested that the majority of cross-cultural research on pathological eating behavior examined popula tions of more Westernized countries in Asia such as Hong Kong, Singapore, China, and Japan (Cummins et al., 2005). Eating pathology continues to be an understudied phenomenon in other Asian groups. Our review of the literature indicated that Vietnamese international college students studying in the United States, Vietnam, and elsewhere, are an understudied population. In fact, only one study has examined eating pathology in Vietnamese col lege students living in Vietnam (Ko et al., 2015). In this rare study, Ko et al. (2015) found that 48.8% of Vietnamese students reported high levels of pathological eating behavior, providing evidence for the presence of eating pathology in this popula tion. Furthermore, there were no studies on eating pathology among international Vietnamese women studying in the United States. Whereas identifying differences between eth nic/cultural groups is important, we acknowledge that recognizing similarities is as relevant in crosscultural research on eating pathology. Notably, past research on pathological eating behavior revealed high rates of body dissatisfaction across multiple ethnic groups in the United States (Eisenberg, Nicklett, Roeder, & Kriz, 2011). A recent survey conducted in several Western countries revealed that 89% of White women were dissatisfied with their bodies (Swami, Tran, Stieger, & Voracek, 2015). Likewise, researchers documented body dis satisfaction as the most commonly reported aspect of eating pathology in Asian Americans (Smart & Tsong, 2014). In a Vietnamese population, Ko et al. (2015) found that 41.4% of Vietnamese under graduates were unhappy with their body shape despite a low prevalence of overweight and obese college students. Given the preceding evidence, it may be beneficial to examine levels of body dis satisfaction in relation to other aspects of eating pathology in both international Vietnamese and

White American undergraduate women. Eating Pathology Eating pathology refers to a range of distinct cognitive and behavioral symptoms that are characteristic of eating disorders (Ciao, Loth, & Neumark-Sztainer, 2014). The DSM-5 identified four categories of eating disorders: anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder (BED), and eating disorder not otherwise specified (EDNOS; APA, 2013). Features of AN include restriction of energy intake that results in significantly low body weight, intense fear of becom ing fat, and distorted self-evaluation (APA, 2013). Defining BN, the DSM-5 stated that individuals with this eating disorder engage in recurrent episodes of binge eating (i.e., a discrete, short period of time in which individuals consume a large amount of food and experience a lack of control over eating) and compensatory behaviors to prevent weight gain (e.g., vomiting, using laxatives or diuretics, fasting, or excessive exercising). Importantly, the DSM-5 specified that individuals with AN may binge eat and use purging behaviors on occasion. Similar to BN, BED is characterized by recurrent episodes of binge eating, but individuals with this disorder do not engage in compensatory behaviors (APA, 2013). Additional diagnostic criteria for BED included in the DSM-5 are depression, guilt, and marked distress. Lastly, EDNOS comprises a diagnostic group of individuals who do not meet criteria for AN, BN, or BED but exhibit characteristics of an eating disorder at a clinically significant level (i.e., symptoms that cause significant distress or impair ment; APA, 2013). A dimensional approach is widely used in assessing abnormal eating behaviors (Forbush et al., 2013). That is, instead of categorizing individuals into different types, assessments address the level of symptom severity across dimensions of eating pathology. Luo, Donnellan, Burt, and Klump (2016) found that a dimensional approach outper formed a categorical approach in assessing eating pathology in a community sample. Researchers have established specific dimensions of eating pathology using popular multidimensional measures such as the self-report Eating Disorder Examination Questionnaire (Fairburn & Beglin, 1994) or the Eating Disorder Inventory-3 (Garner, 2004). Forbush et al., however, criticized the dimensions of eating pathology created on the basis of the pre ceding measures, stating that they were too broad and had low discriminant validity. In addition,

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

223

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

they stated that the preceding assessment tools were developed and validated using small samples exclusively comprised of young women. Using the same multidimensional approach, Forbush et al. established eight aspects of eating pathology using factor analysis to examine data collected from diverse, large, clinical and nonclinical samples. They demonstrated that the eight dimensions of eating pathology (body dissatisfaction, cognitive restraint, binge eating, restricting, purging, exces sive exercising, muscle building, and negative attitudes toward obesity) are conceptually relevant to the framework of eating disorder psychopathol ogy and adequately capture the diagnostic criteria of the DSM-IV-TR (APA, 2000). In the present study, we aimed to examine eating pathology using the multidimensional approach that Forbush et al. identified.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

224

Theoretical Perspective Fiske and Taylor (2017) described social cognition as cognitive processes through which individuals form mental representations of themselves in social situations and in terms of social relations. These representations, or schemata, consist of domain-specific qualities that are important to individuals, such as body image (Fiske & Taylor, 2017). That is, through social interaction, individu als learn what others expect from them and actively construct aspects of their relational self-schemata based on domain-specific expectations of socially significant others in their lives (Andersen & Chen, 2002). During childhood, for example, children form their relational self-schemata in accordance with the expectations of their parents who, at the time, are the most salient source of influence (Fiske & Taylor, 2017). As individuals grow up and encounter new people who take on the role of significant others, such interactions could activate early relational self-schemata and elicit established responses (Andersen & Chen, 2002) or modify them in these novel social contexts (Crocker, Fiske, & Taylor, 1984). Whereas social cognition is a universal cogni tive process, Fiske and Taylor (2017) emphasized that individuals form their self-schemata differ ently depending on their cultural backgrounds. For example, in many Western countries with an individualistic orientation, individuals tend to focus on their individuality and unique character istics that separate them from others. In contrast, individuals from collectivistic backgrounds such as many East Asian, Southern European, and

Latin American cultures view themselves as parts of a larger social group and build their identity around group membership. In addition, Fiske and Taylor stated that these cultural distinctions in cognition manifest themselves in how individuals construct their self-schemata. They suggested that, because the emphasis of most Western cultures is on the individual, this cultural background fosters encouragement for self-promotion. On the other hand, in most East Asian countries, individuals may form their self-schemata around the sociocultural expectation to achieve the greater good. This striving for group achievements, consequently, may create an environment that is conducive to criticism if individuals fail to accomplish group goals (Fiske & Taylor, 2017). Despite past research finding no differences in levels of body dissatisfaction between Asian American women and their White American peers (Grabe & Hyde, 2006), the collectivistic orientation of Asian women (from Asia) may place greater pressure on them to meet a set of beauty standards that are central in their culture. In fact, some Asian women perceived adhering to their cultural expectation of thinness as bringing honor to their family (Smart, Tsong, Mejía, Hayashino, & Braaten, 2011). Vietnam, like most other Asian countries, is a relatively collectivistic society (Parks & Vu, 1994). In the present study, we propose that striving for the socially prescribed idealized body is a domain-specific quality in the self-schemata of both international Vietnamese and White American students. Nonetheless, international Vietnamese women may experience more sociocultural pres sure to be thin due to their relatively collectivistic cultural background. Adding to the preceding framework, Arnett (2000) proposed the concept of emerging adult hood as a transitional period between the ages of 18 to 25. Based on this conceptual perspective, young adults from the late teens to mid-twenties actively explore their identity through meeting new people and experimenting with romantic relation ships. Furthermore, Arnett stated that during this transition, individuals encounter critical changes regarding their relational experiences as many of them have moved out of their home and have started living independently from their parents. A large number of young adults share housing with their friends and/or their romantic partners during their college years (Arnett, 2000). The U.S. Census Bureau (2018) estimated that around 65% of U.S. young adults from 18 to 24 years of

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

age occupied college dormitories, cohabitated with their friends or unmarried partners, or lived alone. Interestingly, Vietnamese college students reported similar rates of living independently from parents, suggesting that the conceptual framework of emerging adulthood may not be unique to the U.S. population of young adults (Diep, Knibbe, Giang, & De Vries, 2013). Given that we studied a group of international Vietnamese undergraduates in the United States, we expected an even higher rate of cohabitation with nonrelatives or friends among this group of students. In the present study, we suggested that, due to significant shifts in residential status in emerging adulthood, young adults might experience greater social influence on the self-schemata from their friends and/or their romantic partners (new forms of social referencing) than from their parents. Alongside changes in residential status, Arnett (2000) proposed that risk-taking is a common feature of emerging adulthood. Risk-taking refers to any behavior that has a high degree of uncertainty and possibly danger associated with its outcomes (Rosenbloom, 2003). Without being certain about the outcome of an action, individuals engage in risk-taking using their subjective perception of whether the benefits overweigh the losses (Burns & Wilde, 1995). Previous research documented a peak of risk-taking behavior (e.g., alcohol and drug use; unprotected sex; risky driving behavior) among the U.S. young adults during their transition to independent living (Bowers, Segrin, & Joyce, 2016; Kypri, McCarthy, Coe, & Brown, 2004). Likewise, Diep et al. (2013) found that living away from home was a strong predictor of alcohol-related harm among Vietnamese college students in Vietnam. Discussing this phenomenon, Arnett stated that less parental control and greater friend influence allowed emerging adults to engage in risk-taking more freely than adolescents. Consistent with Arnett’s (2000) perspective, young adults were more susceptible to risk-taking in the presence of their friends (Gardner & Steinberg, 2005). It is possible that gaining acceptance and meeting the expectation of friends outweigh the undesirable outcomes associated with risk-taking for many young adults (Reniers et al., 2017). We proposed that emerging adults may engage in risky eating behavior in a similar manner. Shisslak, Crago, and Estes (1995) defined risky eating behavior as all manifestations of eating disorder psychopathology (e.g., binge eating, restricting, purging) but at a lower frequency and intensity.

Research evidence has supported the notion that perceived benefits of engaging in risky eating behavior may provide motivation for individuals to carry out the behavior. Specifically, Cruz-Sáez, Pascual, Salaberria, Etxebarria, and Echeburúa (2015) documented that young girls who believed that achieving a certain body weight and shape was important in seeking acceptance from others engaged in higher levels of risky eating behavior than their peers who did not hold this belief. On this basis, we proposed that socially significant sources (e.g., friends and partners) influence risky eating behavior even as they shape the self-schemata of emerging adults. Interpersonal Influences: Primacy of Friends and Partners Friend influence emerged as a strong predictor of eating pathology in both Asian American and White American college women. For example, friend influence, but not parental influence, predicted body dissatisfaction in Asian Americans (Javier & Belgrave, 2015) and a drive for thinness in White American undergraduates (Chang, Yu, & Lin, 2014). Investigating a specific Asian popula tion, researchers reported that Japanese women attributed symptoms of BN to peer pressure and bullying more than to their family dynamics (Dryer, Uesaka, Manalo, & Tyson, 2015). Interestingly, Japa nese women in this study identified pressure from partners as an additional source of interpersonal influence that contributed to their pathological eating behavior. Given the preceding evidence, we examined influences of friends and partners in predicting eating pathology in international Vietnamese and White American college students. Present Study We examined differences in aspects of eating pathology between international Vietnamese and White American undergraduate women, both study ing in the United States. In addition, we investigated differences across aspects of eating pathology within each ethnic/cultural group. Adding to the novelty of the study, we studied interpersonal influences in terms of friends and partners as predictors of pathological eating behavior in both international Vietnamese and White American college students. We hypothesized that (Hypothesis 1) aspects of eating pathology would be greater in international Vietnamese undergraduate women compared to White American undergraduate women, (Hypoth esis 2) body dissatisfaction would be higher than

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

225

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

226

other aspects of eating pathology in both (a) international Vietnamese and (b) White American undergraduate women, and (Hypothesis 3) friend and partner influences would positively predict eat ing pathology in both (a) international Vietnamese and (b) White American undergraduate women.

items examining students’ parental education as an approximation of socioeconomic status. We found that 31.8% of the international Vietnamese students and 27.5% of the White American students were first-generation college students, establishing relative similarity across the two samples.

Method

Measures Demographics. We created a questionnaire asking participants to indicate their age, gender, ethnicity, and credit enrollment to confirm study eligibility. Furthermore, we asked participants to report the highest level of education of their parents to approximate social class representation in this study. International Vietnamese students answered an additional question regarding their duration of stay in the United States (M = 2.98 years, SD = 1.37). Family Histor y of Eating Questionnaire (FHEQ; Moreno & Thelen, 1993). The FHEQ consists of eight items measuring direct and indirect parental influences on eating behavior. Participants rated the items on a 5-point Likert-type scale rang ing from 1 (not at all or never) to 5 (very much or very often). Moreno and Thelen (1993) reported evidence for the construct validity of the FHEQ in their original validation study. Later, Chang, Yu, & Lin (2014) modified items in the original FHEQ to assess direct friend influence on eating behavior in Asian American and White American college women by replacing mother and father with friends. In the same study, internal consistency Cronbach’s α for this subscale was .64 for both Asian American and White American students. In the present study, we used Chang et al.’s FHEQ to examine direct friend influence (e.g., “Have your friends ever encouraged you to go on a diet or lose weight?”) and added to it by using the same question content to assess direct partner influence (e.g., “Has your boyfriend or partner ever encouraged you to go on a diet or lose weight?”). Higher scores on both subscales indicated greater interpersonal influences on eating behavior. We dropped one item from the direct friend influence subscale to improve internal consistency reliabilities. In the present study, Cronbach’s αs of the two subscales were .68 (friend influence) and .80 (partner influence) in international Vietnamese participants. In White Americans, Cronbach’s αs were .61 and .74 for friend and partner influences, respectively. Eating Pathology Symptoms Inventory (EPSI; Forbush et al., 2013). The EPSI consisted of 45 items measuring eight aspects of eating pathol ogy. In the present study, we used seven out of

Participants A total of N = 84 college women participated in this study. Only 41 international Vietnamese, but all 40 White Americans, completed the protocols. There was a ceiling on the number of volunteer partici pants that we could recruit from the community of international Vietnamese undergraduate women studying in the United States. The colleges in our small city have a low representation of international Vietnamese students. Despite having easier access to White American students, we recruited approxi mately the same number of students to match the international Vietnamese group, for comparative purposes. International Vietnamese and White American students who were at least 18 years of age and enrolled in college full-time (at least a 12-credit course load) in the United States were eligible for the study. First-year students represented 37.2% of the international Vietnamese sample, followed by sophomores (23,3%), seniors (20.9%), and juniors (16.3%). Similarly, almost half the White American sample was comprised of first-year stu dents (47.5%), followed by sophomores (32.5%), juniors (15%), and seniors (5%). International Vietnamese students reported an average of 36 months (SD = 15.67) regarding their duration of stay in the United States. Previous research demonstrated that parental education was one of the main indicators of socioeconomic status (Erola, Jalonen, & Lehti, 2016). In the present study, the socioeconomic status of undergraduate women could have shaped residential status during their college years and influenced eating pathology. In fact, Nevonen and Norring (2004) provided evidence for a positive association between socioeconomic status and eating pathology. We recruited White American undergraduates from a public university, where a large proportion of students come from workingclass families. In contrast, a considerable number of our international Vietnamese participants were drawn from private universities in the community, and their family backgrounds may reflect a higher socioeconomic status. On this basis, we included

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

the eight subscales including Body Dissatisfaction (e.g., “I did not like how clothes fit the shape of my body”), Binge Eating (e.g., “I ate when I was not hungry”), Cognitive Restraint (e.g., “I tried to exclude ‘unhealthy’ foods from my diet”), Purging (e.g., “I considered taking diuretics to lose weight”), Restricting (e.g., “People told me that I do not eat very much”), Excessive Exercising (e.g., “I felt that I needed to exercise nearly every day”), and Negative Attitudes Toward Obesity (e.g., “I thought that obese people lack self-control”). We excluded items from the Muscle Building subscale because this aspect of eating pathology is more prevalent in men (Forbush et al., 2013). Participants rated the items on a 5-point Likert-type scale, where 0 indicated never, and 4 indicated very often. We used subscale mean scores for all comparative statistical analyses because the number of items differed across subscales. Forbush et al. validated the EPSI using a large sample of college students and found that the questionnaire had excellent internal consistency, with Cronbach’s αs ranging from .84 to .89 for each subscale, and a mean test-retest reliability coefficient of .73. In another validation study, Forbush, Wildes, and Hunt (2014) found that the EPSI was significantly correlated with other measures of eating pathology, providing evidence for its convergent validity. The Chinese-version of the EPSI had acceptable internal consistency and good convergent validity (Tang, Forbush, & Lui, 2015). In the present study, the EPSI had an internal consistency Cronbach’s α of .91 in both international Vietnamese and White American undergraduates. Cronbach’s αs for all the subscales ranged from .68 to .88 in international Vietnamese students and from .72 to .90 in White American students. Procedure We obtained ethics approval from the institutional review board at our university and informed consent from participants. White American under graduate women were recruited via a departmental research pool. Using the chain-referral method of sampling, we recruited international Vietnamese undergraduates studying in the United States by reaching out to students who were known contacts from universities in the region and their associates who expressed interest in participating in the study. Data collection occurred in person. Instead of using paper and pencil protocols, participants completed the protocol as an online digital survey that we constructed on Qualtrics software. Students

used their own computers or ones that we provided. The protocol consisted of the demographic form, the FHEQ–modified (Moreno & Thelen, 1993), and the EPSI (Forbush et al., 2013). The duration of participation was about 30 minutes. For their participation, White American students received research experience credit as partial fulfillment of a requirement in their 100- and 200-level psychology classes. In addition, international Vietnamese and White American students had the opportunity to enter a voluntary raffle for two $25 gift cards.

Results Descriptive Statistics Descriptive statistics and a correlation matrix for two sources of interpersonal influences and patho logical eating by ethnicity/culture appear in Table 1. Among international Vietnamese, the duration of stay in the United States was not significantly cor related with any of the eating pathology constructs, with significance levels ranging from p = .27 to p = .91. International Vietnamese and White American undergraduates did not differ significantly in terms of first-generation and continuing-generation status, an approximation of social class in this study, t(80) = -.42, p = .674 (two-tailed). Field (2016) as well as Ghasemi and Zahediasl (2012) stated that significance tests are unaffected by the shape of observed data in samples greater than n = 30, because sampling distributions are normally distributed in such instances. Hypothesis 1 Using a one-way Multivariate Analysis of Variance (MANOVA), we examined potential ethnic/cultural differences in aspects of eating pathology between international Vietnamese and White American undergraduates. Box’s test of equality of covariance TABLE 1 Correlations for Study Variables by Ethnicity Variable

M (SD)

1. Friend influence

–

.38**

.44**

6.71 (2.49)

2. Partner influence

.15

–

.24

7.34 (3.49)

3. P athological eating

.25

.49**

–

98.54 (21.16)

M (SD)

4.85 (1.78)

5.65 (2.29)

90.48 (19.65)

Note. The analyses above are one-tailed, Pearson Bivariate Correlations. The correlations for international Vietnamese students are indicated on the top panel, and for White American students on the bottom panel. International Vietnamese, n = 41; White American, n = .40. ** p < .01 .

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

227

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

revealed that the assumption of homogeneity of variance was violated as p = .005, so we used Pillai’s trace test, which is a conservative index of signifi cance (Tabachnick & Fidell, 1989). There was a sig nificant ethnic/cultural difference across aspects of eating pathology, Pillai’s trace = .25, F(7, 73) = 3.41, p = .003, ηp2 = .25, a large effect size (Cohen, 1988), with an observed sample power of .95. Follow-up one-way Analyses of Variance (ANOVAs) revealed that international Vietnamese participants reported significantly higher levels of eating pathology than White American participants regarding binge eating, purging, restricting, and negative attitudes toward obesity, partially supporting Hypothesis 1 (see Table 2). There were no significant differences between the two groups on body dissatisfaction, cognitive restraint, and excessive exercise. Hypothesis 2 We conducted a one-way repeated-measures ANOVA to examine differences across aspects of eating pathology within each ethnic/cultural group. Mauchly’s test of sphericity indicated that the assumption of sphericity was violated in both TABLE 2 Ethnic/Cultural Differences in Aspects of Eating Pathology Aspects of eating pathology Body dissatisfaction Binge eating

Cognitive restraint

Purging

Restricting

Excessive exercise

Negative attitudes toward obesity

Ethnicity

International Vietnamese

2.83

0.81

White American

2.93

0.73

International Vietnamese

2.55

0.77

White American

2.23

0.66

International Vietnamese

2.82

1.01

White American

2.80

1.11

International Vietnamese

1.65

0.85

White American

1.24

0.48

International Vietnamese

2.68

0.77

White American

2.24

0.84

International Vietnamese

2.42

0.82

White American

2.69

1.04

International Vietnamese

2.36

0.88

White American

1.89

0.86

F(1, 79)

η p2

0.36

.551

.01

4.24

.043

.05a

0.01

.929

.00

7.00

.010

.08a

5.98

.017

.07a

1.70

.196

.02

5.94

.017

.07a

Note. M = mean (of subscale mean scores), SD = standard deviation, amedium effect sizes (Cohen, 1988).

228

analyses (international Vietnamese sample, p = .001; White American sample, p = .015). Statistical values were the same for both the Sphericity Assumed and Greenhouse Geisser indices. As a result, we presented ANOVA results using the Sphericity Assumed index. There were significant differences across aspects of eating pathology in international Vietnamese, F(6, 240) = 13.00, p < .001, ηp2 = .25, and in White Americans, F(6, 234) = 28.48, p < .001, ηp2 = .42, both effect sizes were large (Cohen, 1988). Planned pairwise comparisons revealed that international Vietnamese participants rated body dissatisfaction significantly higher than binge eating, purging, restricting, and negative attitudes toward obesity, partially supporting Hypothesis 2a (see Table 3). Furthermore, body dissatisfaction was significantly higher than binge eating, purging, excessive exercise, and negative attitudes toward obesity in White American participants, partially supporting Hypothesis 2b (see Table 3). Body dissatisfaction did not differ significantly from either cognitive restraint or excessive exercise in international Vietnamese, and either cognitive restraint or restricting in White Americans. Hypothesis 3 To test Hypothesis 3, we used multiple regression analyses to examine interpersonal influences as predictors of pathological eating in international Vietnamese and in White American undergradu ate women. Friend influence positively predicted pathological eating in international Vietnamese participants, but partner influence did not, AdjR2 = .15, F(2, 38) = 4.64, p = .02, n = 41, Cohen’s ƒ2 = .18, a medium effect size (Cohen, 1988), partially supporting Hypothesis 3a (see Table 4). On the other hand, partner influence positively predicted pathological eating in White American participants, but friend influence did not, AdjR2 = .23, F(2, 37) = 6.80 p = .003, n = 40, Cohen’s ƒ2 = .30, a mediumlarge effect size (Cohen, 1988), partially supporting Hypothesis 3b (see Table 4).

Discussion Evidence from the literature on pathological eating behavior suggests that Vietnamese women are understudied in this regard (Ko et al., 2015). Addressing the preceding gap, we examined aspects of eating pathology in a sample of international Vietnamese college women in the United States. First, we investigated potential differences in aspects of eating pathology between international Vietnamese undergraduate women and their White

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

American counterparts. Second, we examined differences between aspects of eating pathol ogy within international Vietnamese and White American students. Lastly, we studied the inter personal influences of friends and partners as predictors of pathological eating behavior in both groups of students. We used social cognitive theory (Fiske & Taylor, 2017) to frame our questions and interpret our findings. We found that aspects of eating pathology differed between international Vietnamese and White American undergraduate women (see Table 2). Specifically, international Vietnamese students reported higher levels of binge eating, purging, restricting, and negative attitudes toward obesity than did their White American peers, partially supporting our first hypothesis. Our results were consistent with Ko et al. (2015), who found a high prevalence of pathological eating behavior in Vietnamese undergraduate women studying in Vietnam. Adding to their study, we found evidence illustrating ethnic/cultural differences in eating pathology between international Vietnamese students and their White American peers while also identifying novel aspects of eating pathology in this population, beyond those identified by Ko et al. We found no differences in body dissatisfaction, cognitive restraint, and excessive exercise between the two groups of students. Similarly, a meta-analysis on ethnicity and body image also established no sig nificant difference in body dissatisfaction between Asian Americans and White Americans (Grabe & Hyde, 2006). No studies have addressed differences in stan dards of beauty between international Vietnamese and White American women. Our results extended the literature by providing evidence for potentially differing cultural values with regard to the thin ness ideal between international Vietnamese and White American female undergraduates. Previous cross-cultural research on perceived physical attractiveness revealed that women from Asian countries had a considerably lower idealized body size compared to Western standards (Lee, Leung, Lee, Yu, & Leung, 1996; Rongmuang et al., 2011). Interestingly, Wardle et al. (2006) examined body image and weight control across several countries in Europe, the Mediterranean, Pacific Asia, and the Americas, and found that women from Asian countries had the lowest body mass index but reported the highest levels of weight concerns and perceptions of being overweight. Likewise, in the study by Ko et al. (2015), only 4% of Vietnamese

undergraduate women were overweight or obese, but more than 40% of them believed themselves to be fat. The preceding findings together indicate that the thinness ideal is, indeed, more severe among Asians in general compared to other ethnicities, and among Vietnamese in particular. Addressing cultural variations in social cognition, Fiske and Taylor (2017) stated that individuals with a collectivistic background formed their selfschemata around the sociocultural expectation to adhere to group norms. Our results supported this cultural perspective of social cognition, demonstrat ing that Vietnamese women may internalize the thinness ideal of their culture and perceive this beauty standard as an integral part of their mental self-representation. Supporting our second hypothesis, body dis satisfaction emerged as the predominant aspect of eating pathology reported by the undergraduate women in our study. Body dissatisfaction was significantly higher than binge eating, purging, excessive exercise, and negative attitudes toward obesity in international Vietnamese college students (see Table 3). Similarly, body dissatisfaction was TABLE 3 Pairwise Comparisons Between Body Dissatisfaction and Other Types of Eating Pathology

International Vietnamese (n = 41)

Body dissatisfaction compared to other aspects of eating pathology

F(1, 40)

η p2

F(1, 39)

η p2

Binge eating

7.22

.010

.15b

25.41

.000

.40b

Cognitive restraint

0.00

.979

.00

0.68

.415

.02

105.69

.000

.73b

206.40

.000

.84b

0.66

.423

.02

Excessive exercise

10.01

.003

Negative attitudes toward obesity

11.69

.001

Purging Restricting

White American (n = 40)

21.60

.000

.36b

.20

1.89

.177

.05

.23

51.09

.000

.57b

Note. bLarge effect sizes (Cohen, 1988).

TABLE 4 Regression Analyses for Interpersonal Influence Predicting Pathological Eating Predictor

International Vietnamese β

White American

95% CI

Friend influence

.40*

[0.72, 6.12]

.18

[-1.21, 5.17]

Partner influence

.09

[-1.41, 2.45]

.46**

[1.48, 6.42]

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

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

229

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

230

significantly higher than binge eating, purging, restricting, and negative attitudes toward obesity in White American college students (see Table 3). The preceding findings are consistent with previ ous research that reported a high level of body dissatisfaction in Vietnamese college students (Ko et al., 2015) and White American college students (Eisenberg et al., 2011). Despite the preceding similarities, body dissatisfaction did not differ significantly from restricting among international Vietnamese and from excessive exercise among White Americans. It appears that international Vietnamese students who were high in body dissatisfaction were also high in restricting their food intake. On the other hand, White American students who were high in body dissatisfaction were also high in using excessive exercise. Our findings may reflect a cultural difference in the manner that college students engage in compensatory behaviors regarding their body dissatisfaction. Indeed, examining differing beauty standards between Asian and Western women, previous cul tural research revealed an interesting difference in their preference for body shape. Specifically, studies of women in Western countries indicated that hav ing muscle tone was one of the most commonly desired physical characteristics (Altabe, 1998). This was reflected in our finding that White Americans who were high in body dissatisfaction were also high in excessive exercise. On the other hand, Asian women were more concerned about being thin rather than having a fit body (Kaneko, Kiriike, Ikenaga, Miyawaki, & Yamagami, 1999; Rongmuang et al., 2011). The preceding finding was true in our study as well, where international Vietnamese women who were high in body dissatisfaction also restricted their food intake. This difference in the type of body that is desired illustrates beauty standards that may be culturally shaped. From a social cognitive perspective, having a toned and fit body may be an appearance-based quality on which White American women build their self-schemata. In contrast, if wanting a thin body is a culturally shaped beauty standard among international Vietnamese women, they may be more likely than their White American peers to form self-schemata based on this expectation. In turn, in their efforts to achieve the “fit” cognitive self-schemata, White American women appear to engage in excessive exercising. On the other hand, international Vietnamese women seem to engage in restricting food intake to address body dissatisfaction because this particular compensatory behavior could help

them to achieve the “thin” cognitive self-schemata. Further adding to the literature, our study revealed that friend or partner influence posi tively predicted eating pathology in international Vietnamese and White Americans, respectively (see Table 4). Our results reflected our theoretical underpinning of social cognition in the context of emerging adulthood. Fiske and Taylor (2017) proposed that individuals form self-schemata based on domain-specific expectations (e.g., body image) of socially significant others (e.g., parents and friends) through social interaction. Whereas parents are a salient source of influence during childhood, college-age individuals may be more influenced by their friends and their partners due to major changes in their residential status (Arnett, 2000). As a result, we suggested that undergraduate women may shape their self-schemata using the appearance-based expectations of their friends and/or their partners. Our findings provided evidence for this perspective, demonstrating that college women may engage in pathological eating behavior in order to seek approval from their friends and partners. On this basis, we propose that clinical interventions for college students may be more effective if they address pressure from friends and partners as significant contributors to the development and/or maintenance of eating pathology. Interestingly, however, we identified a dif ferential influence of the type of social influence on eating pathology. That is, we found that friend influence positively predicted pathological eating in international Vietnamese undergraduates, but partner influence did not. In contrast, for White American undergraduates, partner influence positively predicted pathological eating, but their friend influence did not. The preceding differences in significant social influences may inform us about variation in underlying self-schemata in these two groups of women. Because we did not ask students in our study to indicate whether they had a signifi cant partner, our methodology could be insufficient to examine the influence of partners as a predictor of eating pathology. We suspect that only a small number of our international Vietnamese students were in a romantic relationship at the time of our study, which might have contributed to our findings in this group. Regarding transition to independent living in emerging adulthood, Arnett (2000) stated that large numbers of U.S. young adults cohabitate with their partners during their college years rather than with their parents. It is important, however, to

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

note that the data and research that Arnett used to support his theory were drawn from samples in the United States, where cultural norms may be different from those in Vietnam. As in most Asian countries, living with a partner before marriage may be considered inappropriate in the Vietnamese culture. Our international Vietnamese undergradu ate women may be more likely to cohabitate with their friends rather than their romantic partners, thereby experiencing more social influence on their self-schemata from their friends than their partners. Finally, we revisit individualistic and collectivis tic values here, in relation to both the differential pursuit of culturally idealized body types between Asian and Western women as well as variations in social influence on eating pathology. In contrast to individualistic cultures, collectivistic cultures value the needs and goals of the group over personal desires (Triandis, Bontempo, Villareal, Asai, & Lucca, 1988). Markus and Kitayama (1991) estab lished that individuals who live in a collectivistic culture tend to develop a more interdependent self-construal and less independent self-construal. Because interdependent self-construal places a great emphasis on relationships with others, women in collectivistic cultures may experience more sociocultural pressure than their peers from an individualistic background, to meet the cultural expectation of an idealized body. Examining predictors of eating disturbances among Asian American undergraduate women, Chang, Yu, and Kahle (2014) found that individuals who were high in interdependent self-construal and low in independent self-construal reported greater levels of body dissatisfaction and bulimic symptoms. This conceptual distinction between interdependent and independent self-construal is consistent with the theoretical underpinning of social cognition (Fiske & Taylor, 2017). In Asian cultures, achieving the cultural expectation of thinness may bring honor to their family (Smart et al., 2011). Our findings provided evidence regarding the influences of sociocultural context in shaping self-schemata. Vietnamese women may be more likely than their White American counterparts to form their self-schemata based on heightened sociocultural expectations. Furthermore, they may experience more pressure to achieve the culturally shaped thin ness ideal because of their collectivistic values. This may explain the high levels of pathological eating, especially with regards to compensatory behaviors, among international Vietnamese students.

Limitations We acknowledge that our study had several limita tions. First, our sample size was relatively small for a cross-cultural comparative study, which could have contributed to some of our nonsignificant results. In a literature review, Cummins et al. (2005) stated that to achieve a power level of .80, a sample needs at least 193 participants divided equally into two subgroups. Lack of power could result in nonsig nificant findings in many cross-cultural comparative studies (Cummins et al., 2005). Nonetheless, it is worth noting that, despite our small sample size, MANOVA yielded a ηp2 of .25, a very large effect size (Cohen, 1988), and an observed power of .95, regarding differences in eating pathology between international Vietnamese and White American undergraduate women. Additionally, other analyses with significant results in our study had medium effect sizes, in relation to both differences between types of eating pathology as well as interpersonal predictors of eating pathology. Together, these findings extend the literature, despite the small sample sizes. Second, our questionnaires did not account for many factors that may influence our results. For example, when asking international Vietnamese students to indicate the influences of partners on their eating behavior, we did not consider the fact that many of them may not have had a significant partner at the time of the study compared to their White American peers. Furthermore, when we examined social influences as predictors of eating pathology, the subscale assessing direct friend influence yielded low Cronbach’s αs, .68 and .61, in international Vietnamese and White American college women, respectively. Likewise, Chang, Yu, and Lin (2014) reported Cronbach’s αs of .64 for direct friend influence in both Asian American and White American samples. On the other hand, in the subscale developed for the present study, Cronbach’s αs for direct partner influence were adequate in both international Vietnamese (.80) and White American (.74) undergraduates. It appears that direct friend influence was especially inconsistent among White American women, per haps because its salience varied within this sample. All other subscales used in this study demonstrated acceptable reliability. Regarding the construct validity of our mea sures, cultural and language differences between our international Vietnamese and White American undergraduates could have led to different inter pretations of our questionnaires. Nonetheless, the

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

231

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

232

measures used in this study were previously used with Asian samples. For example, the FHEQ was used to study eating pathology in a sample of Asian American undergraduate women (Chan, Yu, & Lin, 2014), and the EPSI was used to study eating pathology in a Chinese population (Tang, Forbush, & Lui, 2015). To account for the language differ ence as much as possible, we recruited international Vietnamese students from institutions that require a score of 80 or higher on the Test of English as a Foreign Language, which serves as an indicator that they are fluent in English. Regarding potential cultural differences, we faced the limitation of using standardized measures; that is, using standardized language across two cultures may not adequately capture equivalence in meaning in relation to the constructs we examined. This, however, would be true for studies that investigate any construct across cultures even within the United States. Given limited access to international Vietnamese students studying in our area of the United States, we did not conduct pilot measures beforehand, which could explain the low Cronbach’s α of the subscale assessing direct friend influence in both groups of participants. Given that our study was the first to examine eating pathology in a sample of international Vietnamese female undergraduates in the United States, it could serve as a pilot for future research. Additionally, we did not adequately examine the effect of Westernization and acculturation on eating pathology among international Vietnamese students because we had concerns about the length of the protocol. Nonetheless, the duration of stay in the United States was not significantly correlated with any of the eating pathology con structs, among international Vietnamese students. Whereas awareness and internalization of Western appearance norms could contribute to eating dis turbances among Asian women (Stark-Wroblewski, Yanico, & Lupe, 2005), some studies revealed that acculturation was, indeed, a protective factor of body dissatisfaction and eating pathology (Lake, Staiger, & Glowinski, 2000; Sussman, Truong, & Lim, 2007). In another study, Jackson, Jiang, and Chen (2016) found that the impact from Asian/ Chinese media was greater than that of Western media in predicting pathological eating behavior among Chinese undergraduate women. We recom mend that future studies examine the influence of Westernization and acculturation as potential correlates or predictors of eating pathology in international Vietnamese students studying in the

United States. The preceding research suggests that accounting for acculturation may increase differences in aspects of eating pathology between international Vietnamese and White American undergraduate women. Finally, international Vietnamese students were recruited from several schools in the community. As a result, we were not able to give them the same research participation credit that the White American students received. We acknowledge that this might have influenced the motivation of international Vietnamese students to complete the protocol (in fact, only 41 out of 44 international Vietnamese students completed the protocol whereas all White American participants completed the protocol). Conclusions To our knowledge, our study was the first to examine cross-cultural distinctions in pathological eating behavior between international Vietnamese and White American undergraduates in the United States. We found that international Vietnamese students reported higher levels of binge eating, purging, restricting, and negative attitudes toward obesity than their White American peers, provid ing evidence for the prevalence of pathological eating behavior in this population. Furthermore, we established the predominance of body dissatis faction within each group of students, and found that the idealized body in each culture might have shaped the compensatory behaviors that they used. Finally, we identified differential social influences in predicting eating pathology: Friends influenced international Vietnamese students, and partners influenced White American students. Together, our findings lend themselves to a social cognitive framework. Whereas international Vietnamese and White American undergraduates may form their self-schemata using the beauty standards that are specific to their culture, Vietnamese women may experience more pressure than their White Ameri can counterparts to achieve the cultural expectation of thinness because of their collectivistic back ground. Given the high prevalence of pathological eating behavior among undergraduates on college campuses (Eisenberg et al., 2011), university coun selors may be interested in developing interventions that target specific aspects of eating pathology. Furthermore, our findings indicated the need for these interventions to address ethnic/cultural variations between international Vietnamese and White American undergraduates.

Nguyen and Soysa | Eating Pathology in Vietnamese and Americans

References Altabe, M. (1998). Ethnicity and body image: Quantitative and qualitative analysis. International Journal of Eating Disorders, 23, 153–159. https://doi. org/10.1002/(SICI)1098-108X(199803)23:2<153::AID-EAT5>3.0.CO;2-J American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., Text Revision). Washington, DC: American Psychiatric Publishing. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing. Andersen, S. M., & Chen, S. (2002). The relational self: An interpersonal socialcognitive theory. Psychological Review, 109, 619–645. https://doi.org/10.1037/0033-295X.109.4.619 Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469–480. https://doi.org/10.1037/0003-066X.55.5.469 Bowers, J. R., Segrin, C., & Joyce, N. (2016). The role of transitional instability, psychological distress, and dysfunctional drinking in emerging adults’ involvement in risky sex. Journal of Social and Personal Relationships, 33, 1097–1119. https://doi.org/10.1177/0265407515617621 Burns, P. C., & Wilde, G. J. S. (1995). Risk taking in male taxi drivers: Relationships among personality, observational data and driver records. Personality and Individual Differences, 18, 267–278. https://doi.org/10.1016/0191-8869(94)00150-Q Cachelin, F. M., Veisel, C., Barzegarnazari, E., & Striegel-Moore, R. H. (2000). Disordered eating, acculturation, and treatment-seeking in a community sample of Hispanic, Asian, Black, and White American women. Psychology of Women Quarterly, 24, 244–253. https://doi.org/10.1111/j.1471-6402.2000.tb00206.x Chang, E. C., Yu, E. A., & Kahle, E. R. (2014). BMI, body discrepancy, and selfconstrual as predictors of eating disturbances in European and Asian American females. Eating Behaviors, 15, 328–330. https://doi.org/10.1016/j.eatbeh.2013.12.001 Chang, E. C., Yu, E. A., & Lin, E. Y. (2014). An examination of ethnic variations in perfectionism and interpersonal influences as predictors of eating disturbances: A look at Asian and European American females. Asian American Journal of Psychology, 5, 243–251. https://doi.org/10.1037/a0034621 Ciao, A. C., Loth, K., & Neumark-Sztainer, D. (2014). Preventing eating disorder pathology: Common and unique features of successful eating disorders prevention program. Current Psychiatry Reports, 16, 453. https://doi.org/10.1007/s11920-014-0453-0 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Crocker, J., Fiske, S. T., & Taylor, S. E. (1984). Schematic bases of belief change. In J. R. Eiser (Ed.), Attitudinal judgment (pp. 197–226). New York, NY: Springer. Cruz-Sáez, S., Pascual, A., Salaberria, K., Etxebarria, I., & Echeburúa, E. (2015). Risky eating behaviors and beliefs among adolescent girls. Journal of Health Psychology, 20, 154–163. https://doi.org/10.1177/1359105313500683 Cummins, L. H., Simmons, A. M., & Zane, N. S. (2005). Eating disorders in Asian populations: Critique of current approaches to the study of culture, ethnicity, and eating disorders. American Journal of Orthopsychiatry, 75, 553–574. https://doi.org/10.1037/0002-9432.75.4.553 Diep, P. B., Knibbe, R. A., Giang, K. B., & De Vries, N. (2013). Alcohol-related harm among university students in Hanoi, Vietnam. Global Health Action, 6, 1–10. https://doi.org/10.3402/gha.v6i0.18857 Dryer, R., Uesaka, Y., Manalo, E., & Tyson, G. (2015). Cross‐cultural examination of beliefs about the causes of bulimia nervosa among Australian and Japanese females. International Journal of Eating Disorders, 48, 176–186. https://doi.org/10.1002/eat.22269 Eisenberg, D., Nicklett, E. J., Roeder, K., & Kirz, N. E. (2011). Eating disorder symptoms among college students: Prevalence, persistence, correlates, and treatment-seeking. Journal of American College Health, 59, 700–707. https://doi.org/10.1080/07448481.2010.546461 Erola, J., Jalonen, S., & Lehti, H. (2016). Parental education, class and income over early life course and children’s achievement. Research in Social Stratification and Mobility, 44, 33–43. https://doi.org/10.1016/j.rssm.2016.01.003 Fairburn, C. G., & Beglin, S. J. (1994). Assessment of eating disorders: Interview or self-report questionnaire? International Journal of Eating Disorders, 16,

363–370. https://psycnet.apa.org/record/1995-15839-001 Field, A. (2016). Assumptions. In An adventure in statistics: The reality enigma (pp. 395–428). London, UK: Sage Publications Ltd. Fiske, S. T., & Taylor, S. E. (2017). Social cognition: From brains to culture (3rd ed.). London, UK: Sage Publications Ltd. Forbush, K. T., Wildes, J. E., & Hunt, T. K. (2014). Gender norms, psychometric properties, and validity for the Eating Pathology Symptoms Inventory. International Journal of Eating Disorders, 47, 85–91. https://doi.org/10.1002/eat.22180 Forbush, K. T., Wildes, J. E., Pollack, L. O., Dunbar, D., Luo, J., Patterson, K., . . . Watson, D. (2013). Development and validation of the Eating Pathology Symptoms Inventory (EPSI). Psychological Assessment, 25, 859–878. https://doi.org/10.1037/a0032639 Garner, D. M. (2004). Eating Disorder Inventory–3 (EDI-3) professional manual. Odessa, FL: Psychological Assessment Resources. Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology, 41, 625–635. https://doi.org/10.1037/0012-1649.41.4.625 Ghasemi A., & Zahediasl, S. (2012). Normality tests for statistical analysis: A guide for non-statisticians. International Journal of Endocrinology and Metabolism, 10, 486–489. https://doi.org/10.5812/ijem.3505 Grabe, S., & Hyde, J. S. (2006). Ethnicity and body dissatisfaction among women in the United States: A meta-analysis. Psychological Bulletin, 132, 622–640. https://doi.org/10.1037/0033-2909.132.4.622 Jackson, T., & Chen, H. (2007). Identifying the eating disorder symptomatic in China: The role of sociocultural factors and culturally defined appearance concerns. Journal of Psychosomatic Research, 62, 241–249. https://doi.org/10.1016/j.jpsychores.2006.09.010 Jackson, T., Jiang, C., & Chen, H. (2016). Associations between Chinese/Asian versus Western mass media influences and body image disturbances of young Chinese women. Body Image, 17, 175–183. https://doi.org/10.1016/j.bodyim.2016.03.007 Javier, S. J., & Belgrave, F. Z. (2015). An examination of influences on body dissatisfaction among Asian American college females: Do family, media, or peers play a role? Journal of American College Health, 63, 579–583. https://doi.org/10.1080/07448481.2015.1031240 Kaneko, K., Kiriike, N., Ikenaga, K., Miyawaki, D., & Yamagami, S. (1999). Weight and shape concerns and dieting behaviours among pre‐adolescents and adolescents in Japan. Psychiatry and Clinical Neurosciences, 53, 365–371. https://doi.org/10.1046/j.1440-1819.1999.00559.x Ko, N., Tam, D. M., Viet, N. K., Scheib, P., Wirsching, M., & Zeeck, A. (2015). Disordered eating behaviors in university students in Hanoi, Vietnam. Journal of Eating Disorders, 3, 18. https://doi.org/10.1186/s40337-015-0054-2 Kypri, K., McCarthy, D. M., Coe, M. T., & Brown, S. A. (2004). Transition to independent living and substance involvement of treated and high risk youth. Journal of Child and Adolescent Substance Abuse, 13, 85–100. https://doi.org/10.1300/J029v13n03_05 Lake, A. J., Staiger, P. K., & Glowinski, H. (2000). Effect of Western culture on women’s attitudes to eating and perceptions of body shape. International Journal of Eating Disorders, 27, 83–89. https://doi.org/10.1002/(SICI)1098108X(200001)27:1<83::AID-EAT9>3.0.CO;2-J Lee, S., Leung, T., Lee, A. M., Yu, H., & Leung, C. M. (1996). Body dissatisfaction among Chinese undergraduates and its implications for eating disorders in Hong Kong. International Journal of Eating Disorders, 20, 77–84. https://doi.org/10.1002/(SICI)1098-108X(199607)20:1<77::AIDEAT9>3.0.CO;2-1 Luo, X., Donnellan, M. B., Burt, S. A., & Klump, K. L. (2016). The dimensional nature of eating pathology: Evidence from a direct comparison of categorical, dimensional, and hybrid models. Journal of Abnormal Psychology, 125, 715–726 https://doi.org/10.1037/abn0000174 Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253. https://doi.org/10.1037/0033-295X.98.2.224 Moreno, A., & Thelen, M. H. (1993). Parental factors related to bulimia nervosa. Addictive Behaviors, 18, 681–689. https://doi.org/10.1016/0306-4603(93)90021-Z Nevonen, L., & Norring, C. (2004). Socio-economic variables and eating disorders: A comparison between patients and normal controls. Eating and Weight Disorders, 9, 279–284. https://doi.org/10.1007/BF03325082

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

233

Eating Pathology in Vietnamese and Americans | Nguyen and Soysa

Parks, C. D., & Vu, A. D. (1994). Social dilemma behavior of individuals from highly individualist and collectivist cultures. Journal of Conflict Resolution, 38, 708–718. https://doi.org/10.1177/0022002794038004006 Pike, K. M., & Walsh, T. (1996). Ethnicity and eating disorders: Implications for incidence and treatment. Psychopharmacology Bulletin, 32, 265–274. Retrieved from https://psycnet.apa.org/record/1996-05703-010 Reniers, R. L. E. P., Beavan, A., Keogan, L., Furneaux, A., Mayhew, S., & Wood, S. J. (2017). Is it all in the reward? Peers influence risk‐taking behaviour in young adulthood. British Journal of Psychology, 108, 276–295. https://doi.org/10.1111/bjop.12195 Roh Ryu, H., Lyle, R. M., & Mccabe, G. P. (2003). Factors associated with weight concerns and unhealthy eating patterns among young Korean females. Eating Disorders: The Journal of Treatment and Prevention, 11, 129–141. https://doi.org/10.1080/10640260390199316 Rongmuang, D., Corte, C., McCreary, L. L., Park, C. G., Miller, A., & Gallo, A. (2011). Salience of physical appearance characteristics among young women in Thailand. Body Image, 8, 396–403. https://doi.org/10.1016/j.bodyim.2011.05.004 Rosenbloom, T. (2003). Risk evaluation and risky behavior of high and low sensation seekers. Social Behavior and Personality: An International Journal, 31, 375–386. https://doi.org/10.2224/sbp.2003.31.4.375 Shisslak, C. M., Crago, M., & Estes, L. S. (1995). The spectrum of eating disturbances. International Journal of Eating Disorders, 18, 209–219. https:// doi.org/ 10.1002/1098-108X(199511)18:3<209::AID-EAT2260180303>3.0.CO;2-E Smart, R., & Tsong, Y. (2014). Weight, body dissatisfaction, and disordered eating: Asian American women’s perspectives. Asian American Journal of Psychology, 5, 344–352. https://doi.org/10.1037/a0035599 Smart, R., Tsong, Y., Mejía, O. L., Hayashino, D., & Braaten, M. E. T. (2011). Therapists’ experiences treating Asian American women with eating disorders. Professional Psychology: Research and Practice, 42, 308–315. https://doi.org/10.1037/a0024179 Stark-Wroblewski, K., Yanico, B. J., & Lupe, S. (2005). Acculturation, internalization of western appearance norms, and eating pathology among Japanese and Chinese international student women. Psychology of Women Quarterly, 29, 38–46. https://doi.org/10.1111/j.1471-6402.2005.00166.x Sussman, N. M., Truong, N., & Lim, J. (2007). Who experiences ‘America the beautiful’?: Ethnicity moderating the effect of acculturation on body image and risks for eating disorders among immigrant women. International Journal of Intercultural Relations, 31, 29–49. https://doi.org/10.1016/j.ijintrel.2006.03.003

Swami, V., Tran, U. S., Stieger, S., & Voracek, M. (2015). Associations between women’s body image and happiness: Results of the YouBeauty.com Body Image Survey (YBIS). Journal of Happiness Studies: An Interdisciplinary Forum on Subjective Well-Being, 16, 705–718. https://doi.org/10.1007/s10902-014-9530-7 Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics. New York, NY: Harper & Row, Publishers, Inc. Tang, X., Forbush, K. T., & Lui, P. P. (2015). Development and validation of the Chinese‐language version of the Eating Pathology Symptoms Inventory. International Journal of Eating Disorders, 48, 1016–1023. https://doi.org/10.1002/eat.22423 Triandis, H. C., Bontempo, R., Villareal, M. J., Asai, M., & Lucca, N. (1988). Individualism and collectivism: Cross-cultural perspectives on self-ingroup relationships. Journal of Personality and Social Psychology, 54, 323–338. https://doi.org/10.1037/0022-3514.54.2.323 Tsai, M. C., Hoerr, R. L., & Song, R. O. (1998). Dieting behavior of Asian college women attending a US university. Journal of American College Health, 46, 163–168. https://doi.org/10.1080/07448489809595604 U.S. Census Bureau. (2018). America’s Families and Living Arrangements: 2018. Retrieved from https://census.gov/data/tables/2018/demo/families/cps-2018.html Wardle, J., Haase, A. M., & Steptoe, A. (2006). Body image and weight control in young adults: International comparisons in university students from 22 countries. International Journal of Obesity, 30, 644–651. https://doi.org/10.1038/sj.ijo.0803050 Author Note. Ngoc Nguyen, https://orcid.org/0000-0002-8265-4846, and Champika K. Soysa, Department of Psychology, Worcester State University, Worcester, MA. This study was partially supported by the Worcester State University Summer Undergraduate Research Grant. Special thanks to Psi Chi Journal reviewers for their support. Correspondence concerning this article should be addressed to Champika K. Soysa, Department of Psychology, Worcester State University, Worcester, MA. E-mail: csoysa@worcester.edu.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

234

https://doi.org/10.24839/2325-7342.JN24.4.235

An Experimental Analysis of Emotion Induction Prior to Reading a Health Narrative on Personal Risk Perception, Health Intentions, and Behavior Aisha L. Udochi, University of Central Arkansas Lindsay A. Kennedy* , Hendrix College Marc A Sestir* , University of Central Arkansas

ABSTRACT. When people find risk for diseases personally relevant, research shows they are more likely to engage in health-protective behaviors (Brewer, Chapman, Rothman, Leask, & Kempe, 2017). One way to make risk information more personally relevant is to present it in narrative form (Dunlop, Wakefield, & Kashima, 2008). Because positive emotions have been linked to broadened social categories (Waugh, & Fredrickson, 2006), improved processing of health information (Raghunathan & Trope, 2002), and better health outcomes (Kok et al., 2013), they may promote engagement in healthier behaviors, as well. Thus, we predicted inducing positive emotions in participants prior to reading a health narrative would lead to increased identification with the narrative figure and, subsequently, increased personal perceived risk for the diseases mentioned in the article and increased health behavioral intentions and behaviors. In the present study, 124 participants (29 men, 94 women, 1 did not report) watched a short video that induced either positive, negative, or neutral emotions, read a health narrative, then completed a packet of surveys measuring for risk perception for certain health diseases and intentions of engaging in healthy behaviors. Participants were then gifted a pedometer app, which they were asked to use for 2 weeks and share data from to assess their actual engagement in healthy behaviors. Results did not support our predictions, but findings suggested positive emotions may decrease perceived seriousness of certain diseases (obesity: p = .01, Îˇ2 = .07; diabetes: p = .04, Îˇ2 = .05). Future implications of findings are discussed. Keywords: mood-as-a-resource hypothesis, health narratives, identification, risk perception, emotions

ith health care costs continually rising (e.g., Claxton, Rae, Levitt, & Cox, 2018), it is important to understand the psychological factors that promote a healthy lifestyle. Risk perception, or the subjective evaluation of the likelihood and severity of potential risk (Ferrer & Klein, 2015), is a factor that is relevant to a wide variety of health decisions. Risk perception has been found to predict health-related intentions *Faculty mentor

and behaviors in areas such as vaccination (Brewer, Weinstein, Cuite, & Herrington, 2004; Brewer et al., 2007), exercise (e.g., Courneya & Hellsten, 2001), and AIDS prevention (van der Velde, Hooykaas, & van der Joop, 1992). Indeed, Sheeran, Harris, and Epton (2014) conducted a meta-analysis among 208 qualifying studies of the effect of experimentally increasing risk perception on health-related behavioral intentions and behavior, finding overall

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

235

Emotions, Narratives, and Risk Perception | Udochi, Kennedy, and Sestir

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

236

effect sizes of d = .31 and d = .23, respectively. Furthermore, risk perception is often a factor represented in health behavior theories due to its relationship with engagement in health behaviors (for a review see Brewer et al., 2007). For example, the health belief model posits that risk perception plays a major role in decisions about engaging in preventative behaviors (e.g., Rosenstock, 1974), and Witte’s extended parallel process model (1992) utilizes perceived risk as a key element of the effectiveness of fear appeals to influence health behavior. Risk perception appears both theoretically and empirically important in encouraging health behavior change; thus, understanding the factors that influence risk perception can be an important step toward increasing the likelihood of individuals engaging in preventative behaviors rather than seeking treatment after health problems arise. One factor that has been consistently found to undermine personal perceptions of risk is unrealis tic optimism (Kim & Niederdeppe, 2016). Unrealistic optimism is the tendency for people to believe they are more likely to experience positive events and less likely to experience negative events than their peers or an objective standard (Weinstein, 1980). This optimistic bias has been explained as being the result of both motivational and cognitive biases, such that individuals seek to maintain a positive image of themselves in order to reduce anxiety and engage in biased recall of relevant personal experiences to separate themselves from stereotypes associated with negative events (Kim & Niederdeppe, 2016; Weinstein, 1980). Regardless of the specific underlying mechanism, this effect is primarily characterized by an underestimation of personal risk. Like other biases in reasoning, researchers have looked for ways to overcome unrealistic optimism and increase perceptions of risk in individuals. One mode of information presentation that has been found to be particularly effective is the use of per sonal narratives (e.g., Kim & Niederdeppe, 2016). Personal narratives conveying health information have been found to be more effective at promoting a sense of personal risk than objective statistics conveying health information, thus resulting in a higher likelihood of engaging in health-protec tive behaviors (de Wit, Das, & Vet, 2008; Kim & Niederdeppe, 2016). Logically, objective statistics should be more persuasive than narratives because they better represent collective trends; however, the reverse is frequently true (e.g., de Wit et al., 2008; Green & Brock, 2000; Yoo, Kreuter, Lai, & Fu,

2014). One key reason for this discrepancy is simply that narratives are, by nature, stories, and stories have the ability to transport the reader into another world, evoke powerful emotions and motivations, and even change the attitudes of the reader (Green & Brock, 2000). Furthermore, narratives allow for readers to identify with a character within the mes sage, increasing the likelihood that the reader will respond to the message as if the events were actually happening to them (Cohen, 2001). Following from these lines of research, Dunlop, Wakefield, and Kashima (2008) proposed a model to address the influence of emotions while reading health narratives on perceptions of personal risk and subsequent health-protective behaviors. To summarize the model, these authors propose that exposure to a message can elicit three different classes of emotional responses: message-referential, plot-referential, or self-referential emotions. Message-referential emotions arise in response to the features of the message itself, such as an image contained within or the person presenting the message. Plot-referential emotions arise in response to a character or situation described within the message, proposed to be a result of transporta tion experienced by the reader. Finally, and most important to their model, the authors describe self-referential emotions as responses to a message that prompt thinking about one’s own life and self. This class of emotional responses is the only one hypothesized to have a direct effect on perceived personal risk (Dunlop et al., 2008). The other two classes indirectly influence perceived personal risk to the extent that they lead to self-referential emotions through character identification (plotreferential) or self-referencing through the message itself (Dunlop et al., 2008). Although not explicitly depicted in the model, the authors restricted their discussion to negative emotions only. They reasoned that, because public health messages are designed to reduce risky behavior, associating positive feelings with such actions, as marketing campaigns do with their products, would establish the wrong perception of risky behavior. Focusing on discrete negative emotions, then—which can highlight the conse quences of risky behavior—is most appropriate for the health domain (Dunlop et al., 2008). Though much research exists regarding the influence of negatively-valenced public health narratives (e.g., de Wit et al., 2008; Kim & Niederdeppe, 2016; Yoo et al., 2014), the impact of positive emotions remains largely unexplored.

Udochi, Kennedy, and Sestir | Emotions, Narratives, and Risk Perception

Considering that positive emotions are associ ated with improved coping ability (Fredrickson & Joiner, 2018), longer and healthier lives (Chida & Steptoe, 2008), reduced chance of contracting certain diseases (e.g., Boehm & Kubzansky, 2012), and better physical and mental health (Mahony, 2000), their contribution to health behavior is an important topic of inquiry. Fredrickson’s (1998; 2001) broaden-and-build theory gives foundational support for the health benefits of positive emotions. According to this theory, positive emotions serve to broaden attention and thought-action repertoires, unlike negative emotions which tend to narrow an individual’s focus. Through this broadened atten tion, positive emotions then help build personal resources, such as interpersonal relationships and even physical skills (Fredrickson, 1998; 2001). Focusing on the broadening property of posi tive emotions, research shows that individuals who viewed a video clip eliciting joy were better able to distinguish among other-race group faces, thus demonstrating a significantly lower level of the ownrace bias than participants who viewed a video clip eliciting negative or neutral emotional responses (Johnson & Fredrickson, 2005). Similarly, Waugh and Fredrickson (2006) found that the experi ence of positive emotions during the first week of college predicted greater feelings of self-other overlap between new roommates. Taken together, these results suggest that one broadening effect of positive emotions is the expansion of one’s social categories. This ability of positive emotions to allow for more inclusive categorization and self-other overlap may make identification with a character within a message more likely to occur, as perceived similarity with a narrative character is considered an important feature of identification (Slater & Rouner, 2002). This increase in identification then, as hypothesized by Dunlop et al. (2008), could impact people’s personal perceived risk. In addition, positive moods and outlooks (i.e., optimism) have been found to have beneficial effects on information processing, even when the information is negative (Das, Vonkeman, & Hartmann, 2012). According to the mood-asa-resource hypothesis, a positive mood increases the likelihood that one will act in the service of long-term, rather than short-term, goals by provid ing the individual with a psychological buffer for coping with self-relevant negative information (Raghunathan & Trope, 2002). In an experimental test of this hypothesis, those induced into a posi tive mood demonstrated greater recall of negative

information, as well as greater behavioral intentions of changing in response to the negative information (Raghunathan & Trope, 2002). Similarly, optimism related to one’s health has been found to increase attention paid to and recall of risk information and decrease defensive responses to such information (Aspinwall & Brunhart, 1996). Taken together, research on positive emotions, positive moods, and optimism suggest the impor tant role that positive emotions may play in health communication—in particular, risk communica tion. Given the relative lack of research regarding the role of positive emotions on the interpretation of health-related narratives, the main purpose of this study was to compare the effects of inducing a negative, positive, or neutral emotional state before reading a health narrative on perceptions of personal risk, behavioral intentions, and selfreported behaviors. We hypothesized that exposure to a positive emotion induction, relative to a negative emotion or neutral emotion induction, prior to reading a health narrative would result in higher levels of perceived personal risk, and a higher likelihood of protective behavioral intentions and healthrelated behaviors (Hypothesis 1). Furthermore, we hypothesized that these beneficial effects of the positive emotion induction would be mediated by either one or both of the following factors: 1) greater levels of identification with the narrative character (following from Johnson & Fredrickson, 2005 and Dunlop et al., 2008); and/or 2) greater recall of information from the narrative (following from Raghunathan & Trope, 2002; Das et al., 2012) (Hypothesis 2).

Method Participants A total of 124 undergraduate students (29 men, 94 women, 1 did not report) were recruited through emails and campus flyers from the general student body of a small liberal arts college in the Southern United States. Because the study utilized a pedometer app from the Apple App Store, participants were required to own an iPhone or iPod Touch; students without such devices did not qualify to participate in the study. If participants were enrolled in a psychology course, they could receive extra credit for their participation in each part of the study at the discretion of their instruc tor. All participants who completed the first part of the study received health brochures and a free download of a pedometer app; all participants who

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

237

Emotions, Narratives, and Risk Perception | Udochi, Kennedy, and Sestir

completed the follow-up portion of the study two weeks later were entered into a drawing for a $25 gift card to a local restaurant. The experimental session took approximately 30 minutes to complete with a 10-minute follow-up survey.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

238

Materials Emotion-induction videos. Three different short videos were used to induce positive, negative, and neutral emotions for the study. A clip from the movie Rudy (Fried, Woods, & Anspaugh, 1993), which portrayed a character who finally achieved his goal of playing football at his dream school while a crowd cheers for him, was used to induce positive emotions. A clip from the movie Vertical Limit (Campbell, King, Nasatir, & Phillips, 2000), was used to induce negative emotions. The clip depicted a scene in which a group of rock climbers have an emergency that results in them hanging off a cliff, forcing them to decide whether to cut the rope on one of the other characters in order to sur vive. An instructional tiling video (studyinghealth, 2013), used to neutrally influence emotions, served as the control and simply provided instructions on how to tile a bathroom or countertop. All videos were similar in length (approximately 5 minutes), contained interpersonal interactions, and were pilot-tested to ensure that they induced the emo tions for which they were intended. Health narratives. A personal health narra tive, titled “Student Voices,” was written for the purposes of this study and presented as a student newspaper publication from another institution. The article was written from the perspective of a college student, Jordan, describing how their diagnosis of prehypertension altered their per spective of engaging in healthier behaviors. The article began with the author reminiscing on their original (lax) view of exercising, then transitioned to them explaining the negative implications of having prehypertension, such as susceptibility to heart problems, type II diabetes, and obesity. The article ended with the author discussing the benefits they received from living a healthier lifestyle and challenging the reader to assume better health habits, as well. The narrative contained a balance of positively- and negatively-valenced information. The gendered pronouns used to introduce the author were matched to the gender of the participant. State emotional experiences. The Modified Differential Emotions Scale (mDES; Fredrickson, Tugade, Waugh, & Larkin, 2003) was used to assess participants’ moods. This 22-item questionnaire

measured the current mood of the participants by requiring them to rate the extent to which different emotions (e.g., “amused, fun-loving, silly;” “sad, downhearted, unhappy”) described their feelings at multiple points in the study. Eleven items each measured for positive emotions (α = .92) and negative emotions (α = .88). Two additional items asked about feelings of boredom and tiredness to assess additional impacts the three videos may have had on feeling states. Items were rated on a 5-point Likert-type scale, ranging from 0 (not at all) to 4 (extremely). Manipulation check. In addition to measuring reactions to the health narrative, the mDES was also used as a manipulation check for the emotioninducing videos. Transportation and identification. Our mea sure for transportation and identification was the “Narrative Questionnaire” created by Green & Brock (2000). Thirteen questions (α = .77) mea sured the participants’ transportation into the health narrative, asking them to rate items such as “I was mentally involved in the article while reading it” on a 7-point Likert-type scale ranging from 1 (not at all) to 7 (very much). Three items (α = .69) measured for identification with the character in the narrative, asking participants to rate statements such as “When good things happened to Jordan, I felt happy, but when negative things happened to Jordan, I felt sad” on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Information recall. Six items were used to measure participants’ ability to recall content from the article. These questions were asked in an open-response format and included questions such as “What are two causes of high blood pressure, obesity, and type II diabetes listed in the article?” Answers were scored as 0 (incorrect) or 1 (correct). These questions were included to assess the extent to which the emotion induction videos influenced participants’ ability to accurately recall information from the health narrative as a means of testing the mood-as-a-resource hypothesis (Raghunathan & Trope, 2002). Risk perception. Participants rated their perceived risk for developing heart disease, type II diabetes, and obesity—the three diseases discussed within the narrative. For each disease, participants responded to five questions that were rated on a 7-point Likert-type scale, ranging from 1 (not at all) to 7 (very much). These items were written to tap into the three dimensions of perceived risk: perceived likelihood (two questions; e.g., “How

Udochi, Kennedy, and Sestir | Emotions, Narratives, and Risk Perception

likely do you think it is that you will be diagnosed with heart disease sometime in the future?”), perceived susceptibility (two questions; e.g., “How susceptible or vulnerable do you think you are to developing heart disease sometime in the near future?”), and perceived severity or seriousness (“How serious a disease do you think heart disease is?”; Brewer et al., 2007). Cronbach’s αs at Contact 1 for the individual heart disease, type II diabetes, and obesity scales were .75, .79, and .80, respectively, with a composite reliability of .86. Cronbach’s αs at Contact 2 for the individual heart disease, type II diabetes, and obesity scales were .74, .83, and .83, respectively, with a composite reliability of .90. Behavioral intentions and self-reported behaviors. Participants rated their intentions to engage in seven different health-protective behaviors (behav iors related to the three health conditions) over the next two weeks on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). An N/A option was also provided. For example, participants rated their agreement with the state ments, “I intend to avoid smoking cigarettes” and “I intend to engage in 30 minutes of moderate exercise 5 times per week.” All behavioral intention questions were written for the purposes of this study. Following from seminal work by Davidson and Jaccard (1979) on the measurement of attitudes about performing behaviors, these questions were written to reference intention to perform a specific behavior (e.g., “I intend to engage in 30 minutes of moderate exercise 5 times per week” versus “I intend to exercise”) within a specific time frame (i.e., two weeks). In the follow-up questionnaire, participants self-reported on their actual behaviors for the past two weeks by answering the same questions (e.g., “I avoided smoking cigarettes.”). Because these measures were intended to cover a broad range of protective behaviors across three distinct health conditions, items were analyzed individually; thus, alphas were not calculated. Health-related behaviors. Participants’ healthrelated behaviors were measured in two ways: (a) health information-seeking behavior was measured by the number of health brochures (out of six) participants took with them at the conclusion of the first part of the study; and (b) physical activity was measured by participants’ use of a pedometer app over the course of two weeks, which recorded the length of time participants used it, the number of days participants used the app, how many steps they took, and their total distance traveled, in miles, in those two weeks.

Procedure Contact 1. Experimentation began following approval by the Hendrix College Human Subjects Review Board (Protocol #F1403). Each session was conducted in a private cubicle. After consenting, participants were randomly assigned to watch one of the three emotion-inducing videos. After the video, participants completed the mDES, which was used as a manipulation check. Participants then read the health narrative and completed the mDES for a second time, followed by the measures of transportation, identification, information recall, risk perception, and behavioral intentions (in that order). The final questions inquired about the gender of the participant, their regular exercising habits, whether they had been diagnosed with heart disease, type II diabetes, or obesity, and how they heard about the study. Upon completion, participants received a partial debriefing statement and instructions for downloading a pedometer app offered as incentive for their participation. Participants were asked to use the pedometer app for the next two weeks, then were allowed to leave the study and take health brochures with them. The number of health brochures each participant took (from 0 to 6) was recorded by the researcher. Contact 2. Two weeks after their participation in the first half of the study, participants were contacted via email and asked to complete a follow-up survey on SurveyMonkey. Participants again completed the measure of risk perception for each health condition before self-reporting on the extent to which they actually engaged in each of the seven health behaviors over the past two weeks. In addition, participants were asked to report how many days they used their pedometer app and exercised, and estimate the number of steps they took and the number of minutes they exercised in total over the past two weeks. Participants were also asked to directly share their pedometer data with the researchers to corroborate these self-report responses. Finally, participants received a complete debriefing on the purpose of the study and were entered into a drawing for a $25 gift card to a local restaurant. Data Analysis Plan To test Hypothesis 1, that exposure to a positive emotion induction prior to reading a health narrative would result in higher perceived risk, greater behavioral intentions, and more engage ment in health-related behaviors than exposure to a negative or neutral emotion induction, we

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

239

Emotions, Narratives, and Risk Perception | Udochi, Kennedy, and Sestir

planned to conduct a series of one-way Analyses of Variance (ANOVAs). Because the mechanisms we proposed to test were not intended to be diseasespecific, focal tests of Hypothesis 1 were planned on composite measures of perceived risk that averaged across heart disease, type II diabetes, and obesity. Hypothesis 2 proposed two potential mediators of any significant effects found in the analysis of Hypothesis 1: identification with the narrative char acter and accuracy of information recalled from the narrative. Thus, we planned to conduct and report on two tests of mediation—one for each proposed mediator—for any significant effects the emotion induction had on our focal dependent measures. Any deviations from this a priori analysis plan are reported in the Results section and additional exploratory analyses are labeled as such.

Results Descriptive Statistics and Sample Characteristics On average, participants reported exercising around 3.69 days a week for 58.71 minutes per exercise session. Across all emotion induction con ditions, no participants were previously diagnosed with diabetes, three participants were diagnosed with heart disease (2.44% of participants), and four participants were diagnosed with obesity (3.25% of participants) prior to their participation in the study. There were no significant differences found between emotion induction conditions on preexisting exercise habits or diagnoses for heart disease, diabetes, or obesity, and no significant gender differences were observed on any primary dependent variable.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

240

Manipulation Checks A series of one-way ANOVAs were conducted to assess differences in state positive and negative emotions, boredom, and tiredness following the video. A statistically significant difference between groups emerged on reported positive emotions, F(2, 121) = 55.33, p < .001, η2 = .48. A Tukey posthoc test found that participants in the positive induction condition (M = 2.38, SD = 0.93) scored significantly higher than both the neutral (M = 0.85, SD = 0.68) and negative (M = 0.97, SD = 0.54) induc tion conditions (both ps < .001). The negative and neutral induction conditions did not significantly differ from one another on reported state positive emotion (p = .75). Likewise, a statistically significant difference emerged between groups on reported state nega tive emotion, F(2, 121) = 85.33, p < .001, η2 = .59; a

Tukey post-hoc test reported that participants in the negative induction condition (M = 1.29, SD = 0.64) scored significantly higher than both the neutral (M = 0.21, SD = 0.29) and positive (M = 0.15, SD = 0.31) induction conditions (both ps < .001). The positive and neutral induction conditions did not significantly differ from one another on reported state negative emotions (p = .82). Furthermore, a statistically significant dif ference emerged between groups on reported boredom, F(2, 121) = 39.51, η2 = .40, and tiredness, F(2, 121) = 15.1, η2 = .20, after viewing the emotion inducting videos (both ps < .001). Post-hoc Tukey tests revealed that participants who viewed the neutral video were significantly more bored (M = 2.1, SD = 1.34) and tired (M = 1.96, SD = 1.46) than participants who viewed the positive (Bored: M = 0.73, SD = 1.12; Tired: M = 0.73, SD = 1.05) and negative (Bored: M = 0.12, SD = 0.40; Tired: M = 0.68, SD = 1.06) induction videos (all ps < .001). Induction Durability Check To examine the durability of the emotion induc tion, repeated measures ANOVAs were conducted using emotion induction condition as the betweenparticipants variable and timing of emotional measurement (immediately after emotion induc tion versus after reading the health narrative) as the within-participants variable. Regarding positive emotions, a statistically significant interaction emerged, such that the differences in positive emo tions found between emotion induction conditions immediately after the emotion induction dissipated after reading the health narrative, F(2, 120) = 32.52, p < .001, ηp2 = .35. Significant main effects also emerged for emotion induction condition, F(2, 120) = 25.30, p < .001, ηp2 = .30, and time, F(1, 120) = 160.00, p < .001, ηp2 = .57. The same pattern of effects emerged for negative emotions, such that the interaction, F(2, 113) = 57.15, ηp2 = .50, main effect of emotion induction condition, F(2, 113) = 33.60, ηp2 = .37, and main effect of time, F(1, 113) = 354.65, ηp2 = .76, were all statistically significant (all ps < .001). The same was also true for boredom (interaction: F(2, 120) = 33.64, ηp2 = .36; main effect of emotion induction condition: F(2, 120) = 20.66, ηp2 = .26; main effect of time: F(1, 120) = 18.30, ηp2 = .13) and tiredness (interaction: F(2, 121) = 8.76, ηp2 = .13; main effect of emotion induction condition: F(2, 121) = 7.95, ηp2 = .12; main effect of time: F(1, 121) = 84.15, ηp2 = .41), with all ps < .001. Follow-up tests (one-way ANOVAs) confirmed that, when emotional states were retested after reading the

Udochi, Kennedy, and Sestir | Emotions, Narratives, and Risk Perception

health narrative, none of the emotion induction conditions differed significantly on any emotion, all Fs < 1.58, all ps > .21, all η2s < .03. Descriptive statistics for all emotion induction conditions at each time point can be found in Table 1. Hypothesis 1 Positive emotion induction, relative to negative or neutral emotion induction, will result in higher levels of perceived personal risk for diseases and subsequent protective behavioral intentions and behaviors. Risk perception. A one-way ANOVA was con ducted to test the effects of the emotion induction on participants’ perceptions of personal risk at Contact 1. Contrary to Hypothesis 1, emotion induction conditions did not differ significantly on perceived risk at Contact 1 when measures for each disease were analyzed collectively, F(2, 117) = 1.12, p = .33, η2 = .02. For exploratory purposes, we also analyzed the effects of the emotion induction on risk percep tion for each disease separately and no significant effects emerged, all Fs < 1.55, all ps > .22, all η2s < .03 (descriptive statistics are provided in Table 2). When examining perceived risk items from Contact 1 individually for exploratory purposes, a significant effect of emotion induction condition on perceived seriousness of diseases emerged, with the positive induction condition perceiving obesity significantly less seriously (M = 5.71, SD = 1.49) than both the negative (M = 6.37, SD = 0.92) and neutral (M = 6.42, SD = 1.05) induction conditions, F(2, 120) = 4.63, p = .01, η2 = .07, and diabetes significantly less seriously (M = 5.85, SD = 1.15) than the neutral induction condition (M = 6.39, SD = 0.89), F(2, 120) = 3.43, p = .04, η2 = .05. No significant differences emerged on any remaining items on perceived risk, all Fs < 1.41, all ps > .25, all η2s < .03. Furthermore, repeated measures ANOVAs were conducted to test the effect of the emo tion induction condition on perceived risk over time, with emotion induction condition as the between-participants variable and timing of the measure (Contact 1 versus Contact 2) as the within-participants variable. These analyses were also exploratory in nature. When risk perception for the three diseases was analyzed collectively, no significant interaction or main effects emerged; however, the main effect of emotion induction condition approached significance, F(2, 51) = 2.67, p = .08, η2 = .10. Tukey tests revealed that overall risk perception across both time points was marginally

lower in the positive induction condition (M = 3.61, SD = 0.97) than the neutral induction condition (M = 4.32, SD = 0.90), p = .07. Neither the positive induction condition nor the neutral induction condition significantly differed from the negative induction condition (M = 4.03, SD = 0.96; both TABLE 1 Descriptive Statistics for Manipulation Checks in Contact 1 and Contact 2

Control Condition

Positive Condition

Negative Condition

Positive

0.85

0.68

2.37

0.93

0.96

0.54

Negative

0.21

0.29

0.15

0.31

1.29

0.64

Bored

2.10

1.34

0.73

1.12

0.13

0.40

Tired

1.96

1.46

0.73

1.05

0.68

1.06

Positive

2.19

0.76

2.47

0.78

2.23

0.73

Negative

1.34

0.34

1.45

0.44

1.48

0.53

Bored

1.41

0.63

1.39

0.70

1.50

0.85

Tired

2.29

1.31

1.95

1.28

1.83

1.12

Emotion Scores After Video

Emotion Scores After Narrative

TABLE 2 Descriptive Statistics for Focal Dependent Measures at Contact 1 and Contact 2

Control Condition

Positive Condition

Negative Condition

Composite

4.28

0.91

3.67

1.06

4.02

0.92

Heart Disease

3.92

1.37

3.91

1.26

3.38

1.05

Diabetes

3.53

1.35

2.89

1.47

3.30

1.31

Obesity

3.53

1.63

2.75

1.32

3.49

1.52

Composite

4.35

0.88

3.54

0.88

4.04

0.99

Heart Disease

4.69

0.96

3.99

0.87

4.04

0.88

Diabetes

4.13

1.04

3.40

0.88

4.00

1.18

Obesity

4.21

1.42

3.24

1.19

4.10

1.19

Perceived Risk Contact 1

Perceived Risk Contact 2

Pedometer Data Days Used

15.40

8.47

Time Used

7:06:52

3:25:26

11:04:18

10.40

4:34:03

2.61

12:49:36

10.83

8:06:57

2.40

Total Steps

41427

23115

68874

27353

75469

49938

Total Distance

17.27

7.58

31.98

13.20

34.13

25.40

Note. Composite Perceived Risk represents the average of all risk items across all diseases. Total Distance was calculated in miles.

241

Emotions, Narratives, and Risk Perception | Udochi, Kennedy, and Sestir

ps > .31). When risk perception for each disease was analyzed separately, there were no significant emotion induction condition x time interactions, all Fs < 2.11, all ps > .13, all η2s < .07, or significant main effects of emotion induction condition, all Fs < 2.26, all ps > .11, all η2s < .08. However, for all three diseases, a significant main effect of time emerged, all Fs > 14.62, all ps < .001, all η2s > .21, such that risk perception increased over time, regardless of emotion induction condition. Descriptive statistics for these effects are presented in Table 2. Behavioral intentions and self-reported behaviors. One-way ANOVAs were conducted to test the effects of the emotion induction videos on participants’ intentions at Contact 1 to engage in a variety of health-protective behaviors over the next two weeks and their self-reported engagement in these same behaviors at Contact 2. Contrary to Hypothesis 1, emotion induction condition did not have significant effects on intentions to engage in any of the provided health behaviors in the follow ing two weeks, all Fs < 0.84, all ps > .40, all η2s < .02. Emotion induction condition also did not have any significant effects on participants’ self-reported engagement in health behaviors, including selfreported pedometer steps, all Fs < 2.63, all ps > .08, all η2s < .09. Health-related behaviors. One-way ANOVAs were conducted to test the effects of the emotion induction videos on participants’ taking of health brochures at Contact 1 (health information-seeking behavior) and their actual pedometer use in the two weeks between Contact 1 and Contact 2. In regard to the health information-seeking behavior of bro chure taking, the positive induction condition (M = 0.54, SD = 1.47) did not significantly differ from the negative (M = 0.23, SD = 1.00) and neutral (M = 0.38, SD = 0.79) induction conditions, F(2, 120) = 0.78, p = .46, η2 = .01. Due to a low rate of providing direct pedometer access in Contact 2, inferential tests were not conducted on these data. Descriptive statistics can be found in Table 2.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

242

Hypothesis 2 The beneficial effects of positive emotion induction would be mediated by either one or both of the following factors: (a) greater levels of identification with the narrative character; and/or (b) greater recall of information from the narrative. Because emotion induction condition did not affect the main outcome variables, we did not conduct tests of mediation. However, exploratory one-way ANOVAs were conducted to test for

differences between emotion induction conditions on identification with the narrative character and accuracy of narrative information recall. Regarding identification with the narrative character, we discovered that separating one of the identification items (i.e., “When you read this article, how often did you feel or react as if the experiences of Jordan were happening to you?”) raised the remaining two items’ Cronbach’s α from .69 to .81. The average of these two items was analyzed separately from the third identification item. One-way ANOVAs determined there were no statistically significant differences between groups on levels of transportation or identification (1- or 2-item measure), all Fs < 0.65, all ps > .52, all η2s < .02. Regarding recall of information from the health narrative, a significant difference in accuracy was found when participants were asked to recall two causes for the diseases included in the narrative, F(2, 121) = 4.00, p = .02, η2 = .06, and a marginally significant difference was found between groups when asked to recall two negative thoughts the narrative character expressed about exercising, F(2, 121) = 2.67, p = .07, η2 = .04. In both cases, the neutral induction condition demonstrated better recall of information (M = 1.00, SD = 0.22; M = 0.86, SD = 0.57, respectively) than both the positive (M = .83, SD = 0.38; M = 0.56, SD = 0.63, respectively) and negative (M = .80, SD = 0.40; M = 0.78, SD = 0.61, respectively) induction conditions. There were no additional statistically significant differences between groups for accurately recalling information from the health narrative, all Fs < 1.91, all ps > .15, all η2s < .04.

Discussion Prior research on risk perception has shown that personal perceived risk for diseases is linked to better health habits (Brewer et al., 2004; van der Velde et al., 1992), and that increasing risk percep tion can increase both health-related behavioral intentions and behaviors (Sheeran et al., 2014). In the present study, we attempted to increase risk perception and, thus, health behaviors, indirectly by inducing an emotional state—positive, negative, or neutral—before exposing our participants to a health-related narrative. To maximize the personal relevance of the health information provided and avoid the defensive information processing that often results from purely statistical information dissemination (de Wit et al., 2008), we utilized narratives to provide participants with both a

Udochi, Kennedy, and Sestir | Emotions, Narratives, and Risk Perception

story to be transported into and a character with whom to identify (Green & Brock, 2000). Further, by manipulating emotional state prior to health information exposure, we attempted to contribute to preexisting literature on the role mood plays in information processing and health behavior intentions. In Contact 2, we hoped to assess the long-term effects of Contact 1’s emotion induction, health information processing, and behavioral intentions by measuring actual engagement in healthy behaviors two weeks after Contact 1. In this study, we did not find a significant relationship between emotion induction condition and risk perception, health intentions, or health behaviors. In short, our hypotheses were largely not supported. Instead, our findings suggested positive emotions may, indeed, decrease people’s perceived seriousness of diseases, such as diabetes and obesity. Thus, our data partially reinforced the relevance of negative emotions in health communication. Our findings also did not support the moodas-a-resource hypothesis (Raghunathan & Trope, 2002), which states that, when confronted with negative health information, positive emotions provide a buffer that allows people to focus more on processing the content of the information, relative to negative or neutral emotions. In contrast to the mood-as-a-resource hypothesis, the only significant or marginally significant differences related to information processing were products of the neutral induction condition exemplifying the best memory recall, suggesting that positive emotions might not lead to improved narrative content interpretation as suggested in the model. Furthermore, our results were not consistent with the broadening effect of positive emotions on social categories (Johnson & Fredrickson, 2005), such that participants in the positive induction condition did not show higher identification with the narra tive character than participants in the negative or neutral induction conditions. These results should be interpreted with several limitations in mind. Importantly, we did not acquire baseline measurements of the participants’ emo tional states prior to administering any procedure, constricting our ability to assess the impact that the video and narrative had on their moods and the efficacy of our random assignment procedures. The lack of baseline measurements for risk perception impeded exploration of possible pre-existing group differences in those measures as well. In addition, our data suggested that participants’ induced state emotions did not persist after reading the health

narrative (see Table 1), leaving little possibility for the emotion induction to impact participants’ perceptions of risk for diseases and health inten tions and behaviors. Thus, future research should measure participants’ emotional states before any manipulations, use a longer-lasting or stronger manipulation, and further explore the relationship between emotions and health behaviors proposed in the mood-as-a-resource hypothesis. Measurement limitations also could have impacted our attempts to explore the distal effects emotions have on behavioral intentions and actual engagement in healthy behaviors. Although each behavior was assessed individually, all were intended to represent behaviors that could protect against all diseases presented in the narrative; however, the items did not intercorrelate well (intentions α = .56, behaviors α = .34), indicating inconsistency within participants’ responses. Upon reviewing the questions asked on behavioral intentions, it is pos sible participants did not associate all of the health behaviors provided with diabetes, obesity, or heart disease because we did not explicitly define the relationship between the behaviors and diseases for participants. This lack of context may have led them to answer the questions in ways inconsistent with the health information we provided, contributing to the low reliability of the measures. Future tests of these relationships should explicitly establish the connections between health problems and their associated preventative health behaviors before presenting participants with a questionnaire on behavioral intentions. Reliability was also low for the 3-item measure of identification with the narrative character (α = .69), suggesting a more general need for stronger measures in follow-up research. Our study also utilized a small convenience sample of college students in the southern United States, which could have been additionally restricted in representativeness by the requirement for par ticipants to own an iPhone and iPod Touch. While smartphone ownership is nearly universal among college-aged individuals (99%) and common among the general population (81%; Pew Research Center, 2019), users of Apple products could have differed substantively from those that used other products or none at all. Our study also did not involve the collection of the age nor race of the participants, which is another limitation, as these factors could have influenced our findings. Further, our sample was comprised of an uneven distribution of gender. Thus, our findings may not be gener alizable to broader populations. Future research

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

243

Emotions, Narratives, and Risk Perception | Udochi, Kennedy, and Sestir

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

244

should utilize materials that work across all major manufacturer devices and types and, if possible, include a broader range of ages and backgrounds to best ensure generalizable findings. Additional methodological adjustments for future research include counterbalancing measures to eliminate any potential order effects, reducing reliance on self-report measures to increase accuracy of data, utilizing a larger sample with more generalizable characteristics, and testing complex hypotheses across multiple smaller studies. The pedometer app we gifted to the partici pants also provided its own limitations because the data we received from the participants was highly variable and incomplete. Contact 2 requested participants email the researchers their pedom eter data, but only 15 participants (12.1% of the participant population) opted to do so, thus restricting our ability to use these data. Providing our participants with a specific incentive to share their pedometer data with us at the end of the study may have improved response rates. Future research on participants’ actual engagement in health behaviors should use more reliable measurement and consider research designs that may improve follow-up response rates. In addition to methodological improvements, future research should consider the effects of emotional inductions on other factors included in theories of behavior change. For example, Ajzen’s theory of planned behavior (1985) has been linked to the effectiveness of health programming at improving health behaviors and outcomes (Wright, Broadbent, Graves, & Gibson, 2016). Assessing factors such as perceived behavioral control or subjective normative environment, in conjunction with emotional inductions, could provide a clearer picture of the relationship between state emotions, health narratives, and subsequent health behaviors. This would allow for the possibility that different predictors of behaviors may benefit from different emotional states (e.g., risk perception may be enhanced by negative moods whereas perceived behavioral control may be enhanced by positive moods). In a similar vein, future research should explore the potential importance of alignment between the valence of emotion induced (i.e., positive or negative) and the valence of the targeted behavior change (e.g., promoting positive versus preventing negative health behavior). Conclusion In summary, our results did not match our initial

predictions and, instead, suggest positive emo tions may reduce personal risk perception and motivation to engage in health behaviors. However, more research—with stronger and longer-lasting emotional inductions—should be conducted to better understand the role positive emotions may play in health communication. Our findings suggest that health communicators hoping to increase the perceived seriousness of their health information might consider the role of emotional tone in their messages, specifically avoiding eliciting positive emotions, as our findings suggest positive mood may reduce perceived seriousness. Health com municators should also consider the longevity of mood induction from their messages, as our study shows that short-lived emotional inductions do not significantly affect health behaviors. Future studies should explore the effects of longer-term emotional differences on health habits and continue to assess optimal emotion induction strategies for increasing perceived seriousness of diseases, inspiring healthy behavioral intentions and actions, and producing better health outcomes.

References Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In Kuhl J., Beckmann, J. (Eds.), Action Control (pp. 11–39). SSSP Springer Series in Social Psychology. Springer, Berlin, Heidelberg. Aspinwall, L. G., & Brunhart, S. M. (1996). Distinguishing optimism from denial: Optimistic beliefs predict attention to health threats. Personality and Social Psychology Bulletin, 22, 993–1003. https://doi.org/10.1177/01461672962210002 Boehm, J. K., & Kubzansky, L. D. (2012). The heart’s content: The association between positive psychological well-being and cardiovascular health. Psychological Bulletin, 138, 655–691. http://dx.doi.org/10.1037/a0027448 Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M., McCaul, K. D., & Weinstein, N. D. (2007). Meta-analysis of the relationship between risk perception and health behavior: The example of vaccination. Health Psychology, 26, 136–145. https://doi.org/10.1037/0278-6133.26.2.136 Brewer, N. T., Chapman, G. B., Rothman, A. J., Leask, J., & Kempe, A. (2017). Increasing vaccination: Putting psychological science into action. Psychological Science in the Public Interest, 18, 149–207. https://doi.org/10.1177/1529100618760521 Brewer, N. T., Weinstein, N. D., Cuite, C. L., & Herrington, J. E. (2004). Risk perceptions and their relation to risk behavior. Annals of Behavioral Medicine, 27(2), 125–130. https://doi.org/10.1207/s15324796abm2702_7 Campbell, M. (Producer/Director), King, R., Nasatir, M., & Phillips, L. (Producers). (2000). Vertical Limit [DVD]. United States: Columbia Pictures. Chida, Y., & Steptoe, A. (2008). Positive psychological well-being and mortality: A quantitative review of prospective observational studies. Psychosomatic Medicine, 70, 741–756. http://dx.doi.org/10.1097/PSY.0b013e31818105ba Claxton, G., Rae, M., Levitt, L., & Cox, C. (2018). How have healthcare prices grown in the U.S. over time? Retrieved from https://www.healthsystemtracker. org/chart-collection/how-have-healthcare-prices-grown-in-the-u-s-overtime/#item-standardized-payment-rates-for-inpatient-hospital-stays-arehigher-and-have-increased-more-rapidly-for-the-privately-insured-thanfor-medicare-or-medicaid Cohen, J. (2001). Defining identification: A theoretical look at the identification of audiences with media characters. Mass Communication and Society, 4, 245–264. https://doi.org/10.1207/s15327825mcs0403_01 Courneya, K. S., & Hellsten, L.-A. M. (2001). Cancer prevention as a source of exercise motivation: an experimental test using protection motivation theory. Psychology, Health and Medicine, 6, 59–64.

Udochi, Kennedy, and Sestir | Emotions, Narratives, and Risk Perception

http://dx.doi.org/10.1080/13548500125267 Das, E., Vonkeman, C., & Hartmann, T. (2012). Mood as a resource in dealing with health recommendations: How mood affects information processing and acceptance of quit-smoking messages. Psychology and Health, 27, 116–127. https://doi.org/10.1080/08870446.2011.569888 Davidson, A. R., & Jaccard, J. J. (1979). Variables that moderate the attitude– behavior relation: Results of a longitudinal survey. Journal of Personality and Social Psychology, 37, 1364–1376. http://dx.doi.org/10.1037/0022-3514.37.8.1364 de Wit, J. B. F., Das, E., & Vet, R. (2008). What works best: Objective statistics or personal testimonial? Assessment of the persuasive effects of different types of message evidence on risk perception. Health Psychology, 27, 110–115. https://doi.org/10.1037/0278-6133.27.1.110 Dunlop, S., Wakefield, M., & Kashima, Y. (2008). Can you feel it? Negative emotion, risk, and narrative in health communication. Media Psychology, 11, 52–75. https://doi.org/10.1080/15213260701853112 Ferrer, R., & Klein, W. M. (2015). Risk perceptions and health behavior. Current Opinion in Psychology, 5, 85–89. http://dx.doi.org/10.1016/j.copsyc.2015.03.012 Fredrickson, B. L. (1998). What good are positive emotions? Review of General Psychology, 2, 300–319. https://doi.org/10.1037//1089-2680.2.3.300 Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56, 218–226. https://doi.org/10.1037//0003-066x.56.3.218 Fredrickson, B. L., & Joiner, T. (2018). Reflections on positive emotions and upward spirals. Perspectives on Psychological Science, 13, 194–199. https://doi.org/10.1177/1745691617692106 Fredrickson, B. L., Tugade, M. M., Waugh, C. E., & Larkin, G. (2003). What good are positive emotions in crises? A prospective study of resilience and emotions following the terrorist attacks on the United States on September 11th, 2001. Journal of Personality and Social Psychology, 84, 365–376. https://doi.org/10.1037//0022-3514.84.2.365 Fried, R. N., Woods, C. (Producers), & Anspaugh, D. (Director). (1993). Rudy [DVD]. United States: TriStar Pictures. Green, M.C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79, 701–721. https://doi.org/10.1037//0022-3514.79.5.701 Johnson, K. J., & Fredrickson, B. L. (2005). ‘We all look the same to me’: Positive emotions eliminate the own-race bias in face recognition. Psychological Science, 16, 875–881. https://doi.org/10.1111/j.1467-9280.2005.01631.x Kim, H. K., & Niederdeppe, J. (2016). Effects of self-affirmation, narratives, and informational messages in reducing unrealistic optimism about alcohol-related problems among college students. Human Communication Research, 42, 246–268. https://doi.org/10.1111/hcre.12073 Kok, B. E., Coffey, K. A., Cohn, M. A., Catalino, L. I., Vacharkulksemsuk, T., Algoe, S. B., Brantley, M., & Fredrickson, B. L. (2013). How positive emotions build physical health: Perceived positive social connections account for the upward spiral between positive emotions and vagal tone. Psychological Science, 24, 1–10. https://doi.org/10.1177/0956797612470827 Mahony, D. L. (2000, Spring). Is laughter the best medicine or any medicine at all? Eye on Psi Chi, 4(3), 18–21. http://dx.doi.org/10.24839/1092-0803.Eye4.3.18 Pew Research Center. (2019). Mobile fact sheet. Retrieved from https://www.pewinternet.org/fact-sheet/mobile/

Raghunathan, R., & Trope, Y. (2002). Walking the tightrope between feeling good and being accurate: Mood as a resource in processing persuasive messages. Journal of Personality and Social Psychology, 83, 510–525. https://doi.org/10.1037//0022-3514.83.3.510 Rosenstock, I. M. (1974). The Health Belief Model and preventive health behavior. Health Education Monographs, 2, 354–386. http://dx.doi.org/10.1177/109019817400200405 Sheeran, P., Harris, P. R., & Epton, T. (2014). Does heightening risk appraisals change people’s intentions and behavior? A meta-analysis of experimental studies. Psychological Bulletin, 140, 511–543. http://dx.doi.org/10.1037/ a0033065 Slater, M. D., & Rouner, D. (2002). Entertainment education and elaboration likelihood: Understanding the processing of narrative persuasion. Communication Theory, 12, 173–191. http://dx.doi.org/10.1111/j.1468-2885.2002.tb00265.x studyinghealth. (2013). Tiling (Neutral Induction - 5 and a half minutes) [Video file]. Retrieved from https://www.youtube.com/watch?v=0z8VdZIPIa4&feature=youtu.be van der Velde, F., Hooykaas, C., & van der Joop, P. (1992). Risk perception and behavior: Pessimism, realism, and optimism about AIDS-related health behavior. Psychology & Health, 6, 23–28. http://dx.doi.org/10.1080/08870449208402018 Waugh, C. E., & Fredrickson, B. L. (2006). Nice to know you: Positive emotions, self-other overlap, and complex understanding in the formation of a new relationship. Journal of Positive Psychology, 1, 93–106. https://doi.org/10.1080/17439760500510569 Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806–820. https://doi.org/10.1037//0022-3514.39.5.806 Witte, K. (1992). Putting the fear back into fear appeals: The Extended Parallel Process Model. Communication Monographs, 59, 329–349. https://doi.org/10.1080/03637759209376276 Wright, R. R., Broadbent, C., Graves, A., & Gibson, J. (2016). Health behavior change promotion among latter-day saint college students. Psi Chi Journal of Psychological Research, 21, 200–215. https://doi.org/10.24839/2164-8204.JN21.3.200 Yoo, J. H., Kreuter, M. W., Lai, C., & Fu, Q. (2014). Understanding narrative effects: The role of discrete negative emotions on message processing and attitudes among low-income African American women. Health Communication, 29, 494–504. https://doi.org/10.1080/10410236.2013.776001 Author Note. Aisha L. Udochi, Department of Psychology and Counseling, University of Central Arkansas; Lindsay A. Kennedy, https://orcid.org/0000-0002-5632-3189, Department of Psychology, Hendrix College; Marc A. Sestir, https://orcid.org/0000-0001-6014-6464, Department of Psychology and Counseling, University of Central Arkansas. Special thanks to Psi Chi Journal reviewers for their support. Correspondence concerning this article should be addressed to Aisha L. Udochi, Department of Psychology and Counseling, University of Central Arkansas, Conway, AR, 72035. E-mail: aisha.udochi@yahoo.com

245

https://doi.org/10.24839/2325-7342.JN24.4.246

The Effects of Conflicting Dietary Information on Dieting Self-Efficacy and Motivation Vivian Wei Lin Leung , Elizabeth J. Krumrei-Mancuso* Pepperdine University

, and Janet Trammell*

ABSTRACT. There are many different popular diet trends, each with different recommendations. Such a proliferation of often conflicting information is likely overwhelming, may lead to confusion about how to eat healthily, and may influence self-efficacy and motivation to do so. In the current study, relationships among conflicting dietary information, dieting self-efficacy, motivation, and gender, were examined in 194 undergraduate college students (M age = 19.25, SD = 1.37). Participants randomly received information on either one recommended diet (MyPlate) or on multiple conflicting diets. We hypothesized that (a) people’s level of motivation and self-efficacy for eating healthily would be lower when presented with conflicting information about healthily eating than when presented with consistent information about eating healthily, and (b) this difference would be larger in men than in women. Contrary to the hypothesis, consistency of diet information did not affect self-efficacy (p = .81, η p2 < .01) or motivation (p = .75, ηp2 = .001). However, men showed lower intrinsic motivation to eat healthier than women (p = .02, ηp2 = .63), regardless of the consistency of information. Interestingly, those who viewed only one recommended diet (MyPlate) reported feeling more overwhelmed than those who viewed multiple conflicting diets, p = .001, ηp2 = .06, which could be explained by a lack of familiarity with MyPlate. These findings suggest that receiving conflicting information does not detrimentally affect motivation and self-efficacy to eat healthily, but that familiarity is an important variable for future consideration. Keywords: diet, conflicting information, self-efficacy, motivation, diet, eating

T WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

246

he United States is increasingly becoming more health conscious. In fact, 54% of people in the United States pay more attention to eating healthy foods today compared to 20 years ago (DeSilver, 2016). In particular, collegeaged students report the intention to eat healthful foods (Heckler, Gardner, & Robinson, 2010) and the intention to eat a variety of foods to establish healthy eating habits (Davy, Benes, & Driskell, 2006). Yet, despite this increased attention on and intention to engage in healthier eating, more than half of Americans claim that most days they should be eating healthier than they are (DeSilver, 2016),

suggesting that there are barriers to eating healthily. One such potential barrier may be the specific nature and quantity of available health information about dieting. Historically, information about healthy eating has been consistent and simple. In the early 1990s, the U.S. Department of Agriculture (USDA) introduced a food pyramid suggesting the optimal number of servings to be eaten each day from various basic food groups (USDA, 2018). In 2011, this was updated to MyPlate, which was meant to help people focus more on food variety, amount, nutrition, and appropriate levels of saturated fat,

*Faculty mentor

Leung, Krumrei-Mancuso, and Trammell | Conflicting Information on Self-Efficacy and Motivation

sodium, and added sugars (USDA, 2018). Although other dietary information outside of governmental guidelines has existed for decades, in recent years, many diet trends have been popularized and are quite distinct from the governmental guidelines, such as the ketogenic diet, the Paleolithic diet, the Whole 30 diet, the Mediterranean diet, and intermittent fasting diet, to name a few. All of these diets have unique requirements and all claim to help individuals eat healthily; yet, their recom mendations contradict one another as well as the government MyPlate suggestions. For instance, the ketogenic diet recommends a high fat and low carbohydrates diet, and the paleolithic diet discourages intake of legumes and grains, yet the MyPlate guidelines encourage balanced intake of dairy, vegetables, grains, and protein. The current study examined how the availability of such a vast amount of contradictory diet information might impact people’s self-efficacy and motivation when it comes to eating healthily. Self-efficacy, individuals’ beliefs about their capabilities to produce the desired effects through actions (Bandura, 1998), has been established as important for improving eating behavior (Prestwich et al., 2014; Schneider, O’Leary, & Agras, 1987). Parkinson et al. (2016) proposed that high levels of self-efficacy were attributable to people’s drives to change their behaviors, and perhaps their ability to maintain such behavior over time, which could positively influence people’s intentions to engage in healthy eating behavior. In support of this idea, dietary self-efficacy has been shown to be a good predictor of dietary self-care and self-reported adherence to dietary self-care activities in diabetic patients (Senécal & Nouwen, 2000) and to positively predict recovery and control of bulimic behaviors (Schneider, O’Leary, & Agras, 1987). Thus, selfefficacy is a crucial predictor of behaviors related to engagement in healthy eating habits. However, one potential barrier to dietary selfefficacy is perceived difficulty of eating healthily; that is, the intention of engaging in a healthy eating behavior depends on the perception of difficulty of engaging in such behavior (Povey & Conner, 2000). The increasingly conflicting diet information available in U. S. society could negatively impact self-efficacy because the information may challenge individuals’ beliefs in their capability to engage in a correct diet trend, among the extensive choices that are available. In addition to self-efficacy, intrinsic motivation to engage in behavioral change, in this case healthier

eating, is an essential construct in explaining success at healthy eating. Intrinsic motivation is a predictor of the intention to make plans for a healthy diet and the act of achieving it (Ridder, Wit, & Adriaanse, 2009), is positively related to healthy eating and exercising behaviors (Naughton, McCarthy, & McCarthy, 2015), and is predictive of increased physical fitness (Kato et al., 2013). Another predictor of healthy eating is identified regulation, which was previously described as a construct of extrinsic motivation, but Ryan and Deci (2000) had explained it as an autonomous composite of motivation because identified regulation is related to the integration or assimilation of regulations and self. Although identified regulation does not directly relate to one’s inherent enjoyment (i.e., intrinsic motivation), it is still related to the ability to focus on one’s thoughts for personal success (Guay, Delise, Fernet, Julian, & Senecal, 2008). Relating to healthier dieting, intrinsic motivation is the intention to engage in healthy eating, and identified regulation is one’s thought to set healthy eating as a personal desire. The current research presumed that having contradicting recommendations for a multitude of different diets could be overwhelming to people and aimed to examine the effects of contradicting information on self-efficacy and motivation to engage in healthy eating. The effects of conflicting dietary information have not been examined in previous literature with regard to healthy eating, however, a body of literature has examined the effects of contradictory information on healthcare choices. For example, research has found that receiving contradictory information about medica tion from different sources (e.g., about dosage, effects, duration, side-effects, and severity) is associ ated with less medication adherence (Carpenter et al., 2010). Elstad, Carpenter, Devellis, and Blalock (2012) found that, when patients with arthritis were presented with conflicting information, most relied on trial and error when the risk was minimal (e.g., minor side-effects of medication), and other strategies (e.g., heuristic processing, and weighing benefits and risks) when the risk involved was greater (e.g., heart problems). Conflicting information seems to pose negative effects because individuals are not able to engage with or believe in discrepant information from different sources (Carpenter, 2016). On these bases, we expected that presenting conflicting dietary information may alter one’s decision-making or will to diet, thus affecting self-efficacy and motivation for healthy

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

247

Conflicting Information on Self-Efficacy and Motivation | Leung, Krumrei-Mancuso, and Trammell

eating habits. Eating behaviors tend to differ among gen ders, and these differences include amount of food intake, the choice of nutrition source, and awareness of health benefits of eating healthy. For example, Wardle et al. (2004) studied the gender differences in food choices and found that women reported avoiding high-fat foods, eating more fruit and fiber, and limiting salt compared to men. Women also reported attaching greater importance to eating and dieting healthily (Wardle et al., 2014). In addition, a number of studies have found that gender was a predictor of self-efficacy for changing eating behaviors (Bruce, Beech, Thrope, Mincey, & Griffith, 2017; Horacek et al., 2002; Stephens et al., 2017). Horacek et al. (2002) found that men had lower confidence in eating more fruits and vegetables than women, which was associated with men having inadequate fruit and vegetable intake. Robles et al. (2014) found that women had higher self-efficacy in reading nutrition facts labels, and gender was shown to be a strong predictor of healthy eating behaviors. For these reasons, we examined gender differences in self-efficacy and motivation to eat healthily. The Current Study Despite the established relationship in the litera ture between motivation, self-efficacy, and healthy eating, it is unclear whether conflicting dietary information interferes with individuals’ motivation to eat healthy or beliefs in their capability to eat healthy. The current study aimed to investigate the role of conflicting dietary information in dieting self-efficacy and motivation, with an examination of potential gender differences. We anticipated that receiving conflicting information about multiple diets would make participants feel overwhelmed, which is a factor that negatively influences inten tions to engage in healthy eating (e.g., Povey & Con ner, 2000). We further anticipated that receiving conflicting information about healthy dieting would decrease dietary self-efficacy and motivation. Finally, we hypothesized that men would report lower selfefficacy and motivation scores than women when presented with conflicting dietary information.

Method WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

248

Participants The sample consisted of 194 undergraduate students, 140 (72%) women and 54 (28%) men (Mage = 19.25, SD = 1.37). See Table 1 for a detailed description of the sample.

Design We employed a between-subjects experimental design in which participants were randomly assigned to one of two conditions that differed in reading material presented. The first condi tion consisted of a one-page description about the MyPlate guidelines suggested by the U.S. Government. This was presented as consistent, nonconflicting information as all information was coherently presented from one source. The second condition consisted of a one-page description about six different and conflicting diet trends (MyPlate, Mediterranean diet, ketogenic diet, paleolithic diet, intermittent fasting, and juice cleanse), with one to two paragraphs describing each. This was presented as conflicting diet information because the diet trends were different one another in their recommendations. The one-page length was decided based on providing the most succinct information to partici pants on various food groups. Both conditions were equated for total length and included information on the suggested portion of food groups (i.e., dairy, fiber, carbohydrate, protein, and fats) for each diet. Measures and Materials After reverse coding relevant items, all items were summed to create a total score for each of the three scales, including the Nutrition Self-Efficacy Scale (Schwarzer & Renner, 2009), the Eating Habits Confidence Survey (Sallis et al., 1988), and the Regulations of Eating Behavior Scale (Pelletier, Dion, Slovinec-D’Angelo, & Reid, 2004), before arriving to the exact measures for the analyses. Descriptive characteristics. Seven questions addressed the age, gender, race, household income, height, and weight of participants. Height and weight were used to calculate participants’ Body Mass Index (BMI). Consistent dietar y information. The first condition consisted of a one-page description, 652 words in total, with eight short paragraphs about the MyPlate guidelines suggested by the U.S. Government. This was presented as consistent, non conflicting information because all information was coherently presented from one source. Appendix A consists of the full text that was presented in the consistent dietary information condition. Conflicting dietary information. The second condition consisted of a one-page description, 700 words in total, with six different and conflicting diet trends (MyPlate, Mediterranean diet, ketogenic diet, paleolithic diet, intermittent fasting, and juice

Leung, Krumrei-Mancuso, and Trammell | Conflicting Information on Self-Efficacy and Motivation

cleanse), with one to two paragraphs describing each. This was presented as conflicting diet infor mation because the diet trends were different from one another in their recommendations. Appendix B consists of the full text that was presented in the conflicting dietary information condition. Self-reported preexisting knowledge of diets. Before reading the assigned dietary description, which was either the consistent or conflicting dietary information, participants answered two questions assessing their familiarity and knowledge of the six diet trends: “How well do you know each of the following diet trends?” and “How healthy do you think each of the following diet trend is?” Responses were recorded on a scale of 1 (not at all) to 10 (to a great extent). Although a true measure of participants’ knowledge on each diet was not addressed, this question assessed their self-reported preexisting knowledge of the diet trends. Self-efficacy for healthy eating habits. Two scales were used to measure participants’ selfefficacy specific to eating healthily. The first measure of healthy eating was the Nutrition SelfEfficacy Scale. Four items of the 5-item Nutrition Self-Efficacy Scale (Schwarzer & Renner, 2009) were used to assess self-efficacy to overcome certain barriers when attempting to stick to healthful foods (Schwarzer & Renner, 2009). All items formed a response to the stem: “I can manage to stick to healthful foods…” A sample item is “I can manage to stick to healthful foods, even if I need a long time to develop the necessary routines.” Participants’ responses were rated on a 4-point Likert-type scale from 1 (very uncertain) to 4 (very certain). One item (“I can manage to stick to healthful foods, even if I do not receive a great deal of support from others when making my first attempts”) from the original scale was eliminated because it assessed a social factor of self-efficacy, which was irrelevant to the current study. In this study, internal consistency was acceptable (Cronbach’s α = .78); this value was calculated with the sum of each scale (i.e., Nutri tion Self-Efficacy Scale, Eating Habits Confidence Survey, and the Intrinsic Motivation subscale and Identified Regulation subscale of the Regulations of Eating Behavior Scale). In addition, the second measure of healthy eating was a score adapted from 15 items of the Eating Habits Confidence Survey (Sallis et al., 1988) to assess self-efficacy for healthy eating habits. The original scale focused on salt and fat intake, but for the purpose of our study, we altered the content to reflect healthy eating in general.

For instance, we changed “stick to your low fat, low salt foods when you feel depressed, bored, or tense,” to “stick to your healthy diet when you feel depressed, bored, or tense.” Items were rated on a 6-point Likert-type scale from (0 = “does not apply,” 1 = “I know I cannot,” and 5 = “I know I can”). The internal consistency in the current study was good (Cronbach’s α = .83). Motivation. We used the Intrinsic Motivation (4 items) and Identified Regulation (4 items) subscales of the Regulations of Eating Behavior Scale (Pelletier, Dion, Slovinec-D’Angelo, & Reid, 2004) to measure motivation for the regulation of eating behaviors. Participants were asked to indicate TABLE 1 Descriptive Characteristics of the Sample Variable

Age

19.24

Body Mass Index

21.67 n

Min

Max

1.37

3.99

15.75

37.32

Gender Female

140

72.0

Male

28.0

0.5

European American

44.1

Mixed Race

15.4

Asian

23.6

Latino/a

6.7

White Hispanic

3.1

Native American

3.1

African American

2.1

Other

1.0

Less than $25,000

6.2

$25,000–$34,999

5.1

$35,000–$49,999

5.6

$50,000–$74,999

6.2

$75,000–$99,999

8.2

$100,000–$149,999

15.9

$150,000–$199,999

5.6

$200,000 or more

19.0

Unknown

28.2

Prefer not to say Race

Annual Household Income

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

249

Conflicting Information on Self-Efficacy and Motivation | Leung, Krumrei-Mancuso, and Trammell

motivation level for regulating eating behavior in response to the question, “Why are you regulating your eating behaviors?” Intrinsically motivated behaviors are defined as behaviors for the pleasure, interest, and satisfaction derived from participation (Pelletier et al., 2004). A sample item from the intrinsic motivation subscale is: “I take pleasure in fixing healthy meals.” Identified regulated behav iors are activities that fit a person’s values and goals through the integration of identified regulations and self (Pelletier et al., 2004). A sample item from the identified regulation subscale is: “It is a way to ensure long-term health benefits.” Items were rated on a 7-point Likert-type scale from 1 (does not correspond at all) to 7 (corresponds exactly). In the current study, internal consistency was good for the intrinsic motivation subscale (Cronbach’s α = .87) and the identified regulation subscale (Cronbach’s α = .88). Manipulation check. A question assessing whether the conditions caused participants to feel overwhelmed was presented immediately after participants read the diet information and before completing the scales for self-efficacy and motivation. Participants were asked to rate how they felt after reading the description of the diet trend(s), (1 = “not at all overwhelmed,” and 10 = “very overwhelmed”).

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

250

Procedure This study was approved by the Seaver College Institutional Review Board, and all participants provided informed consent prior to participat ing. Participants were recruited either through a research participation portal for undergraduate students enrolled in psychology classes, or the uni versity’s social media platform page, from October to December of 2018. Participation took place online. First, self-reported familiarity knowledge for the six diet trends (MyPlate, Mediterranean diet, ketogenic diet, paleolithic diet, intermittent fasting, and juice cleanse) were rated. Next, participants were randomly assigned to read either consistent or conflicting dietary information. Then, participants answered questions assessing their levels of selfefficacy and motivation. Demographics information was assessed in the end of the survey. Students enrolled in psychology classes received credit within their psychology course for participating. Other students were offered a chance to enter a raffle for one of three $15 visa gift cards.

Results Prior to conducting analyses, three cases were

deleted listwise for missing data (n = 1) or not meeting the inclusion criteria of being at least 18 years old (n = 2). We conducted preliminary analyses to examine links between key descriptive characteristics and study variables. Age (r = .18, p = .01), BMI (r = -.24, p = .001), and gender (r = .17, p = .02) were correlated with intrinsic motivation to eat healthy. Therefore, these variables were controlled in analyses of motivation. In addition, level of familiarity with the various diet trends was significantly related to intrinsic motivation (r = .27, p ≤ .001), identified regulation (r = .23, p = .001), and self-efficacy for healthy eating habits (r = .22, p = .002), therefore, this factor was controlled in all the analyses. The average duration to complete the survey was 8 minutes 35 seconds, with a SD of 1 minute and 42 seconds. Self-Efficacy for Healthy Eating Habits Results of a two-way Analysis of Variance (ANOVA) for the Nutrition Self-Efficacy Scale showed no main effect for gender, F(1, 190) = 0.02, p = .90, ηp2 < .01, or conflicting versus consistent diet information, F(1, 190) = 0.06, p = .81, ηp2 < .01. There was also no interaction between gender and the nature of diet information presented, F(1, 190) = 0.22, p = .64, ηp2 = .001. Results of a two-way ANOVA for the Eating Habits Confidence Survey also showed no significant main effect for gender, F(1, 189) = 3.16, p = .08, ηp2 = .02, or conflicting versus consistent diet information, F(1, 189) = 0.42, p = .52, ηp2 = .002. There was also no interaction between gender and the type of diet information presented, F(1, 189) = 0.001, p = .98, ηp2 < .01. Motivation Results of a two-way ANOVA for intrinsic motiva tion showed a significant main effect for gender, F(1, 190) = 5.35, p = .02, ηp2 = .03, such that men showed lower intrinsic motivation (M = 16.83) to eat healthy than women (M = 18.91), regardless of type of information. There was no main effect for type of information, F(1, 190) = 0.10, p = .75, ηp2 = .001. There was also no interaction between gender and the type of diet information presented, F(1, 190) = 0.04, p = .85, ηp2 < .01. Results of a two-way ANOVA for identified regulation showed no main effect for gender, F(1, 189) = 0.12, p = .73, ηp2 = .001, or for type of information, F(1, 189) = 1.24, p = .27, ηp2 = .007. There was also no interaction between gender and the type of diet information presented, F(1, 189) = 0.14, p = .71, ηp2 = .001.

Leung, Krumrei-Mancuso, and Trammell | Conflicting Information on Self-Efficacy and Motivation

Manipulation Check Results from a two-way ANOVA showed no main effects for gender, F(1, 189) = 0.19, p = .67, ηp2 = .001, or interaction between gender and the type of information, F(1, 189) = 0.40, p = .53, ηp2 = .002. However, there was a significant effect for type of information, F(1, 189) = 10.96, p = .001, ηp2 = .055. In contrast to the hypothesis, participants presented with consistent, nonconflicting diet information showed greater levels of feeling overwhelmed (M = 5.78), compared to participants who were presented with conflicting diet information (M = 18.91).

Discussion The current study evaluated the effects of conflict ing diet information on college students’ levels of self-efficacy and motivation to eat healthy. Our goal was to increase understanding about how the wide variety of diet trends, which contradict one another, affect people’s self-efficacy and motivation to eat healthy. We had expected that individuals presented with conflicting diet information would experience lower levels of self-efficacy and motivation to eat healthy compared to individuals presented with consistent diet information. Furthermore, we predicted men to have lower scores on self-efficacy and motivation than women. One of our hypotheses was supported. Although no main effect was observed for con flicting versus consistent diet information in levels of self-efficacy or motivation to eat healthy, men displayed lower levels of intrinsic motivation to eat healthy than women, which was in line with some previous research (Bruce et al., 2017; Stephens et al., 2017). This finding was in the same direction as previous research on gender and self-efficacy (Bruce et al., 2017; Horacek et al., 2002; Stephens et al., 2017). Although specific reasons are unknown, it seems men do not take as much pleasure, interest, or satisfaction in healthy eating as women. Future research may look into the many social, cultural, and environmental differences that may help explain why men have lower intrinsic motivation to eat healthy than women. A potential explanation of the insignificant findings for conflicting versus consistent diet information is that the experimental manipulation in this study was not effective. We had expected the conflicting dietary information condition to result in participants feeling more overwhelmed than the consistent dietary information condition. However, we observed the opposite. Given that the two condi tions were matched for overall length, more details

were provided about the MyPlate guidelines within the consistent condition than about any one of the diets in the conflicting condition. Perhaps this amount of information about one dietary recom mendation was more overwhelming to participants. Nevertheless, the differences observed between the two conditions for how overwhelmed participants felt did not translate to differences in levels of selfefficacy or motivation for eating healthy. It is worth considering how our unique sample might have impacted the findings. First, our sample had a mean Body Mass Index (BMI) score of 21.69, in comparison to the national mean of 26 BMI for U.S. adults (Centers for Disease Control and Prevention, 2004). This suggests that participants fell within a healthy weight-to-height ratio and therefore, may not have been in need of the dietary information to the same extent we would expect from individuals in the general population, where average BMI reflects the overweight category. In addition, our sample was less familiar with the presented diets than we had expected, which might have limited our ability to observe an effect in the current study. Specifically, 54.9% (n = 107) participants scored 1 for familiarity with the pre sented description of the MyPlate guidelines. In fact, except for juice cleanse, for which participants scored a mean of 6.7 out of 10 for familiarity, mean familiarity scores for all the other diets were below the midpoint of the scale (see Table 2). The unexpected unfamiliarity of these university students with the MyPlate recommendations may suggest that college students have not learned about the governmental dietary guidelines. However, another potential explanation is that participants were not familiar with the MyPlate label. Given the age range of the sample body, participants TABLE 2 Mean Scores of Familiarity on Each Diet Trend Diet Trend

MyPlate Guidelines

3.12

3.05

Mediterranean

3.47

2.65

Ketogenic

3.59

3.11

Paleolithic

4.15

2.94

Intermittent Fasting

5.52

3.16

Juice Cleanse

6.73

2.75

Note. Familiarity was based on one Likert-type item ranging from 1 (not at all) to 10 (to a great extent).

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

251

Conflicting Information on Self-Efficacy and Motivation | Leung, Krumrei-Mancuso, and Trammell

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

252

could have potentially been more familiar with the former versions of the MyPlate (food pyramid or MyPyramid) because the MyPlate guidelines weren’t implemented until 2011. Regardless, it is alarming to note that familiarity with the government suggested diet guidelines was substantially lower than that of any other diet trend assessed, in which juice cleanse had the highest familiarity rating of all. Juice cleanse is not known to be a healthy diet trend, as it was mentioned by the National Center for Complementary and Integrative Health (NCCIH) as a program that may potentially cause diarrhea, dehydration, and electrolyte imbalances (NCCIH, 2017). This sug gests a need to reevaluate the modalities by which emerging adults are educated about healthy eating. Further work is needed to evaluate young adults’ knowledge of standard, government suggested dietrelated information. In addition, given that juice cleanse was ranked highest for familiarity, future research may examine how popular diet trends are promoted and compare this to the methods used to promote governmental diet guidelines. Consideration can be given to marketing strategies and the platforms used to promote diet guidelines. A limitation of this study is that the sample was not representative of the general population in the areas of gender, age, household income, race, and BMI. In addition, future research should utilize a stronger experimental manipulation. Providing one page of information per condition might not have been sufficient to create an effect in the current study. Future research may consider study designs with stronger manipulation of information. In addition, because the current study seemed to have measured self-rated familiarity instead of actual knowledge of the diet trends, which was how this study labeled the measure, future research may benefit from assessing participants’ actual preexist ing knowledge of the various diets presented, rather than assessing self-rated familiarity. Despite the fact that the current study was unable to address links between conflicting infor mation and levels of self-efficacy and motivation for eating healthy, our results yield potential concern about young adults’ knowledge about healthy diet. Although our sample would be considered healthy according to their BMI scores, and financially stable according to mean household income, participants averaged below a score of 5 out of 10 on familiarity for most of the diets presented, except for juice cleanse. Notably, among all diets assessed, the MyPlate guideline had the lowest

rating for familiarity. This raises questions about how a lack of familiarity with government suggested food guidelines among some emerging adults may impact eating behaviors. It seems greater, or differ ent, efforts are needed to promote evidence-based information about healthy food choices among college students.

References Bandura, A. (1998). Health promotion from the perspective of social cognitive theory. Psychology & Health, 13, 623–649. https://doi.org/10.1080/08870449808407422 Bruce, M. A., Beech, B. M., Thorpe, R. J., Mincey, K., & Griffith, D. M. (2017). Racial and gender disparities in sugar consumption change efficacy among firstyear college students. Appetite, 109, 33–39. https://doi.org/10.1016/j.appet.2016.11.017 Carpenter, D. M., DeVellis R. F., Fisher E. B., DeVellis B. M., Hogan S. L., & Jordan J. M. (2010). The effect of conflicting medication information and physician support on medication adherence for chronically ill patients. Patient Education and Counseling, 81, 169–176. https://doi.org/10.1016/j.pec.2009.11.006 Carpenter, D. M., Geryk, L. L., Chen, A. T., Nagler, R. H., Dieckmann, N. F., & Han, P. K. J. (2016). Conflicting health information: A critical research need. Health Expectations, 19, 1173–1182. https://doi.org/10.1111/hex.12438 Centers for Disease Control and Prevention (2004). Healthy weight, overweight, and obesity among U.S. adults (Publication No. 03-0260). Retrieved from https://www.cdc.gov/nchs/data/nhanes/databriefs/adultweight.pdf Davy, S. R., Benes, B. A., & Driskell, J. A. (2006). Sex differences in dieting trends, eating habits, and nutrition beliefs of a group of midwestern college students. Journal of the American Dietetic Association, 106, 1673–1677. https://doi.org/10.1016/j.jada.2006.07.017 DeSilver, D. (2016). What’s on your table? How America’s diet has changed over the decades. Pew Research Center. Retrieved from www.pewresearch. org/fact-tank/2016/12/13/whats-on-your-table-how-americas-diet-haschanged-over-the-decades/ Elstad, E., Carpenter, D. M., Devellis, R. F., & Blalock, S. J. (2012). Patient decision making in the face of conflicting medication information. International Journal of Qualitative Studies on Health and Well-being, 7, 1–11. https://doi.org/10.3402/qhw.v7i0.18523 Guay, F., Delisle, M. N., Fernet, C., Julian, E., & Senecal, C. (2008). Does taskrelated identified regulation moderate the sociometer effect? A study of performance feedback, perceived inclusion, and state self-esteem. Social Behavior and Personality, 36, 239–254. Heckler, E. B., Gardner, C. D., & Robinson, T. N. (2010). Effects of a college course about food and society on students’ eating behaviors. American Journal of Preventive Medicine, 38, 543–547. https://doi.org/10.1016/j.amepre.2010.01.026 Horacek, T. M., White, A., Betts, N. M., Hoerr, S., Georgiou, C., Nitzke, S., . . . Greene, G. (2002). Self-efficacy, perceived benefits, and weight satisfaction discriminate among stages of change for fruit and vegetable intakes for young men and women. Journal of the American Dietetic Association, 102, 1466–1470. https://doi.org/10.1016/S0002-8223(02)90325-1 https://doi.org/10.2224/sbp.2008.36.2.239 Kato, Y., Iwanaga, M., Roth, R., Hamasaki, T., & Greimel, E. (2013). Psychometric validation of the Motivation for Healthy Eating Scale (MHES). Scientific Research, 4, 136–141. https://doi.org/10.4236/psych.2013.42020 National Center for Complementary and Integrative Health (2017). ‘Detoxes’ and ‘cleanses.’ Retrieved from https://nccih.nih.gov/health/detoxes-cleanses Naughton, P., McCarthy, S. N., & McCarthy, M. B. (2015). The creation of a healthy eating motivation score and its association with food choice and physical activity in a cross sectional sample of Irish adults. The International Journal of Behavioral Nutrition and Physical Activity, 12:74. https://doi.org/10.1186/s12966-015-0234-0 Parkinson, J., David, P., & Rundle‐Thiele, S. (2017). Self‐efficacy or perceived behavioural control: Which influences consumers’ physical activity and healthful eating behaviour maintenance? Journal of Consumer Behaviour, 16, 413–423. https://doi.org/10.1002/cb.1641 Pelletier, L. G., Dion, Angelo, S. C., Slovinec-D’Angelo, M., & Reid, R. (2004). Why

Leung, Krumrei-Mancuso, and Trammell | Conflicting Information on Self-Efficacy and Motivation

do you regulate what you eat? Relationships between forms of regulation, eating behaviors, sustained dietary behavior change, and psychological adjustment. Motivation and Emotion, 28, 245–277. https://doi.org/10.1023/B:MOEM.0000040154.40922.14 Povey, R., Conner, M., Sparks, P., James, R., & Shepherd, R. (2000). Application of the theory of planned behaviour to two dietary behaviours: Roles of perceived control and self-efficacy. British Journal of Health Psychology, 5, 121–139. https://doi.org/10.1348/135910700168810 Prestwich, A., Kellar, I., Parker, R., MacRae, S., Learmonth, M., Sykes, B., . . . Castle, H. (2014). How can self-efficacy be increased? Meta-analysis of dietary interventions. Health Psychology Review, 8, 270–285. https://doi.org/10.1080/17437199.2013.813729 Ridder, D. D., Wit, J. D., & Adriaanse, M. A. (2009). Making plans for healthy diet: The role of motivation and action orientation. European Journal of Social Psychology, 39, 622–630. https://doi.org/10.1002/ejsp.560 Robles, B., Smith, L. V., Ponce, M., Piron, J., & Kuo, T. (2014). The influence of gender and self-efficacy on healthy eating in a low-income urban population affected by structural changes to the food environment. Journal of Obesity, 2014, 1–12. https://doi.org/10.1155/2014/908391 Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68–78. https://doi.org/10.1037/0003-066X.55.1.68 Sallis, J. F., Pinski, R. B., Grossman, R. M., Patterson, T. L., & Nader, P. R. (1988). The development of self-efficacy scales for health-related diet and exercise behaviors. Health Education Research, 3, 283–292. https://doi.org/10.1093/her/3.3.283 Schneider, J. A., O’Leary, A., & Agras, W. S. (1987). The role of perceived selfefficacy in recovery from bulimia: A preliminary examination. Behaviour Research and Therapy, 25, 429–432. https://doi.org/10.1016/0005-7967(87)90020-9 Schwarzer, R., & Renner, B. (2009). Health-specific self-efficacy scales. Freie

Universität Berlin. Retrieved March 14, 2019. Senécal, C. & Nouwen, A. (2000). Motivation and dietary self-care in adults with diabetes: Are self-efficacy and autonomous self-regulation complementary or competing constructs? Health Psychology, 19, 452–457. https://doi.org/10/1037//0278-6133.19.5.452 Stephens, J. D., Althouse, A., Tan, A., & Melnyk, B. M. (2017). The role of race and gender in nutrition habits and self-efficacy: Results from the young adult weight loss study. Journal of Obesity, 1–6. https://doi.org/10.1155/2017/5980698 United States Department of Agriculture. (2018, January 26) MyPlate. Retrieved February 18, 2018, from https://www.choosemyplate.gov/MyPlate U.S. Department of Health and Human Services & U.S. Department of Agriculture. (2015, December). 2015–2020 Dietary Guidelines for Americans (8). Retrieved February 20, 2018, from http://health.gov/dietaryguidelines/2015/guidelines/ Wardle, J., Haase, A. M., Steptoe, A., Nillapun, M., Jonwutiwes, K., & Bellisie, F. (2004). Gender differences in food choice: The contribution of health beliefs and dieting. Annals of Behavioral Medicine, 27, 107–116. https://doi.org/10.1207/s15324796abm2702_5 Author Note. Vivian W. L. Leung, https://orcid.org/0000-0002-7005-3584, Department of Psychology, Pepperdine University; Elizabeth J. KrumreiMancuso, https://orcid.org/0000-0001-6151-7845 Department of Psychology, Pepperdine University; Janet Trammell, https://orcid.org/0000-0002-0304-6974, Department of Psychology, Pepperdine University. Special thanks to Psi Chi Journal reviewers for their support. Correspondence concerning this article should be addressed to Vivian W. L. Leung, Department of Psychology, Pepperdine University, Malibu, CA 90263. E-mail: vivian.lwleung@gmail.com

APPENDIX A Reading Material for Consistent Information Conditions Regulated by the United States Department of Agriculture, the MyPlate guideline replaced MyPyramid in June 2011. MyPlate is derived from the Dietary Guidelines for Americans for the purpose of helping Americans eat healthfully, and for consumers to make better food choices. In order to help Americans to find a balanced diet, the guidelines specifically highlight the importance of focusing on food variety, amount, and nutrition, as well as choosing foods and beverages with less saturated fat, sodium, and added sugars. Starting with small changes, the guidelines hope to aid in building healthier eating styles and supporting healthy eating for everyone. Some small changes recommended include “make half your grains whole grains,”“move to low-fat and fat-free milk or yogurt,”“vary your protein routine,” and “make half your plate fruits and vegetables.” There are five food groups in MyPlate, including dairy, vegetables, grains, fruits, and protein. According to sex, age, and level of physical activity, the intake of these food groups may vary. Generally, it is suggested to divide one’s diet to approximately 30% of grains, 40% of vegetables, 10% of fruits, and 20% of protein, with small portions of dairy. Below is the recommended portion for each food group. Dairy Both women and men are recommended to have no more than 3 cups of dairy products per day. Examples of what counts as a cup in the dairy group are 1 cup of milk, 1 cup of yogurt, 2 cups of cottage cheese and 1.5 cups of ice cream. Some examples of common portions and cup equivalents are 1 small container of yogurt being ¾ cup, 1 slice of hard cheese being ½ cup of milk, and 1 scoop of ice cream being 1/3 cup of milk. Vegetables Women may have no more than 2.5 cups of vegetables, and men no more than 3 cups per meal. Examples of what counts as a cup of vegetables are 1 cup of cooked spinach, 1 large baked sweet potato, 1 cup of cooked beans, 1 cup of green peas, and 1 cup of chopped onions. Grains Women may have up to 6 ounces equivalents of grains, and men up to 8 ounces equivalents. The daily minimum amount of whole grains for women is 3 ounce equivalent, and for men is 3 to 4 ounces equivalents. Examples of what counts as one ounce-equivalent of grains are 1 mini bagel, 1 regular slice of bread, 5 whole wheat crackers, half a cup of oatmeal, and 3 cups of popcorn. Some examples of common portions and ounce-equivalents are 1 large bagel being 4 ounce-equivalents, 2 regular slices being 2 ounce-equivalents, 1 cup of cooked rice being 2 ounce-equivalents, and 1 cup of cooked pasta being 2 ounce-equivalents. Fruits Women may have up to 1.5 to 2 cups of fruits, and men up to 2 cups. Examples of what counts as a cup of fruit are 1 small apple, 1 cup of mixed fruit, 1 whole cup of grapes, 1 cup of sliced plum, 1 large orange, and 1 cup of 100% fruit juice. Protein Women may have up to 5 to 5.5 ounces of equivalents, and men up to 5.5 to 6.5 ounces equivalents. Examples of one ounce-equivalent are 1 ounce cooked lean beef, 1 ounce cooked chicken without skin, ¼ cup of tofu, and 1 ounce cooked fish or shell fish. Some common portions and ounce-equivalents are 1 small steak being 3.5-4 ounce-equivalents, half of a small chicken breast being 3 ounce-equivalents, 1 cup of lentil soup being 2 ounce-equivalents, and 1 small trout being 3 ounce-equivalents. Oil The daily allowance of oil is 5-6 teaspoons maximum for women, and 6-7 teaspoons for men. Examples of foods rich in oils are margarine with ~3 tsp of oil in 1 tablespoon, peanut butter with ~4tsp of oil in 2 tablespoons, half of an avocado with ~3tsp of oil, and margarine with ~2.5tsp of oil in 1 tablespoon.

253

Conflicting Information on Self-Efficacy and Motivation | Leung, Krumrei-Mancuso, and Trammell

APPENDIX B Reading Material for Consistent Information Conditions Paleolithic Diet (Caveman diet) The author of the book The Paleo Diet, Dr. Cordain, has spent more than 20 years researching the nutrient composition of diets. The paleolithic diet aims to recreate the diet cavemen used to have, in order to reach the perfect balance of nutrients. People have reported feeling better mentally and physically after trying the diet. The following food items are allowed in a paleolithic diet: Grass-fed meats, fish/seafood, fresh fruits, fresh vegetables, eggs, nuts, seeds, healthy oils (olive, walnut, flaxseed, macadamia, avocado, coconut). The following food items are not allowed in a paleolithic diet: Cereal grains, legumes (including peanuts, chickpeas, beans), dairy, refined sugar, potatoes, processed foods (e.g. hotdogs, spam), overly salty foods, refined vegetable oils, candy/junk/processed food. Mediterranean Diet The Mediterranean diet is based on the diet habits of the people in Crete, Greece. Research has shown that the Mediterranean diet is associated with a reduced risk of Alzheimer’s disease and increased longevity. This diet mostly includes the foods that are typically available in the named region. The three main daily meals should follow these basic elements: 1. Cereals: 1–2 servings per meal in the form of bread, pasta, rice, couscous or others. (preferably whole grain) 2. Vegetables: For lunch and dinner. Two or more servings per meal, and at least one of the serving should be raw. 3. Fruits: 1–2 servings per meal. 4. Dairy products: Preferably low-fat yogurt. 5. Olive oil: Should be principle source of fat. 6. Alcohol: moderate amount. Ketogenic Diet The ketogenic diet has been widely researched for medical purposes. It is considered a low carbs high fat diet (LCHF). The diet aims to lower blood sugar and insulin levels, while boosting the body’s metabolism by shifting from carbs to fat and ketones. By lowering the intake of carbs and maximizing the intake of fats, the body is introduced to the ketosis state, which helps produce a compound from the breakdown of body fats. Only 20–30g of net carbs is suggested for an everyday intake. Overall, the daily caloric intake should only include 5% carbohydrates, with 70% fats and 25% protein. Intermittent Fasting Although still a rather new trend, intermittent fasting has been acclaimed to help endurance levels for athletes. Intermittent fasting involves two states: the fed state (high insulin) and the fasted state (low insulin). This method suggests there will be no net weight gain if the two states are balanced. There is no specific food that is allowed or not allowed, but there are various ways to do the fasting schedule. Examples: 16:8 (fast for 16 hours per day, eat all meals within 8 hours per day), 5:2 (5 regular eating days per week and 2 fasting days per week with a maximum of 500 calories per fasting day), and 24-hour fast (only eating once per day at the same time, frequency of fasting around 2–3 times per week) Juice Cleanse A lot of people have claimed to benefit from juice cleanses not only for weight loss and detox, but also for the journey of controlling one’s mindset and body. There are many different brands of juice cleanses available nowadays. A juice cleanse typically lasts for no more than seven days. However, there are cases where people have tried to do a juice cleanse for as long as a month. Despite the varied durations, a common guideline for the cleanse is having three juices a day to replace breakfast, lunch, and dinner. MyPlate Guidelines MyPlate was introduced to replace the food pyramid by the United States Department of Agriculture. The guidelines are meant to help Americans eat healthfully, and for consumers to make better food choices. There are five food groups in MyPlate, including dairy, vegetables, grains, fruits, and protein. It suggests that women may have up to 5 ounces of protein per meal, and 6 ounces for men; an equivalent of 3 ounces of protein is a small chicken breast in half. For grains, women may have up to 3 ounces per meal and men may have up to 3-4 ounces; an equivalent of 4 ounces of grain is a large bagel. For vegetables, women may have 2 cups per meal and men may have 3 cups.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

254

https://doi.org/10.24839/2325-7342.JN24.4.255

Young Women's Sexist Beliefs and Internalized Misogyny: Links With Psychosocial and Relational Functioning and Political Behavior Adrian J. Dehlin and Renee V. Galliher* Utah State University

ABSTRACT. The current study examined links among sexism, psychosocial functioning, and political behavior in 210 young women from the United States. Participants completed a survey including the Ambivalent Sexism Inventory, Revised Religious Fundamentalism Scale, Attitudes Toward Women Scale, Internalized Misogyny Scale, and Revised Dyadic Adjustment Scale. Higher religious fundamentalism was associated with lower relationship quality, mediated by internalized misogyny, traditional gender roles, and hostile sexism. Although mental health outcomes were also collected, associations with sexist attitudes were nonsignificant. The intersection of sexist attitudes and internalized misogyny with political affiliation and voting behavior was also explored. Participants who voted for Clinton/Kaine reported lower levels of internalized misogyny when compared to those who voted for Trump/Pence. In addition, Democrat and Independent individuals reported significantly lower levels of internalized misogyny and hostile sexism when compared to Republican and Not Affiliated individuals. Keywords: sexism, relationship quality, religious fundamentalism, Trump, internalized misogyny

exism is defined as a belief, practice, or system that supports the notion that men are intrinsically superior to women (Anderson, 2010; Borrell et al., 2011). Past studies have found sexism to be a prevalent form of prejudice that most women experience on a weekly and sometimes daily basis (Berg, 2006; Swim, Hyers, Coher, & Ferguson, 2001). In this study, we explored young women’s endorsement of sexist ideology, as it relates to a number of important socialization experiences and psychosocial outcomes. Sexism is a ubiquitous experience in the lives of young women in the United States. For example, Berg (2006) reported that all 382 women in her sample reported experiencing sexism, and 25% said they felt it happened “a lot” (p. 975). In another study, participants were asked to record the number of sexist incidents they observed over a span of 7

*Faculty mentor

to 13 days (Swim et al., 2001). Participant records indicated that incidents of sexism occurred at least once per week, with some participants reporting sexist experiences daily. Thus, sexism is common, and often a daily occurrence for many women. Perhaps the most well-known conceptualiza tion of modern sexism is the ambivalent sexism framework proposed by Glick and Fiske (1996). Ambivalent sexism refers to sexist beliefs that fit in to two major categories, hostile and benevo lent (Anderson, 2010; Huang, Davies, Sibley, & Osborne, 2016). Hostile sexism aims to validate “…male power, traditional gender roles, and men’s exploitation of women as sexual objects through derogatory characterizations of women” (Anderson, 2010, p. 151). Although hostile sexism can be easily identified, benevolent sexism has a tendency to go unnoticed (Huang et al., 2016). Benevolent

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

255

Sexist Beliefs and Internalized Misogyny | Dehlin and Galliher

sexism “relies on kinder and gentler justifications of male dominance and prescribed gender roles; it recognizes men’s dependence on women and takes a romanticized view of heterosexual relationships” (Anderson, 2010, p. 151). Because it is subtler by nature, people are much less likely to be held accountable when conveying benevolent sexism in comparison to hostile sexism (Barreto & Ellemers, 2005). In addition to the sexist attitudes individuals confront externally on a day-to-day basis, these beliefs can be internalized. According to Spengler (2014), internalized misogyny is made up of two main elements: self-objectification and passive acceptance of gender roles. These components are linked to a plethora of negative outcomes includ ing psychological distress, disordered eating, and mental illness. Given the omnipresent nature of misogynistic and sexist messages received by women in patriarchal societies, the internalization of sexist ideology is often automatic and unnoticed. One study found that women conveyed dialogic practices of internalized sexism (i.e., invalidating, derogat ing, or objectifying women in everyday language) on average 11 times per 10-minute increment of conversation (Bearman, Korobov, & Thorne, 2009). This rate of frequency illustrated how extensive internalized sexism truly is within society.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

256

Religious and Political Context of Sexism Religious and political contexts are powerful and overlapping socialization forces related to attitudes about gender. We hypothesized that conservative religious and political affiliations support adher ence to traditional, rigid gender attitudes. In the following section, we outline the relevance of religious and political contexts for embracing sexist ideology. Religion. A number of authors have identified conservative and traditional religious belief systems as an important socializing context for attitudes about women and gender role expectations. For example, Mikolajczak and Pietrzak (2014) found that higher levels of religiosity in a sample of Catholic Polish women were related to higher endorsement of benevolent sexism, mediated by their adherence to values for conservatism. Burn and Busso (2005) found that benevolent sexism was linked positively to internal and external reli giosity and scriptural literalism in a sample of U.S. Christian college students. Maltby, Hall, Anderson, and Edwards (2010) observed a significant relation ship between the “protective paternalism” (p. 619)

component of benevolent sexism and Christian Orthodoxy in a sample of evangelical Christian col lege students, although only for men. Finally, Glick, Sakalh-Uğurlu, Akbaş, Orta, and Ceylan (2016) examined the association between honor beliefs—a strict set of rules for women that typically include compliance to men, sexual purity, and religious adherence—and two correlates, religiosity and sex ism. In their large Turkish sample, men were more apt to report endorsing honor beliefs than women. Hostile and benevolent sexism were positively cor related with religiosity. Furthermore, hostile sexism, benevolent sexism, and religiosity were positively correlated with honor belief acceptance. The body of research related to religiosity and sexism has operationalized religiosity using a wide variety of definitions and measures. In this study, we introduced the concept of religious fundamentalism as a potentially relevant aspect of religious belief sys tems (Alderdice, 2010). Defined as the belief in the absolute authority and unquestionable superiority of a particular sacred text or set of religious teach ings, religious fundamentalism can be endorsed to varying degrees within any faith community. In general, more conservative faith traditions—those which espouse literal interpretations of their sacred texts or more rigid expectations for maintaining spiritual morality—are more closely associated with religious fundamentalism. However, across faith contexts, we hypothesized that the more dog matic, morally rigid or narrow religious beliefs and attitudes associated with religious fundamentalism would be associated with sexist attitudes (i.e., more rigid, morally driven attitudes about gender), and would link to psychosocial outcomes through their relationship with sexist beliefs (i.e., mediation). U.S. political climate. Following the presiden tial election of 2016, discussion of internalized sexism, or women holding beliefs that support their own oppression, was in the mainstream conscious ness (Bialik, 2017; Fenton & Lopez, 2016; Moore, 2016). Conversations surrounding gender equality were a central component of ongoing divisive dialogue. Bock, Byrd-Craven, and Burkley (2017), in a sample of college students from a southwestern university in the United States, observed that those who affiliated with the Republican political party and voted for Donald Trump in the 2016 U.S. presidential election reported higher levels of benevolent and hostile sexism and greater adher ence to traditional gender roles, compared to Democrats and those who voted for Hilary Clinton. Similarly, Blair (2017) capitalized on the timing of

Dehlin and Galliher | Sexist Beliefs and Internalized Misogyny

her study of general attitudes regarding a number of social issues (e.g., LGBTQ rights, racism and Islamophobia, sexism). As a follow-up to original data collection, she assessed voting intentions just prior to the 2016 presidential election and found ambivalent sexism to be a strong predictor of inten tion to vote for Donald Trump or a third party/ undecided relative to Hilary Clinton. However, because of the inclusion criteria for the original, larger study, Blair’s sample was 95% male, and 60% resided in the state of Utah. In the current study, we examined self-reported voting behavior and political affiliation in a nationwide sample of young women. Psychosocial Correlates of Sexism A number of recent studies have examined links between sexism (both externally experienced and internalized) and the psychosocial functioning of victims. Experiencing prejudice and discrimination has been found to result in a wide range of nega tive mental health and well-being outcomes, and influence dynamics within romantic relationships. Berg (2006) assessed associations among gender-related stressors, frequency of experienced sexist events, and PTSD symptoms in a large com munity sample of women. A significant positive correlation emerged between experienced levels of everyday sexism and PTSD scores. This relationship was found to be especially strong when individuals reported “recent sexist degradation” (p. 984). Similarly, Borrell and colleagues (2011), with a sample of over 10,000 women, found that individu als who reported experiencing sexism had poorer overall mental health when compared to those who did not perceive sexism. The same was true when researchers looked at the prevalence of specific types of mental illness. Pervasiveness of depression and anxiety was highest among survey participants who perceived sexism. Although research on internalized misogyny is still developing, some studies have assessed out comes associated with women’s internalization of sexist beliefs. Szymanski, Gupta, Carr, and Stewart (2009) examined relationships between sexist events and psychological distress in a sample of college women. Internalized misogyny moderated, and “intensified” (p. 101), the relationship between sexism and distress. Subsequently, Szymanski and Henrichs-Bech (2014) more directly assessed links between psychological distress and internalized misogyny, and observed a direct association between internalized misogyny and psychological distress.

As research on sexism continues to expand, its role in romantic relationships has also emerged as a significant area of study. Lee, Fiske, Glick, and Chen (2010) examined endorsement of benevolent and hostile sexism and the traits American men and women preferred/selected in romantic relation ships. Women and men who endorsed benevolent sexism ideals were more likely to select for a “tra ditional gender partner” (p. 590) when compared to those who did not endorse benevolent sexism ideals. For women, “traditional” was characterized by selecting traits such as “strong” and “traditional male,” while discarding traits like “feminine” (p. 590). For men, “traditional” was characterized by selecting traits such as “warm and traditional female” and discarding “not traditional” (p. 590). However, Casad, Salazar, and Macina (2015) found that engaged women who reported higher levels of benevolent sexism reported lower relationship satisfaction and self-assurance, suggesting that endorsement of traditional roles in relationships may come with a relational and personal cost. Overall, Sibley, and Tan (2011) used obser vational methodology to examine links between sexism and relationship conflict in a sample of 99 heterosexual couples. Higher levels of self-reported hostile sexism in men was related to lower levels of openness during a recorded conflict interaction in both men and women, as rated by couple members while they reviewed the recording. Further, men’s hostile sexism was linked to higher observer ratings of hostile communication by both men and women. Men’s hostile sexism was also indirectly linked to lower couple members’ ratings of the success of the discussion in bringing about their desired change, via its effect on openness in the interaction. Interestingly, when men endorsed higher levels of benevolent sexism, men were more open and less hostile in the interaction. However, if women endorsed benevolent sexism and their husbands did not, women were more hostile, less open, and perceived their discussions as less successful. Thus, sexism emerges as a dyadic process that unfolds in a relational context. Summary and Research Questions In sum, a review of the literature highlights a number of psychosocial and relational outcomes that have been linked consistently to sexist belief systems. We highlight fundamentalist, dogmatic religious beliefs as a potential socializing context for the development of sexist attitudes. We also note that the broader political climate appears to

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

257

Sexist Beliefs and Internalized Misogyny | Dehlin and Galliher

be tightly connected to contemporary attitudes about women. Within this context, we posed the following questions. First, how are sexist attitudes and beliefs related to psychosocial health (i.e., anxiety, depression, and self-esteem) and relationship quality? We hypothesized that endorsement of sexist attitudes would be associated with lower self-reported mental health and more relationship distress. Second, do sexist attitudes mediate links between religious fundamentalism and psychosocial health or relationship quality? We expected that the indirect effect of religious fundamentalism would be deleterious, through its effect on sexist beliefs. And third, is there a relationship between endorse ment of sexist attitudes and political affiliation or behavior? We expected participants with a higher endorsement of sexist attitudes to report more conservative political affiliations, with an increased likelihood to have voted for Donald Trump.

Method Study Design This study was approved and monitored by the authors’ institutional review board for the protec tion of human research participants. A correlational design examined relationships among internalized misogyny/sexism, psychological health, relation ship quality, religious fundamentalism, and political behavior. Participants Our sample included 210 women, ages 18–25 (M = 22, SD = 2.33). This age restriction ensured that individuals were able to answer questions about their voting behavior and political affiliation, focus ing on the young adult population in particular. Table 1 provides a summary of demographic data.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

258

Measures Demographic information. Items assessed age, gen der identity, sexual orientation, religious affiliation, socioeconomic status, educational status, relation ship status, political affiliation, voting behavior, and ethnicity/race. Ambivalent Sexism Inventory. (Glicke & Fiske, 1996). This measure consists of 22 items rated on a Likert scale from 0 (disagree strongly) to 5 (agree strongly). Items are divided into two types, hostile and benevolent sexism. Following reverse scoring adjustments, the ambivalent sexism total can be calculated by taking the average of the hostile and benevolent sexism scores. Glicke and Fiske reported evidence of convergent and discriminate validity

over six different measure development samples (1996). Cronbach’s αs reported by Glicke and Fiske ranged from .62 to .86. In the current study, hostile sexism yielded an α of .86, and benevolent sexism yielded an α of .80. Internalized Misogyny Scale. (Piggott, 2004). This measure consists of 17 items rated on a Likerttype scale from 1 (strongly disagree) to 7 (strongly agree). Totals range from 17 to 119, with higher scores indicating higher levels of internalized misogyny. Piggott (2004) reported significant positive correlations with the Body Image scale and Modern Sexism scale, and Cronbach’s αs of .87 and .88. In the current study, Cronbach’s α = .93. Attitudes Toward Women Scale. (Spence & Hahn, 1997; Spence, Helmreich, & Stapp, 1973). This measure consists of 12 items assessing endorsement of traditional gender roles, rated on a Likert-type scale from 1 (strongly agree) to 4 (strongly disagree). Following reverse scoring adjust ments, higher scores indicate stronger adherence to traditional gender roles. Spence and colleagues have confirmed a single factor structure in multiple samples. The scale also showed acceptable testretest reliability, and αs in the mid .80s or higher (Spence & Hahn, 1997). In the current study, Cronbach’s α = .89. Revised Religious Fundamentalism Scale. (Altemeyer & Hunsberger, 2004). This measure assesses rigid, dogmatic religious attitudes, and consists of 12 items rated on a 9-point scale. One total score is calculated as the mean across all items, with higher scores indicating greater fundamental ism. Altemeyer and Hunsberger (2004) reported a correlation of .68 with right-wing authoritarianism, and Cronbach’s α of .92. In the current study, this scale yielded an α of .88. Generalized Anxiety Disorder 7-item Scale (GAD-7). (Spitzer, Kroenke, Williams, & Löwe, 2006). This measure assesses day-to-day experiences of anxiety, and consists of 7 items rated on a Likerttype scale from 0 (not at all) to 3 (nearly every day). GAD-7 scores were significantly related to declines in functioning, self-reported disability, number of clinic visits, and level of difficulty attributed to symptoms in a clinical sample (Spitzer et al., 2006). Spitzer and colleagues (2006) reported a Cronbach’s α of .92, with a test-retest reliability of 0.83. In the current study, α = .92. Rosenberg Self-Esteem Scale (RSES). (Rosenberg, 1965). This measure consists of 10 items rated on a Likert-type scale from 0 (strongly disagree) to 4 (strongly agree). Following reverse

Dehlin and Galliher | Sexist Beliefs and Internalized Misogyny

scoring adjustments, higher scores indicate higher levels of self-esteem. Evidence of good construct validity for the RSES has been reported by multiple studies (Tinakon, & Nahathai, 2012). Rosenberg (1965) reported test-retest reliability of .85, and a Cronbach’s α of .92. In the current study, this scale demonstrated an α of .84. Center for Epidemiological Studies-Depression Scale (CES-D). (Radloff, 1977). This measure assesses symptoms of depression over the past week, and consists of 20 items rated on a Likert-type scale from 1 (rarely or none of the time) to 4 (most or all of the time). Following reverse scoring adjustments, higher scores indicate higher levels of depression. Radloff (1977) found that the scale effectively dis criminated between depressive and nondepressive cases. Radloff reported αs ranging from .85 to .90, and test-retest correlations between .45 and .70. In the current study, the CES-D yielded an α of .90. Revised Dyadic Adjustment Scale (RDAS). (Busby, Christensen, Crane, & Larson, 1995). This measure consists of 14 items rated on a 5 or 6-point Likert-type scale. Items are divided into three scales: consensus, satisfaction, and cohesion. Consensus refers to the level of reported agreement between romantic partners in their decision making, leisure, values, and affection. Satisfaction is defined by sta bility and conflict—assessed through participants’ divorce consideration, marriage regret, quarrel frequency, and overall annoyance with partner. Cohesion assesses the degree to which romantic partners have activities in common and engage in a “stimulating exchange of ideas” (Busby et al., 1995, p. 296). Items 1–6 assess with consensus, items 7–10 assess satisfaction, and items 11–14 assess cohesion. Higher scores indicate higher relationship quality. The measure has been found to effectively discern between satisfied and distressed relationships (Busby et al, 1995). Busby and colleagues (1995) reported an α of .90 for the overall scale. In the current study, αs were calculated for each subscale: Relationship Consensus = .83, Satisfaction = .83, and Cohesion = .82. Procedure Participants were recruited through a Qualtrics survey panel. Eligibility requirements included identifying as a woman, being able to complete the survey in English, residence in the United States, and age between 18 and 25. Qualtrics representa tives worked with researchers to prepare an online Qualtrics survey, and then coordinated with study panel partners to recruit a prearranged number

TABLE 1 Demographic Information Variable

Sexual orientation Heterosexual Gay/Lesbian

149

71.0

3.8

Bisexual

18.1

Queer/Pansexual/Questioning

5.7

1.4

27.1

3.3

Asexual Ethnicity* African American American Indian/Alaska Native Asian/Asian American

7.6

Latinx

13.3

Middle Eastern

1.0

Native Hawaiian/Pacific Islander

0.5

122

58.1

0.5

Single/Not Dating

35.2

Dating Nonexclusively

9.5

In a Committed Relationship

36.2

Engaged/Married

17.6

1.5

4.8

White/European American Other Relationship Status

Separated/Divorced Education Level Less than High School Completed High School

36.2

100

47.6

10.5

1.0

Less Than $20,000

35.2

$20,000–$39,999

31.0

$40,000–$59.000

17.1

$60,000–$79,000

7.6

$80,000–$99.000

2.9

6.2

Some College/Vocational Training Bachelor's Degree Graduate/Professional Degree Income

$100,000 and above Religion/Faith Tradition Agnostic/Atheist

19.5

Christian

124

59.0

Buddhist

2.9

Muslim

1.9

Jewish Other/None

1.0

15.7

Note. *Participants could select more than one ethnicity, so that total exceeds 100%.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

259

Sexist Beliefs and Internalized Misogyny | Dehlin and Galliher

of participants. Survey participants were existing members of the survey panels, and were offered the opportunity to participate in the survey through standardized e-mail notifications. If they chose to participate in the survey, they were compensated by the survey panel partner in accordance with the panel guidelines. Survey participants were typically compensated in the form of airline miles, gift cards, cash, merchandise, or coupons. Complete, cleaned participant data were delivered to researchers in an anonymous format.

Results Research Question 1: Bivariate Correlations Table 2 presents means and standard deviations for all study variables, along with bivariate correlations between measures of sexist attitudes and psychoso cial functioning. Measures of sexist attitudes were roughly normally distributed around the midpoints of the scales. Measures of mental health were also roughly normally distributed, but mean scores for relationship quality were near the high end of the scales. The most consistent patterns of significant bivariate correlation were with relationship quality. Hostile sexism, internalized misogyny, and endorse ment of traditional gender roles were all linked to lower relationship quality across the three RDAS scales. Higher levels of religious fundamentalism were associated with lower levels of self-esteem and with lower levels of relationship satisfaction. Additionally, higher levels of religious fundamen talism were associated with higher scores across three measures: internalized misogyny, ambivalent sexism, and attitudes toward women. There were, however, no significant correlations between mea sures of sexist attitudes and measures of mental health (i.e., depression, anxiety, or self-esteem).

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

260

Research Question 2: Test of Indirect Effects Table 3 presents a summary of the mediation models. Primary mediation analyses were con ducted using the PROCESS macro in SPSS (Hayes, 2013). The PROCESS macro utilizes bootstrapping techniques and ordinary least square regression to calculate direct effects of the independent variable (religious fundamentalism) on the dependent variables (relationship quality), as well as the indi rect effects of religious fundamentalism through the sexism variables. Based on the patterns of bivariate correlation, mediation analyses were not conducted for the mental health outcomes because there was no indication that either sexist attitudes

or religious fundamentalism consistently linked to mental health. However, mediation models were tested using religious fundamentalism as the independent variable, the four measures of sexist attitudes as mediators (in four separate models), and the three RDAS scales as dependent variables in separate models. Significant mediation (indirect effects) are indicated by confidence intervals that do not include zero. Across all models, there was no significant direct effect of religious fundamentalism on relationship quality. However, religious funda mentalism was strongly related to higher scores on all four measures of sexist ideology. Internalized Misogyny, Attitudes Toward Women, and Hostile Sexism all consistently demonstrated negative direct effects on RDAS scales. And finally, significant indirect effects of religious fundamentalism on all three RDAS scales emerged through Internalized Misogyny, Attitudes Toward Women, and Hostile Sexism. Higher levels of religious fundamentalism linked to higher endorsement of sexist attitudes and traditional gender roles, which in turn linked to lower relationship quality. Research Question 3: Political Behavior Means and standard deviations for all groups for all measures of sexist ideology are presented in Table 4. A series of one-way Analyses of Variance (ANOVAs) were conducted to examine differences in sexist attitudes among participants in terms of political affiliation and voting behavior. Four groups were compared with regard to political affiliation (Republican, Democrat, Independent, not affili ated). Four groups were also compared with regard to voting behavior in the 2016 presidential election (voted for Trump, voted for Clinton, registered but did not vote, and not registered). Political affiliation. All four ANOVAs examin ing differences among the political affiliation groups were statistically significant: internalized misogyny, F(3, 198) = 6.76, p < .001, h = .09; hostile sexism, F(3, 198) = 11.83, p < .001, h = .15; benevolent sexism, F(3, 198) = 4.03, p = .008, h = .06; and attitudes toward women, F(3, 198) = 3.63, p = .014, h = .05. Scheffe post-hoc pairwise comparisons were conducted among the four groups for each ANOVA. Table 3 illustrates signifi cant mean differences between groups in terms of political affiliation. Participants who identified as Democrat or Independent reported significantly lower internalized misogyny and hostile sexism when compared to Republican and Not Affiliated participants. Republican participants reported

Dehlin and Galliher | Sexist Beliefs and Internalized Misogyny

significantly higher levels of benevolent sexism when compared to Independent participants. Not Affiliated participants reported stronger adherence to traditional gender roles when compared to Independent participants. Voting behavior. Three of the four ANOVAs examining differences among the voting groups were significant; internalized misogyny, F(3, 194) = 5.56, p = .001, h = .08; hostile sexism, F(3, 194) = 10.04, p < .001, h = .13; and benevolent sexism, F(3, 194) = 3.90, p = .010, h = .06. There were no differences among the voting groups on attitudes toward women, F(3, 194) = 1.07, p = .361, h = .02. Table 3 displays the results of Scheffe post-hoc tests. Participants who voted for Trump/Pence reported significantly higher levels of internalized misogyny when compared to participants who voted for Clinton/Kaine or participants who were registered, but did not vote. Participants who voted for Trump/ Pence or were not registered to vote reported significantly higher hostile sexism scores than those who voted for Clinton/Kaine and those who were registered but did not vote. Participants who voted for Trump/Pence also reported significantly higher levels of benevolent sexism when compared to those who voted for Clinton/Kaine.

Discussion This study supports several conclusions regarding the connection between sexism and internalized misogyny and a variety of psychosocial and political factors. As hypothesized, sexist beliefs were consis tently linked to both relational functioning and political behavior. However, we did not observe the hypothesized links to mental health variables. Associations Between Sexism and Relationship Functioning in the Context of Religion Internalization of sexist beliefs was consistently sig nificantly related to relationship quality, although effect sizes were relatively small. Furthermore, an indirect pathway emerged from higher religious fundamentalism to lower relationship quality, through internalized misogyny, endorsement of traditional gender roles, and hostile sexism. Past studies have identified religiosity as an important variable for unpacking the context of sexism (e.g., Burn & Busso, 2005; Mikolajczak and Pietrzak, 2014; Tasdemir & Sakalli-Ugurlu, 2009). Similarly, in this sample, we observed sig nificant and strong relationships between religious fundamentalism and all forms of sexist attitudes and traditional gender roles. Our measure of

religious fundamentalism does not map exactly on to previously used measures of religiosity. For example, Mikolajczak and Pietrzak (2014) assessed religious affiliation (Catholic vs. non-Catholic) and religious participation. Tasdemir and Sakalli-Ugurlu (2009) assessed the extent to which participants endorsed essential components of Islamic beliefs. Rather than assessing religious activity or specific religious ideology, religious fundamentalism, as a construct, captures inflexible, dogmatic religious attitudes, via items such as “To lead the best, most meaningful life, one must belong to the one, fun damentally true religion” and “The fundamentals of God’s religion should never be tampered with, or compromised with others’ beliefs.” (Altemeyer & Hunsberger, 2004, p. 130). Endorsement of the items on the religious fundamentalism appear to link more closely to all forms of sexism, including more overt and intense hostile sexism, and may be particularly relevant in more conservative, dogmatic religious contexts. When we examined the relationships between religious fundamentalism and relationship quality as mediated by sexist beliefs, benevolent sexism was the only variable not found to be a significant mediator. Consistently, across all three scales of the RDAS, religious fundamentalism related to higher TABLE 2 Descriptive Statistics and Correlations Among Study Variables Variable

Internalized Misogyny

Generalized Anxiety Disorder-7 Scale

-.01

Center for Epidemiological StudiesDepression Scale

Ambivalent Sexism Inventory – Hostile

Ambivalent Sexism Inventory – Benevolent

Attitudes Toward Women Scale

Religious Fundamentalism

-.05

-.04

-.12

-.09

0.24(0.90)

-.01

-.09

-.10

-.03

-.09

2.27(0.63)

Revised Religious Fundamentalism Scale

.32**

.38**

.61**

.31**

–

4.57(1.66)

Rosenberg SelfEsteem Scale

-.04

-.09

-.11

-.03

-.17*

2.36(0.61)

RDAS– Consensus

-.25**

-.20**

-.10

-.29**

-.04

4.35(1.02)

RDAS– Satisfaction

-.29**

-.22**

-.13

-.41**

-.15*

4.41(1.15)

*.15*

-.10

4.57(1.66)

1.76(0.63)

4.57(1.66)

RDAS–Cohesion

-.11

-.16*

M (SD)

3.30(1.32)

3.04(1.05)

.01 3.49(0.91)

M (SD)

–

Note. RDAS = Revised Dyadic Adjustment Scale.

261

Sexist Beliefs and Internalized Misogyny | Dehlin and Galliher

TABLE 3 Tests of Mediation Effect

coeff

LLCI

ULCI

Direct Effects: Same for All Models Religious Fundamentalism > Internalized Misogyny

.25

.05

4.83

< .001

.15

.36

Religious Fundamentalism > Attitudes Toward Women

.11

.03

4.43

< .001

.06

.16

Religious Fundamentalism > Hostile Sexism

.24

.04

5.76

< .001

.16

.32

Religious Fundamentalism > Benevolent Sexism

.34

.03

11.02

< .001

.28

.40

.10

Direct and Indirect Effects for Separate Models RDAS: Cohesions Direct Effects Religious Fundamentalism > RDAS: Cohesion

.02

.04

.41

.68

-.06

Internalized Misogyny > RDAS: Cohesion

-.08

.05

-1.67

.09

-.18

.02

Attitudes Toward Women > RDAS: Cohesion

-.22

.11

-2.16

.03

-.44

-.02

Hostile Sexism > RDAS: Cohesion

-.16

.07

-2.49

.02

-.29

-.04

Benevolent Sexism > RDAS: Cohesion

.03

.09

0.33

.74

-.15

.21

Through Internalized Misogyny

-.02

.02

-.05

.00

Through Attitudes Toward Women

-.03

.02

-.06

-.01

Through Hostile Sexism

-.04

.02

-.08

-.01

Through Benevolent Sexism

.01

.03

-.05

.08

Religious Fundamentalism > RDAS: Satisfaction

-.05

.05

-0.95

.35

-.14

-.05

Internalized Misogyny > RDAS: Satisfaction

-.23

.06

-3.78

<.001

-.35

-.11

Attitudes Toward Women > RDAS: Satisfaction

-.73

.12

-6.02

<.001

-.10

-.49

Hostile Sexism > RDAS: Satisfaction

-.21

.08

-2.56

.01

-.37

-.05

Benevolent Sexism > RDAS: Satisfaction

-.07

.11

-0.65

.52

-.29

.15

Indirect Effects of Religious Fundamentalism on RDAS: Cohesion

RDAS: Satisfaction Direct Effects

Indirect Effects of Religious Fundamentalism on RDAS: Satisfaction Through Internalized Misogyny

-.06

.02

-.11

-.02

Through Attitudes Toward Women

-.08

.02

-.14

-.04

Through Hostile Sexism

-.05

.02

-.09

-.02

Through Benevolent Sexism

-.02

.03

-.10

.04

Religious Fundamentalism > RDAS: Consensus

.03

.04

0.63

.52

-.05

.12

Internalized Misogyny > RDAS: Consensus

-.20

.06

-3.68

<.001

-.31

-.09

Attitudes Toward Women > RDAS: Consensus

-.49

.11

-4.36

<.001

-.72

-.27

Hostile Sexism > RDAS: Consensus

-.21

.07

-2.89

.004

-.35

-.07

Benevolent Sexism > RDAS: Consensus

-.13

.10

-1.37

.17

-.32

.07

RDAS: Consensus Direct Effects

Indirect Effects of Religious Fundamentalism on RDAS: Consensus Through Internalized Misogyny

-.05

.02

-.10

-.02

Through Attitudes Toward Women

-.06

.02

-.11

-.02

Through Hostile Sexism

-.05

.02

-.10

-.02

Through Benevolent Sexism

-.04

.03

-.11

.02

Note. LLCI = Lower level confidence interval. UCLI = Upper level confidence interval. RDAS = Revised Dyadic Adjustment Scale.

262

levels of internalized misogyny, hostile sexism, and endorsement of traditional gender roles. In turn, all of those variables linked to more negative relationship qualities. Fundamentalist religious belief systems tend to emphasize the importance of family and marriage—which is viewed as a critical part of members’ spiritual lives and development. Our data suggests that embracing these beliefs is associated with gender-related attitudes that are linked to poorer relationship quality. Because of this, it may be that some religious communi ties are socializing members in ways that are counterproductive to their own goals. Our lack of significant findings related to benevolent sexism speak to the complicated nature of this particular form of sexist ideology. Overall et al. (2011) also observed complex patterns of association between benevolent sexism and observed romantic relation ship interactions, suggesting that the impact of benevolent sexism may depend on the way that it manifests within couples. Sexist Beliefs and Political Behavior Participants who identified as Republican/not affiliated or voted for Trump/were not registered to vote had the highest levels of sexist beliefs and internalized misogyny overall. Participants who identified as Democrats/Independents or voted for Clinton/were registered but did not vote had lower sexist beliefs overall. Group differences were more pronounced in internalized misogyny and hostile sexism, whereas differences were less pronounced in terms of benevolent sexism and traditional gender roles. The majority of participants had more liberal ideologies, and half of participants did not vote. Of those who voted, 80% voted Democrat. The election of Trump in the 2016 United States presidential election served as a catalyst for increased dialogue surrounding the impacts of sexism. Following the release of voting demograph ics, sources across various ideologies reported that White women were the second largest group responsible for Trump’s election—with White men being the first (“Exit polls”, 2016; “Fox News exit polls”, 2016; Huang, Jacoby, Strickland, & Lai, 2016). This was particularly shocking when considering the release of a recorded conversation between Trump and an Access Hollywood inter viewer that occurred pre-election (Fahrenthold, 2016). In the recording, Trump is heard relaying a variety of misogynistic sentiments—with the most quoted being, “grab ‘em by the pussy” (“Transcript”, 2016, para. 22). This elicited passionate public

Dehlin and Galliher | Sexist Beliefs and Internalized Misogyny

conversation surrounding the presence and influ ence of sexism and internalized misogyny within the realm of politics. This translated into the writing and publication of media articles that hypothesized the role of internalized misogyny in the election (“A Vote”, 2016; “How Unconscious”, 2017). Although peer-reviewed research on the subject is relatively scarce, there are a few publications in existence. One such study examined the relation ship between sexism and participation in the 2016 presidential election (Bock et al., 2017), in which participants who reported higher scores on Hostile Sexism and Attitudes Toward Women measures were more likely to have voted for Trump. Our results are consistent with these findings. Because the present sociopolitical climate in America has greatly impacted public awareness, perceptions, and behavior regarding sexism, research on the topic is even more relevant. We also observed that participants who were registered but did not vote, and participants who were not registered to vote reported distinctly different response patterns. There is a good chance that those who were regis tered but did not vote felt disillusioned by the 2016 election in particular. In contrast, individuals who were not registered to vote were completely disen gaged from the political process. This difference might lead to a unique set of responses. Strengths and Limitations This study used a nationwide sample that was demo graphically representative of the United States to further understand complex relationships between variables associated with internalized sexism. In terms of strengths, results provide individuals with the opportunity to be better informed about conscious and unconscious forces that influence young women as they navigate patriarchal socializa tion contexts. In addition, our correlational data lay a foundation from which researchers can make predictions and identify areas in need of additional exploration. Finally, because of our broad sampling strategy, findings can be generalized within the given demographic constraints. There are also limitations that should be taken into account. A range of psychosocial variables were selected that we anticipated would correlate with sexist beliefs, internalized misogyny, and adher ence to traditional gender roles. As mentioned, the present study identified no links with mental health outcomes. This finding is not consistent with existing literature, which suggests the need for a more comprehensive assessment of mental health. Although this study included anxiety and

depression measures, there are some variables we did not include (i.e., PTSD symptoms). This provides an opportunity for future research. In addition, although we had reasons for constraining participant age, we cannot apply study findings to other developmental stages. This also provides a great opportunity for future research. Conclusions Overall, we observed associations between socializa tion contexts, internalized sexist beliefs, and psy chosocial functioning that were largely consistent with hypotheses. It is important to understand the meaning and application of our findings to the lived experience of young women. The specific types of religious messages inherent in funda mentalist religious ideology, and the rigid gender expectations that correlate so strongly with funda mentalist beliefs, should be framed within a larger values context. We observed associations between fundamentalism, sexism, and relationship quality that may not align with the goals of religiously fundamentalist communities. On the other hand, we observed links to political behavior that may align very closely with the values in more conserva tive communities. Thus, our findings may be quite relevant in religious and educational contexts, as we continue to grapple with issues of gender equity and gender role definition as a larger society. TABLE 3 Means and Standard Deviations for Political Groups Variable

Internalized Misogyny

Hostile Sexism

Benevolent Sexism

Attitudes Toward Women

Political Affiliation Democrat (n = 75)

M(SD)

3.01(1.41)a

2.67(1.09)a

3.37(0.90)

1.76(0.72)

Republican (n = 27)

3.94(1.15)b

3.67(0.83)b

3.89(0.98)a

1.77(0.56)

Independent (n = 43)

2.92(1.27)a

2.75(0.92)a

3.25(1.05)b

1.52(0.48)a

Not Affiliated (n = 57)

3.71(1.14)b

3.42(0.88)b

3.68(0.71)

1.94(0.59)b

Donald Trump/Mike Pence (n = 21)

4.29(1.22)a

3.97(0.68)ac

4.04(0.72)a

1.98(0.58)

Hilary Clinton/Tim Kaine (n = 79)

3.06(1.44)b

2.71(1.12)b

3.30(1.07)b

1.75(0.69)

I am registered, but did not vote (n = 50)

3.17(1.22)b

3.07(0.91)bc

3.58(0.81)

1.69(0.60)

I am not registered to vote (n = 48)

3.50(1.08)

3.26(0.89)ac

3.52(0.73)

1.81(0.63)

2016 Vote

Note. Significant differences among groups noted with superscripts.

263

Sexist Beliefs and Internalized Misogyny | Dehlin and Galliher

References

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

264

Alderdice, J. (2010). On the psychology of religious fundamentalism. In P. J. Verhagen, H. M. van Praag, J. J. López-Ibor Jr., J. L. Cox, & D. Moussaoui (Eds.), Religion and psychiatry: Beyond boundaries. (pp. 305–317). WileyBlackwell. https://doi-org.dist.lib.usu.edu/10.1002/9780470682203.ch17 Altemeyer, B., & Hunsberger, B. (2004). A revised religious fundamentalism scale: The short and sweet of it. The International Journal for the Psychology of Religion, 14, 47–54. http://dx.doi.org/10.1207/s15327582ijpr1401_4 Anderson, K. J. (2010). Benign bigotry: The psychology of subtle prejudice. New York, NY: Cambridge University Press. Barreto, M., & Ellemers, N. (2005). The burden of benevolent sexism: How it contributes to the maintenance of gender inequalities. European Journal of Social Psychology, 35, 633–642. http://dx.doi.org/10.1002/ejsp.270 Bearman, S., Korobov, N., & Thorne, A. (2009). The fabric of internalized sexism. Journal of Integrated Social Sciences, 1, 10–47. Retrieved from www.jiss.org Berg, S. H. (2006). Everyday sexism and posttraumatic stress disorder in women: A correlational study. Violence Against Women, 12, 970–988. https://doi.org/10.1177/1077801206293082 Bialik, C. (2017, January 21). How unconscious sexism could help explain Trump’s win. FiveThirtyEight. Retrieved from https://fivethirtyeight.com/features/ how-unconscious-sexism-could-help-explain-trumps-win/ Blair, K. L. (2017). Did Secretary Clinton lose to a ‘basket of deplorables’? An examination of Islamophobia, homophobia, sexism and conservative ideology in the 2016 US presidential election. Psychology and Sexuality, 8, 334–355. https://doi.org/10.1080/19419899.2017.1397051 Bock, J., Byrd-Craven, J., & Burkley M. (2017). The role of sexism in voting in the 2016 presidential election. Personality and Individual Differences, 119, 189–193. https://doi.org/10.1016/j.paid.2017.07.026 Borrell, C., Artazcoz, L., Gil-González, D., Perez, K., Pérez, G., Vives-Cases, C., & Rohlfs, I. (2011). Determinants of perceived sexism and their role on the Association of Sexism With Mental Health. Women and Health, 51, 583–603. https://doi.org/10.1080/03630242.2011.608416 Burn, S. M., & Busso, J. (2005). Ambivalent sexism, scriptural literalism, and religiosity. Psychology of Women Quarterly, 29, 412–418. https://doi.org/10.1111/j.1471-6402.2005.00241.x Busby, D. M., Christensen, C., Crane, D. R., & Larson, J. H. (1995). A revision of the Dyadic Adjustment Scale for use with distressed and nondistressed couples: Construct hierarchy and multidimensional scales. Journal of Marital and Family Therapy, 21, 289–308. http://dx.doi.org/10.1111/j.1752-0606.1995.tb00163.x Casad, B. J., Salazar, M. M., & Macina, V. (2015). The real versus the ideal: Predicting relationship satisfaction and well-being from endorsement of marriage myths and benevolent sexism. Psychology of Women Quarterly, 39, 119–129. http://dx.doi.org/10.1177/0361684314528304 Exit polls. (2016, November 23). CNN Politics. Retrieved from https://edition.cnn.com/election/2016/results/exit-polls Fahrenthold, D. A. (2016, October 8). Trump recorded having extremely lewd conversation about women in 2005. The Washington Post. Retrieved from https://www.washingtonpost.com/politics/trumprecorded-having-extremely-lewd-conversation-about-women-in2005/2016/10/07/3b9ce776-8cb4-11e6-bf8a-3d26847eeed4_story.html Fenton & Lopez, C. (2016, Dec. 20). A vote against self-interest: Trump and internalized misogyny. Girls’ Globe. Retrieved from https://girlsglobe. org/2016/12/20/a-vote-against-self-interest-trump-and-internalizedmisogyny/ Fox news exit polls. (2016). Fox News. Retrieved from http://www.foxnews.com/politics/elections/2016/exit-polls Glick, P., & Fiske, S., T. (1996). The ambivalent sexism inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70, 491–512. http://dx.doi.org/10.1037/0022-3514.70.3.491 Glick, P., Sakalh-Uğurlu, N., Akbaş, G., Orta, I.M., & Ceylan, S. (2016). Why do women endorse honor beliefs? Ambivalent sexism and religiosity as predictors. Sex Roles, 75, 543–554. http://dx.doi.org/10.1007/s11199-015-0550-5 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford. Huang, J., Jacoby, S., Stirckland, M., & Lai, R. (2016, November 8). Election 2016: Exit Polls. The New York Times. Retrieved from https://www.nytimes.com/ interactive/2016/11/08/us/politics/election-exit-polls.html Huang, Y., Davies, P. G., Sibley, C. G., & Osborne, D. (2016). Benevolent sexism, attitudes toward motherhood, and reproductive rights: A multi-study longitudinal examination of abortion attitudes. Personality and Social

Psychology Bulletin, 42, 970–984. http://dx.doi.org/10.1177/0146167216649607 Lee, T. L., Fiske, S. T., Glick, P., & Chen, Z. (2010). Ambivalent sexism in close relationships: (Hostile) power and (benevolent) romance shape relationship ideals. Sex Roles, 62, 583–601. https://doi.org/10.1007/s11199-010-9770-x Maltby, L. E., Hall, M. E. L., Anderson, T. L., & Edwards, K. (2010). Religion and sexism: The moderating role of participant gender. Sex Roles, 62, 615–622. http://dx.doi.org/10.1007/s11199-010-9754-x Mikolajczak, M., & Pietrzak, J. (2014). Ambivalent sexism and religion: Connected through values. Sex Roles, 70, 387–399. https://doi.org/10.1007/s11199-014-0379-3 Moore, S. (2016, Nov. 16). Why did women vote for Trump? Because misogyny is not a male-only attribute. The Guardian. Retrieved from https://www. theguardian.com/lifeandstyle/commentisfree/2016/nov/16/why-didwomen-vote-for-trump-because-misogyny-is-not-a-male-only-attribute Overall, N. C., Sibley, C. G., & Tan, R. (2011). The cost and benefits of sexism: Resistance to influence during relationship conflict. Journal of Personality and Social Psychology, 101, 271–290. https://doi.org/10.1037/a0022727 Piggot, M. (2004). Double jeopardy: Lesbians and the legacy of multiple stigmatized identities (Unpublished undergraduate thesis). Swinburne University of Technology, Australia. Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401 http://dx.doi.org/10.1177/014662167700100306 Rosenberg, M. (1965). Society and adolescent child. Princeton, NJ: Princeton University Press. Spence, J. T., & Hahn, E. D. (1997). The Attitudes Toward Women scale and attitude change in college students. Psychology of Women Quarterly, 21, 17–34. https://doi.org/10.1111/j.1471-6402.1997.tb00098.x Spence, J. T., Helmreich, R., & Stapp, J. (1973). A short version of the Attitudes toward Women Scale (AWS). Bulletin of the Psychonomic Society, 2, 219–220. http://dx.doi.org/10.3758/BF03329252 Spengler, R., (2014). Evil woman: An examination of internalized misogyny (Doctoral dissertation). Retrieved from ProQuest Dissertations & Theses Full Text. (Order No. 3664321) Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder. Arch Intern Med. 166, 1092–1097. https://doi.org/10.1001/archinte.166.10.1092 Swim, J. K., Hyers, L. L., Coher, L. L., & Ferguson, M. J. (2001). Everyday sexism: Evidence for its incidence, nature, and psychological impact from three daily diary studies. Journal of Social Issues, 57, 31–53. http://dx.doi.org/10.1111/0022-4537.00200 Szymanski, D. M., Gupta, A., Carr, E. R., & Stewart, D. (2009). Internalized misogyny as a moderator of the link between sexist events and women’s psychological distress. Sex Roles, 61, 101–109. https://doi.org/10.1007/s11199-009-9611-y Szymanski, D. M., & Henrichs-Beck, C. (2014). Exploring sexual minority women’s experiences of external and internalized heterosexism and sexism and their links to coping and distress. Sex Roles, 70, 28–42. http://dx.doi.org/10.1007/s11199-013-0329-5 Tasdemir, N., & Sakalli-Ugurlu, N. (2010). The relationships between ambivalent sexism and religiosity among Turkish university students. Sex Roles, 62, 420–426. http://dx.doi.org/10.1007/s11199-009-9693-6 Tinakon, W., & Nahathai, W. (2012). A comparison of reliability and construct validity between the original and revised versions of the Rosenberg selfesteem scale. Psychiatry Investigation, 9, 54–58. https://doi.org/10.4306/pi.2012.9.1.54 Transcript: Donald Trump’s taped comments about women. (2016, October 8). The New York Times. Retrieved from https://www.nytimes.com/2016/10/08/ us/donald-trump-tape-transcript.html Author Note. Adrian J. Dehlin, https://orcid.org/0000-0002-7250-1037, Department of Psychology, Utah State University; Renee V. Galliher, Department of Psychology, Utah State University. This study was supported by the Utah State University Office of Research and Graduate Studies, provided by the Undergraduate Research and Creative Opportunities Grant. Special thanks to Psi Chi Journal reviewers for their support. Correspondence concerning this article should be addressed to Adrian J. Dehlin or Dr. Renee V. Galliher, Department of Psychology, Utah State University, Logan, UT, 84322-2810. E-mail: anna.dehlin@aggiemail.usu.edu

https://doi.org/10.24839/2325-7342.JN24.4.265

Social and Cognitive Effects of Smartphone Use in Face-to-Face Verbal Interactions Timothy J. Johnson, Marygrace Y. Kaiser*, and Alexander B. Swan* Eureka College

ABSTRACT. Smartphone possession is on the rise in the United States, and their presence may have detrimental effects in face-to-face (f2f) interactions. Two potential theories were explored: expectation violation theory, which suggests that smartphone presence in a conversation could lead to negative social evaluations, and cognitive load theory, which suggests that smartphone use during f2f conversations could be a distraction. Participants engaged in a scripted dyadic conversation scenario with a confederate, where the confederate did or did not appear to use a smartphone during the conversation, and/or the participant did or did not use a smartphone for a simultaneous texting task. In each condition, participants were asked to socially evaluate their confederate partner’s conversation behavior and perform a conversation recognition task. As predicted, participants had more negative social evaluations for behaviors of confederate partners using their smartphones (ηp² = .07) and scored lower on the recognition task (ηp² = .22) when engaged in their own smartphone use. Results of this study suggest social and cognitive rationales for not using smartphones in f2f interactions and encourage continued research on the effects smartphones may have in different f2f interactions.

Open Data and Open Materials badges earned for transparent research practices. Data and materials are available at https://osf.io/fd6ue/

Keywords: smartphones, face-to-face, cognitive load, expectation violation

rom 2011 to 2018, the percentage of U.S. adults possessing smartphones has rapidly doubled from 35 to 77%. This number is expected to grow, considering that 94% of individuals aged 18–29 have a smartphone (Pew Research Center, 2018). A 2014 survey of eight European countries suggested that 69% of children use a mobile phone, indicating 10 years old as the modal age for first phone ownership (GSMA, 2014). The impact that the growing rate of mobile phone possession has on relationships and human interactions continues to be explored, with the current focus on smartphone use. Although some *Faculty mentor

consider smartphones as a tool for maintaining relationships (Borae & Peña, 2010; Lundquist, Lefebvre, & Garramone, 2014) by allowing connections with far-away friends or sustained constant communication with others, some suggest that smartphones, or even just mobile phones, might be a detriment to more intimate human interactions (Lundquist et al., 2014) or that they may change the dynamics of face-to-face (f2f) interactions completely (Humphreys, 2005). Many studies on smartphones’ interactions with human relationships have focused on individuals’ beliefs about smartphone use in their relationships

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

265

Smartphones in Face-to-Face Interactions | Johnson, Kaiser, and Swan

as a whole, or through field observations of behaviors when smartphones are present in f2f interactions. These studies have found that people generally think their conversations with others are less satisfying, and that they feel less connected and less close to their conversation partners when a smartphone is merely present in the vicinity (Misra, Chenge, Genevie, & Yuan, 2014; Przybylski & Weinstein, 2013). However, it appears there have been no attempts to measure the actual practical effects smartphones may have in an interpersonal interaction scenario. Social Impact of Smartphones It is possible that the mere existence of negative social beliefs on smartphone presence might have harmful effects on social evaluation in conversation. Individuals may perceive their partners as being less engaged, or their conversation behavior as less desirable, if their partners are using a smartphone. Thus, the mere sight of a conversation partner using a smartphone, whether or not it actually distracts the partner from the conversation, could violate a cultural expectancy of undivided attention in a conversation. Developed by Burgoon (1978) to understand individuals’ reactions to personal space violations, expectation violation theory describes communica tion violations in many different situations (e.g., romantic relationships: Bevan, 2003; education: Frisby & Sidelinger, 2013; first dates: Morr & Mon geau, 2004). Burgoon (1993) defined expectancies as “an enduring pattern of anticipated behavior” specific to a particular individual or relationship (p. 31). An expectancy violation occurs when an individual’s expectancy for anticipated, typical behavior in a particular relationship (e.g., paying for someone’s food on a first date or personal space when talking with a stranger) is violated by the relationship partner’s behavior (e.g., no offer to pay for a meal on a first date, being uncomfortably close when speaking with a stranger). A violation of an individual’s typical expectations consequently leads to more awareness of the violating behavior and a sequence of evaluation where the individual interprets the unexpected behavior (Afifi & Metts, 1998; Bevan, 2003; Burgoon & Walther, 1990). When an individual evaluates a partner’s atypical behavior, Afifi and Metts (1998) suggested that the sequence of evaluation is summarized by three distinct aspects. The first is violation valence, which refers to the desirability (or undesirability) of a violating behavior. Valence refers to whether 266

a behavior is positive or negative (e.g., do I find a stranger’s violation of my personal space as a good thing or a bad thing?). Second, violation expectedness refers to how far a violating behavior differs from what is otherwise normally expected (e.g., how much was I expecting that stranger to enter my personal space?). Last, violation importance refers to the impact a behavior has on the relationship of the people in the interaction (e.g., will I want to interact with that stranger again in the future? What do I think of them as a person?). By applying the principles of expectancy described by Burgoon (1993) to a conversation scenario, there is a common expectancy of undi vided attention when speaking in f2f interactions. Recently, Miller-Ott and Kelly (2015) extended expectation violation theory to describe a violation in expectation caused by mobile phone use during times when undivided attention is expected in romantic relationships, such as when two people are getting acquainted on a first date or are out on a formal date. These particular situations represent moments when undivided attention during verbal communication is desired but violated by the use of a smartphone. A smartphone’s visibility during conversation has been shown to decrease partner ratings of closeness, connection, relationship quality, conversation quality, and satisfaction in f2f conversation settings (Misra et al., 2014; Przybylski & Weinstein, 2013), suggesting that this expectancy violation might have a clear effect on f2f interaction evaluations. An individual appearing to use a phone in an f2f interaction will be evaluated as violating the undivided attention behaviors expected of conversation (violation expectancy) and their deviant behavior is evaluated in a negative light (violation valence). Cognitive Impact of Smartphones Perhaps there is an additional, but separate, cognitive explanation for the social belief that smartphones negatively affect relationships. The use of a phones may overload an individual’s ability to attend and remember details of a conversation, thus making them a negative conversation feature. This would not be surprising, considering that mobile phone use can be a cognitive distraction in a variety of settings (e.g., education: Gingerich & Lineweaver, 2013; driving performance: Horberry, Anderson, Regan, Triggs, & Brown, 2005; walk ing: Hyman, Boss, Wise, Mckenzie, & Caggiano, 2010). Although research on the distracting effects of smartphone use on verbal communication

Johnson, Kaiser, and Swan | Smartphones in Face-to-Face Interactions

specifically has yet to be conducted, to the authors’ knowledge, its capability to distract individuals from a task at hand could implicate similar distractions in verbal communication. Cognitive load theory, which states that work ing memory is limited in the number of stimuli or schemas it can contain simultaneously (Sweller, 1988), might explain possible distracting effects of smartphones. Heavy cognitive load can cause errors or interference with a current cognitive task (Cooper & Sweller, 1987; Sweller, 1988). It is plausible that stimuli present from smartphone use (e.g., texting, skimming articles, mobile games) could place a load on working memory. People who are engaged in the primary task of conversing with another person would have their working memory burdened by tasks or distractions of a simultaneous secondary task of smartphone use (Drago, 2015). Consequently, the increased load on working memory would cause difficulty with cognitive tasks demanded in f2f communication, such as attentive listening (Kirschner, 2002). Current Study In the present study, we explored the negative (rather than positive) beliefs about smartphones. There are two pieces to explore in an ecologically valid laboratory experiment: the effects of the smartphone’s use by a conversation partner and the effects of the smartphone’s use on the person using it. Thus, the investigation was social cognitive in nature. To assess how a person views a conversation partner’s phone use, we investigated whether this violates social expectations (Burgoon, 1978), and in order to assess the impact of the phone use on the person internally, we sought to determine if the phone use is an attentional and memory distraction (e.g., Sweller, 1988). Both social evaluation and working memory overload were investigated as causes of potential smartphone related negative effects in f2f inter actions. Specifically, expectancy violation and cognitive load theories were assessed by measures of partner behavior evaluation and conversation recall, respectively, in a conversation scenario with a confederate where smartphones were present or absent. The hypotheses for the present study are as follows: 1. Behavior expectancy: Participants will view their partners’ behavior as violating typical conversation behavior more if the partners used their phones during the conversation than if not.

2. Behavior desirability: Participants will rate their partners’ conversation behaviors as more desirable if the partner was not using a smartphone. 3. Interaction of expectancy and phone use: If both the confederate and participant were using their phones, it was predicted the participants would rate their partner’s behaviors as violating typical conversation behavior less than if only the confederate used the phone. 4. Interaction of desirability and phone use: Partner behavior will be rated more desirable by the participant if both individuals were using their phone than if only the confederate used it. 5. Cognitive load: Retrieval performance of information from the conversation would be lower when the participants used a smartphone for a texting task than when they did not use it.

Method Participants The participants for this study were from the student population of a small liberal arts college (ages ranged from 18–22 years old). Because the design for this study combined both cognitive and social evaluation measurements into one methodol ogy, the number of participants required for this study was based on effect sizes of previous studies involving expectancy violation. These studies had smaller effect sizes than studies involving cognitive load theory. For a moderate multivariate effect size (f2) of .06, as indicated by Burgoon and Le Poire (1993), with power = .80, a total sample size of 116 participants was needed across all experimental groups. Due to the difficulties of recruitment from such a small population, only 108 participants were originally recruited. Ultimately, 87 participants (44 women and 43 men) completed the study due to the criterion that excluded participants who had prior familiarity with the confederates. Each partici pant was compensated $5 USD. Before conducting this study, Institutional Review Board approval by Eureka College was granted (#2018-01). Design and Materials1 A 2 (confederate phone use) x 2 (participant phone use) between-subjects design was used in order to determine if the two independent variables of confederate phone use and participant phone use demonstrated a change in the dependent variables of behavior expectancy and behavior desirability (measuring social evaluation) and conversation For the full set of materials, please see our open materials and data: https://osf.io/fd6ue/ 1

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

267

Smartphones in Face-to-Face Interactions | Johnson, Kaiser, and Swan

retention (measuring the effects of cognitive load). Participants engaged in a partially scripted conversation with a same-gender confederate in one of four dyadic conditions: (a) the participant conversed with a confederate with both phones absent; (b) the participant conversed with a confederate who appeared to be using a phone, while the participant’s phone was absent; (c) the participant used a phone to text the researcher and conversed with a confederate who was not using a phone; (d) the participant used a phone to text the researcher and conversed with a confederate who also appeared to use a phone. The use of a confederate as the conversation partner was deliberate. This allowed us to isolate a single perspective in a conversation dyad to reduce noise and confounds. One major issue in previous partner f2f interaction studies (Misra et al., 2014; Przybylski & Weinstein, 2013) is that the participants selected their partners and likely knew each other before the study. Not only might this have caused a bias in partner ratings, but also it would have led to more variation due to the differ ent partners. Additionally, participants were only selected to participate if they did not have a prior relationship with the confederate (to reduce bias). Another difference of this study compared to previ ous methodologies (Misra et al., 2014; Przybylski & Weinstein, 2013) is that we introduced deception to the actual purpose of the study prior to their interaction with their conversation partner. This approach rules out the possibility that the hypoth esized negative effects would bias participants. Social evaluation. Participants completed two 5-question, 7-point Likert-type scale questionnaires. These scales were adapted from Burgoon and Le Poire’s (1993) measures of violation expectancy and violation valence. The first measured the extent to which expectations of normal conversation were violated by partner behavior (Expectancy). For example, participants were asked whether they agree or disagree with a statement such as “My con versation partner behaved appropriately during our conversation.” The second questionnaire asked for the participant’s evaluation of the desirability of the confederate’s conversation behavior (Desirability). For example, participants were asked to agree or disagree with statements such as “My conversation partner interacted in a way that most people would find enjoyable.” Cognitive load. To test cognitive load effects, a basic adaption to existing task-switching methodolo gies (e.g., Chandler & Sweller, 1996; Sweller 1998) 268

was adopted. The task-switching method measures the cognitive load induced by a secondary task by measuring performance on a memory test related to a primary task. Participants were assigned the primary task of conversing with the confederate in an icebreaker question exercise. Both the participant and confederate were asked to “try to listen to and retain” their partner’s answers to the questions. In the two participant phone use condi tions, participants were asked to keep their phone out and at strategic intervals, the researcher texted the participants additional personal questions to answer, representing the secondary task. The effect that this secondary task had on cognitive load, and consequently listening attention, was measured by recognition accuracy scores on a post conversation multiple choice test. Participants answered several questions about the confederate that were men tioned in the icebreaker activity (e.g., confederate’s hometown, leisure activities). Their raw score out of the total questions was used in all analyses. Procedure Participants arrived at the study location and were told that they would be participating in a research study on the effectiveness of icebreaker questions with a stranger. The cover story included a situation where the participants discussed five icebreaker questions with a partner. This partner was believed to be a second participant, but it was a confederate working on behalf of the researcher. Before the icebreaker activity, participants were told to either place their smartphones in a box outside the room or to use their smartphones in a text messaging task (these were the two conditions of the participant phone use variable). During the icebreaker session, the participant and confederate took turns answer ing the questions. The researcher was in a separate room during this time. The confederate answered the questions according to a pre-established script. The icebreaker session lasted approximately 5–10 minutes. Confederates were also instructed before each study (separately from the participant) whether to appear to use their phone or to keep it away when the participant was answering questions (these were the two conditions of the confederate phone use variable). Participants instructed to use their phones received unrelated background questions from the researcher. An example of this kind of questioning was “How many people would you consider to be in your closest friendship circle, how many in your acquaintance circle, and how many in your enemy

Johnson, Kaiser, and Swan | Smartphones in Face-to-Face Interactions

circle?” They were instructed that these messages should be read and replied to while their partner was answering questions. There were five separate text message questions in total, and they were sent and received strategically when the participant was answering an icebreaker question. This was done so participants would attend to the text and their partner simultaneously. Thus, this created two com peting question-and-answer tasks at specific times during the icebreaker scenario. In the confederate phone use condition, the confederate appeared to be doing a similar texting task (in actuality, they texted the researcher random information). After the icebreaker conversation ended, the researcher directed the participant to a different room. Participants completed the social evaluation scales and then completed the multiple choice test. Participants were debriefed, compensated, and excused. After data collection was completed on campus, all participants received an email fully debriefing them on the purposes of the study and rationale for the use of deception.

Results Eighty-seven participants initially completed the study. Eight participants were removed because they did not use their phone as instructed. After removal, 79 participants remained in the analysis. Thirty-six of the remaining participants were instructed to use their phones during the study and 43 were not. Forty-one participants interacted with a confederate who appeared to be using a phone, and 38 participants interacted with a confederate who did not appear to use it. Three separate 2 x 2 Analyses of Variance were conducted to determine if the phone usage of the participant and confeder ate produced differences on measures of social evaluation and cognitive load. Social Evaluation Analyses To measure social evaluation, the behavior expec tancy and behavior desirability scales were analyzed by taking scale averages. Means and standard deviations are available in Table 1. The first scale measured the confederate’s conformity to the participant’s expectations for typical conversation behavior. The possibility for a participant’s average questionnaire score ranged from 1 (highest violation of expectations) to 7 (expected conversation behavior). Overall, participants rated their confederate partners as highly conforming to expected con versation behavior across conditions. The second scale measured the participants’ evaluation of the

desirability of their confederate partners’ conversa tion behavior. The possibility for a participant’s average questionnaire score ranged from 1 (least desirable) to 7 (most desirable). Overall, participants rated their confederate partners’ conversation behaviors as highly desirable across conditions. Hypothesis 1 was supported. A main effect for confederate phone presence on expectancy was found, F(1, 75) = 5.25, p = .03, ηp² = .07. As illustrated in Figure 1, participants rated their confederate partner lower on expected conversation behavior when the confederates appeared to be using their phone (M = 6.13, SD = 0.93) than when they did not use their phones (M = 6.61, SD = 0.93). Hypothesis 2 was supported. A main effect for confederate phone use on behavior desirability was observed, F(1, 75) = 6.21, p = .02, ηp² = .08. As illustrated in Figure 2, participants rated their confederate partners’ conversation behavior as more desirable when the confederates were not using their phones (M = 6.51, SD = 0.93) than when they appeared to be using their phones (M = 5.99, SD = 0.93). Hypotheses 3 and 4 were not supported. No interactions were observed between the variables of confederate and participant phone use on measures of expectation violations. Cognitive Load Analyses To measure the effects of phone use on cognitive load, participant scores on a 21-question conversa tion-recall questionnaire were analyzed. Although the questionnaire originally contained 27 questions, six questions were removed from the analysis due to confederate script errors. Participants received one point for each correct answer on the questionnaire and received an overall score from 0 (none correct) to 21 (all correct). Overall, participants answered approximately 76% of the questionnaire correctly across conditions. Means and standard deviations are available in Table 1. Hypothesis 5 was supported. Scores on the recall questionnaire were significantly lower when the participants were using their phones (M = 14.14, SD = 2.76) than when they were not (M = 16.93, SD = 2.75), F(1, 75) = 19.86, p < .001, ηp² = .21 (see Figure 3 for overall effect).

Discussion The findings of the present study indicate that smartphones indeed may have negative effects on both social evaluation and attention in dyadic f2f interactions. The behavior of an individual who

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

269

Smartphones in Face-to-Face Interactions | Johnson, Kaiser, and Swan

appears to use a smartphone when conversing is rated as less desirable and more in violation of normal conversation behavior expectations than the behavior of someone not using a phone dur ing a conversation. Additionally, individuals who use a smartphone to text in an f2f conversation TABLE 1 Means (Standard Deviation) for Behavior Expectancy, Behavior Desirability, and Recognition Test Score for Each Phone Condition Behavior Expectancy

Behavior Desirability

Recognition Score

Phone Condition

M(SD)

No phones used (n =21)

6.54(0.50)

6.54(0.61)

16.24(2.76)

Confederate phone use; no participant phone use (n = 22)

5.93(1.23)

5.77(1.29)

17.59(2.24)

No confederate phone use; participant phone use (n = 17)

6.68(0.48)

6.48(0.55)

14.12(3.08)

Both use phones (n = 19)

6.34(1.16)

6.21(0.97)

14.16(2.99)

Total sample (N = 79)

6.35(0.95)

6.24(0.96)

15.66(3.08)

FIGURE 1 Participant Phone Yes

Participant Phone No

Behavior Expectancy Rating

7 6 5 4 3 2 1

Yes

Confederate Phone Use Figure 1. The average ratings that participants gave regarding the expectancy of their confederate partner’s behavior depending on confederate/participant phone use. There is a main effect of confederate phone use. Error bars represent standard errors.

FIGURE 2 Participant Phone Yes

Participant Phone No

Behavior Desirability Rating

7 6 5 4 3 2 1

Yes

Confederate Phone Use Figure 2. The average evaluation ratings participants gave regarding their confederate partners depending on confederate/participant phone use. There is a main effect for confederate phone use. Error bars represent standard errors.

270

remember less from a conversation than those who do not use a phone. Practically, these findings build on previous research to suggest a strong argument for putting down smartphones when interacting with other individuals in an f2f context. The social findings using the expectation violation theory complement previous studies (Misra et al., 2014; Przybylski & Weinstein, 2013), which focused mostly on partner evaluations rather than partner behavior evalua tions. This study demonstrated that the behavior of phone use itself is commonly viewed as undesirable and a violation of expected conversation behaviors. This socially unexpected and undesirable behavior may be the root of lower ratings of partner close ness, connection, relationship quality, conversa tion quality, and satisfaction found in previous smartphone-influenced f2f interaction studies (Misra et al., 2014; Przybylski & Weinstein, 2013). Additionally, this study successfully extended previous findings of mobile phones’ distracting power in many contexts (Gingerich & Lineweaver, 2013; Horberry et al., 2005; Hyman et al., 2010) to the particular situation of f2f conversations. Phones in a conversation are distracting and compete for the limited attentional and memory resources needed to be a good conversation partner. This represents a step forward in the ecological validity of previous findings to the particular context of conversation and human relationships. One area where the hypothesis for this study was not supported was the interaction between participant phone use and confederate phone use on social evaluation of confederate behaviors. It was expected that participants would rate their partners’ behavior more negatively when only the partner used a phone, rather than when both individuals used phones. This prediction was based on expectation violation theory and assumed that the confederate’s use of the phone would not violate expectations if both individuals had been instructed to use their phone by the researcher. The lack of an interaction actually builds upon the argument against using smartphones in f2f interactions. Essentially, the findings suggest that smartphone use in f2f conversations is unexpected and undesirable, possibly even if both individuals are using their phone. These results point toward a further examination of the social beliefs surround ing smartphone use in f2f interactions. One interesting, unanticipated finding (albeit anecdotal) was that participants and confederates were both observably reluctant to use their phones

Johnson, Kaiser, and Swan | Smartphones in Face-to-Face Interactions

Limitations These findings should not be over-generalized We thank an anonymous reviewer for this analysis suggestion. 2

FIGURE 3 20

Conversation Recall Score (Adj)

during the icebreaker scenario. Both confederates expressed to the researcher that the particular condition where they had to use their phones, but not the participant, was their “least favorite.” One confederate consistently groaned whenever he was told a data collection session would be for that condition. Post-hoc contrast analyses were con ducted to investigate this anecdotal interpretation.2 On the expectancy measures, behavior violation and behavior desirability, the cell mean for the confederate phone use only was compared to the mean of the other three cells in a simple contrast test. For behavior violation, participants rated confederates significantly lower in the confederate phone use condition than the average of the other three conditions, (5.93 vs. 6.51) t(21) = -2.27, p = .04, d = .48. Additionally, for behavior desirability, participants rated confederates significantly lower in that condition than the average of the other three conditions, (5.77 vs. 6.41) t(21) = -2.32, p = .03, d = .50. These contrasts suggest that phone use was the most salient when the participant was not allowed to use their phone at the directive of the researcher but the confederate had no similar sanctions. Similarly, several participants were excluded from analyses because they neglected to use their phone in the conversation when instructed. Although they listened to the researcher and expressed an understanding of the directions, many participants did not use their phones for the texting task unless prompted by the confederate to be at ease with remarks like: “Did he tell you to respond to questions or something? I honestly don’t mind if you do—I know you’re listening.” Some, even after prompting by both the researcher and confederate, refused to use their phones. In debriefing sessions, participants were probed informally; some gave answers such as “I don’t like it when people use their phone when talking to me, so I didn’t” or “It just seemed too rude.” Some participants also verbally expressed their difficulty in the recognition task because “it was hard to focus on the texts and listen at the same time.” These anecdotal observa tions built on the data to suggest that participants and confederates both were aware of social expecta tions and possible cognitive ramifications regarding smartphones in f2f interactions.

Confederate Phone Yes

Confederate Phone No

18 16 14 12 10 8 6 4 2 0

Yes

Participant Phone Use Figure 3. The overall scores of participants on the conversation recall task depending on confederate/participant phone use. There is a main effect for participant phone use. Error bars represent standard errors.

to phones in all relationships and interactions, because it merely represents three carefully selected cognitive and social effects that phones have in the specific instance of a dyadic, same-gender, f2f, casual verbal interaction between two strangers. The methodology also limits generalization to dif ferent uses of smartphones in conversation because it only examined the specific texting activity of individuals responding to the researcher’s ques tions. It is possible that there could be differences in behavior and evaluation if the participants were having a voluntary text conversation with another individual, or using their phone in some other manner. Furthermore, participants for this study were all undergraduate college students from a small liberal arts college. We caution that the findings from such a specific population should not be generalized to all individuals of all ages. One dif ference between the college population sampled and the population at large that is significant for this particular study is age. Because the age group sampled (roughly 18–22 years) have likely grown up using smartphone technology, it is possible that we might observe different social and cognitive effects in f2f interactions between older individu als. Another difficulty with sampling on the small campus was a lack of an ability for random or strati fied sampling. Instead, participants were selected via convenience sampling in different classrooms and common areas on campus. It is possible that a more outgoing, engaged sample of the student population selected in this way would likely differ from a random sample. There were a few inconsistencies with the methodology of the study that might have added minor variation to the results. These include:

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

271

Smartphones in Face-to-Face Interactions | Johnson, Kaiser, and Swan

researcher script/procedure errors (e.g., asking to place phone away at different times or confederate arriving after participant), confederate behavior variance (e.g., accidentally forgetting something on the script, transitioning between questions differently, different clothing/appearance, or dif ferences between style of speech between the two confederates), and participant phone use variance (e.g., some participants only used their phones halfway through, or used them to do other things than instructed). Although these methodological inconsistencies slightly reduced the internal validity of the study, the significant results found despite them are worth exploring. Real-life conversations are not predictable and include many minor varia tions between individuals. It is also important to take into account a possible ceiling effect in the two social evaluation variables. Across conditions, participants rated their confederate partners’ behavior as conforming to conversation expectations and desirable. Although there was a significant difference in these variables when the confederates had a phone than when they did not, in both cases their behavior was rated highly. It can be argued that this ceiling effect is due to smartphones having an unexaggerated effect, but it is more likely due to the familiarity effect. Moreland and Zajonc (1980) found that a more familiar face is rated as more desirable and similar to an individual than an unfamiliar one. Being on a campus of fewer than 600 individuals, most participants would have interacted with the confederate at some point outside of the study, and crossed paths somewhat often. It is very likely that the familiarity effect would have an important role to play in how typical and desirable the participants rated their partners’ behavior. Future Research This data on smartphones’ effects on an individual’s cognitive load and social evaluation should establish a basis for more research on how phone use might measurably affect cognitive conversational abilities, social expectations, and behavior evaluations in different instances. It would be helpful to build on this basis by incorporating a measure of the effects of smartphone use on the social evaluation of the partners themselves, rather than simply their behaviors. This could demonstrate that the negative evaluation of smartphone use behaviors has implications for personal social desirability. Future studies could assess similar effects in mixed-gender pairs, parent-child interactions, small 272

groups, among close friends, and in romantic inter actions. It is possible that different relationships and contexts beyond stranger-stranger interactions and icebreaker interactions might lead to different findings. Conclusion Although the limitations for the study should qualify these conclusions somewhat, they certainly form a basis and set a direction for future research on the effects of smartphones in f2f interactions. This study successfully demonstrated three significant, ecologi cally valid, negative effects that smartphone use in same-gender, dyadic, f2f interactions may cause on a college campus population. First, an individual using a smartphone in an interaction may be viewed as violating expectations for typical conversation behavior more than an individual who does not. Second, the smartphone-using individual’s behavior may be viewed as less desirable than someone not using a phone. Third, individuals who use their smartphones while listening will be often less capable of attending and remembering details from a conversation than others who do not use it. These results provide some rationale to rethink smartphone use in common f2f interactions.

References Afifi, W. A., & Metts, S. (1998). Characteristics and consequences of expectation violations in close relationships. Journal of Social and Personal Relationships, 15, 365–392. https://doi.org/10.1177/0265407598153004 Bevan, J. L. (2003). Expectancy violation theory and sexual resistance in close, cross-sex relationships. Communication Monographs, 70, 68–82. https://doi.org/10.1080/715114662 Borae, J., & Peña, J. F. (2010). Mobile communication in romantic relationships: Mobile phone use, relational uncertainty, love, commitment, and attachment styles. Communication Reports, 23, 39–51. https://doi.org/10.1080/08934211003598742 Burgoon, J. K. (1978). A communication model of personal space violations: Explication and an initial test. Human Communication Research, 4, 129–142. https://doi.org/10.1111/j.1468-2958.1978.tb00603.x Burgoon, J. K. (1993). Interpersonal expectations, expectation violations, and emotional communication. Journal of Language and Social Psychology, 12, 30–48. https://doi.org/10.1177/0261927X93121003 Burgoon, J. K., & Le Poire, B. A. (1993). Effects of communication expectancies, actual communication, and expectancy disconfirmation on evaluations of communicators and their communication behavior. Human Communication Research, 20, 67–96. https://doi.org/10.1111/j.1468-2958.1993.tb00316.x Burgoon, J. K., & Walther, J. B. (1990). Nonverbal expectancies and the evaluative consequences of violations. Human Communication Research, 17, 232–265. https://doi.org/10.1111/j.1468-2958.1990.tb00232.x Chandler, P. & Sweller, J. (1996). Cognitive load while learning to use a computer program. Applied Cognitive Psychology, 10, 151–170. https://doi. org/10.1002/(SICI)1099-0720(199604)10:2<151::AID-ACP380>3.0.CO;2-U Cooper, G., & Sweller, J. (1987). Effects of schema acquisition and rule automation on mathematical problem-solving transfer. Journal of Educational Psychology, 79, 347–362. http://dx.doi.org/10.1037/0022-0663.79.4.347 Drago, E. (2015). The effect of technology on face-to-face communication. Elon Journal of Undergraduate Research in Communications, 6, 2–21. Retrieved from http://www.inquiriesjournal.com/a?id=1137 Frisby, B. J., & Sidelinger, R. J. (2013). Violating student expectations:

Johnson, Kaiser, and Swan | Smartphones in Face-to-Face Interactions

Student disclosures and student reactions in the college classroom. Communication Studies, 64, 241–258. https://doi.org/10.1080/10510974.2012.755636 Gingerich, A. C., & Lineweaver, T. T. (2014). OMG! Texting in class = U fail :( Empirical evidence that text messaging during class disrupts comprehension. Teaching of Psychology, 41, 44–51. https://doi.org/10.1177/0098628313514177 GSMA. (2014). Children’s use of mobile phones: A special report. Retrieved from https://www.gsma.com/publicpolicy/wp-content/uploads/2012/03/ GSMA_Childrens_use_of_mobile_phones_2014.pdf Horberry, T., Anderson, J., Regan, M. A., Triggs, T. J., & Brown, J. (2005). Driver distraction: The effects of concurrent in-vehicle tasks, road environment complexity and age on driving performance. Accident Analysis and Prevention, 38, 185–191 https://doi.org/10.1016/j.aap.2005.09.007 Humphreys, L. (2005). Cellphones in public: Social interactions in a wireless era. New Media Society, 7, 810–833. https://doi.org/10.1177/1461444805058164 Hyman I., Boss, S. M., Wise, B., Mckenzie, K., Caggiano, J. (2010). Did you see the unicycling clown? Intattentional blindness while walking and talking on a cell phone. Applied Cognitive Psychology, 24, 597–607. https://doi.org/10.1002/acp.1638 Kirschner, P. A. (2002). Cognitive load theory: Implications of cognitive load theory on the design of learning. Learning and Instruction, 12, 1–10. https://doi.org/10.1016/S0959-4752(01)00014-7 Lundquist, A. R., Lefebvre, E. J., & Garramone, S. J. (2014). Smartphones: Fulfilling the need for immediacy in everyday life, but at what cost? International Journal of Humanities and Social Science, 4, 80–89. Miller-Ott, A., & Kelly, L. (2015). The presence of cell phones in romantic partner face-to-face interactions: An expectancy violation theory approach. Southern Communication Journal, 80, 253–270. https://doi.org/10.1080/1041794X.2015.1055371 Misra, S., Cheng, L., Genevie, J., & Yuan, M. (2014). The iPhone effect: The quality of in-person social interaction in the presence of mobile devices. Environment and Behavior, 48(2), 1–24. https://doi.org/10.1177/0013916514539755 Moreland, R. L., & Zajonc, R. B. (1980). Exposure effects in person perception: Familiarity, similarity, and attraction. Journal of Experimental Social

Psychology, 18, 395–415. https://doi.org/10.1016/0022-1031(82)90062-2 Morr, M. C., & Mongeau, P. A. (2004). First date expectations: The impact of sex of initiator, alcohol consumption, and relationship type. Communication Research, 31, 3–35. https://doi.org/10.1177/0093650203260202 Pew Research Center. (2018). Mobile Fact Sheet. Retrieved from http://pewinternet.org/fact-sheet/mobile/ Przybylski, A. K., & Weinstein, N. (2012). Can you connect with me now? How the presence of mobile communication technology influences face-to-face conversation quality. Journal of Social and Personal Relationships, 30, 237–246. https://doi.org/10.1177/0265407512453827 Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12, 257–285. https://doi.org/10.1016/0364-0213(88)90023-7 Author Note. Timothy J. Johnson, Marygrace Y. Kaiser, and Alexander B. Swan, https://orcid.org/0000-0001-8171-6534, Social Science and Business Division, Eureka College. This research was supported by the 2018 Spring Research Grant provided by the Psi Chi, the International Honor Society in Psychology. The authors would like to acknowledge Haley Joseph and Samuel Monroe for their assistance during data collection, and Prabhu Venkataraman, Ann Fulop, and several anonymous reviewers for their helpful comments and suggestions throughout the idea formation, data collection process, and earlier versions of this manuscript. Correspondence concerning this article should be addressed to Alexander B. Swan, Social Science and Business Division, 300 E College Ave, Eureka College, Eureka, IL, 61530. E-mail: aswan@eureka.edu.

273

https://doi.org/10.24839/2325-7342.JN24.4.274

Helicopter Parenting and Emotion Regulation in U.S. College Students Susan J. Wenze* Lafayette College

, Anna B. Pohoryles

, and Jennifer M. DeCicco*

ABSTRACT. Most research has suggested that helicopter parenting is associated with negative outcomes. Few studies have explored underlying mechanisms. The present study examined the mediating role of emotional processing. Participants (n = 104 U.S. college students) completed measures of helicopter parenting behaviors, emotional processing, depression, and anxiety. Relationships between helicopter parenting and depression (95% BCa CI = .01 to .33; R2 = .23, f2 = .30) and helicopter parenting and anxiety (95% BCa CI = .01 to .13; R2 = .14, f2 = .17) were mediated by experiential avoidance. The relationship between autonomy support and depression was mediated by expressive suppression (95% BCa CI = -.27 to -.01; R2 = .08, f2 = .08), cognitive reappraisal (95% BCa CI = -.36 to -.03; R2 = .09, f2 = .10), psychological flexibility (95% BCa CI = -.59 to -.04; R2 = .46, f2 = .86), and experiential avoidance (95% BCa CI = -.39 to -.03; R2 = .24, f2 = .31). The relationship between autonomy support and anxiety was mediated by cognitive reappraisal (95% BCa CI = -.17 to -.02; R2 = .09, f2 = .10), psychological flexibility (95% BCa CI = -.28 to -.02; R2 = .49, f2 = .97), and experiential avoidance (95% BCa CI = -.16 to -.01; R2 = .14, f2 = .17). This was one of the first studies to identify mechanisms underlying links between helicopter parenting and negative outcomes. Results have implications for parent education and psychotherapy with college students. Keywords: helicopter parenting, college students, depression, anxiety, emotional processing

H WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

274

elicopter parenting refers to a distinct form of parental control (Padilla-Walker & Nelson, 2012) best characterized by a tendency to hover over one’s child(ren), resolve potential problems for them, and rescue them from difficulties and challenges in a developmentally inappropriate way (Cline & Fay, 1990). Helicopter parents generally have benevolent intentions. As such, they have a parenting style that is typically (but not always) high on warmth and support but which is also universally high on control (Nelson, PadillaWalker, & Nielson, 2015) and low on granting autonomy (Padilla-Walker & Nelson, 2012). In other words, helicopter parents not only hover but also do not support (or even actively discourage) autonomous behaviors in their offspring. Vinson

(2013) argued that American ideas about what constitutes “good parenting” might have shifted in recent decades such that hovering and discouraging autonomy are now the accepted norms, and parents who resist this trend are sometimes judged negatively by other parents. Indeed, helicopter parenting behaviors have increased in the United States in recent years due to a variety of factors including increased safety concerns, ubiquitous technology permitting constant contact with one’s children, greater economic insecurity, and rising costs of college tuition (Vinson, 2013). Although some have argued that helicopter parenting can begin during pregnancy and con tinue through graduate school and beyond (Vinson, 2013), academic research on the prevalence and

*Faculty mentor

Wenze, Pohoryles, and DeCicco | Helicopter Parenting

correlates of helicopter parenting has largely focused on traditional (i.e., 18 to 22-year-old) college students. This may in part be because of the unique problems that this parenting style can pose for young adults. That is, the major develop mental tasks of this period (i.e., exploring identity, developing a sense of independence, assuming adult responsibilities; Tanner, 2006) may become complicated if one’s parents seek to maintain control and do not promote autonomy. Further, compared with their noncollege-enrolled peers, young adults who are attending college may be at elevated risk for helicopter parenting for financial or logistical reasons (e.g., parents are contributing to college tuition, students live at home or return home during summer or winter breaks).1 Some studies have identified adaptive cor relates of helicopter parenting for college students under certain conditions. For example, Nelson et al. (2015) found that, when parents also showed high warmth, helicopter parenting was associated with fewer risk-taking behaviors (e.g., marijuana use, recreational prescription medication use, shoplifting) in undergraduate students. Other stud ies have found that specific aspects of helicopter parenting may be associated with positive outcomes. For example, Luebbe et al. (2018) reported that, when other facets of helicopter parenting (i.e., academic and personal management, direct inter ventions, and autonomy-limiting behaviors) were not present, parents’ information-seeking behaviors (e.g., keeping tabs on one’s children and helping them make decisions) were associated with better decision-making and better academic functioning in college student offspring. However, most published research has sug gested largely maladaptive correlates of helicopter parenting. For example, helicopter parenting has been associated with lower life satisfaction (Schiffrin et al., 2014), worse physical health (Reed, Duncan, Lucier-Greer, Fixelle, & Ferraro, 2016), a sense of entitlement (Segrin, Woszidlo, Givertz, Bauer, & Murphy, 2012), maladaptive academic motivations (Schiffrin & Liss, 2017), low self-efficacy, and poor peer relationships (van Ingen et al., 2015) in college students. It is interesting to note that the adaptive outcomes that have been linked with helicopter parenting are mostly behavioral in nature, whereas the maladaptive outcomes are largely psychologi cal. Although one cannot assume cause-and-effect relationships from correlational studies, perhaps In this manuscript, we use “college student” to signify traditional age, residential college students. 1

there are specific conditions under which helicop ter parenting, information-seeking in particular, fosters positive behaviors in college students, while simultaneously creating longer term psychological challenges. One area that has received less research atten tion (but where the literature is perhaps most consistent) concerns links between helicopter par enting behaviors and poor mental health outcomes in college students. College students who report that their parents engage in high levels of helicopter parenting endorse more symptoms of generalized anxiety (Darlow, Norvilitis, & Schuetze, 2017; Reed et al., 2016), social anxiety (Kouros, Pruitt, Ekas, Kiriaki, & Sunderland, 2017), depression (Darlow et al., 2017; Reilly & Semkovska, 2018; Schiffrin et al., 2014), as well as greater psychotropic medica tion use for anxiety and depression (LeMoyne & Buchanan, 2011). However, only three published studies have examined potential mediators of these associations, meaning that the mechanisms by which helicopter parenting may be associated with poor mental health outcomes for college students have not been largely explored. Reed et al. (2016) found that helicopter parenting and autonomy support have indirect relationships with anxiety and depression through lower self-efficacy. Schiffrin et al. (2014) reported that relationships between helicopter parenting/autonomy support and depression were mediated by college students’ perceived violation of their need for autonomy and competence. Finally, Reilly and Semkovska (2018) found that the relationship between perceived helicopter parenting and depression was explained by decreased resilience. Lower self-efficacy, poor resilience, and thwarted psychological needs (e.g., need for autonomy, competence, independence) are prob ably not the only mechanisms that help explain relationships between helicopter parenting behav iors and negative mental health outcomes in college students. In particular, when considering depres sion and anxiety, one might expect that difficulties in processing and regulating one’s emotions are relevant; such difficulties are fundamental to the development and maintenance of internalizing disorders. Emotion regulation has been defined as, “the extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions . . . to accomplish one’s goal” (Thompson, 1994, pp. 27–28). Maladaptive emotion regulation strategies have robust associations with depression and anxiety. For example, lower use of cognitive

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

275

Helicopter Parenting | Wenze, Pohoryles, and DeCicco

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

276

reappraisal (i.e., re-interpreting a potentially emotionally challenging event in a way that alters the emotional impact; Gross & John, 2003) and higher use of expressive suppression (i.e., inhibiting expressions of emotion; Gross & John, 2003) have been linked with lower positive emotion, lower well-being, and higher negative emotion (Gross & John, 2003), and with more symptoms of depression and anxiety (Dryman & Heimberg, 2018; Loevaas et al., 2018). Psychological flexibility, “the ability to contact the present moment more fully as a conscious human being, and to change or persist in behavior when doing so serves valued ends” (Hayes, Luoma, Bond, Masuda, & Lillis, 2006, p. 7), is a related construct that has also been strongly linked with depression and anxiety. Those who score lower on measures of psychological flexibility show greater symptoms of depression and anxiety (e.g., Roemer, Salters, Raffa, & Orsillo, 2005). Experi ential avoidance (i.e., rigid attempts to alter the form, frequency, or intensity of unwanted internal experiences; Hayes, Strosahl, & Wilson, 1999) is one aspect of psychological inflexibility that may play a particularly central role in the development and maintenance of depressive and anxiety disorders. For example, trait-level experiential avoidance prospectively predicts diagnoses of major depres sion, persistent depressive disorder, and generalized anxiety disorder (Shallcross, Troy, Boland, & Mauss, 2010; Spinhoven, Drost, Rooij, van Hemert, & Penninx, 2014), and reducing experiential avoid ance can lead to improvement in depressive symptoms (Berking, Neacsiu, Comtois, & Linehan, 2009). Importantly, because of the conceptual similarities between emotion regulation and psychological flexibility and the fundamental role that both play in the etiology of depression and anxiety, some have proposed a unified treatment for emotional disorders that focuses on increasing cognitive reappraisal, decreasing emotional avoid ance, and facilitating action that is not associated with the dysregulated emotions (i.e., targeting emotion regulation and psychological flexibility as the key treatment components for depression and anxiety; Barlow, Allen, & Choate, 2016). The importance of adaptive emotion regulation skills and psychological flexibility for reducing risk of depression and anxiety is clear. A parenting style that is over-controlling and discourages autonomy (i.e., high on helicopter parenting behaviors) might prevent offspring from encountering and success fully coping with adversity, thereby robbing the

child of experiences that foster healthy emotionprocessing tendencies. Indeed, some have argued that parental over-involvement is driven by a desire to prevent child distress (Creswell, O’Connor, & Brewin, 2008) but that it has the effect of hinder ing development of coping mechanisms, leaving offspring feeling anxious, less competent, and more vulnerable to stress (Bronson & Merryman, 2009; Creswell et al., 2008; Gibbs, 2009; Hofer & Moore, 2010; Marano, 2008). Thus, the relationship between helicopter parenting behaviors and poor mental health outcomes in college students may be mediated by poor emotion processing skills. In the current study, we aimed to add to the small but growing body of literature on the mecha nisms by which helicopter parenting behaviors are associated with negative mental health outcomes in college students by examining the potential mediat ing role of emotional processing. Specifically, we tested the mediating role of expressive suppression, cognitive reappraisal, psychological flexibility, and experiential avoidance on the relationships between helicopter parenting/autonomy support and depression/anxiety symptoms. We expected that the (negative) relationships between autonomy support and anxiety/depression would be mediated by better emotion regulation (i.e., lower expressive suppression, lower experiential avoidance, higher cognitive reappraisal, higher psychological flex ibility). In contrast, we expected that the (positive) relationships between helicopter parenting and anxiety/depression would be mediated by worse emotion regulation (i.e., higher expressive suppres sion, higher experiential avoidance, lower cognitive reappraisal, lower psychological flexibility).

Method Participants Participants were 104 undergraduate students at Lafayette College, a private, mid-Atlantic, small American liberal arts college, including 81 (77.88%) women and 23 (22.12%) men. No par ticipants identified as gender nonbinary. The most commonly identified major course of study was psychology (n = 28, 26.92% of the sample). Thirtyeight participants (36.54%) were first-year students, 32 (30.77%) were sophomores, 23 (22.12%) were juniors, 10 (9.62%) were seniors, and 1 (0.96%) was beyond their fourth (i.e., senior) year but had not yet graduated. The average age was 19.15 (SD = 1.12) years, and the average self-reported GPA was 3.36 (SD = 0.41) on a 4-point scale. Racial identification was 64.42% White, 21.15%

Wenze, Pohoryles, and DeCicco | Helicopter Parenting

Asian/Asian American, 9.62% Black or African American, and 2.88% biracial or another race. Just over 10% (10.58%) of the sample identified as Hispanic. Two participants did not report their racial identity and 2 participants did not report their ethnic identity. Procedure The current study reflects secondary analyses of an earlier dataset, collected to examine questions unrelated to the present work (Wenze, Gaugler, Sheets, & DeCicco, 2018). None of these data have been previously published. Sample size was deter mined by the previous study but guidelines suggest 115 participants would be needed to achieve .80 power for our models, given small-to-medium-sized a paths and medium-sized b paths, while 53 would be needed for medium and large paths, respectively (Fritz & MacKinnon, 2007). No data were excluded from analyses. There were 5 missing data points, for which we substituted the average item score for the relevant (sub)scale. One participant did not complete the Helicopter Parenting Behaviors measure (HPB). The Lafayette College Institutional Review Board approved all study procedures (proposal # AY1516-66, first author served as study PI). Partici pants were recruited via daily emailed announce ments about a range of campus-wide events, as well as posts to the College’s first-year, sophomore, junior, and senior Facebook pages. The project was described as a study of mood, stress, and coping. All procedures for the current paper were com pleted individually, in person, in paper and pencil format, in the first author’s lab. After the informed consent process, participants completed measures of helicopter parenting, emotion regulation, psychological flexibility, experiential avoidance, depression, and anxiety, as well as other procedures not relevant to this report. Compensation for this part of the study was $10 cash, which participants received after they completed all study procedures. Participants were informed in the study advertise ment and during informed consent procedures that they would be compensated. Measures We used the HPB (Schiffrin et al., 2014), a 15-item self-report scale, to assess perceived helicopter parenting behaviors. Intended respondents are college students, rating their mother’s behaviors. The HPB yields two subscale scores, one measuring helicopter parenting (9 items, 1 = strongly disagree

to 6 = strongly agree) and one assessing autonomy support (6 items, 1 = strongly disagree to 6 = strongly agree). Higher scores reflect higher levels of the measured constructs. Scale developers reported acceptable reliability (e.g., helicopter parenting Cronbach’s α = .77, autonomy support Cronbach’s α = .71) and validity (e.g., helicopter parenting was positively correlated with depression and negatively correlated with life satisfaction, autonomy, and competence; Schiffrin et al., 2014). We used the Emotion Regulation Question naire (ERQ; Gross & John, 2003), a 10-item self-report scale, to assess emotion regulation. The ERQ yields two subscale scores, one measur ing expressive suppression (4 items, 1 = strongly disagree to 7 = strongly agree) and one measuring cognitive reappraisal (6 items, 1 = strongly disagree to 7 = strongly agree). Higher scores reflect higher levels of the measured constructs. Scale developers reported acceptable reliability (expressive sup pression Cronbach’s α = .73, cognitive reappraisal Cronbach’s α = .79) and validity (e.g., reappraisal was positively related to reinterpretation, ß = .43, and negatively related to rumination, ß = -.29; suppression was positively related to rumination, ß = .19, and negatively related to venting, ß = -.43; Gross & John, 2003). We u s e d t h e A c c e p t a n c e a n d A c t i o n Questionnaire – II (AAQ-II; Bond et al., 2011), a 7-item self-report measure (1 = never true to 7 = always true), to assess psychological flexibility. Higher scores reflect higher levels of psychological flexibility. The AAQ-II has good internal consistency (Cronbach’s αs ranged from .86 to .89) and cor relates with the longer scale from which the AAQ-II was derived at .97 (Bond et al., 2011). We used the Brief Experiential Avoidance Questionnaire (BEAQ; Gamez et al., 2014), a 15-item self-report measure (1 = strongly disagree to 6 = strongly agree), to assess experiential avoidance. Higher scores reflect higher levels of experiential avoidance. The BEAQ has good reliability (mean Cronbach’s α across 3 samples = .86) and moderate to high average correlations (r = -.39 to r = .83) with the subscales of the longer scale from which the BEAQ was derived (Gamez et al., 2014). We used the Center for Epidemiologic Studies Depression Questionnaire (CESD; Radloff, 1977), a 20-item self-report measure (0 = rarely or none of the time, <1 day to 3 = most or all of the time, 5–7 days), to assess symptoms of depression over the past week. Higher scores reflect more symptoms of depression. Radloff (1977) reported good reliability

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

277

Helicopter Parenting | Wenze, Pohoryles, and DeCicco

(Cronbach’s α = .85, test-retest reliability ranging from 0.48 to 0.67) and validity (correlations with clinician-administered depression ratings in the .69 to .75 range). We used the Generalized Anxiety Disorder 7-item scale (GAD-7; Spitzer, Kroenke, Williams, & Löwe, 2006), a 7-item self-report measure (0 = not at all to 3 = nearly every day), to assess symptoms of anxiety over the past 2 weeks. Higher scores reflect more symptoms of anxiety. The GAD-7 has strong reliability (Cronbach’s α = .92, one-week test-retest reliability = .83) and validity (intraclass correlation = .83; Spitzer et al., 2006). Overview of Analytic Approach We used SPSS version 25 for all analyses. For media tion analyses, we used a nonparametric bootstrap ping approach (Preacher & Hayes, 2004, 2008). This approach is preferred to earlier methods for examining mediation (e.g., Baron & Kenny, 1986) because it does not require distributional assump tions about the data, it is appropriate for smaller samples such as ours (see “Procedure,” above), it formally tests whether the difference between the c and c’ paths is significant, and it simultaneously tests the significance of the a and b paths. Media tion is significant if the confidence intervals for the a*b values (derived from 5,000 datasets created by randomly drawing participants and sampling them with replacement) do not contain zero. See Figure 1 for a conceptual model of mediation and the a, b, c, and c’ pathways. Of note, the significance of the a and b paths is inconsequential in this approach, nor must a total effect or direct effect of X on Y be dem onstrated. Instead, an indirect effect is significant, and mediation is demonstrated if the confidence interval for the a*b product does not include zero (Hayes & Rockwood, 2017; Zhao, Lynch, & Chen, 2010). In the current study, we examined whether expressive suppression, cognitive reappraisal, FIGURE 1 X (Predictor or independent variable)

c Path (Total Effect)

Y (Outcome or dependent variable)

M (Mediator or intervening variable) a Path

X (Predictor or independent variable)

b Path

c' Path (Direct Effect)

Y (Outcome or dependent variable)

Figure 1. Conceptual model of mediation. a*b = c – c’ = indirect effect, or the measure of the amount of mediation.

278

psychological flexibility, and experiential avoidance (Ms) mediated relationships between helicop ter parenting and autonomy support (Xs) and depression and anxiety (Ys). Although there is no consensus about the optimal measure of effect size for mediation, the mediation ratio (P M) is com monly reported and the index of mediation (abcs) offers certain statistical and conceptual advantages (Preacher & Kelley, 2011). Therefore, in addition to R2 and Cohen’s f2, we report PM and abcs for all significant effects.

Results Table 1 presents means, standard deviations, and Cronbach’s αs of all study variables, as well as intercorrelations between variables. Table 2 presents bootstrapped regression results for the mediation of the effect of helicopter parenting on depression symptoms. The relationship between helicopter parenting and depression was mediated by experiential avoidance (95% bias-corrected and accelerated [Bca] confidence interval [CI] = .01 to .33, PM = 15, abcs = .09, R2 = .23, f2 = .30). Higher helicopter parenting was associated with higher experiential avoidance and, in turn, higher depres sion symptoms. Table 3 presents bootstrapped regression results for the mediation of the effect of helicopter parenting on anxiety symptoms. The relation ship between helicopter parenting and anxiety was mediated by experiential avoidance (95% Bca CI = .01 to .13, P M = 1, ab cs = .07, R 2 = .14, f2 = .17). Higher helicopter parenting was associated with higher experiential avoidance and, in turn, higher anxiety symptoms. Table 4 presents bootstrapped regression results for the mediation of the effect of autonomy support on depression symptoms. The relationship between autonomy support and depression was mediated by expressive suppression (95% Bca CI = -.27 to -.01, PM = .30, abcs = .06, R2 = .08, f2 = .08), cognitive reappraisal (95% Bca CI = -.36 to -.03, PM = .41, abcs = .08, R2 = .09, f2 = .10), psychological flex ibility (95% Bca CI = -.59 to -.04, PM = .76, abcs = .15, R2 = .46, f2 = .86), and experiential avoidance (95% Bca CI = -.39 to -.03, PM = .46, abcs = .09, R2 = .24, f2 = .31). Higher autonomy support was associated with lower expressive suppression and experiential avoidance, and higher cognitive reappraisal and psychological flexibility; in turn, these factors were associated with lower depression. Table 5 presents bootstrapped regression results for the mediation of the effect of autonomy

Wenze, Pohoryles, and DeCicco | Helicopter Parenting

TABLE 1 Means, Standard Deviations, Reliability, Coefficients, and Intercorrelations Between Study Variables Intercorrelations Sum total M(SD)

Average item score M(SD)

1. Helicopter parenting (HPB)

20.20 (7.00)

2. Autonomy support (HPB)

27.26 (5.93)

3. Expressive suppression (ERQ) 4. Cognitive reapprasial (ERQ)

2.24(0.78)

.74

–

-.11

-.04

-.08

-.18

.19

.01

.08

4.54(0.99)

.71

–

-.27**

.32**

.22*

-.20*

-.13

13.99 (5.20)

3.50(1.30)

.76

–

-.19

-.38***

.42***

.24*

.23*

29.54 (5.31)

4.92(0.89)

.72

–

.32**

-.13

-.27**

-.30**

5. Psychological Flexbility (AAQ)

35.68 (8.80)

5.10(1.26)

.89

–

-.47***

-.67***

-.70***

6. Experiential Avoidance (BEAQ)

46.32 (9.31)

3.09(0.62)

.80

–

.48***

.38***

7. Depression symptoms (CESD)

17.35 (11.18)

0.87(0.56)

.92

–

.76***

8. Anxiety symptoms (GAD-7)

6.76 (5.02)

0.97(0.72)

.89

–

Note. HPB = Helicopter Parenting Behaviors measure; ERQ = Emotion Regulation Questionnaire; AAQ = Acceptance and Action Questionnaire – II; BEAQ = Brief Experiential Avoidance Questionnaire; CESD = Center for Epidemiologic Studies Depression Scale; GAD-7 = Generalized Anxiety Disorder 7-item scale. *p < .05. **p < .01. ***p < .001.

TABLE 2

TABLE 3

Bootstrapped Regression Results for the Mediation of the Effect of Helicopter Parenting on Depression Symptoms

Bootstrapped Regression Results for the Mediation of the Effect of Helicopter Parenting on Anxiety Symptoms

Pathway

Estimate

LLCI

ULCI

Pathway

Estimate

.95

-.31

.33

HP → Anx (c)

.05

.07

Mediator = Experiential Avoidance

LLCI

ULCI

.45

-.09

.20

Mediator = Experiential Avoidance

HP → Depr (c)

.01

.16

HP → EA (a)

.25

.13

.06

-.01

.51

HP → EA (a)

.25

.13

.06

-.01

.51

EA → Depr (b)

.59

.11

< .001

.38

.80

EA → Anx (b)

.20

.05

< .001

.10

.30

HP → Depr (c')

-.14

.14

.34

-.42

.15

HP → Anx (c')

.004

.07

.95

-.13

.14

Indirect effect (a*b)

.15*

.08

.01

.33

Indirect effect (a*b)

.05*

.03

.01

.12

-.31

.33

HP → Anx (c)

.05

.07

.45

-.09

.20

Mediator = Psychological Flexibility

HP → Depr (c)

.01

.16

HP → PsyFlex (a)

-.22

.12

.07

-.47

.02

HP → PsyFlex (a)

-.22

.12

.07

-.47

.02

PsyFlex → Depr (b)

-.90

.09

< .001

-1.08

-.71

PsyFlex → Anx (b)

-.41

.04

< .001

-.49

-.33

HP → Depr (c')

-.19

.12

.11

-.43

.05

HP → Anx (c')

-.04

.05

.49

-.14

.07

Indirect effect (a*b)

.20

.11

-.02

.43

Indirect effect (a*b)

.09

.05

-.01

.19

.95

Mediator = Expressive Suppression

HP → Depr (c)

.01

.16

.95

-.31

.33

HP → Anx (c)

.05

.07

.45

-.09

.20

HP → ExpSupr (a)

-.03

.07

.70

-.18

.12

HP → ExpSupr (a)

-.03

.07

.70

-.18

.12

ExpSupr → Depr (b)

.53

.21

.01

.11

.94

ExpSupr → Anx (b)

.23

.09

.02

.04

.41

HP → Depr (c')

.02

.16

.87

-.28

.33

HP → Anx (c')

.06

.07

.39

-.08

.20

Indirect effect (a*b)

-.02

.05

-.15

.07

Indirect effect (a*b)

-.01

.02

-.06

.03

HP → Depr (c)

.01

.16

.95

-.31

.33

HP → Anx (c)

.05

.07

.45

-.09

.20

HP → CogReapr (a)

-.06

.07

.43

-.20

.09

HP → CogReapr (a)

-.06

.07

.43

-.20

.09

CogReapr → Depr (b)

-.61

.21

.005

-1.03

-.19

CogReapr → Anx (b)

-.29

.09

.002

-.48

-.11

HP → Depr (c')

-.03

.15

.87

-.33

.28

HP → Anx (c')

.04

.07

.59

-.10

.17

Indirect effect (a*b)

.04

.05

-.04

.14

Indirect effect (a*b)

.02

-.02

.07

Mediator = Cognitive Reappraisal

Note. * Significant at the .05 level. LLCI = lower bound of a 95% confidence interval; ULCI = upper bound of a 95% confidence interval; HP = helicopter parenting; Depr = depression symptoms; EA = experiential avoidance; PsyFlex = psychological flexibility; ExpSupr = expressive suppression; CogReapr = cognitive reappraisal.

Note. * Significant at the .05 level. LLCI = lower bound of a 95% confidence interval; ULCI = upper bound of a 95% confidence interval; HP = helicopter parenting; Anx = anxiety symptoms; EA = experiential avoidance; PsyFlex = psychological flexibility; ExpSupr = expressive suppression; CogReapr = cognitive reappraisal.

279

Helicopter Parenting | Wenze, Pohoryles, and DeCicco

TABLE 4

TABLE 5

Bootstrapped Regression Results for the Mediation of the Effect of Autonomy Support on Depression Symptoms

Bootstrapped Regression Results for the Mediation of the Effect of Autonomy Support on Anxiety Symptoms

Pathway

Estimate

LLCI

ULCI

Pathway

Estimate

Mediator = Experiential Avoidance

LLCI

ULCI

Mediator = Experiential Avoidance

AS → Depr (c)

-.37

.18

.05

-.74

-.003

AS → Anx (c)

-.11

.08

.20

-.27

.06

AS → EA (a)

-.32

.15

.04

-.62

-.01

AS → EA (a)

-.32

.15

.04

-.62

-.01

EA → Depr (b)

.55

.11

< .001

.33

.76

EA → Anx (b)

.20

.05

< .001

.10

.30

AS → Depr (c')

-.20

.17

.25

-.53

.14

AS → Anx (c')

-.05

.08

.58

-.20

.11

Indirect effect (a*b)

-.17*

.09

-.39

-.02

Indirect effect (a*b)

-.06*

.04

-.16

-.01

-.74

-.003

AS → Anx (c)

-.11

.08

-.27

.06

Mediator = Psychological Flexibility AS → Depr (c)

.05

Mediator = Psychological Flexibility

-.37

.18

AS → PsyFlex (a)

.33

.14

.02

.05

.62

AS → PsyFlex (a)

.33

.14

.02

.05

.62

PsyFlex → Depr (b)

-.86

.10

<.001

-1.05

-.67

PsyFlex → Anx (b)

-.41

.04

< .001

-.49

-.32

AS → Depr (c')

-.09

.14

.55

-.37

.20

AS → Anx (c')

.03

.06

.66

-.10

.15

Indirect effect (a*b)

-.28*

.14

-.59

-.04

Indirect effect (a*b)

-.13*

.07

-.28

-.02

AS → Depr (c)

-.37

.18

.05

-.74

-.003

AS → Anx (c)

-.11

.08

.20

-.27

.06

AS → ExpSupr (a)

-.24

.08

.01

-.41

-.07

AS → ExpSupr (a)

-.24

.08

.005

-.41

-.07

ExpSupr → Depr (b)

.44

.21

.04

.02

.87

ExpSupr → Anx (b)

.21

.10

.04

.01

.40

AS → Depr (c')

-.26

.19

.17

-.64

.11

AS → Anx (c')

-.06

.09

.51

-.23

.11

Indirect effect (a*b)

-.11*

.07

-.27

-.01

Indirect effect (a*b)

-.05

.04

-.13

.001

AS → Depr (c)

-.37

.18

.05

-.74

-.003

AS → Anx (c)

-.11

.08

.20

-.27

.06

AS → CogReapr (a)

.28

.08

.001

.11

.44

AS → CogReapr (a)

.28

.08

.001

.11

.44

CogReapr → Depr (b)

-.53

.22

.02

-.96

-.09

CogReapr → Anx (b)

-.29

.10

.005

-.48

-.09

AS → Depr (c')

-.22

.19

.24

-.60

.15

AS → Anx (c')

-.03

.09

.74

-.20

.14

Indirect effect (a*b)

-.15*

.08

-.36

-.03

Indirect effect (a*b)

-.08*

.04

-.17

-.02

Mediator = Expressive Suppression

Mediator = Cognitive Reappraisal

Note. * Significant at the .05 level. LLCI = lower bound of a 95% confidence interval; ULCI = upper bound of a 95% confidence interval; AS = autonomy support; Depr = depression symptoms; EA = experiential avoidance; PsyFlex = psychological flexibility; ExpSupr = expressive suppression; CogReapr = cognitive reappraisal.

support on anxiety symptoms. The relationship between autonomy support and anxiety was medi ated by cognitive reappraisal (95% Bca CI = -.17 to -.02, PM = .73, abcs = .09, R2 = .09, f2 = .10), psycho logical flexibility (95% Bca CI = -.28 to -.02, PM = 1.18, abcs = .15, R2 = .49, f2 = .97), and experiential avoidance (95% Bca CI = -.16 to -.01, PM = .55, abcs = .07, R2 = .14, f2 = .17). Higher autonomy support was associated with lower experiential avoidance and higher cognitive reappraisal and psychological flexibility; in turn, these factors were associated with lower anxiety. WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

280

.20

Supplemental Analyses As one way to correct for the possibility of Type I error, we reran mediational analyses using 99% confidence intervals. The relationship between

autonomy support and depression was mediated by cognitive reappraisal (99% Bca CI = -.43 to -.002, PM = .39, abcs = .08, R2 = .09, f2 = .10). The relation ship between autonomy support and anxiety was also mediated by cognitive reappraisal (99% Bca CI = -.19 to -.003, PM = .72, abcs = .09, R2 = .09, f2 = .10). The other mediational relationships were no longer significant. Another way to correct for Type I error, while also allowing us to examine the unique effects of each proposed mediator on these relationships, is to re-run analyses of the effects of helicopter parent ing and autonomy support (Xs) on depression and anxiety (Ys), including all 4 proposed mediators in each of these four analyses. Controlling for expressive suppression, cognitive reappraisal, and psychological flexibility, experiential avoidance

Wenze, Pohoryles, and DeCicco | Helicopter Parenting

mediated the relationship between helicopter parenting and depression (95% Bca CI = .01 to .20, PM = 7.72, abcs = .05, R2 = .53, f2 = 1.12); higher helicopter parenting was associated with higher experiential avoidance, which predicted greater depression. None of the proposed variables independently mediated the relationship between helicopter parenting and anxiety. Controlling for expressive suppression and cognitive reappraisal, psychological flexibility and experiential avoidance mediated the relationship between autonomy sup port and depression (95% Bca CI = -.52 to -.05, PM = .72, abcs = .13 and 95% Bca CI = -.25 to -.01, PM = .26, abcs = .05, R2 = .51, f2 = 1.02); higher autonomy support was independently associated with higher psychological flexibility and lower experiential avoidance, both of which in turn predicted lower depression. Controlling for expressive suppression, cognitive reappraisal, and experiential avoidance, psychological flexibility mediated the relationship between autonomy support and anxiety (95% Bca CI = -.26 to -.02, PM = .77, abcs = .15, R2 = .51, f2 = 1.04); higher autonomy support was associated with higher psychological flexibility, which predicted lower anxiety.

Discussion The current study was one of the first to examine mechanisms behind the link between helicopter parenting behaviors and negative mental health outcomes in college students. Such analyses are important for a full understanding of the psy chological correlates of this parenting style, and crucial for considering ways to educate parents and maximize college student well-being. In the current sample, helicopter parenting was not directly related to depression or anxiety symptoms. Likewise, autonomy support was not correlated with anxiety symptoms. However, autonomy sup port was negatively associated with depression symptoms, and even for those total effects that were not significant, we found support for the notion that aspects of emotional processing mediate links between helicopter parenting/autonomy support ive behaviors and depression/anxiety symptoms. In particular, psychological flexibility and experiential avoidance emerged as uniquely and independently important in explaining these relationships. It is important to underscore that the current study was correlational in nature, so cause-andeffect relationships cannot be presumed. Indeed, a college student who is depressed or anxious (and who demonstrates the poor emotional processing that often accompanies these states) might prompt

helicopter parenting behaviors in his or her parents. However, it is also possible that, without experience independently managing stressful events and their (oftentimes difficult) emotional sequelae, one does not develop optimal coping strategies. Helicopter parenting and discouraging autonomous behaviors might prevent offspring from having direct contact with such events and mastering adaptive coping skills. For example, a mother who immediately intervenes at the first sign of her daughter’s struggles in a difficult course or who does not encourage her daughter to resolve interpersonal conflicts on her own prevents her daughter not only from solving these problems herself, but also from figuring out how to manage the anxiety, sadness, or frustration that might come from such experiences. Further, the mother models for her daughter that difficult experiences and negative emotions are dangerous, intolerable, and to be avoided or suppressed because she probably cannot handle them. Conversely, the mother also misses an opportunity to model persistence in the service of valued life goals and positive reappraisal of adversity. If the daughter develops emotional coping strategies and habits in line with these ideas (e.g., experiential avoidance, psychological inflexibility, expressive suppression, inability to cognitively reappraise negative emotions and triggers), she is at elevated risk for depression and anxiety (Gross & John, 2003; Hayes et al., 2006; Woodruff et al., 2014). Implications Our findings have implications for counseling col lege students. First, they underscore the importance of assessing and discussing students’ perceptions of their parents’ styles of parenting. For students who are struggling with depression or anxiety and who perceive low autonomy support from their parents, techniques drawn from Emotion Regulation Therapy (Renna, Quintero, Fresco, & Mennin, 2017) or Acceptance and Commitment Therapy (Hayes et al., 2006) might be helpful; such approaches aim to improve emotion regulation and increase psychological flexibility. Importantly, the current findings should also be disseminated to parents of college students. Having a child leave home can be a challenging experience for parents (Bouchard, 2014), and some may have a difficult time encouraging their college student’s developing independence. Knowing that autonomy-supportive behaviors are associated with better emotional processing and lower risk of depression and anxiety may help parents adjust their behavior.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

281

Helicopter Parenting | Wenze, Pohoryles, and DeCicco

Limitations This study was not without limitations. Data were based solely on self-report, which introduces the potential for self-presentation biases and assumes that respondents’ statements reflect reality. In the current study, participants might have been hesitant to report psychiatric symptoms or maladaptive emotion regulation strategies, and their percep tions of their parents’ behaviors might differ from the judgements that an impartial, trained observer might make. Demographic restrictions (especially regarding sex and ethnicity) and the nonclinical nature of our sample may limit generalizability of the findings. Similarly, our results might not extend to other subsamples of emerging adults. For example, college students who elect to attend larger universities or graduate degree-granting institutions might have better coping skills, desire more independence, or have been parented in a manner that is more supportive of autonomy than students who are drawn to the individualized atten tion and smaller class sizes often touted by small liberal arts colleges. Emerging adults who are not in college may be parented differently than their college-enrolled peers. Indeed, young adults who attend college tend to come from higher socioeco nomic status households (Pell Institute, 2005), and socioeconomic status has, in turn, been linked with parenting style (Hoff, Laursen, & Tardif, 2002). Of note, our measure of helicopter parenting is specific to the college experience, with questions asking about parent involvement in students’ choice of major and courses, interactions with professors and roommates, and grades. Finally, the present findings may not extend to non-U.S. samples. Social norms for amount and type of parental involvement vary by culture, as may the impact of such involve ment on children’s development (McElhaney & Allen, 2012). In comparison to their Western peers, for example, Asian students may rely more on parental support when transitioning to college (Dmitrieva et al., 2008). Future studies should address these limitations and build upon this work by testing cause-and-effect relationships between helicopter parenting, emotional processing, and mental health in U.S. college students.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

282

Conclusion This study added to a small but growing body of literature on the mechanisms by which helicopter parenting behaviors may be associated with poor mental health outcomes in U.S. college students. We found that expressive suppression, low cog nitive reappraisal, psychological inflexibility,

and experiential avoidance explained various relationships between helicopter parenting/low autonomy support, and depression/anxiety symp toms. Psychological inflexibility and experiential avoidance emerged as independent mediators of helicopter parenting-mental health links. Our findings have implications for counseling college students who are struggling with depression and anxiety. Future studies should build on this work by testing targeted treatment strategies, using other methods of assessment (e.g., behavioral measures or clinical interviews), and examining these processes in non-U.S. and/or noncollege samples.

References Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35, 205–230. https://doi.org/10.1016/S0005-7894(04)80036-4 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, 1173–1182. https://doi.org/10.1037//0022-3514.51.6.1173 Berking, M., Neacsiu, A., Comtois, K. A., & Linehan, M. M. (2009). The impact of experiential avoidance on the reduction of depression in treatment for borderline personality disorder. Behaviour Research and Therapy, 47, 663–670. https://doi.org/10.1016/j.brat.2009.04.011 Bond, F. W., Hayes, S. C., Baer, R. A., Carpenter, K. M., Guenole, N., Orcutt, H. K., . . . Zettle, R. D. (2011). Preliminary psychometric properties of the Acceptance and Action Questionnaire-II: A revised measure of psychological inflexibility and experiential avoidance. Behavior Therapy, 42, 676–688. https://doi.org/10.1016/j.beth.2011.03.007 Bouchard, G. (2014). How do parents react when their children leave home? An integrative review. Journal of Adult Development, 21, 69–79. https://doi.org/10.1007/s10804-013-9180-8 Bronson, P., & Merryman, A. (2009). Nurture shock: New thinking about children. New York, NY: Twelve. Cline, F. W., & Fay, J. (1990). Parenting with love and logic: Teaching children responsibility. Colorado Springs, CO: Pinon Press. Creswell, C., O’Connor, T. G., & Brewin, C. R. (2008). The impact of parents’ expectations on parenting behaviour: An experimental investigation. Behavioural and Cognitive Psychotherapy, 36, 483–490. https://doi.org/10.1017/S1352465808004414 Darlow, V., Norvilitis, J. M., & Schuetze, P. (2017). The relationship between helicopter parenting and adjustment to college. Journal of Child and Family Studies, 26, 2291–2298. http://dx.doi.org/10.1007/s10826-017-0751-3 Dmitrieva, J., Chen, C., & Greenberger, E. (2008). Whither the ‘Whiz Kids’ went: Asian American students’ transition to college. In G. Li & L. Wang (Eds.), Model minority myths revisited: An interdisciplinary approach to demystifying Asian American education experiences (pp. 233–251). Charlotte, NC: Information Age. Dryman, M. T., & Heimberg, R. G. (2018). Emotion regulation in social anxiety and depression: A systematic review of expressive suppression and cognitive reappraisal. Clinical Psychology Review, 65, 17–42. http://dx.doi.org/10.1016/j.cpr.2018.07.004 Fritz, M. S., & MacKinnon, D. P. (2007). Required sample size to detect the mediated effect. Psychological Science, 18, 233–239. https://doi.org/10.1111/j.1467-9280.2007.01882.x Gamez, W., Chmielewski, M., Kotov, R., Ruggero, C., Suzuki, N., & Watson, D. (2014). The Brief Experiential Avoidance Questionnaire: Development and initial validation. Psychological Assessment, 26, 35–45. https://doi.org/10.1037/a0034473 Gibbs, N. (2009, November). The growing backlash against overparenting. Time Magazine. Retrieved from http://content.time.com/time/subscriber/article/0,33009,1940697-1,00.html Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology, 85, 348–362.

Wenze, Pohoryles, and DeCicco | Helicopter Parenting

https://doi.org/10.1037/0022-3514.85.2.348 Hayes, A. F., & Rockwood, N. J. (2017). Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. Behaviour Research and Therapy, 98, 39–57. http://dx.doi.org/10.1016/j.brat.2016.11.001 Hayes, S. C., Luoma, J. B., Bond, F. W., Masuda, A., & Lillis, J. (2006). Acceptance and commitment therapy: Model, processes and outcomes. Behaviour Research and Therapy, 44, 1–25. https://doi.org/10.1016/j.brat.2005.06.006 Hayes, S. C., Strosahl, K., & Wilson, K. G. (1999). Acceptance and commitment therapy: An experiential approach to behavior change. New York, NY: Guilford Press. Hofer, B. K., & Moore, A. S. (2010). IConnected parent. New York, NY: Free Press. Hoff, E., Laursen, B., & Tardif, T. (2002). Socioeconomic status and parenting. In M. H. Bornstein (Ed.), Handbook of Parenting (2nd ed., pp. 231–252). Mahwah, NJ: Erlbaum. Kouros, C. D., Pruitt, M. M., Ekas, N. V., Kiriaki, R., & Sunderland, M. (2017). Helicopter parenting, autonomy support, and college students’ mental health and well-being: The moderating role of sex and ethnicity. Journal of Child and Family Studies, 26, 939–949. https://doi.org/10.1007/s10826-016-0614-3 LeMoyne, T., & Buchanan, T. (2011). Does ‘hovering’ matter? Helicopter parenting and its effect on well-being. Sociological Spectrum, 31, 399–418. https://doi.org/10.1080/02732173.2011.574038 Loevaas, M. E. S, Sund, A. M., Patras, J., Martinsen, K., Hjemdal, O., Neumer, S.-P., . . . Reinfjell, T. (2018). Emotion regulation and its relation to symptoms of anxiety and depression in children aged 8–12 years: Does parental gender play a differentiating role? BMC Psychology, 6, 42. https://doi.org/10.1186/s40359-018-0255-y Luebbe, A. M., Mancini, K. J., Kiel, E. J., Spangler, B. R., Semlak, J. L., & Fussner, L. M. (2018). Dimensionality of helicopter parenting and relations to emotional, decision-making, and academic functioning in emerging adults. Assessment, 25, 841–857. https://doi.org/10.1177/1073191116665907 Marano, H. E. (2008). A nation of wimps: The high cost of invasive parenting. New York, NY: The Crown Publishing Group. McElhaney, K. B., & Allen, J. P. (2012). Sociocultural perspectives on adolescent autonomy. In P. Kerig, M. Shulz, & S. T. Hauser (Eds.), Adolescence and beyond (pp. 161–176). Oxford: Oxford University Press. Nelson, L. J., Padilla-Walker, L. M., & Nielson, M. G. (2015). Is hovering smothering or loving? An examination of parental warmth as a moderator of relations between helicopter parenting and emerging adults’ indices of adjustment. Emerging Adulthood, 3, 282–285. https://doi.org/10.1177/2167696815576458 Padilla-Walker, L. M., & Nelson, L. J. (2012). Black hawk down?: Establishing helicopter parenting as a distinct construct from other forms of parental control during emerging adulthood. Journal of Adolescence, 35, 1177–1190. https://doi.org/10.1016/j.adolescence.2012.03.007 Pell Institute. (2005). Indicators of opportunity in higher education. Washington, DC: Pell Institute. Retrieved from http://www.pellinstitute.org/downloads/ publications-Indicators_of_Opportunity_in_Higher_Education_2005_ Status_Report.pdf Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36, 717–731. https://doi.org/10.3758/BF03206553 Preacher, K. J., & Hayes, A. F. (2008). Asymptotic resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behaviour Research Methods, 40, 879–891. https://doi.org/10.3758/BRM.40.3.879 Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: Quantitative strategies for communicating indirect effects. Psychological Methods, 16, 93–115. https://doi.org/10.1037/a0022658 Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. https://doi.org/10.1177/014662167700100306 Reed, K., Duncan, J. M., Lucier-Greer, M., Fixelle, C., & Ferraro, A. J. (2016). Helicopter parenting and emerging adult self-efficacy: Implications for mental and physical health. Journal of Child and Family Studies, 25, 3136–3149. https://doi.org/10.1007/s10826-016-0466-x Reilly, S., & Semkovska, M. (2018). An examination of the mediatory role of resilience in the relationship between helicopter parenting and severity of depressive symptoms in Irish university students. Adolescent Psychiatry, 8, 32–47. http://dx.doi.org/10.2174/2210676608666180508130224 Renna, M. E., Quintero, J. M., Fresco, D. M., & Mennin, D. S. (2017). Emotionregulation therapy: A mechanism-targeted treatment for disorders of distress. Frontiers in Psychology, 8, 1–14.

https://doi.org/10.3389/fpsyg.2017.00098 Roemer, L., Salters, K., Raffa, S. D., & Orsillo, S. M. (2005). Fear and avoidance of internal experiences in GAD: Preliminary tests of a conceptual model. Cognitive Therapy and Research, 29, 71–88. http://dx.doi.org/10.1007/s10608-005-1650-2 Schiffrin, H. H., & Liss, M. (2017). The effects of helicopter parenting on academic motivation. Journal of Child and Family Studies, 26, 1472–1480. https://doi.org/10.1007/s10826-017-0658-z Schiffrin, H. H., Liss, M., Miles-McLean, H., Geary, K. A., Erchull, M. J., & Tashner, T. (2014). Helping or hovering? The effects of helicopter parenting on college students’ well-being. Journal of Child and Family Studies, 23, 548–557. https://doi.org/10.1007/s10826-013-9716-3 Segrin, C., Woszidlo, A., Givertz, M., Bauer, A., & Murphy, M. T. (2012). The association between overparenting, parent-child communication, and entitlement and adaptive traits in adult children. Family Relations, 61, 237–252. https://doi.org/10.1111/j.1741-3729.2011.00689.x Shallcross, A. J., Troy, A. S., Boland, M., & Mauss, I. B. (2010). Let it be: Accepting negative emotional experiences predicts decreased negative affect and depressive symptoms. Behaviour Research and Therapy, 48, 921–929. https://doi.org/10.1016/j.brat.2010.05.025 Spinhoven, P., Drost, J., de Rooij, M., van Hemert, A. M., & Penninx, B. W. (2014). A longitudinal study of experiential avoidance in emotional disorders. Behavior Therapy, 45, 840–850. https://doi.org/10.1016/j.beth.2014.07.001 Spitzer, R. L., Kroenke, K., Williams, J. B., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of Internal Medicine, 166, 1092–1097. https://doi.org/10.1001/archinte.166.10.1092 Tanner, J. L. (2006). Recentering during emerging adulthood: A critical turning point in life span human development. In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 21–55). Washington, DC: American Psychological Association. https://doi.org/10.1037/11381-002 Thompson, R. A. (1994). Emotion regulation: A theme in search of definition. Monographs of the Society for Research in Child Development, 59, 25–52. https://doi.org/10.2307/1166137 van Ingen, D. J., Freiheit, S. R., Steinfeldt, J. A., Moore, L. L., Wimer, D. J., Knutt, A. D., . . . Roberts, A. (2015). Helicopter parenting: The effect of an overbearing caregiving style on peer attachment and self-efficacy. Journal of College Counseling, 18, 7–20. https://doi.org/10.1002/j.2161-1882.2015.00065.x Vinson, K. E. (2013). Hovering too close: The ramifications of helicopter parenting in higher education. Georgia State University Law Review, 29, 423–452. https://doi.org/10.2139/ssrn.1982763 Woodruff, S. C., Glass, C. R., Arnkoff, D. B., Crowley, K. J., Hindman, R. K., & Hirschhorn, E.W. (2014). Comparing self-compassion, mindfulness, and psychological inflexibility as predictors of psychological health. Mindfulness, 5, 410–421. https://doi.org/10.1007/s12671-013-0195-9 Wenze, S. J., Gaugler, T. L., Sheets, E. S., & DeCicco, J. M. (2018). Momentary experiential avoidance: Within-person correlates, antecedents, and consequences and between-person moderators. Behaviour Research and Therapy, 107, 42–52. https://doi.org/10.1016/j.brat.2018.05.011 Zhao, X., Lynch Jr., J. G., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of Consumer Research, 37, 197–203. https://doi.org/10.1086/651257 Author Note. Susan J. Wenze, https://orcid.org/0000-0002-4180-8642, Department of Psychology, Lafayette College; Anna B. Pohoryles, https://orcid.org/0000-0003-4938-2073, Department of Psychology, Lafayette College; Jennifer M. DeCicco Department of Psychology, Lafayette College Jennifer M. DeCicco is now at the Department of Psychology at Holy Family University. This study was supported by internal funds from Lafayette College. We thank Deniz Bengi, Natalie Cardenas, Jessica Fields, Charlotte Jamieson, and Danielle Kats for assistance with data collection. Special thanks to Psi Chi Journal reviewers for their support. Corresponding concerning this article should be addressed to Susan J. Wenze, Department of Psychology, Lafayette College, Oechsle Hall of Psychology, 350 Hamilton Street, Easton, PA 18042, United States of America. E-mail: wenzes@lafayette.edu.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

283

https://doi.org/10.24839/2325-7342.JN24.4.284

Invited Editorial: Psychological Science Plays a Critical Role in Addressing the Environmental Crisis Ethan A. McMahan Western Oregon University

ABSTRACT. Psychology involves the study of human behavior and the application of knowledge to address real-world problems that involve human behavior. Psychologists are, therefore, problem solvers. The environmental crisis is perhaps the most critical problem facing humanity at this point in history, and there is scientific consensus that observed problematic climate changes are the result of human behavior. Given the nature of our field, psychology is uniquely and well-positioned to assist in addressing this issue by describing and understanding how psychological factors contribute to the degradation of the environment and by encouraging environmentally responsible and sustainability oriented thinking and behavior. Indeed, the field has a responsibility to do so. Keywords: psychology, environmental, climate, problem solving

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

284

sychologists are problem solvers. Solving problems is fundamental to psychology, so much so that we have included this general activity in most of our accepted definitions of the field. For example, in one definition, psychology is described as the scientific study of the mind and behavior, as well as the application of knowledge accumulated from this science toward solving practical problems (Weiten, 2007). The American Psychological Association (APA), the largest professional association in psychology, includes problem-solving in its mission statement, indicating that a central goal of the APA is to apply psychological science and knowledge toward benefiting society and improving lives. Indeed, Psi Chi has dedicated a large amount of its resources toward supporting problem solving by, for example, funding psychological research that examines several key issues currently facing society (e.g., group inequality, discrimination). In short, the application of psychological knowledge and expertise toward addressing the major problems facing humanity is not a small part of our enterprise, but rather a definitive aspect of the field in general. Notably, psychology has been remarkably suc cessful at addressing many real-world problems in

many different domains. Our work has produced positive outcomes in a diversity of areas including those related to the treatment of mental illness, education, the legal system, work environments, and health care, among others. The broad appli cability of psychology to many different areas is a result of the nature of our field. We are able to address many different problems because our field is itself diverse, with numerous areas of emphasis. This diversity is increased when considering the sub fields of psychology that have emerged as a result of integration with other disciplines (e.g., health psychology, engineering psychology, sport psychol ogy). Given the multiple domains within which psychology has played a role, our field, through its focus on problem solving, has made significant contributions to the betterment of society. Despite this positive track record, our work is conspicuously absent in one key area, and it is perhaps time for us to turn in earnest to solving one of the (if not “the”) most significant chal lenges facing humanity at this point in history: the environmental crisis. At first glance, it may seem odd to suggest that psychology can make significant contributions toward addressing the environmental crisis, and one might assume that this should be left

McMahan | Psychology's Role in the Environmental Crisis

to policy makers and the scientific disciplines most closely associated with studying environmental and climate change (e.g., climatologists, meteorologists, biologists). However, as detailed below, I argue that psychology is well-suited and uniquely positioned to address this particular problem. In fact, we psychologists might be some of the best people for the job.

Hard Problems There are a lot of hard problems. Optimizing the U.S. health care system, reducing juvenile drug use, selecting appropriate immigration policies, and effectively treating degenerative diseases (e.g., Alzheimer’s disease) are all examples of hard problems. Hard problems, or what have at times been called “grand challenges” (Kazdin, 2009) or “wicked problems” (Horn & Weber, 2007), do not have easy solutions because these problems are often the result of multiple interdependent events or trends; frequently include numerous secondary embedded problems; impact many stakehold ers who may have different perspectives on the problem; are defined differently depending on cultural, political, and value orientations; and are characterized by incomplete information regarding the causes, characteristics, and consequences of the problem. To make matters worse, the condi tions that characterize hard problems are dynamic and ever-changing, making potentially effective strategies aimed at addressing these conditions time- and context-dependent. Because of the above, ultimate and complete solutions to hard problems are unlikely. Addressing the environmental crisis is a hard problem. The events and conditions that led us to this point are multiple and interdependent (e.g., population growth, industrialization). There are a multitude of subproblems associated with the environmental crisis, some of which are themselves hard (e.g., rising sea levels, drought, habitat loss). The environmental crisis impacts all people (in the world…), but these impacts and our understand ing of these impacts vary depending on location, nationality, socioeconomic status, and so on. And, the environmental crisis is not a static state of affairs but instead a constantly changing global phenomenon. Consider a key characteristic of the environmental crisis: climate change…it’s right there in the name…change. Finally, we are unlikely to fully “solve” the environmental crisis because many the crisis-related events that have transpired cannot be reversed or undone, and many of the

outcomes of the environmental crisis have yet to emerge and cannot be anticipated. Effective strate gies for addressing the environmental crisis will thus need to be numerous, diverse, broadly applicable at times but specific at others, subject to evaluation and amenable to change, and focused on mitigation (i.e., damage control) as well as adaptation.

Psychology: Hard Problem Solver Psychology is well-suited and uniquely positioned to address hard problems such as the environmental crisis. Although, as noted above, a complete solu tion to the environmental crisis is not possible, we as psychologists have much to contribute toward improving both current and future states of affairs. To illustrate this point, let’s focus narrowly on the embedded issue of climate change. The vast major ity of experts agree that human activity, particularly over the last 100 years, has contributed substantially to the warming of the planet (see Intergovern mental Panel on Climate Change, 2018). In other words, climate change was/is caused, at least in part, by human behavior. Given this, we can and perhaps should reframe our understanding of climate change. Climate change does not reflect a problem of the environment but rather a problem of behavior. More specifically, climate change is the result of a long-term and maladaptive pattern of human behavior. Importantly, psychology is the central discipline tasked with describing, under standing, predicting, and addressing maladaptive patterns of behavior. When thought of this way, psychology is not only one of the disciplines that should contribute to addressing climate change, we should take the lead. But, climate change is itself a hard problem, and the complexity of addressing this problem goes beyond just individual behavior. Within the sticky issue of climate change, one has to take into account the many domains in which climate change operates and levels of analysis at which climate change is to be understood. The issue has to be addressed at multiple levels and from multiple perspectives. Here too is reason for psychology to be a key figure in solving the problem. As noted above, psychology is diverse, interdisciplinary, and applicable in many domains. We study multiple aspects of human behavior (e.g., decision-making, problem-solving, intergroup relations) in multiple contexts (e.g., home, work, school), in multiple domains (e.g., health, education), and at several lev els of analysis (e.g., intrapersonal, individual, group, population). Who can address how to encourage

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

285

Psychology's Role in the Environmental Crisis | McMahan

environmentally responsible decision-making? What field has the expertise to assess the effective ness of environmental education on middle-school students’ knowledge of climate change? Who can figure out which incentives are most effective at encouraging sustainability-oriented behavior? Which discipline is well-positioned to investigate how differences in values, beliefs, and goals might impact understanding of environmental issues? Is there someone out there who can tell me how prejudices, discrimination, and intergroup dynam ics might impact (and be impacted by) group-level responses to the inequities that will emerge as a result of climate change-related resource scarcity? Yes, there is. That person is a psychologist.

The Roles and Responsibilities of Psychology Critically, psychology is not the only field that can contribute effective solutions toward addressing the environmental crisis. Many other, if not all, disciplines can play a role. But, given the nature of psychology, its strengths, its applicability, the ease with which we integrate with other disciplines, and the nature of the problem itself, our absence among those leading the charge to address what is perhaps the hardest problem facing humanity right now is an oversight at best and, if we continue in this fashion, an abdication of responsibility at worst. There have been numerous calls for psychol ogy to take on a more central role in addressing the environmental crisis and combatting climate change (e.g., Clayton et al., 2016; Kazdin, 2009; Oskamp, 2000), and we have made great strides. There now exists several subfields within psychology that directly address environmental issues (e.g., con servation psychology, ecopsychology, environmental psychology), several environmentally oriented professional organizations within psychology (e.g., the Society for Environmental, Population, and Conservation Psychology), and publications dedi cated to the communication of research in this area (e.g., Ecopsychology). But, given the significance of this particular problem, psychology can and should continue to develop in this area. And, there is much opportunity for those interested in developing it.

Many relevant research topics are unexplored, and many questions are yet to be answered. Many opportunities for collaboration both within and outside of psychology are possible. It should also be noted that federal research dollars are currently being focused toward projects that have real-world significance, and environmental research has clear practical applications. Through rigorous scientific research and dedi cated application, the field of psychology is capable of improving human welfare and transforming society for the better. Indeed, to this end, much progress has been made. But, many challenges still exist, many hard problems are yet to be solved, and much work is still to be done. Collectively, psycholo gists and our colleagues in related disciplines are powerful enough to guide the direction of human ity, but with this power comes responsibility. We need to fully accept this responsibility and tackle the most significant problems. By doing so, our work can benefit not only individual people and society, but also the larger world to which we belong.

References Clayton, S., Devine-Wright, P., Swim, J., Bonnes, M., Steg, L., Whitmarsh, L., & Carrico, A. (2016). Expanding the role for psychology in addressing environmental challenges. American Psychologist, 71, 199–215. https://doi.org/10.1037/a0039482 Horn, R. E., & Weber, R. P. (2007). New tools for resolving wicked problems: Mess mapping and resolution mapping processes. Strategy Kinetics LLC. Intergovernmental Panel on Climate Change. (2018). Summary for policy makers. In Masson-Delmotte et al. (Eds.), Global Warming of 1.5°C: An IPCC Special Report on the impacts of global warming of 1.5°C above preindustrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty. Available at https://www.ipcc.ch/sr15/ Kazdin, A. E. (2009). Psychological science’s contributions to a sustainable environment: Extending our reach to a grand challenge of society. American Psychologist, 64, 339–356 . https://doi.org/10.1037/a0015685 Oskamp, S. (2000). A sustainable future for humanity? How can psychology help? American Psychologist, 55, 496–508. https://doi.org/10.1037//0003-066X.55.5.496 Weiten, W. (2007). Psychology: Themes and variations. Cengage Learning. Author Note. Ethan A. McMahan, Department of Psychological Sciences, Western Oregon University. Special thanks to Debi Brannan and Jon Grahe for their encouragement and support. Correspondence regarding this article should be addressed to Ethan A. McMahan, Behavioral Sciences Division, Western Oregon University, Monmouth, OR, 97361. E-mail: mcmahane@mail.wou.edu.

WINTER 2019 PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

286

LOOKING FOR COLLABORATIVE RESEARCH EXPERIENCE? Join the Psi Chi CROWD! Students and faculty within the United States and beyond are invited to participate in the CROWD, which is Psi Chi’s annual, guided cross-cultural research project. Specific benefits of joining the CROWD include • a reduced burden of having to solicit large numbers of participants, • increased diversity of student samples, • accessible materials and protocols for participating researchers, and • a convenient platform to engage students in the scientific research process. Contributing to the CROWD provides unique data collection and publication experiences that can be used to strengthen any student’s CV. WINTER 2019

® ®

For more information, visit https://www.psichi.org/Res_Opps or contact the NICE Chair, Megan Irgens, at nicechair@psichi.org.

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH

287

Gain Valuable Research Experience With Psi Chi! Students and faculty are invited to visit Psi Chi’s free Conducting Research online resource at www.psichi.org/page/ConductingResearch. Here are three ways to get involved:

Join a Collaborative Research Project

Recruit Online Participants for Your Studies

Explore Our Research Measures Database

www.psichi.org/page/Res_Opps

www.psichi.org/page/study_links

With Psi Chi’s Network for International Collaborative Exchange (NICE), you can join the CROWD and answer a common research question with researchers internationally. You can also CONNECT with a network of researchers open to collaboration.

Psi Chi is dedicated to helping members find participants to their online research studies. Submit a title and a brief description of your online studies to our Post a Study Tool. We regularly encourage our members to participate in all listed studies.

www.psichi.org/page/ researchlinksdesc

288

This database links to various websites featuring research measures, tools, and instruments. You can search for relevant materials by category or keyword. If you know of additional resources that could be added, please contact research.director@psichi.org

289

Publish Your Research in Psi Chi Journal Undergraduate, graduate, and faculty submissions are welcome year round. Only the first author 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

Add Our Journal to Your Library Ask your librarian to store Psi Chi Journal issues in a database at your local institution. Librarians may also e-mail to request notifications when new issues are released. Contact PsiChiJournal@psichi.org for more information.