psi chi journal of undergraduate research - fall 2001

PSI CHIJournal of Undergraduate Research

Editorial Office

Stephen F. Davis, PhDManaging Editor

Psi Chi Journal of Undergraduate Research649 Hideaway Lane EastLindale, TX 75771-5301

E-mail: [email protected]

Psi Chi National Council

Peter J. Giordano, PhDNational PresidentBelmont University

Elizabeth Yost Hammer, PhDPresident-Elect

Loyola University New Orleans

Jesse E. Purdy, PhDPast-President

Southwestern University

Robert A. Youth, PhD Alvin Y. Wang, PhDEastern Regional Vice-President Southeastern Regional Vice-President

Dowling College University of Central Florida

Joseph J. Palladino, PhD John M. Davis, PhDMidwestern Regional Vice-President Southwestern Regional Vice-President

University of Southern Indiana Southwest Texas State University

Carla J. Reyes, PhD Christopher Koch, PhDRocky Mountain Regional Vice-President Western Regional Vice-President

University of Utah George Fox University

Kay WilsonExecutive Officer

Psi Chi National Office

Publication and Subscription Office

Daniel P. BockertDirector of PublishingPsi Chi National Office

P.O. Box 709Chattanooga, Tennessee 37401- 0709

Telephone: (423) 756-2044Fax (toll-free): 1-877-PsiChi3 (1-877-774-2443)

E-mail: [email protected]


FALL 2001 VOLUME 6, NUMBER 3

91 Sex Differences in the Contributions of Appearance-Related Messages toBody Esteem and Perceptions of ThinnessCarrie L. Giant, Rhonda M. Passino, and Lesa Rae VartanianIndiana University–Purdue University Fort Wayne

101 Effects of High and Low Saturation of Red Hue on Long-TermMemory PerformanceAni G. Nikolova and Jill L. QuiliciCalifornia State University, Northridge

109 Effects of DNA and Eyewitness Evidence on Juror DecisionsNaomi J. Freeman and Diana L. D. PunzoEarlham College

118 Part-Set Cuing is Due to Strong, Not Weak, List CuesMonica L. Boice and Gary J. GarganoSaint Joseph’s University

123 Human Sex Differences in Aggression Within an Evolutionary ModelBarbara J. Hagenah, Christopher Heaps, Eugene Gilden, and Michael RobertsLinfield College

133 Reviewers 134 Psi Chi Research Awards and Grants 136 Subscriptions

Published quarterly by Psi Chi, The National Honor Society in Psychology

90 PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001


Fall 2001Volume 6, Number 3

Published quarterly byPsi Chi, The National HonorSociety in Psychology

Journal Purpose StatementThe twofold purpose of the Psi Chi

Journal of Undergraduate Research is tofoster and reward the scholarly efforts ofundergraduate psychology students aswell as to provide them with a valuablelearning experience. The articles pub-lished in this journal represent primarilythe work of the undergraduate student(s).Faculty supervisors, who deserve recog-nition, are identified by an asterisk next totheir name or on a separate byline.

Since the articles in this journal areprimarily the work of undergraduatestudents, the reader should bear in mindthat: (1) the studies are possibly lesscomplex in design, scope, or samplingthan professional publications and (2)the studies are not limited to significantfindings. The basis for accepting papersfor publication is the agreement amongthree professional reviewers that theproject, hypothesis, and design are wellresearched and conceived for someonewith an undergraduate level of compe-tence and experience.

Instructions for ContributorsThe Psi Chi Journal of Undergraduate

Research encourages undergraduatestudents to submit manuscripts for con-sideration. Submissions are accepted forreview on an ongoing basis. Althoughmanuscripts are limited to empiricalresearch, they may cover any topical areain the psychological sciences.1. The primary author of a submitted

manuscript must be an undergraduatestudent who is a member of Psi Chi.Manuscripts from graduate studentswill be accepted only if the work wascompleted as an undergraduate stu-dent. Additional authors other thanthe primary author may include non–Psi Chi students as well as the facultymentor or supervisor. Membershipverification information for the pri-mary author must be included.

2. Only original manuscripts (not pub-lished or accepted for publicationelsewhere) will be accepted.

3. All manuscripts must be preparedaccording to the Publication Manual ofthe American Psychological Association(5th ed.).

4. What to submit:a. Four copies of the complete manu-

script. Near-letter-quality print isrequired on all copies. Should youdesire a masked (blind) review,make sure that identifying names,

affiliations, etc. appear only on thetitle page and nowhere else on themanuscript; i.e., manuscripts shouldbe reasonably free of clues to theidentity of the authors. Footnotesthat identify the author(s) shouldappear on a separate page. You mustrequest masked review.

b. An e-mail address so that receiptof your manuscript can beacknowledged.

c. A self-addressed envelope withsufficient postage for the return ofyour manuscripts when an editorialdecision has been reached.

d. A sponsoring statement from thefaculty supervisor who attests: (1)that the research adhered to APAethical standards; (2) that the super-visor has read and critiqued themanuscript on content, method,APA style, grammar, and overallpresentation; and (3) that the plan-ning, execution, and writing of themanuscript represent primarily thework of the undergraduate student.

Submit all manuscripts to:Dr. Stephen F. Davis, Managing EditorPsi Chi Journal of Undergraduate Research649 Hideaway Lane EastLindale, TX 75771-5301

The Psi Chi Journal of UndergraduateResearch (ISSN 1089-4136) is publishedquarterly in one volume per year by PsiChi, Inc., The National Honor Society inPsychology, P.O. Box 709, Chattanooga,TN 37401-0709.

Subscriptions are available on acalendar-year basis only (Spring–Winter).U.S. rates are as follows (four issues):Individual $20; Institution $40. For inter-national rates or other information contact:Daniel P. Bockert, Director of Publishing,Psi Chi National Office, P.O. Box 709,Chattanooga, TN 37401-0709; telephone(423) 756-2044; fax (toll-free) 1-877-774-2443; e-mail journal@ psichi.org. Printed inthe USA. Periodicals postage paid atChattanooga, TN, and additional mailingoffices.

Statements of fact or opinion arethe responsibility of the authors aloneand do not imply an opinion on the partof the officers or members of Psi Chi.Copyright 2001 by Psi Chi, The NationalHonor Society in Psychology. Postmaster:Send address changes to Psi Chi Journal ofUndergraduate Research, P.O. Box 709,Chattanooga, TN 37401-0709.

About Psi ChiPsi Chi is the National Honor Society

in Psychology, founded in 1929 for thepurposes of encouraging, stimulating,and maintaining excellence in scholar-ship, and advancing the science of psy-chology. Membership is open to graduateand undergraduate men and women whoare making the study of psychology oneof their major interests and who meet theminimum qualifications. Psi Chi is amember of the Association of CollegeHonor Societies (ACHS) and is anaffiliate of the American PsychologicalAssociation (APA) and the AmericanPsychological Society (APS). Psi Chi’ssister honor society is Psi Beta, the na-tional honor society in psychology forcommunity and junior colleges.

Psi Chi functions as a federation ofchapters located at more than 990 seniorcolleges and universities. The Psi ChiNational Office is located in Chatta-nooga, Tennessee. A National Council,composed of psychologists who are PsiChi members and who are elected by thechapters, guides the affairs of the organi-zation and sets policy with the approvalof the chapters.

Psi Chi serves two major goals—oneimmediate and visibly rewarding to theindividual member, the other slower andmore difficult to accomplish, but offeringgreater rewards in the long run. The firstof these is the Society’s obligation toprovide academic recognition to itsinductees by the mere fact of member-ship. The second goal is the obligation ofeach of the Society’s local chapters tonurture the spark of that accomplish-ment by offering a climate congenial toits creative development. For example,the chapters make active attempts tonourish and stimulate professionalgrowth through programs designed toaugment and enhance the regular cur-riculum and to provide practical experi-ence and fellowship through affiliationwith the chapter. In addition, the na-tional organization provides programs tohelp achieve these goals, including na-tional and regional conventions heldannually in conjunction with the psycho-logical associations, research award andgrant competitions, certificate recognitionprograms, national and regional chapterawards, and national service projects.

For more information about Psi Chi,contact the Psi Chi National Office, P.O.Box 709, Chattanooga, TN 37401-0709;telephone (423) 756-2044; www.psichi.org.

PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001 91Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 91–100 / ISSN 1089 -4136). *Faculty Supervisor

Sex Differences in the Contributions ofAppearance-Related Messages to BodyEsteem and Perceptions of Thinness

Past research has indicated that satisfaction with one’s body shape is negativelycorrelated with mass media influence and appearance-related feedback fromfamily and peers. The present research investigated sex differences in the contri-butions of those sociocultural factors to body esteem and perceptions of thinness.One hundred sixty-six women and 113 men completed questionnaires measuringbody esteem, perceptions of current body shape compared to perceptions of desiredand perceived ideal body shapes, extent of susceptibility to mass media modelinfluence, frequency of teasing from family and peers, and frequency of body-related discussions with peers. Correlations among these variables indicated sexdifferences, and reinforced the importance of sociocultural influences on percep-tions of the body for both men and women.

CARRIE L. GIANT

RHONDA M. PASSINO

LESA RAE VARTANIAN*

Indiana University–Purdue University Fort Wayne

THE OCCURRENCE OF LOW BODY ESTEEM SEEMS

to have increased over the past 20 years(Rieves & Cash, 1996). Low body esteem can

lead to excessive dieting and exercise, which in turncan lead to chronic weight problems, growth distur-bances, and eating disorders (Davis, Kennedy,Ravelski, & Dionne, 1994; Pugliese, Lifshitz, Grad,Fort, & Marks-Katz, 1983). In fact, as the prevalenceof eating disorders is on the rise (Nagel & Jones,1992), the “ideal” body portrayed in the media hasbeen getting thinner (Wiseman, Gray, Mosimann, &Ahrens, 1992). Because of these patterns, research-ers have been actively investigating factors that maycontribute to low body esteem.

Most of the studies regarding body esteem andbody image only include women as participants. Onereason for the wide exclusion of men is that womenusually have lower body esteem than do men (e.g.,Grogan, Williams, & Conner, 1996; Miller et al., 2000)and perceive their own body shape as more dissimilarto their desired body shape than do men (McKinley,1998; Silberstein, Striegel-Moore, Timko, & Rodin,1988). Additionally, women outnumber men when itcomes to eating disorders by a ratio of 10 to 1 (Crisp,1980). For women in the United States, thinness has

become almost synonymous with beauty (Thompson,1990). Furthermore, women are more popularcandidates for body esteem research than are menbecause researchers believe that women experiencegreater societal pressures to be a particular sizeand shape, which is believed to affect their bodyesteem negatively (Delaney, O’Keefe, & Skene, 1997;Ogletree, Williams, Raffeld, Mason, & Fricke, 1990;Rothblum, 1990; Silverstein, Perdue, Peterson, &Kelly, 1986).

Sociocultural factors are thought to be the stron-gest influences on the development of low bodyesteem (Thompson, 1992). Researchers have beeninvestigating the influence of the mass media as a

Author note. Carrie L. Giant and Rhonda M. Passino, undergradu-ate students, Department of Psychology; Lesa Rae Vartanian,Assistant Professor, Department of Psychology.

Portions of these findings were presented at the 14th BiennialConference on Human Development (Memphis, Tennessee, April14–16, 2000), and the annual meeting of the Midwestern Psycho-logical Association (Chicago, Illinois, May 4–6, 2000).

Correspondence concerning this article should be addressedto Lesa Rae Vartanian, PhD, Department of Psychology, IndianaUniversity–Purdue University Fort Wayne, Neff Hall 380D, 2101Coliseum Boulevard East, Fort Wayne, IN 46805. Electronic mailmay be sent to [email protected], and faxes may be sent to (219)481-6972.


APPEARANCE-RELATED MESSAGES AND BODY ESTEEM � Giant, Passino, and Vartanian

sociocultural factor that contributes to low bodyesteem. Societal messages portrayed in the massmedia help to define physical attractiveness andunattractiveness for the general public (Cash &Pruzinsky, 1990). However, these societal messagesoften portray an ideal that is unachievable (Fallon,1990). Research has shown that individuals comparetheir own bodies to mass media models (Grogan etal., 1996). Sixty percent of girls regularly read at leastone magazine that they consider to be an importantsource for information about ideal shape, diet, fitness,and beauty (Levine, Smolak, & Hayden, 1994). Inaddition, 22% of girls who regularly read magazinesreport high interest in imitating the fashion models(Levine et al., 1994). Researchers posit that the massmedia restricts the range of physical beauty, promotesslenderness as the ideal for beauty and fitness, pro-motes slenderness as the path to social, sexual, andoccupational success, and portrays fat as negative(Levine & Smolak, 1996). In a recent study, womenviewed photographs of “ideal” (i.e., slender, attrac-tive) female models to determine if viewing thephotographs had an effect on their self-evaluations(Henderson-King & Henderson-King, 1997). Resultsindicated that after viewing the models in the photo-graphs, thinner women evaluated their sexual attrac-tiveness more positively, whereas heavier womenevaluated their attractiveness more negatively, particu-larly when their body shape was more dissimilar tothe “ideal” body shape. Additional research indicatedthat women who viewed fashion magazines contain-ing images of the thin ideal body immediately feltless satisfied with their own bodies and desired tobe more thin, than did women who viewed news-magazines (Turner, Hamilton, Jacobs, Angood, &Dwyer, 1997). Similar research has shown that womenwho view magazines that contain more images of thethin ideal body are more dissatisfied with their currentbodies (Stice, Schupak-Neuberg, Shaw, & Stein, 1994).

Other factors that contribute to low body esteemamong women are teasing about weight or bodyshape and other appearance-related feedback fromfamily and friends (e.g., Levine et al., 1994). Nega-tive appearance-related comments from family andfriends are thought to be internalized and thus torelate to negative perceptions of one’s body. Appear-ance-related criticism as recalled during childhoodis significantly related to increased dysfunctional bodyimage in adulthood for women (Rieves & Cash, 1996).Women most often identified “peers in general” asthe criticizers (62%), followed by friends (47%) andbrothers (41%). In a similar study, researchers inves-tigated three sociocultural influences—parents,peers, and magazines—as possible factors associated

with girls’ body satisfaction (Levine et al., 1994).Results showed that direct parental pressure frommother and father to be thin correlated positively withweight concerns and dieting behaviors. Additionally,criticism from family members about weight or shape,body-related discussions with peers, and susceptibil-ity to magazine model influence significantly contrib-uted to body dissatisfaction and dieting behaviors.Results indicated that 42% of the girls reported talk-ing with their friends about weight, shape, and diet-ing at least sometimes. In fact, some researchers sug-gest that peers may have greater influence on bodyesteem than mass media models, perhaps becauseverbal feedback from peers has a more direct impactthan viewing mass media models (e.g., Cash, Cash, &Butters, 1983). Thus, exposure to appearance-relatedfeedback from family members and peers and sus-ceptibility to mass media model influence appear toplay significant roles in the development of girls’ andwomen’s body esteem.

Although researchers have had empirical supportfor focusing on women, new findings suggest that menshould also be included in studies on body esteemand body image. The prevalence of anabolic steroiduse suggests that men are also dissatisfied with theirbodies. One out of 15 male high school seniors reporttaking anabolic steroids, and 27% of those youngmen report taking steroids to improve their physicalappearance (Fackelmann, 1988). Additionally, the1999 “Monitoring the Future” study by the NationalInstitute of Drug Abuse (NIDA) indicated that ana-bolic steroid use among young men has increasedfrom 1.8% in 1991 to 2.7% in 1999 (NIDA, 2000).Grogan et al. (1996) claimed that men should alsobe included in studies on body esteem because West-ern attitudes about the male body are changing, andmen are becoming increasingly concerned with bodyimage. Whereas thinness appears to be vogue forwomen, men appear to be concerned with being toothin (i.e., “scrawny”). For instance, one study indi-cated that whereas 35% of men desired to be thin-ner, 43% desired to be heavier (Silberstein et al.,1988). Although recent trends indicate that men in-creasingly have body image concerns, some research-ers posit that men’s body esteem would not be affectedby portrayals of idealized male bodies because menare less likely to make appearance comparisons(Heinberg & Thompson, 1992).

In order to develop and strengthen the emerg-ing evidence that men also experience body imageconcerns, some recent studies have included men intheir samples. For instance, Grogan et al. (1996) ex-amined the effects of viewing same-sex photographicmodels on women’s and men’s body esteem. In this

Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 91–100 / ISSN 1089 - 4136).

PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001 93


study, two groups viewed one of two types of photo-graphs: attractive models or landscapes. The femalemodels were thin and well proportioned, whereas themale models were well toned and muscular. The par-ticipants also completed a body esteem scale beforeand after viewing the photographs. Body esteemscores before viewing the photographs revealed thatwomen, on average, had lower body esteem than men.However, after viewing the attractive models, all par-ticipants demonstrated a significant decrease in bodyesteem. In other words, this study demonstrated thatboth women’s and men’s body esteem is affected simi-larly by viewing attractive models. Subsequently, asGrogan et al. also point out, these findings do notsupport Heinberg & Thompson’s (1992) notion thatmen’s body esteem would not be negatively affectedby exposure to mass media models. Grogan et al. offerone study to the body esteem literature on men, andit only focuses on the influence of mass media mod-els. The subsequent step is to further examine men’sbody image concerns by investigating the contribu-tions of multiple sociocultural factors to men’s bodyesteem.

The purpose of the present study was to extendprevious research on body esteem by further inves-tigating sex differences in body esteem and theirrelation to susceptibility to mass media models andappearance-related feedback from family and peers.Past research has shown that when considering bodyesteem alone women tend to have lower body esteemthan men (e.g., Grogan et al., 1996). Thus, a predic-tion of the current study is that women’s body esteemwill be lower on average than men’s. In addition,past research has shown that women’s body esteemis negatively affected when women are exposed tomass media models or to body-related criticism fromfamily and peers (e.g., Levine et al., 1994; Stice et al.,1994; Turner et al., 1997), and that men’s body esteemis negatively affected when men are exposed to massmedia models (Grogan et al., 1996). Recent findingshave also indicated that men are increasingly con-cerned about their bodies and that men’s andwomen’s body esteem is similarly affected by expo-sure to mass media models (Grogan et al., 1996).Thus, another prediction of this study is that a highsusceptibility to mass media models, as well as ahigh rate of appearance-related feedback fromfamily and peers, will be similarly associated with lowerbody esteem for both men and women. Additionally,a high influence of sociocultural factors is predictedto be associated with both men and women similarlyperceiving their own body shape as more dissimilarto their desired and perceived societal ideal bodyshapes.

MethodParticipants

Participants for this study were 279 studentsenrolled in introductory psychology courses at amidsized midwestern commuter campus. The sampleconsisted of 166 women and 113 men with a meanage of 22.6 years (range: 17 to 48 years, SD = 6.85).The sample was predominantly Caucasian (88.1%),with 5.4% African Americans, 2.9% Hispanic Ameri-cans, 2.2% multiracial, 1.0% Asian Americans, and0.3% Native Americans. The researchers recruitedparticipants through posted sign-up sheets that indi-cated the topic under investigation, the procedureto be used, and possible risks of completing theproject. The researchers treated the participants inaccordance with APA guidelines for ethical treatmentof human participants (APA, 1992).

MeasuresFor the purposes of this study, the researchers

created a Sociocultural Factors Questionnaire. Al-though most items were adapted from subscale itemscreated by Levine et al. (1994), the researchers addednew items in order to obtain specific informationregarding physical fitness, same-sex versus opposite-sex peer teasing and dieting, and childhood versuscurrent appearance-related feedback. The originalsubscales by Levine et al. were Peer Diet (3 items),Peer Teasing (1 item), Family Teasing (5 items), Maga-zine Information (7 items), and Model (8 items). Thefinal measure contained 42 items, which assessed (a)the extent to which appearance-related mass mediamessages influence participants’ notions about bodyimage, (b) the extent to which participants receiveappearance-related compliments or criticism fromfamily and peers, and (c) the extent to which partici-pants’ peers express body image concerns. Becausenew items were added to preexisting subscales, theresearchers completed a new factor analysis. Thisexploratory factor analysis supported the existenceof three distinct factors among 31 items that demon-strated acceptable internal consistency: Mass MediaSusceptibility (α = .85), Family and Peer Teasing (α =.84), and Peers with Body Concerns (α = .73).Appendix A displays those items within each factor.Depending on the nature of the item, the participantchooses how many individuals he or she knows whofit the question asked (from 1 = none to 5 = all ), howoften the item occurs (from 1 = never to 5 = all thetime), or how much he or she agrees with the item(from 1 = disagree to 5 = agree). Higher scores on eachsubscale reflect a greater susceptibility to mass mediamodel influence, more appearance-related teasingfrom family and peers, and a higher frequency of



exposure to discussions with peers about body con-cerns, respectively.

The researchers administered the Body EsteemScale for Adolescents and Adults (Mendelson, White,& Mendelson, 1997) to obtain individuals’ percep-tions and feelings about their bodies. The question-naire contains 20 items, which Mendelson et al.(1997) found to have strong test–retest reliability andconvergent validity. Examples of those items are “I’mhappy about the way I look” and “I think about waysto change my weight.” Participants choose how oftenthey feel the way the item describes (from 1 = never to5 = all the time). Three of the original 20 items weredropped in the present study to maximize internalconsistency. One average score was utilized from theremaining 17 items. The scale’s internal consistency,as measured by the alpha coefficient, was .95. Higherscores reflect more positive feelings about one’s body,or in other words, higher body esteem.

For the purposes of this study, the researcherscreated a Body Shape Questionnaire to assess picto-rially individuals’ body shape perceptions in terms ofthinness and roundness (see Appendix B). The re-searchers selected the male and female figure draw-ings from a larger continuum of figure drawings aspresented by Stunkard, Sorensen, and Schulsinger(1983). Participants selected from a series of five fig-ure drawings differing in thinness: (a) their perceivedcurrent body shape, (b) their desired body shape, and(c) the body shape they believe the media, family,and friends expect them to have (i.e., perceived idealbody shape). The actual number of the figure cho-sen for each item was retained as the score for thatitem, and thus could range from 1 to 5. A higher scorereflects that the participant selected a thinner figure,whereas a lower score reflects that the participantselected a rounder figure. In order to assess the ex-tent to which participants’ desired and perceived idealbody shapes differed from their current body shape,two difference scores were created (c.f. McKinley,1998; Silberstein et al., 1988). These scores were cal-culated by subtracting the participants’ own perceivedcurrent body shape figure from their desired bodyshape figure and their perceived current body shapefigure from their perceived ideal body shape figure.The absolute values of the resulting numbers werethen obtained so that discrepancies in one directionof desiring to be thinner did not cancel out discrep-ancies in the other direction of desiring to be rounder(c.f. Silberstein et al., 1988). The absolute values re-flect the degree, or amount, of body change desired.Therefore, larger absolute values reflect greater de-sired body change, whereas smaller absolute valuesreflect less desired body change. The following analy-

ses only utilize the absolute values of the two discrep-ancy scores.

ProcedureParticipants completed the packet of question-

naires in groups of 10 to 30 individuals. The re-searcher informed the participants that the studyconcerned how individuals feel and think about theirbodies, how individuals feel about and perceive massmedia models, and how often family and friends pro-vided appearance-related comments. The researcherdistributed blank envelopes containing the abovethree questionnaires, presented in random order.When the packets were completed, the researcherinstructed the participants to put all materials backinto the envelope, close the envelope, and turn it in.Thus, participants’ responses were anonymous. Ses-sions lasted approximately 45 min.

ResultsSex Differences in Mean Body Esteem and BodyShape Discrepancy Scores

A t test for independent samples was performedon men’s and women’s body esteem scores in orderto determine the existence of sex differences. Thetest indicated a significant difference between thebody esteem means for men (M = 3.39, SD = .64) ver-sus women (M = 2.91, SD = .85), t(273) = 4.99, p <.001. In addition, t tests for independent samples werecompleted for both body shape discrepancy scores.The test for the absolute value of the desired-currentdiscrepancy indicated no significant difference be-tween the means of men (M = .57, SD = .58) versuswomen (M = .67, SD = .63), t(279) = –1.38, p > .05.Likewise, for the absolute value of the perceived ideal-current discrepancy, the test also indicated no signifi-cant difference between the means of men (M = .77,SD = .63) versus women (M = .84, SD = .77), t(279) =–.78, p > .05. Therefore, men and women report re-quiring a similar amount of body change to matchtheir desired and perceived ideal body shapes.

The Contributions of Sociocultural Factors toBody Esteem and Perceptions of Thinness by Sex

Correlations between body esteem and the socio-cultural factors were investigated separately for menand women. Both men’s and women’s body esteemwas found to be significantly negatively correlatedwith susceptibility to mass media model influence,appearance-related teasing, and exposure to discus-sions with peers about body concerns (see Table 1for correlations by sex). Although both sets of corre-lations for men and women were significant and inthe same direction, Fisher’s r to z transformations




(Daniel, 1995) were utilized to determine if theirmagnitudes differed significantly. A significant sexdifference was found between body esteem and sus-ceptibility to mass media model influence, z = –3.29,p < .05. However, no sex differences were indicatedbetween body esteem and appearance-related teasing,z = –1.26, p > .05, or between body esteem and expo-sure to discussions with peers about body concerns,z = –1.76, p > .05. Thus, the current findings indi-cated that appearance-related teasing and exposureto discussions with peers about body concerns simi-larly affect men’s and women’s body esteem. How-ever, the findings did indicate that women’s body es-teem is related more strongly to susceptibility to massmedia model influence than is men’s body esteem.

In order to examine whether the relations be-tween the sociocultural factors considered and theabsolute values of the two body shape discrepancyscores differed for men versus women, zero-ordercorrelations among the two body shape discrepancyscores and the three sociocultural factors were con-ducted separately for men and women (see Table 1for the correlations by sex). The analysis yielded onlytwo pairs of correlations indicating that men andwomen might be similarly affected—those correla-tions between appearance-related teasing from fam-ily and peers and the absolute value of the desired-current discrepancy (for men, r[111] = .23, p < .05;for women, r[165] = .42, p < .001) and the absolutevalue of the perceived ideal-current discrepancy (formen, r[109] = .31, p < .05; for women, r[165] = .37,p < .001). Thus, the higher degree of teasing menand women experience, the greater amount of bodyshape change necessary to match their desired and

perceived ideal body shapes. Although these four cor-relations for men and women were statistically sig-nificant and in the same direction, Fisher’s r to z trans-formations (Daniel, 1995) were again utilized to de-termine significant differences among the magni-tudes. A significant sex difference was indicated forthe relation between appearance-related teasing andthe absolute value of the desired-current discrepancy,z = –2.22, p < .05. Thus, the magnitude of the relationfor women is stronger than for men. However, no sexdifference was indicated for the relation betweenappearance-related teasing and the absolute value ofthe perceived ideal-current discrepancy, z = –.70, p >.05, indicating that the perceived ideal-current dis-crepancy scores of both sexes were affected similarlyby appearance-related teasing. All other correlationsfor men and women indicated that women’s discrep-ancy scores were affected by the three socioculturalfactors, whereas the men’s discrepancy scores werenot. This sex difference was revealed by the non-significant correlations for men and the significantcorrelations for women. Therefore, a greater extentof susceptibility to mass media model influence andexposure to discussions with peers about body con-cerns was related to women needing a greater amountof body change to match their desired and perceivedideal body shapes. However, higher levels of the samesociocultural influences were not related in a similarway for men.

DiscussionThe present study attempted to replicate and

extend previous research findings by investigating theimpact of multiple sociocultural factors simulta-


TABLE 1

Zero-Order Correlations Among the Variables of Interest Separately for Men and Women

Frequency of Frequency of Degree ofappearance-related discussions with peers susceptibility to

teasing about body concerns media models

Body measures Men Women Men Women Men Women

Body esteem –.40** –.48** –.30* –.42** –.32** –.53**

Desired–currentdiscrepancy scorea .23* .42** .04 .27** .03 .26*

Perceived ideal–currentdiscrepancy scorea .31* .37** .06 .24* –.01 .20*

aRepresents the absolute value of the discrepancy score.*p < .05, two-tailed. **p < .001, two-tailed.



neously on men’s and women’s body esteem and onperceptions of current body thinness versus percep-tions of desired and perceived ideal body thinness.First, we predicted that women’s body esteem wouldbe lower on average than men’s body esteem. Thishypothesis was affirmed as the current analysis indi-cated that women’s body esteem was, in fact, loweron average than men’s. These present results con-cerning sex and body esteem replicated past research(Grogan et al., 1996; Miller et al., 2000). Althoughmen are becoming increasingly concerned with theirbody image, as shown by increases in anabolic ste-roid use (NIDA, 2000) and by recent research por-traying negative effects on men’s body esteem due toviewing attractive models (Grogan et al., 1996), mencontinue to score higher, on average, on measures ofbody satisfaction. However, although a sex differenceexisted for overall body esteem, no sex differenceswere found among the absolute values of the bodyshape discrepancy scores. Because men desire a simi-lar amount of body change with respect to percep-tions of thinness as do women in order to match theirdesired and perceived ideal body shapes, the notionthat men have increasing body image concerns wasreinforced. Men desire to change their bodies just asmuch as women, but men do not report as low a bodyesteem as do women. This result can probably beexplained by the notion that societal pressures onwomen to be a particular size and shape are stillmore pronounced than those pressures on men(Rothblum, 1990), and that women feel more pres-sure to conform to those societal ideals (Ogletree etal., 1990). Therefore, when a woman believes herbody shape does not conform to her desired or per-ceived ideal body shapes, she is likely to feel moredissatisfied with her body than would a man who hasthe same beliefs. Furthermore, there may be a widerrange of acceptable body shapes for men, which mayenable them to feel more satisfied with their bodieseven when their current body shape does not matchtheir desired or perceived ideal body shapes.

Second, the researchers hypothesized that a highsusceptibility to mass media model influence and ahigh rate of appearance-related feedback from fam-ily and peers would be similarly associated with lowerbody esteem for both men and women. The currentresults affirmed this hypothesis, as lower body esteemwas associated with higher levels of susceptibility tomass media models, exposure to discussions withpeers about body concerns, and appearance-relatedteasing from family and peers for both men andwomen. These results are consistent with past researchthat has indicated that women report lower bodyesteem when they also report higher levels of suscep-


tibility or exposure to sociocultural influences(Henderson-King & Henderson-King, 1997; Levineet al., 1994; Rieves & Cash, 1996; Stice et al., 1994;Turner et al., 1997) and that men report lower bodyesteem after exposure to attractive models (Groganet al., 1996). Thus, there is strong evidence support-ing the proposition that sociocultural factors play asignificant role in individuals’ perceptions and feel-ings about their bodies. It appears that individualsare internalizing the criticisms and other negativeappearance-related feedback from close familymembers and friends, as well as mass media messagesabout the ideal body shape. Although all three socio-cultural factors negatively affected both men’s andwomen’s body esteem, women’s body esteem wasmore strongly related to susceptibility to mass mediamodel influence. This result is consistent with thenotion that women face more societal pressures tobe thin (Ogletree et al., 1990). However, this findingis inconsistent with past findings indicating that men’sand women’s body esteem is affected similarly byportrayals of ideal body shapes (Grogan et al., 1996),as the present findings indicated that women wereaffected more profoundly than men. However, men’sand women’s body esteem was affected similarly byappearance-related teasing from family and friendsand by exposure to discussions with peers about bodyconcerns. Although low body esteem may not appearon the surface to be a problematic issue for men, theirbody esteem is significantly affected by these socio-cultural influences that idealize a thin body. AlthoughHeinberg & Thompson (1992) suggested that men’sbody esteem would not be negatively affected by ex-posure to body-related sociocultural influences, thepresent results indicated that men’s body esteem wasinversely related to all three sociocultural factors.With respect to exposure to peers discussing body con-cerns and appearance-related teasing, the notion thatmen’s body esteem is affected similarly to women’sbody esteem by sociocultural factors was replicatedin the present study. These body image concerns ofmen may play a significant role in the recent increaseof anabolic steroid use among men (Fackelmann,1988; NIDA, 2000). Men also appear to have bodyimage concerns and should be included in futurestudies on body esteem and body image.

Lastly, to further extend the findings of past re-search, this study also investigated socioculturalinfluences of discrepancies between men’s andwomen’s perceptions of current body thinness ver-sus perceptions of desired body thinness and the bodyshape they believe most people want them to have interms of thinness (i.e., perceived ideal body shape).We predicted that a high susceptibility to mass media



model influence and a high rate of appearance-related feedback from family and peers would beassociated with both men and women perceiving theirown body thinness as more dissimilar to their desiredand perceived ideal body shapes. However, the dataonly partially affirmed this hypothesis; just one set ofcorrelations indicated that men and women were simi-larly affected. Higher rates of appearance-related teas-ing from family members and friends was the onlysociocultural factor associated with greater discrep-ancies between both men’s and women’s current bodyshapes and their desired and perceived ideal bodyshapes. This finding is consistent with the Cash et al.(1983) notion that appearance-related informationfrom peers has a greater effect on body dissatisfac-tion than do mass media models. Apparently, theappearance-related criticisms from family membersand peers have a very direct and salient impact onone’s internalized perceptions about the body. Theremaining correlations among the socioculturalfactors and the body shape discrepancy scores indi-cated the following sex differences: Women’s desired-current discrepancies were more profoundly affectedby high rates of appearance-related teasing than weremen’s desired-current discrepancies. Furthermore,whereas women’s desired-current and perceived ideal-current discrepancies were significantly related tohigh rates of exposure to discussions with peers aboutbody concerns and susceptibility to mass media modelinfluence, men’s body shape discrepancies were not.These current findings indicate that women desirechanges in body thinness when confronted withexposure to mass media models and body-concernedpeers, whereas men do not.

These results suggest that low body esteem andbody image concerns remain problematic for womento a greater extent than for men. However, recentresearch (e.g., Grogan et al., 1996) and the presentresults also indicate that men have some body imageconcerns when considered in relation to certainsociocultural influences (i.e., exposure to pictures of“ideal” male bodies, appearance-related teasing). Thepresent findings reinforce the prevalent notion thatsociocultural factors strongly affect body esteemand body shape preferences, especially for women.Finally, although women’s body concerns appear tobe associated to a greater extent with socioculturalinfluences, men seem to have increasing concernsabout body image and body shape.

The present study is not without limitations. Forexample, the lack of cultural diversity in the predomi-nately Caucasian sample made it impossible for theresearchers to adequately examine ethnic or racialdifferences. Ethnic or racial differences should be

considered when possible, as some research has indi-cated that ethnic traditions are important elementsof beliefs about body shapes (e.g., Collins, 1991;Mumford, Whitehouse, & Platts, 1991; Rucker & Cash,1992). Furthermore, additional research might con-sider body image and body esteem along other di-mensions of body shape besides thinness (i.e., physi-cal fitness, muscularity) that may be related to socio-cultural factors. Despite these limitations, the presentfindings replicate past research indicating thatwomen’s body esteem is negatively affected by mul-tiple sociocultural influences and also extend pastresearch by demonstrating that men’s body esteem isalso affected by multiple sociocultural influences.Whereas some sex differences exist among the asso-ciations of body esteem and body image with socio-cultural influences, in some respects men and womenare affected similarly, particularly when it comes tothe influence of interpersonal factors, such as teas-ing and exposure to peers’ dissatisfactions with theirown bodies. Thus, a better understanding of howindividuals view and feel about their physical selveswill likely result from studies that include men as wellas women.

ReferencesAmerican Psychological Association. (1992). Ethical principles of

psychologists and code of conduct. American Psychologist, 47,1597–1611.

Cash, T. F., Cash, D. W., & Butters, J. W. (1983). “Mirror, mirror,on the wall . . .?”: Contrast effects and self-evaluations of physi-cal attractiveness. Personality and Social Psychology Bulletin, 9,351–358.

Cash, T. F., & Pruzinsky, T. (Eds.). (1990). Body images: Develop-ment, deviance, and change. New York: Guilford Press.

Collins, M. E. (1991). Body figure perceptions and preferencesamong preadolescent children. International Journal of EatingDisorders, 10, 199–208.

Crisp, A. H. (1980). Anorexia nervosa: Let me be. New York: Grune& Stratton.

Daniel, W. W. (1995). Biostatistics: A foundation for analysis in thehealth sciences (6th ed.). New York: Wiley.

Davis, C., Kennedy, S. H., Ravelski, E., & Dionne, M. (1994). Therole of physical activity in the development and maintenanceof eating disorders. Psychological Medicine, 24, 957–967.

Delaney, M. E., O’Keefe, L. D., & Skene, K. M. L. (1997). Devel-opment of a sociocultural measure of young women’s experi-ences with body weight and shape. Journal of Personality Assess-ment, 69, 63–80.

Fackelmann, K. A. (1988, December 17). Male teenagers at riskof steroid abuse. Science News, 134, 391.

Fallon, A. (1990). Culture in the mirror: Sociocultural determi-nants of body image. In T. F. Cash & T. Pruzinsky (Eds.), Bodyimages: Development, deviance, and change (pp. 80–109). NewYork: Guilford.

Grogan, S., Williams Z., & Conner M. (1996). The effects of view-ing same-gender photographic models on body-esteem. Psy-chology of Women Quarterly, 20, 569–575.

Heinberg, L. J., & Thompson, J. K.(1992). Social comparison:Gender, target importance ratings, and relation to body imagedisturbance. Journal of Social Behavior and Personality, 7, 335–344.

Henderson-King, E., & Henderson-King, D. (1997). Media effects




on women’s body esteem: Social and individual differencefactors. Journal of Applied Social Psychology, 27, 399–417.

Levine, M.P., & Smolak, L. (1996). Media as a context for thedevelopment of disordered eating. In L. Smolak, M. P. Levine,& R. Striegel-Moore (Eds.), The developmental psychopathology ofeating disorders: Implications for research, prevention, and treatment(pp. 235–257). Mahwah, NJ: Erlbaum.

Levine, M. P., Smolak, L., & Hayden H. (1994). The relation ofsociocultural factors to eating attitudes and behaviors amongmiddle school girls. Journal of Early Adolescence, 14, 471–490.

McKinley, N. M. (1998). Gender differences in undergraduates’body esteem: The mediating effect of objectified body con-sciousness and actual/ideal weight discrepancy. Sex Roles, 39,113–123.

Mendelson, B. K., White, D. R., & Mendelson, M. J. (1997). Manualfor the Body Esteem Scale for Adolescents and Adults. ResearchBulletin, 16(2). (Available from the Centre for Research inHuman Development, 7141 Sherbrooke Street West, Montreal,Quebec, Canada H4B 1R6). [EDITOR’S NOTE—Although thelisting above refers to an internal research bulletin, this scalehas also been recently published: Mendelson, B. K., Mendel-son, M. J., & White, D. R. (2001). The Body-Esteem Scale forAdolescents and Adults. Journal of Personality Assessment, 76,90–106.]

Miller, K. J., Gleaves, D. H., Hirsch, T. G., Green, B. A., Snow, A.C., & Corbett, C. C. (2000). Comparisons of body image di-mensions by race/ethnicity and gender in a university popu-lation. International Journal of Eating Disorders, 27, 310–316.

Mumford, D. B., Whitehouse, A. M., & Platts, M. (1991). Socio-cultural correlates of eating disorders among Asian school-girls in Bradford. British Journal of Psychiatry, 158, 222–228.

Nagel, K. L., & Jones, K. H. (1992). Sociological factors in thedevelopment of eating disorders. Adolescence, 27, 107–113.

National Institute of Drug Abuse. (2000). Anabolic steroid abuse(NIH Publication No. 00-3721) [On-line]. Available: www.nida.gov/ResearchReports/Steroids/AnabolicSteroids.html

Ogletree, S. M., Williams, S. W., Raffeld, P., Mason, B., & Fricke,K. (1990). Female attractiveness and eating disorders: Dochildren’s television commercials play a role? Sex Roles 22, 791–797.

Pugliese, M. T., Lifshitz, F., Grad, G., Fort, P., & Marks-Katz, M.

(1983). Fear of obesity: A cause of short stature and delayedpuberty. The New England Journal of Medicine, 309, 513–518.

Rieves, L., & Cash, T. F. (1996). Social developmental factors andwomen’s body-image attitudes. Journal of Social Behavior andPersonality, 11, 63–78.

Rothblum, E. D. (1990). Women and weight: Fad and fiction. Jour-nal of Psychology, 124, 5–24.

Rucker, C. E., & Cash, T. F. (1992). Body images, body-size per-ceptions, and eating behaviors among African-American andWhite college women. International Journal of Eating Disorders,12, 291–299.

Silberstein, L. R., Striegel-Moore, R. H., Timko, C., & Rodin J.(1988). Behavioral and psychological implications of body dis-satisfaction: Do men and women differ? Sex Roles, 19, 219–232.

Silverstein, B., Perdue, L., Peterson, B., & Kelly, E. (1986). Therole of the mass media in promoting a thin standard of bodilyattractiveness for women. Sex Roles, 14, 519–532.

Stice, E., Schupak-Neuberg, E., Shaw, H. E., & Stein, R. I. (1994).Relation of media exposure to eating disorder symptoma-tology: An examination of mediating mechanisms. Journal ofAbnormal Psychology, 103, 836–840.

Stunkard, A. J., Sorensen, T., & Schulsinger, F., (1983). Use of theDanish adoption register for the study of obesity and thinness.In S. S. Kety, L. P. Rowland, R. L. Sidman, & S. W. Mathysse(Eds.), The genetics of neurological and psychiatric disorders. NewYork: Raven Press.

Thompson, J. K. (1990). Body image disturbance: Assessment and treat-ment. Elmsford, NY: Pergamon Press.

Thompson, J. K. (1992). Body image: Extent of disturbance, asso-ciated features, theoretical models, assessment methodologies,intervention strategies, and a proposal for a new DSM IV diag-nostic category—body image disorder. In M. Hersen, R. M.Eisler, and P. M. Miller (Eds.), Progress in behavior modification(Vol. 28, pp. 3–54). Sycamore, IL: Sycamore Publishing.

Turner, S. L., Hamilton, H., Jacobs, M., Angood, L. M., & Dwyer,D. H. (1997). The influence of fashion magazines on the bodyimage satisfaction of college women: An exploratory analysis.Adolescence, 32, 603–614.

Wiseman, C. V., Gray, J. J., Mosimann, J. E., & Ahrens, A. H. (1992).Cultural expectations of thinness in women: An update. Inter-national Journal of Eating Disorders, 11, 85–89.




APPENDIX A

Subscales and Items from the Sociocultural Factors Questionnaire

Subscale—Teasing/Criticism From Others (α = .84)

1. When you were younger, how often did same-sex peers tease you about being too fat or too thin? 2. When you were younger, how often did same-sex peers tease you about not being muscular and/or

physically fit enough? 3. Now in your life, how often do same-sex peers tease you about being too fat or too thin? 4. Now in your life, how often do same-sex peers tease you about not being muscular and/or physically fit

enough? 5. When you were younger, how often did opposite-sex peers tease you about being too fat or too thin? 6. When you were younger, how often did opposite-sex peers tease you about not being muscular and/or

physically fit enough? 7. Now in your life, how often do opposite-sex peers tease you about being too fat or too thin? 8. Now in your life, how often do opposite-sex peers tease you about not being muscular and/or physically

fit enough? 9. How often does your brother (or brothers) tease you about your weight or body shape?10. How often does your sister (or sisters) tease you about your weight or body shape?11. How often do you get criticized about your weight or body shape by an adult female family member?12. How often do you get complimented about your weight or body shape by an adult female family

member?13. How often do you get criticized about your weight or body shape by an adult male family member?14. How often do you get complimented about your weight or body shape by an adult male family member?

Subscale—Mass Media Influence (α = .85)

1. I try to look like some of the models I see by purchasing the products they advertise. 2. If I diet or try to build muscle, I look to the models in magazines for inspiration (e.g., put their pictures on

a wall in your room). 3. I compare myself to the models I see. 4. I would like to look like the models I see in magazine and TV ads. 5. I feel bad about my own body after seeing attractive models in magazines and on TV. 6. I am motivated to exercise or work out after seeing attractive models in magazines and on TV. 7. Magazines/TV influence my idea of the perfect body. 8. I think the models in magazines/on TV are confident and happy. 9. I think the models are among the best looking people I have ever seen.10. When I read a magazine or watch TV, I enjoy looking at the models in the ads.11. How often do you look at or read magazines?

Subscale—Peers Expressing Body Image Concerns (α = .73)

1. Of the same-sex peers you know, how many would like to be thinner? 2. Of the same-sex peers you know, how many would like to be more muscular/physically fit? 3. Of the same-sex peers you know, how many are dieting/trying to lose weight? 4. Of the same-sex peers you know, how many talk about weight, weight loss, and/or building muscle? 5. Of the opposite-sex peers you know, how many would like to be more muscular/physically fit? 6. How often have you changed your appearance to look like an attractive, same-sex peer?



100 PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 91–100 / ISSN 1089 - 4136).


APPENDIX B

Body Shape Questionnaire

DIRECTIONS: This questionnaire asks for your opinions about the way your own body looks in terms ofthinness/roundness, and how thin/round you think others think people’s bodies should look. Please answerthe questions by looking at the drawings of male and female bodies below, and by circling the number thatgoes with the body that best shows your opinion (even if not exactly).

Questions 1–3 are for WOMEN ONLY:

1. Which of the female bodies above do you think you look like right now?

1 2 3 4 5

2. Which of the female bodies above would you like to look like?

1 2 3 4 5

3. Which of the female bodies above do you think everyone ( friends, TV, magazines) wants you to look like?

1 2 3 4 5

Questions 4–6 are for MEN ONLY:

1. Which of the male bodies above do you think you look like right now?

1 2 3 4 5

2. Which of the male bodies above would you like to look like?

1 2 3 4 5

3. Which of the male bodies above do you think everyone ( friends, TV, magazines) wants you to look like?

1 2 3 4 5


Effects of High and Low Saturation ofRed Hue on Long-Term MemoryPerformance

Little is known about the mechanism of our physiological reactions to color and itsimpact on cognition. Despite that knowledge deficit, many researchers have useddifferent colors as an environmental setting in studies on study and recall contextvariations and memory. The goal of this study was to examine whether manipula-tion of color properties, manipulation of color context, or both factors have aneffect on memory recall. Each of the 2 independent variables—color presented atstudy and color presented at test—had 3 levels: high saturation of red hue, lowsaturation of red hue, and neutral (gray) hue. A total of 180 students wereassigned to 1 of the 9 conditions in a 3 × 3 randomized-groups design. Analysis ofvariance revealed a significant effect of color saturation on memory recall indicat-ing that high saturation, both at time of study and time of test, resulted in betterrecall than a neutral hue or low saturation. The results relating color context andrecall were nonsignificant, demonstrating that a change in color context fromstudy to test did not result in greater forgetting. This finding clearly indicatesthat the manipulation of color saturation is the primary cause for the outcome inmemory recall in this study. Implications for use of color to improve memory in avariety of learning situations are discussed.

ANI G. NIKOLOVA

JILL L. QUILICI*

California State University,Northridge

WE PURPOSELY INCORPORATE COLOR IN MANY

elements of our visual world, and yet wehave not fully examined its potential to

trigger physiological and psychological reactions inindividuals. Researchers have found that the differ-ent properties of color can affect the direction of ouremotional reactions in terms of feelings ( Jacobs &Suess, 1975; Weller & Livingston, 1988), anxiety(Profusek & Rainey, 1987), and arousal or level ofalertness (Batra, Urvashi, & Muhar, 1998; Jacobs &Hustmyer, 1974; Mikellides, 1990; Valdez & Mehra-bian, 1994; Wilson, 1966). Studies have also estab-lished color reactions as a function of personality andpsychopathology (Frank, 1976), reactions to color con-cepts (Adams & Osgood, 1973), and behavior changesdue to color properties (Damhorst & Reed, 1986).However, we know little about the possible effect ofcolor on various cognitive processes such as learn-ing, memory performance, reasoning, and problemsolving.

Why Color Saturation?

Experimenters have typically confounded satu-ration and brightness properties of color in studies

investigating the various effects of color, thus leadingto methodologically weak and sometimes contradic-tory results. In most previous studies (Ainsworth,Simpson, & Cassell 1993; Jacobs & Suess, 1975;Profusek & Rainey, 1987; Wilson, 1966), researchershave studied color by manipulating color hue (i.e.,wavelength) solely and have failed to adequately con-

Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 101–108 / ISSN 1089 - 4136). Faculty Supervisor: Arlene K. Sprague

Author note. Ani G. Nikolova, undergraduate student at the Col-lege of Social and Behavior Sciences, California State University,Northridge (CSU Northridge; now a graduate student in general–experimental psychology); Jill L. Quilici, Department of Psychol-ogy, CSU Northridge.

The current research was conducted as an independent studyby Ani Nikolova at the Department of Psychology of CSUNorthridge. The study was presented at the following professionalmeetings: Psi Chi 6th Annual Psychology Undergraduate ResearchCompetition, CSU Northridge, December 1998 (awarded firstplace); 8th Annual Psychology Undergraduate Research Confer-ence, UCLA, May 1999; 4th Annual Student Research and Cre-ative Activity Symposium, CSU Northridge, November 1999; PsiChi/Western Psychological Association Convention, April 2000(awarded a Regional Research Award).

Correspondence concerning this article should be addressedto Ani G. Nikolova, 18448 Arminta Street Apt #205, Reseda, CA91335, and Jill L. Quilici, Department of Psychology, CaliforniaState University, Northridge, 18111 Nordhoff Street, Northridge,CA 91330-8255. Electronic mail may be sent to [email protected] and [email protected].


trol for its other two properties, brightness (i.e., black-to-white quality) and saturation or chrome (i.e.,purity or vividness). Unlike these studies, Valdez andMehrabian (1994), in their research on the effects ofcolor on emotions, found significant emotional reac-tions (arousal, pleasure, and dominance feelings) toboth brightness and saturation variations of color. Inthis study, they addressed various methodological andconceptual issues by using a wide range of color sam-ples and by providing experimental control in inves-tigating the separate effects of color hue, saturation,and brightness. They used 76 color stimuli of eachhue level, varying as much as possible in brightnessand saturation, and used 250 participants to rate thesamples. Valdez and Mehrabian used Mehrabian’s(1978) verbal report scales for pleasure–displeasure,arousal–nonarousal, and dominance–submissivenessmeasures to assess emotional reactions to the differentproperties of color. The arousal–nonarousal scalecontained eight word pairs (e.g., “frustrated–sad”,“troubled–dull”) that differed with respect to arousalbut were equal on dominance and pleasure measures.The results of their study indicated that less brightand more saturated colors elicit significantly higherarousal reactions in individuals than more brightand less saturated colors, regardless of color hue.Surprisingly, results relating hue and arousal werenonsignificant.

Mikellides (1990) also examined the effects of colorproperties on arousal and obtained similar results.In contrast to Valdez and Mehrabian (1994), Mikel-lides measured arousal by means of EEG and EKGrecordings and manipulated color stimuli by usingrooms painted in red and blue variations of satura-tion. Mikellides also found that chromatic strength(saturation) was the property of color that most af-fected our perceptions of how exciting or calming acolor was. Again, there was no significant relationfound between color hue and physiological arousal.

Profusek and Rainey (1987) examined a similartheme by looking at the effects of rooms painted inred or pink on human anxiety level. They found thatrooms painted in pink were significantly less psycho-logically arousing than rooms painted in red. How-ever, more recently, Ainsworth et al. (1993), who alsoexamined the effects of colored rooms on states ofanxiety and on work performance in an office envi-ronment, came to the conclusion that color, in par-ticular red and blue-green, does not have an effecton the variables studied. These results are not verysurprising because, even though neither of these ex-periments controlled for brightness and saturationdimensions of color, pink hue generally tends to beof higher brightness and lower saturation than the

red and blue-green hues. Evidently, we could explainthe contradiction of these studies by an existing cor-relation between levels of arousal and color bright-ness and/or saturation but not between arousal andhue variations.

If Color and Arousal are Significantly Correlated,Can We Hypothesize That Color and MentalPerformance are Correlated as Well?

Psychologists who have studied the amount ofmental effort required to perform a task have foundthat the amount of mental capacity available varieswith the level of arousal (Kahneman, 1973; Yerkes &Dodson, 1908). In particular, they came to the con-clusion that more mental capacity is available whenarousal is moderately high. At the same time, perfor-mance suffers when arousal level is either too highor too low for the specific task. Consequently, we canassume that if certain variations of color brightnessor saturation have an effect on the state of arousal inindividuals, then color should have an effect on theamount of mental capacity available for performingvarious cognitive tasks. Thus, if we look at the rela-tion between memory performance and color satu-ration, we would expect that memory outcome willfluctuate in response to saturation variations. How-ever, because the relation between hue and arousalis weak (Mikellides, 1990; Valdez & Mehrabian, 1994),we would not anticipate a significant effect of huealterations on memory performance.

Is Memory Performance Affected by ColorProperties or by the Memory of Color Context?

We know little about the exact mechanism of ourphysiological reactions to color and its impact onmemory performance. Despite that lack of knowl-edge, many researchers who have conducted studieson the effect of environmental changes on recall(Dulsky, 1935; Weiss & Margolius, 1954) and recog-nition (Murnane & Phelps, 1993) have used back-ground color as an environmental context. Thus, byvarying the color saturation and brightness of thecards on which paired associates were placed, Dulsky(1935) found that participants recalled less informa-tion when he changed the color background pre-sented in the learning situation for the recall test ascompared to when the color background during re-call remained the same as the background duringstudy. Dulsky interpreted the results as due to thecolor context factor alone (i.e., color backgroundduring learning vs. color background during recall),omitting the possibility of any color property effects.

Weiss and Margolius (1954) found similar resultsin their study. They defined same context as colored

SATURATION OF RED HUE AND MEMORY PERFORMANCE � Nikolova and Quilici


PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001 103Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 101–108 / ISSN 1089 - 4136).

background during both study and recall, andaltered context as either colored background duringstudy and gray color during recall or gray back-ground at study and colored context at test. Accord-ing to their results, recall was better when the col-ored backgrounds remained the same compared towhen they were changed. They also found thatmemory performance was significantly better whenthey had employed color contexts, compared to thecondition in which participants viewed a gray back-ground during study or recall. Although we couldattribute the significance of these results to thestimulating or arousing effect of the colored back-ground compared to the effect of neutral (gray)background, Weiss and Margolius concluded thatforgetting increased due to the change of environ-mental context.

More recently, Murnane and Phelps (1993) foundthat memory recognition also decreases with changesin color context. Thus, they concluded that memoryperformance declines when we test participants inan environmental context (i.e., color context) that isdifferent from the study context, not only when wemeasure memory by recall, but also when we testmemory by recognition.

However, in none of these three studies haveresearchers specified the exact color characteristicsin terms of brightness and saturation. More than that,researchers have not considered the possible stimu-lating or inhibiting physiological effects of certaincolor characteristics on recognition or recall. Appar-ently, they have failed to consider the possible effectsof color properties on mental performance whenchanging the environmental contexts and, thus, havemost likely confounded the results of their studies.Whether performance on memory tasks is affectedby changes in the environmental context or by theproperties of color itself is unclear.

Summary of Research Goals

The main goal of this study was to examinewhether manipulation of color properties, manipu-lation of color context, or both factors have an effecton memory recall. The experiment built on the hy-pothesis that if manipulation of color saturation canincrease or decrease arousal reactions (Mikellides,1990; Valdez & Mehrabian, 1994) and if levels ofarousal and memory performance are significantlycorrelated (Yerkes & Dodson, 1908), then we ex-pected that the alteration of saturation of the envi-ronmental color during study or test would affect theoutcome in memory recall.

Based on previous research, we made the follow-ing hypotheses:

1. There would be significant main effects forcolor at study on memory performance as well as colorat test on memory performance. In particular, we ex-pected that participants viewing the highly arousingcolor stimulus (red hue of high saturation) at the timeof material study would perform significantly betteron the recall task than participants viewing the low-arousing (red hue of low saturation) or neutral(medium gray) color stimulus during study, regard-less of the color at test. We expected the same pat-tern of results for the main effect of the color at test:we hypothesized that students viewing the high-satu-rated color at the time of recall would score signifi-cantly higher on the memory test than studentslooking at the low-saturated or neutral color duringrecall, regardless of the properties of color duringstudy. Simultaneously, because the neutral (gray)color has no saturation and no hue (Munsell ColorSystem, available from the Macbeth division ofKollmorgen Corporation), we anticipated that thelow-saturated color, by decreasing the level of arousal(Valdez & Mehrabian, 1994), would also lowermemory performance. Thus, we hypothesized that therecall scores of students viewing the low-saturatedcolor at study or test would be significantly lower thanthe recall scores of students viewing the neutral color.Because of the lack of saturation value of the neutralcolor, we assumed that its properties would not affectthe initial level of arousal, and thus when looking atthe main effects of color we considered participantspresented with this stimulus during study or test torepresent control groups.

2. There would be no significant contextualinteraction between color at study and color at test.Thus, we hypothesized that there would be no signi-ficant differences found between memory perfor-mance of participants presented with the same colorstimuli during study and test and performance ofparticipants presented with different color stimuliduring study and test.

Color hue was not a variable in this experimentbecause, as we mentioned above, Valdez and Mehrabian(1994) have clearly demonstrated that hue itself doesnot have a significant effect on arousal. Because webuilt our hypotheses on the assumption that memoryperformance and states of arousal are significantlycorrelated, if hue does not have an effect on arousal,then the relation between hue and memory shouldbe weak as well. We did not consider the results ofsome of the studies mentioned above (Murnane &Phelps, 1993; Weiss & Margolius, 1954), which im-plied that color hue accounted for significant changesin memory, because these researchers failed to con-trol for brightness and saturation of the color stimuli.



We chose randomly the red hue from the variationof basic colors (hues) of the visual spectrum.

MethodParticipants

We tested a randomly drawn sample of 180 students.All participants were female undergraduates fromCalifornia State University, Northridge, who partially ful-filled an introductory psychology course requirementby voluntarily participating in the Psychology Depart-ment Participant Pool. We included only female par-ticipants in the study because women show a significantlystronger pattern of reactions to the different proper-ties of color than men do (Valdez & Mehrabian, 1994).

DesignWe randomly assigned participants to one of the

nine treatment combinations, according to a 3 × 3randomized-groups design. Each of the two indepen-dent variables—color presented at the time of studyand color presented at the time of test—had threelevels: high saturation of red hue, low saturation ofred hue, and neutral (medium-gray) hue. The de-pendent variable was memory recall.

Materials and SettingFor each color variation, there were two color-

sample boards of the size 30 cm × 45 cm. We tookall color stimuli from the Munsell System for colornotation (available from the Macbeth division ofKollmorgen Corporation), and we chose them to beof equal brightness. We used the following specifi-cations of color stimuli: Munsell red with saturation(chrome) of 14 and brightness (value) of 4 (high-saturated red), Munsell red with saturation of 4 andbrightness of 4 (low-saturated red), and Munsell graywith brightness of 4 (N4) and no saturation (neutralcolor). We placed the two sample boards for eachtreatment condition on the wall and on the surfaceof a desk, each one at the background of a medium-gray poster. The laboratory room contained no win-dows and was illuminated with fluorescent tubes,which approximated daylight condition.

We used 30 common English words, which werebetween four and six letters in length and containedeither one or two syllables, to create our study list forthe dependent variable. The first author recorded thewords from the study list with an equal 2- s intervalbetween each word by pronouncing all words withthe same volume and tension of voice. Also, we usedpseudoisochromatic plates (Hardy, Rand, & Rittler,1957) for the purpose of detecting defective colorvision in the participants. The color vision test con-sisted of six screening plates. Each plate consisted of

a maximum of two colored symbols or no symbols,located at different positions of the plate. The se-quence of the six plates contained the following com-binations of shapes: (a) a circle and an X, (b) a tri-angle and an X, (c) a circle, (d) no symbols, (e) acircle and an X, and (f) a circle and a triangle.

ProcedureWe assigned participants to one of the conditions

based on their order of appearance and tested themindividually. We seated each student at a desk, ap-proximately 75 cm from the color board positionedon the wall. We kept the environment quiet through-out the experiment and held the room temperatureconstant at 21° C.

Prior to the experiment, we tested all participantsfor defective color vision. The first author presentedeach participant with the six screening plates in se-quence and asked the participants to indicate thenumber, shape, and location of all colored symbolson the plate. First, the experimenter asked the par-ticipant how many colored symbols she saw. Second,the experimenter instructed the participant to indi-cate the shapes of the colored symbols by allowingthe participant to use her own names (e.g., a circlecould be labeled as “zero,” “circle,” etc.). Third, theexperimenter asked the participant to point to thelocation of the symbol. An error consisted of failureto see all symbols, or to name incorrectly any symbol,or to point to an incorrect location of any symbol.We identified no color-blind students during the test;all students gave correct responses to all questionsreferring to the six plates.

The experiment consisted of three parts.1. Listening task (2.5 min). Upon arrival, we asked

each student to look at the color in front of her for aperiod of 1 min. The purpose of this procedure wasto get the person’s eye photoreceptors adapted to thecolor stimulus of the surrounding, which is an im-portant condition for the validity of the experiment.Next, we asked the student to listen to the word list,played by a tape recorder (1.5 min), and to try toremember as many of the words as possible. Thewords were presented auditorily so that students couldkeep their eyes focused on the color-sample boards.

2. A short math task (50 s). During the secondpart of the experiment, we asked each participant tostart counting aloud, backward, from 100 to 0, eachtime subtracting three numbers (e.g., 100, 97, 94).The purpose of this procedure was to prevent stu-dents from rehearsing the words. Students did notlook at the color-sample boards while carrying outthis task. While students were counting, we changedcolor-sample boards when it was necessary.




3. Memory recall (2.5 min). We used a recall testto assess the effects of color manipulations andchange of color context on long-term memory per-formance. We informed the participants that the timeallowed for completing the task would be 2.5 min.We also instructed each student to pronounce aloudall of the words remembered, while looking at thecolor in front of her.

Results

We used a 3 × 3 randomized-groups analysis ofvariance to examine the effects of color at study (high,low, neutral) and color at test (high, low, neutral) onrecall. The number of correct words recalled by eachindividual served as data for the corresponding con-dition. Incorrect responses and words recalled morethan one time were excluded from the data.

Does color saturation at study have an effect onmemory recall? As predicted, the main effect of colorat study was significant, F (2, 171) = 4.46, p < .05.Figure 1 illustrates the means and standard errors forthis effect. Furthermore, analysis of main compari-sons for color at study (Factor A) revealed significantdifferences between the recall scores of students inthe high-saturated group and students in the low-satu-rated group, FAcomp(1, 171) = 7.33, p < .01, andbetween the recall scores of the high-saturated andneutral groups, FAcomp(1, 171) = 6.03, p < .03. Thus,participants presented with the high-saturated red atthe time of material study scored significantly higheron the recall task than participants presented withthe low-saturated red and participants viewing theneutral stimulus. However, contrary to what wehypothesized, the analysis of the main comparisonbetween memory performance of students in the low-saturated group and students in the neutral groupwas not significant, FAcomp(1, 171) = 0.063, p > .05.Thus, the recall scores of students looking at the low-saturated color during study were not significantlydifferent from the recall scores of students lookingat the neutral stimulus.

Does color saturation at test have an effect onmemory recall? Likewise, the main effect of color attest was also significant, F(2, 171) = 14.01, p < .01.Analysis of main comparisons for color at test (FactorB) revealed significant differences between the memoryperformance of participants in the high-saturatedgroup and the performance of participants in the low-saturated group, FBcomp(1, 171) = 27.86, p < .01, as wellas between the recall scores of the high-saturated andneutral groups, FBcomp(1, 171) = 5.66, p < .03. Thus,similar to the pattern of results for Factor A, studentstested in the presence of the high-saturated stimulusscored significantly higher on memory recall thanstudents tested in the presence of the low-saturatedstimulus and students presented with the neutral colorduring recall. As predicted, memory performance ofstudents viewing the gray color at test was significantlybetter than the memory performance of studentslooking at the low-saturated red, FBcomp(1, 171) = 8.41,p < .01. (Refer to Figure 2 to see the means and stan-dard errors for the main effect of color at test.)


FIGURE 1

Memory recall as a function of study color.

Red (High) Red (Low) Gray

Color at Study

Nu

mb

er o

f Wo

rds

Rec

alle

d

15

10

5

0

FIGURE 2

Memory recall as a function of test color.

Red (High) Red (Low) Gray

Color at Test

Nu

mb

er o

f Wo

rds

Rec

alle

d

15

10

5

0


Does a change in color context from time of studyto time of test reduce memory recall? The presentresults revealed no significant interaction effect be-tween the two factors on memory recall, F (4, 171) =2.37, p > .05. As hypothesized, memory recall of stu-dents viewing the same color stimuli during study andduring test did not differ significantly from the recallof students viewing a different color stimulus duringstudy than they viewed at test. Table 1 illustrates themean recall scores and standard errors as a functionof color at study and color at test.

DiscussionThe present study provides evidence for the

hypothesis that manipulating color saturation of theenvironment during study or test, while holding othercolor properties (hue and brightness) constant, canaffect human memory performance. As hypothesized,the current results indicate that material learned inthe presence of a high-saturated color is recalledsignificantly more than when the same material islearned in the presence of a low-saturated or a neu-tral (gray) color. The same results apply when thesecolors are presented during the memory test, regard-less of the environmental color at study. At the sametime, the findings demonstrate that scores on memoryrecall are significantly lower when a low-saturatedcolor stimulus is presented during the test comparedto when a neutral stimulus is presented. These re-sults are consistent with Valdez and Mehrabian’sfindings (1994), showing that high-saturated colorselicit significantly higher arousal in individuals thanlow-saturated colors, regardless of hue and brightnessvalues.

Furthermore, the finding that there was asignificant relation between the low-saturated red andthe neutral gray color during testing (recall washigher for participants viewing the gray color), butnot between the same colors during study, is quiteinteresting. Because too much or too little arousalcan hinder performance on tasks requiring mentaleffort (Yerkes & Dodson, 1908), we can infer that thelow-saturated color of testing, by inducing low arousal,decreases memory performance. Yet, according to thepresent results, the same does not apply to color atstudy. Whether these findings could be explainedby the fact that study and recall are two distinct in-formation-processing stages generated by differentphysiological mechanisms, or by a possible inter-active effect of type of information processing andlevel of saturation of a given hue, should be furtherinvestigated.

The results also provide evidence for the hypoth-esis stating that change of color context does not have

a significant effect on memory recall. Prior studieshave established that changing the environmentalcontext between learning and retrieval plays animportant role in many explanations of forgetting(Godden & Baddeley, 1975; Smith, Glenberg, & Bjork,1978). Apparently, this theory does not apply to colorsaturation context, even though many researchershave used color in studies on context manipulationand memory. Based on our results, researchers mustraise questions about the cause of context-dependentmemory effects found by manipulating the colorof environmental context. It is possible that changesin color saturation rather than changes in environ-mental context caused these findings. Color satura-tion at study and color saturation at test have inde-pendent (not interactive) effects on memory per-formance. To illustrate, we found that memory perfor-mance of participants viewing the high-saturatedcolor during both study and test was not significantlydifferent from the memory performance of partici-pants looking at the low-saturated color during studybut with a high-saturated stimulus at test. However,when we looked at the separate effects of color at studyand color at test we encountered significant results.Thus, we found that students viewing the high-satu-rated color at study scored significantly higher onrecall than students presented with a low-saturatedcolor during study, regardless of the color propertiesat test. The same pattern of the results was observedfor color at test—participants presented with the high-saturated color during recall performed significantlybetter in memory than participants tested in a low-saturated environment.

Evidently, manipulating the properties of colorseems to play a crucial role in the outcome onmemory recall tests. Clearly, researchers should con-sider the stimulating and inhibiting effects of colorcharacteristics in studies on memory and manipula-tion of color contexts. Furthermore, because ourknowledge of the effects of color brightness and hue



TABLE 1

Mean Recall Scores (and Standard Errors) as aFunction of Color at Study and Color at Test

Color at study

Color at test Red (high) Red (low) Gray

Red (high) 14.20 (0.56) 13.00 (0.51) 13.70 (0.62)

Red (low) 12.05 (0.50) 9.55 (0.64) 11.10 (0.54)

Gray 13.35 (0.88) 12.85 (0.67) 11.00 (0.70)


on memory is limited, researchers should use carewhen using colors to examine the effects of environ-mental changes and memory performance.

Apparently, color has the potential to triggerphysiological reactions in humans. The mechanismleading to these reactions, though, is beyond thescope of this paper. One possible explanation of theresults might be that color’s potential to triggerarousal is responsible for the changes in our mentalcapacity, or the amount of mental effort available forprocessing information. In other words, the directionof our arousal reactions to the different propertiesof color shapes the outcome of our memory perfor-mance. Because there is a strong relation betweenmental capacity and level of alertness (Kahneman,1973; Yerkes & Dodson, 1908), we infer that the po-tential of color to affect our mental activation mayaccount for its effect on memory recall. However,further research should examine whether there is adirect relation between color, arousal, and memory.

A potential criticism of this study must be ad-dressed—namely, the lack of male participants, whichlimits the generalizability of our findings. Becausewomen tend to be significantly more sensitive, interms of arousal and other emotional reactions, tosaturation and brightness variations of color than men(Valdez & Mehrabian, 1994), we decided not to in-clude male participants in order to avoid further com-plexity of the experimental design. Future researchshould investigate the effects of color properties onmemory of male participants, as well as any possiblesex differences in memory performance due to dif-ferent patterns of emotional reactions to color prop-erties of men and women. Because men tend to showsignificantly weaker arousal reactions to color prop-erties than women, if there is a significant correla-tion between levels of color saturation (brightness)and levels of physiological arousal, we would expectthat color saturation (brightness) would have asignificantly stronger effect on the memory of womenthan on the memory of men.

Another criticism of the study could be that itsfindings are primarily limited to the auditory presen-tation of study material and does not address visualpresentation of information, displayed separately orembedded in the color stimulus. If we had presentedparticipants with study materials in a written formaton a separate piece of paper, participants would havevisually focused on the study list and not as much onthe color stimulus of the environment. The imple-menting of such a procedure would have interferedwith the primary goals of our study. Yet, a critic mightargue that if we had presented the list of words aspart of the color stimulus, the outcome of the recall

results could have been different. However, recentresearch suggests that color and form are representedseparately in memory, are assessed independently,and can be integrated if the task requires it, but thatsuch integration is a deliberate strategic response, notthe natural consequence of encoding (Hanna &Remington, 1996). Therefore, if we had presentedthe test material visually as part of the color stimulusinstead of auditorily, we would have increased thepossible flaw due to individual differences in theknowledge and use of memory strategies. Althoughauditory presentation does not guarantee us that par-ticipants did not use any memory strategies in theirattempt to perform well on the memory task, becausewe randomly assigned the participants to the condi-tions based on the order of their appearance, we haveminimized the possibly confounding effect stemmingfrom differences between the participants due to theirdegree of knowledge and use of memory strategies.

Further studies should also examine the possibleeffect of color on other cognitive processes, such asreasoning, problem solving, and performance on vari-ous tasks requiring mental effort. If high color satu-ration increases arousal, which, in turn, increasescapacity for mental effort, then we would expect thathigh-saturated colors would significantly increaseperformance on problem solving, reasoning, andother similar cognitive processes compared to the low-saturated colors. Concurrently, researchers shouldinvestigate the effect of different percentage varia-tions of color brightness and saturation on memory,as well as on other mental processes. As we mentionedearlier, similar to the effect of color saturation onarousal, different variations of color brightnesssignificantly affect the levels of emotional arousal inhumans. According to Valdez and Mehrabian (1994),less bright colors elicit significantly higher arousal re-actions in individuals than more bright colors. Basedon their findings and on the current results, we wouldexpect that less bright colors would increase memoryrecall and more bright colors would decrease recall.Thus, we would expect a different pattern of resultsrelating color brightness and memory compared tothe results of the current study on color saturationand memory; while high-saturated colors tend to im-prove memory performance, as supported by thepresent results, we would expect that very brightcolors would hinder memory performance.

Because we have purposely incorporated colorin many elements of the human world, researchersshould pursue a theoretical explanation of the sharedreactions to color properties. Such knowledge wouldbe of great application in many environmental set-tings of the academic and industrial world where




learning and memory performance is crucial andvaluable. Choosing wisely the colors of the schoolsetting, such as the color of the school classrooms,libraries, and other learning facilities, might improvesignificantly students’ academic performance. Cur-rently, the walls of most classrooms and libraries arepainted in white color, which tends to be of highbrightness and low saturation. Our findings supportthe idea that students who learn or recall informa-tion in the presence of a low-saturated color settingscore significantly lower on a memory test than stu-dents who learn or recall information in the presenceof a high-saturated color setting. Therefore, in theschool environment, it is important that we imple-ment high-saturated colors and avoid the use oflow-saturated color stimuli in order to optimize theconditions for students’ academic success. Concur-rently, an appropriate use of color in the work envi-ronment could enhance the performance of officeworkers carrying out mentally demanding tasks byoptimizing their performance in memory. However,because the present findings are limited to the ef-fects of saturation levels of red hue with mediumbrightness, and because we examined its effects dur-ing a brief period of study (1.5 min) and test (2.5min) time, future studies should include other levelsof saturation and brightness variations of differenthues and investigate the relation between color prop-erties and memory for a longer interval of time. Ifthe color of the environment has the potential toaffect our productivity by inducing higher levels ofalertness and, consequently, higher attention andbetter performance on mental tasks, then furtherexploration and utilization of the findings in this areais worthwhile.

ReferencesAdams, F. M., & Osgood, C. E. (1973). A cross-cultural study of

the affective meanings of color. Journal of Cross-CulturalPsychology, 4, 135–156.

Ainsworth, R. A., Simpson, L., & Cassell, D. (1993). Effects of threecolors in an office interior on mood and performance. Percep-tual and Motor Skills, 76, 235–241.



Batra, P., Urvashi, & Muhar, I. S. (1998). Hue and variation in CFF.Journal of the Indian Academy of Applied Psychology, 24, 83–86.

Damhorst, M. L., & Reed, J. A. P. (1986). Clothing color valueand facial expression: Effects on evaluations of female jobapplicants. Social Behavior and Personality, 14, 89–98.

Dulsky, S. G. (1935). The effect of a change in background onrecall and relearning. Journal of Experimental Psychology, 18, 725–740.

Frank, G. (1976). On the validity of hypotheses derived from theRorschach: I. The relationship between color and affect.Perceptual and Motor Skills, 43, 411–427.

Godden, D. R., & Baddeley, A. D. (1975). Context-dependentmemory in two natural environments: On land and under-water. British Journal of Psychology, 66, 325–331.

Hanna, A., & Remington, R. (1996). The representation of colorand form in long-term memory. Memory and Cognition, 24, 322–330.

Hardy, L. H., Rand, G., & Rittler, M. C. (1957). AO H-R-R Pseudo-isochromatic Plates. Buffalo, NY: American Optical.

Jacobs, K. W., & Hustmyer, F. E., Jr. (1974). Effects of four psycho-logical primary colors on GSR, heart rate and respiration rate.Perceptual and Motor Skills, 38, 763–766.

Jacobs, K. W., & Suess, J. F. (1975). Effects of four psychologicalprimary colors on anxiety state. Perceptual and Motor Skills, 41,207–210.

Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ:Prentice-Hall.

Mehrabian, A. (1978). Measures of individual differences intemperament. Educational and Psychological Measurement, 38,1105–1117.

Mikellides, B. (1990). Color and physiological arousal. Journal ofArchitectural and Planning Research, 7, 13–20.

Murnane, K., & Phelps, M. P. (1993). A global activation approachto the effect of changes in environmental context on recogni-tion. Journal of Experimental Psychology: Learning, Memory, andCognition, 19, 882–894.

Profusek, P. J., & Rainey, D. W. (1987). Effects of Baker-MillerPink and red on state anxiety, grip strength, and motor preci-sion. Perceptual and Motor Skills, 65, 941–942.

Smith, S. M., Glenberg, A., & Bjork, R. A. (1978). Environmentalcontext and human memory. Memory and Cognition, 6, 342–353.

Valdez, P., & Mehrabian, A. (1994). Effects of color on emotions.Journal of Experimental Psychology: General, 123, 394–409.

Weiss, W., & Margolius, G. (1954). The effect of context stimulion learning and retention. Journal of Experimental Psychology,48, 318–322.

Weller, L., & Livingston, R. (1988). Effects of color of question-naire on emotional responses. Journal of General Psychology, 115,433–440.

Wilson, G. D. (1966). Arousal properties of red versus green.Perceptual and Motor Skills, 23, 947–949.

Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength ofstimulus to rapidity of habit-formation. Journal of ComparativeNeurology and Psychology, 18, 459–482.

PSI CHI JOURNAL OF UNDERGRADUATE RESEARCH � Fall 2001 109Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 109–117 / ISSN 1089 -4136). *Faculty Supervisor

Effects of DNA and EyewitnessEvidence on Juror Decisions

The present experiment investigated whether DNA evidence, eyewitness evidence,or a combination of both was more persuasive to mock jurors. The study alsoexplored whether varying the credibility of the testimony affected persuasiveness.The researchers hypothesized that the combination of eyewitness and DNA evi-dence would be the most persuasive to jurors. Eyewitness evidence would be morepersuasive than DNA evidence, and credible testimony would be more persuasivethan discredited testimony. One hundred-fifty college students read an excerpt froma court transcript describing a first-degree murder trial. Participants rendered averdict and answered questions concerning confidence, understanding, reliabil-ity, and persuasiveness of testimony. The analyses partly supported the hypoth-eses. DNA evidence was more persuasive, reliable, influential, and less likely to beviewed as wrong than eyewitness evidence, regardless of whether it was discred-ited. DNA evidence elicited more guilty verdicts, and jurors were more confident intheir verdict. We discuss these results in light of the 2-routes-to-persuasion theory.

NAOMI J. FREEMAN

DIANA L. D. PUNZO*

Earlham College

THE PERSUASIVENESS OF DIFFERENT TYPES OF

evidence in criminal cases has been an on-going issue in the field of forensic psychology.

For years, researchers have studied the influence ofaccuracy, confidence, and reliability of eyewitness tes-timony on juror decisions (Brigham & Bothwell, 1983;Cutler, Penrod, & Martens, 1987; Loftus, 1980;Wrightsman, 1991). Eyewitness testimony and its ef-fect on jurors are topics of interest not only to thepsychological world, but also to the legal communitybecause “the evidence produced by an eyewitnessoften makes the difference between an unsolvedcrime and a conviction” (Wrightsman, 1991, p. 157).Often, however, eyewitnesses are inaccurate in theiridentification of perpetrators (Wrightsman, 1991).Mistakes in identification can occur from the timewhen the crime is committed up to the time whenthe witness testifies at the trial. In addition, new ex-periences, leading questions, or police lineups canalter memories. Despite the unreliability of eyewit-ness identification, research has found that jurors areheavily persuaded by this form of evidence (Loftus,1980; Wrightsman, 1991).

Many theories have attempted to explain whyeyewitness testimony is such a powerful form of evi-dence. Elizabeth Loftus (1980) has suggested that

people do not have a general understanding of theworking of human memory. People tend to believethat memory works similar to a videotape recorder.However, Loftus has suggested that, in a way, peopleactually construct their memories. Loftus also arguesthat in everyday life truly precise memories are veryrarely demanded of people. The common errors inrecollection that occur daily go unnoticed becausethey are not particularly important. Because peopledo not understand the exact workings of humanmemory and tend to trust their own memories, theyalso tend to trust the memories of others. Therefore,jurors are more likely to trust the memory of an eye-witness, especially a confident eyewitness, than otherforms of evidence. Other witnesses, such as experts,can provide only a small piece of information; an eye-witness, however, can describe events that occurred

Author note. Naomi J. Freeman, Earlham College, and Diana L. D.Punzo, PhD, Earlham College Department of Psychology.

The author would like to thank Katharine S. Milar, PhD,Earlham College, for her assistance in preparing the manuscriptfor publication, and Lynda Hall, PhD, Ohio Wesleyan University,for her assistance with the statistical analysis. The author wouldalso like to thank Martin Lewis, JD, for his assistance in the devel-opment of the court transcripts used in the study.

Correspondence concerning this article should be addressedto Naomi J. Freeman at [email protected].


in detail to the jurors, thus providing a great deal ofinformation for the mental construction of eventsin jurors’ minds. Loftus suggested that it might beeasier for jurors to work with a smooth recollectionof the events, even if they are inaccurate, than tohave to piece together fragments of the event on theirown.

To test her theory that people will be more per-suaded by eyewitness than other forms of evidence,Loftus (1980) asked participants to read a case inwhich a bad check had been written to purchase atelevision. Participants participated in one of four evi-dence conditions: (a) an eyewitness (the store clerk)who positively identified the defendant as the manwho wrote the check, (b) a polygraph expert whotestified that a polygraph test stated the defendantwas lying when he claimed he did not write the check,(c) a fingerprint expert who examined prints left onthe counter of the store by the person who wrote thecheck and the fingerprints matched the defendant’sprints, and (d) a handwriting expert who examinedthe check and claimed that it matched the defen-dant’s handwriting. Results indicated that convictionswere highest in the eyewitness condition (78%), fol-lowed by fingerprints (70%) and polygraph test(53%), and lowest in the handwriting condition(34%). These results confirm that eyewitness evidenceis one of the most persuasive kinds of evidence thatcan be presented.

Research on whether juries believe discreditedeyewitnesses has produced equivocal results. Loftus(1980) has reported that jurors place a great deal ofweight on eyewitness testimony, even if that testimonyis suspect or discredited. McCloskey and Egeth (1983)conducted a meta-analysis of eyewitness research anddetermined that “jurors” often disregard eyewitnesstestimony that has been convincingly discredited.Kennedy and Haygood (1992) have argued thatjurors are aware of the unreliability of eyewitnesses.The findings on the persuasiveness of a discreditedeyewitness are, thus, inconclusive.

The persuasiveness of other types of evidence,such as deoxyribonucleic acid (DNA) evidence, hasbeen less extensively investigated compared to eye-witness evidence. DNA profiling is a technique thatallows an individual to be identified by comparing asample of DNA material from the crime scene, suchas hair, blood, or semen, to the DNA of a suspect( Joyce, 1990). A forensic examiner often uses statis-tical language to present DNA evidence. For example,a DNA forensic examiner is likely to state that theblood found at a crime scene matches the defendant’sblood and that there is only a 2% chance that arandomly selected person from the population would

have that DNA pattern or that 1 out of every 170million to 1.2 billion people would have that DNApattern.

Although DNA evidence is extremely reliable, itis often ignored by jurors (Goodman, 1992; Smith,Penrod, Otto, & Park, 1996; Wells, 1992). Goodman(1992) studied the extent to which prospectivejurors comprehend and use probabilistic evidencewhen rendering a verdict. Participants read a courttranscript describing a homicide. Goodman variedthe frequency with which DNA from blood found atthe crime scene matched the defendant. Resultsindicated that jurors’ estimates of guilt were lowestwhen the likelihood of a random match between theblood found at the crime scene and the blood typeof the defendant was greatest (Goodman, 1992). Shealso found that the jurors recognized that thescientific evidence was significantly valuable, butthey “failed to make fine distinctions between prob-ability evidence that were mildly incriminating, mod-erately incriminating, and strongly incriminating”(Goodman, 1992, p. 371).

When Goodman (1992) compared the weightmock jurors placed on the statistical evidence to aBayesian analysis, which predicted how much weightjurors should have given that evidence, she found thatmock jurors underused the scientific evidence inreaching a verdict. Additional research by Wright,MacEachern, Stoffer, and MacDonald (1996) sup-ported Goodman’s findings and suggested thatpeople do not understand statistical evidence and,therefore, are reluctant to use it when rendering averdict.

In order to discover if people are reluctant tomake guilty decisions when the evidence is based onlyon statistical evidence, Wells (1992) conducted a se-ries of five experiments. Wells concluded that evenstudents who have a general understanding of statis-tical information disregard statistical evidence whenit is presented.

In their review of jurors’ use of scientific evidence,Cooper, Bennett, and Sukel (1996) argued that mostjurors are not able to understand or comprehendlegal concepts or scientific data. People who wouldlikely understand this evidence, such as college stu-dents, lawyers, and doctors, are often exempt fromserving on juries. Theories of persuasion (Petty &Cacioppo, 1981) suggest that when evidence is com-plex or difficult to understand, jurors will take theperipheral route to persuasion and base their deci-sion on factors other than the strength of the argu-ment, such as the expert’s credibility. Cooper et al.(1996) tested this hypothesis and discovered thatwhen mock jurors evaluated extremely complex


EFFECTS OF DNA AND EYEWITNESS � Freeman and Punzo


scientific evidence, they based their verdict on othercues besides the evidence, such as credibility of thewitness. The more complex the evidence, the morejurors relied on the credibility of the witness to makea decision in the case. Taken together, these resultssuggest that when jurors view complex scientific tes-timony, such as DNA evidence, factors other than thecontent of the testimony will influence believability.

The purpose of the present experiment was toexamine the relative importance of eyewitness evi-dence and DNA evidence to mock jurors. Participantsread an excerpt of a court transcript describing a first-degree murder trial. The transcript contained eithereyewitness evidence, DNA evidence, or both types ofevidence. We also manipulated the credibility of thetestimony (either credible or discredited). After theparticipants read the transcript, they rendered a ver-dict in the case and answered several questions con-cerning understanding of testimony, confidence inverdict, credibility of witness, and persuasiveness oftestimony.

Based on the results of previous research (Good-man, 1992; Loftus, 1980; Wells, 1992) and the theoryof persuasion espoused by Petty and Cacioppo (1981),we hypothesized that the combination of the eyewit-ness and DNA evidence would be the most persua-sive to jurors. Eyewitness evidence would be more per-suasive than DNA evidence, and credible testimonywould be more persuasive than discredited testimony.

MethodParticipants

One hundred fifty college students (107 women,43 men) participated in one of six different experi-mental conditions. Participants, ages 18 to 23, at-tended either Ohio Wesleyan University or EarlhamCollege. Most participants received course extracredit in return for their participation.

Materials and ProcedureThe researcher told participants only that the

experiment was about juror’s perception of criminalcases. Participants completed the experiment ingroups ranging from 5 to 40 participants. Participantsreceived excerpts from a court transcript and wereinstructed to read the transcript carefully and toimagine they were real jurors serving on the case. Anattorney assisted with the writing of the court tran-scripts, which were written specially for this study.There were two independent variables: (a) the typeof evidence that was being presented, either eye-witness evidence, DNA evidence, or both types ofevidence; and (b) the credibility of the testimony,either credible testimony or one that had been dis-

credited through a cross-examination. In the com-bined condition, in which participants evaluated bothtypes of evidence, the order was counterbalanced; halfthe participants read the DNA evidence first, whereasthe other half read the eyewitness evidence first. Thus,there were six experimental conditions to which par-ticipants in each experimental session were randomlyassigned. Each experimental condition received a dif-ferent color court transcript. For example, partici-pants in the eyewitness credible condition received apurple transcript, whereas participants in the com-bined discredited condition read a blue transcript.Participants read a direct examination as well as across-examination in all conditions.

In the eyewitness credible condition, the pros-ecution presented evidence that an eyewitness sawthe defendant running from the scene of the crimeholding a knife. In addition, this eyewitness pickedthe defendant out from a lineup, and during the trial,identified the defendant, David K. Johnson, as theperpetrator. In the cross-examination, little damagewas done to the witness’s credibility. The eyewitnesstestified that he did not need glasses at night, thatthere was plenty of light in the area, and that he got agood look at the defendant.

In the eyewitness discredited condition, partici-pants read the same prosecution evidence used inthe eyewitness credible condition, but in addition,they received a defense attorney’s cross-examinationof the eyewitness in which he challenged the cred-ibility of the eyewitness. The defense attorney pointedout that the eyewitness had very poor eyesight andwas not wearing his glasses the night of the murder.The defense also presented evidence that stated itwas dark the night of the murder and there was poorlighting in the area, which would have made identifi-cation of the perpetrator extremely difficult.

In the DNA credible condition, the prosecutionpresented a DNA forensic examiner who explainedwhat DNA is and the testing process. The DNA ex-pert then stated that he tested two items for this case:blood retrieved from the sidewalk outside the victim’shouse and samples from under the fingernails of thevictim. The DNA expert testified that the profile fromthe two DNA items matched that of the defendant,David K. Johnson. The results from the DNA testingwere presented in frequencies. Jurors read that thesepatterns occur in 1 out of every 170 million to 1.2billion people. In the cross-examination, the foren-sic examiner testified to the number of hours put intothe case and that the lab had no previous errors onthe state proficiency tests.

In the DNA discredited condition, participantsviewed the direct examination used in the DNA cred-





ible condition, but were then presented with a cross-examination that questioned the reliability of DNAtesting. Previous lab errors were presented as well aspossibilities that the DNA evidence was contaminated.

In the combined credible condition, participantsread a combined court transcript of the eyewitnesscredible condition and the DNA credible condition.Participants in the discredited combined conditionviewed the same materials used in the eyewitness dis-credited condition combined with the materials usedin the DNA discredited condition.

After they had read the designated court tran-script, participants gave the researcher their transcriptbefore receiving the questionnaire to ensure that theydid not refer back to the transcripts while answeringquestions. After the transcript was collected, partici-pants completed a questionnaire. Participants ren-dered a verdict, either guilty or not guilty, and thenindicated their confidence in that verdict. Other ques-tions concerned understanding of evidence, credibil-ity of witness, accuracy and reliability of the testimony,and persuasiveness, importance, and influence of tes-timony. Participants answered questions on a Likert-type scale ranging from 1 to 9 in which 1 was less and9 was more of the quality being asked. The participantsin the eyewitness and DNA conditions answered 18questions, whereas participants in the combined con-ditions answered 32 questions. Each participant com-pleted the questionnaire individually without knowl-edge of other participants’ responses. After partici-pants completed all parts of the measures, they weredebriefed and thanked for their time.

ResultsPreliminary Analyses

In the combined condition, composite scores foreach dependent variable were obtained by taking theaverage of the eyewitness and DNA scores. Eachcondition’s court transcript was on different colorpaper; therefore, in order to determine whether thetranscript color was a confounding variable, 48 addi-tional participants were administered the eyewitnesscredible transcript, with 6 participants receiving eachof the eight colors. A one-way analysis of variance(ANOVA) compared the effect of color of the courttranscript on verdicts, confidence in verdict, and de-gree of guilt. No significant effects were discovered,indicating that color of transcript had no effect onthese dependent variables.

A preliminary analysis also compared the re-sponses of men and women on several of the depen-dent variables. No significant differences were re-vealed; therefore, this variable was not examinedfurther.

A two-way ANOVA was conducted as a manipula-tion check to determine if type of evidence andcredibility of testimony affected witness credibility.This analysis ensured that, in fact, there was a sig-nificant difference between the credible and discred-ited conditions. A significant interaction was revealed,F (2, 144) = 3.127, p < .05, indicating that the cred-ibility of testimony interacted with type of evidenceto influence credibility of witness. The manipula-tion check indicated that the credibility manipulationworked only in the eyewitness condition. Neither theDNA nor the combined condition revealed a signi-ficant difference in mean witness credibility betweenthe credible and discredited conditions. Althoughthere was a main effect for credibility of testimony,F(1, 144) = 23.515, p < .01, this main effect was dueentirely to the eyewitness condition. It appears thatone cannot discredit a DNA witness. It should also bementioned that the main effect for type of evidencewas significant, F(2, 144) = 38.556, p < .01. Mock ju-rors rated the DNA evidence as more credible thanthe combined or the eyewitness evidence. Participantsviewed the DNA evidence as more credible regard-less of whether the testimony was discredited or not.

All data were analyzed using two-way ANOVAs todetermine if type of evidence and credibility of testi-mony had an effect on the dependent variables. Posthoc analyses were conducted with least significantdifference (LSD) tests.

Guilty Verdicts and Confidence RatingsTable 1 shows the mean scores and standard de-

viations for degree of guilt and confidence in verdictfor each condition. A significant main effect for typeof evidence was found for confidence in verdict anddegree of guilt of defendant, F(2, 144) = 7.028, p <.01, and F(2, 138) = 2.83, p < .01, respectively. Partici-pants in the eyewitness conditions rated the defen-dant as less guilty (p < .05) and were less confident(p < .05) in their verdict than participants in eitherthe DNA only or combined conditions, which did notsignificantly differ.

Tables 2 and 3 show the results for the variableof guilty verdicts. We conducted chi-square tests ofindependence to determine if type of evidence andcredibility of testimony had an effect on verdict. Asignificant effect for type of evidence was discovered,χ2(5, N = 150) = 29.76, p < .01, indicating that theDNA condition produced more guilty verdicts thanwas expected by chance, whereas the eyewitness con-dition yielded fewer guilty verdicts than was expectedby chance. Results confirmed a significant effect forcredibility of testimony, χ2(1, N = 150) = 3.86, p < .05.Participants in the credible conditions were more



TABLE 1

Means and Standard Deviations of Ratings of Guilt and Confidence as aFunction of Type of Evidence and Credibility of Testimony

Eyewitness DNA Combined

Credible Discredited Credible Discredited Credible Discredited

Dependent variables M SD M SD M SD M SD M SD M SD

Degree of guilt 5.43 1.61 4.97 2.12 7.09 1.50 6.97 1.63 7.18 1.69 6.06 1.63

Confidence in verdict 5.08 2.14 5.29 1.92 6.96 1.31 5.29 2.04 6.17 1.93 6.12 1.83

Note. Higher scores indicate more of the attribute.

TABLE 2

Verdict as a Function of Type of Evidence

Type of evidence

Verdict Eyewitness DNA Combined Total

Guilty 19 45 36 100

Not guilty 30 5 15 50

Total 49 50 51 150

TABLE 3

Verdict as a Function of Credibilityof Testimony

Credibility of testimony

Verdict Credible Discredited Total

Guilty 57 43 100

Not guilty 20 30 50

Total 77 73 150

likely to vote guilty than participants in the discred-ited conditions.

Persuasiveness and Importance of EvidenceTable 4 shows the mean scores and standard de-

viations for persuasiveness of evidence and impor-tance of evidence. Results revealed a significant maineffect for credibility of testimony on the variable ofpersuasiveness, F(1, 144) = 5.582, p < .05, suggestingthat the more credible the testimony, the more per-suasive it was to prospective jurors. A significant maineffect for type of evidence was also discovered, F(2,144) = 6.137, p < .01. Post hoc tests showed that eye-witness evidence was less persuasive than the com-bined and DNA conditions (p < .05); however, therewas no significant difference between the DNA andthe combined conditions.

The two measures, importance of evidence to theverdict and influence of testimony, were combinedby adding the two scores to make a composite scorefor the variable of “importance.” A significant maineffect for type of evidence was revealed for the vari-able “importance,” F(2, 144) = 10.318, p < .01. Posthoc analyses showed that the DNA evidence was

TABLE 4

Means and Standard Deviations of Ratings of Persuasiveness and Importanceof Evidence as a Function of Type of Evidence and Credibility of Testimony




Persuasive 5.88 2.03 5.46 2.13 6.73 1.28 6.46 1.14 6.58 1.61 5.96 1.53

Importance 13.04 4.23 12.50 4.93 15.62 2.74 15.38 2.55 13.85 2.06 12.60 2.59



significantly more important (p < .05) to the verdictthan the eyewitness evidence or the combined evi-dence. No significant difference was found betweenthe eyewitness and combined conditions.

Understanding of EvidenceTable 5 shows mean scores and standard devia-

tions for the variables understanding of evidence,technicality of evidence, and difficulty of evidence. Asignificant main effect for type of evidence was foundfor the variables understanding of evidence, F(2, 144)= 11.321, p < .01, technicality of evidence, F(2, 144) =15.277, p < .01, and difficulty of evidence, F(2, 144),p < .01. Post hoc LSD tests revealed that eyewitnesstestimony yielded significantly (p < .05) more under-standing of evidence than both the combined andDNA evidence; however, DNA evidence did not dif-fer from the combined condition. In addition, theDNA evidence was viewed as significantly (p < .05)more technical than both the eyewitness evidence andthe combined evidence. The fact that no significantdifference was found between the eyewitness andcombined evidence implies that mock jurors did not

believe the combined evidence to be any more tech-nical than the eyewitness evidence. Consistent withthe previous findings, participants found the DNAevidence more difficult to understand than the com-bined evidence (p < .05), and eyewitness evidence,which was the least difficult to understand.

Accuracy of Testimony and WitnessTable 6 shows the means and standard deviations

for the variables likelihood of errors and reliabilityof evidence. A significant main effect for type of evi-dence on likelihood of errors was found, F(1, 144) =13.421, p < .01. Participants believed that there was agreater chance that the eyewitness was wrong in hisidentification than that the lab had made errors dur-ing the DNA testing.

We created the variable, reliability of evidence,by adding together the scores for the ratings of reli-ability of testimony and believability of witness.ANOVAs yielded significant main effects for credibil-ity of testimony, F(1, 144) = 14.383, p < .01, and typeof evidence, F(2, 144) = 25.837, p < .01, on the vari-able of reliability of evidence. Participants viewed the



TABLE 5

Means and Standard Deviations of Ratings of Understanding of Evidenceas a Function of Type of Evidence and Credibility of Testimony




Understanding 8.08 1.68 7.83 1.79 6.73 1.25 6.38 1.66 6.98 1.24 7.20 1.17

Technicality 5.48 2.53 5.25 2.12 7.42 1.24 6.75 1.59 5.28 1.25 5.13 1.47

Difficulty 2.04 1.01 1.88 .78 4.99 2.23 4.59 1.96 3.70 1.46 3.22 1.43


TABLE 6

Means and Standard Deviations of Ratings of Accuracy of Testimony and Witnessas a Function of Type of Evidence and Credibility of Testimony




Errors 5.44 2.16 5.58 2.13 3.62 1.81 3.63 1.64 4.48 1.98 4.30 1.56

Reliability of evidence 11.52 3.89 9.10 3.56 15.58 1.92 13.75 2.95 13.19 2.48 11.84 2.91



credible testimony as more reliable than the discred-ited testimony. Participants thought the DNA evi-dence was more reliable than the combined evidenceand the eyewitness evidence. The eyewitness was theleast reliable form of evidence (p < .05). It appearsthat adding the eyewitness evidence to the DNA evi-dence to produce the combined condition reducesthe judgments of reliability of the DNA evidence.

Aspects of the WitnessTable 7 shows the means and standard deviations

for the variables witness intelligence and confidence.The dependent variable intelligence of witness re-vealed a significant main effect for type of evidence,F(2, 138) = 8.444, p < .01. Post hoc analyses showedthat the witness was seen as significantly (p < .05) moreintelligent in the DNA conditions than in the com-bined or eyewitness conditions. Jurors viewed the wit-ness in the combined conditions as significantly (p <.05) more intelligent than the witness in the eyewit-ness conditions.

The variable confidence of the witness yieldeda significant main effect for credibility of testimony,F(1, 138) = 10.262, p < .01. This main effect meansthat mock jurors believed the witness was more con-fident in the credible testimony condition than in thediscredited testimony condition.

DiscussionThe purpose of the present experiment was to

investigate whether certain types of evidence—DNA,eyewitness, or a combination of both—are more per-suasive to mock jurors. The experiment also exploredwhether varying the credibility of the testimony af-fected persuasiveness. Results indicated that DNAevidence was more persuasive to mock jurors regard-less of whether or not it was discredited. DNA evi-dence received the most guilty verdicts and partici-pants were more confident in their verdict in the DNA



conditions than in the combined or eyewitness con-ditions. The DNA evidence alone was more convinc-ing to jurors than when it was added to the eyewit-ness evidence. Overall, it appeared that adding theeyewitness evidence onto the DNA evidence to cre-ate the combined conditions actually reduced thepersuasiveness of the DNA evidence.

The results supported previous research (Kennedy& Haygood, 1992; McCloskey & Egeth, 1983), whichindicated that discrediting a witness weakens theinfluence of that witness in the juror’s mind. Theresults, however, contradicted the findings of Loftus(1980), who suggested that jurors overestimate theaccuracy of eyewitness evidence, even when it hasbeen discredited. The results also contradicted pre-vious DNA evidence research, which suggested thatjurors ignore DNA and scientific evidence because itis too complex (Goodman, 1992; Smith et al., 1996;Wells, 1992).

The results of the present research seem to sug-gest that either jurors are aware of the inconsisten-cies and unreliability that surround eyewitness evi-dence and, therefore, do not place a great deal ofweight on it when rendering a verdict or that jurorshave more faith in DNA evidence than eyewitnessevidence. Mock jurors believed the DNA expert wit-ness more than they believed the eyewitness. In addi-tion, participants viewed the eyewitness as less reli-able, less accurate, less confident, less intelligent, andless credible than the DNA expert witness. Jurors alsoperceived the eyewitness evidence to be less influ-ential and important to their decision and to be lesspersuasive than the DNA evidence or the combinedevidence. The findings also suggest that the eyewit-ness was more likely to be wrong in his/her identifi-cation than the lab was to have errors in the DNAtesting process. Mock jurors thought to a greaterdegree that the defendant was guilty in both thecredible and discredited DNA conditions followed by

TABLE 7

Means and Standard Deviations of Ratings of Aspects of the Witness as aFunction of Type of Evidence and Credibility of Testimony




Intelligence of witness 6.47 1.50 6.38 1.56 7.94 1.04 7.25 1.23 7.30 1.17 6.74 1.14

Confidence of witness 7.71 1.58 7.25 1.38 8.19 1.06 6.84 1.71 7.42 1.48 6.70 1.39




both the credible and discredited combined condi-tions, than both eyewitness conditions.

The results of this study could have profoundimplications not only for the psychological world, butfor the legal community as well. It may be that jurorswill believe an expert DNA witness regardless ofwhether the testimony is suspect. If this is the case,then a lawyer only needs to present the DNA evidenceand it would be believed regardless of its validity.Furthermore, the more credible and intelligent thewitness appears to the jurors, the more likely theevidence is to be believed.

Contrary to previous findings, eyewitness testi-mony does not appear to be persuasive to jurors.Jurors seem to be aware of the inconsistency andunreliability of eyewitness recollection. The resultsof the current research suggest that eyewitness evi-dence is not a powerful type of evidence, whichcontradicts previous research findings (Loftus, 1980).In fact, because jurors are able to take the centralroute to persuasion when examining eyewitness evi-dence, the evidence may be scrutinized more thanDNA evidence, which allows the juror to take theperipheral route. When jurors take the central route,they examine and analyze the evidence, therefore,jurors would notice the unreliability and inconsisten-cies that surround eyewitness identification. Becauseof the technicality of the evidence in the DNA condi-tions, the mock jurors, as in the study conducted byCooper et al. (1996), were forced to take the periph-eral route to persuasion and, therefore, perceived theDNA expert witness to have greater expertise andcredibility than the eyewitness.

There were several limitations of this study. Dueto time constraints and lack of funding, written courttranscripts were used. The use of a live courtroomsetting or the use of a video recording of a trial sothat jurors could actually see and hear the testimonywould improve this study. The responses of thejurors to reading the evidence might be different thanthe responses from hearing and seeing the evidencebeing presented by a real person. It may be that par-ticipants are better able to process technical infor-mation when reading it than when hearing it from awitness.

In addition, the education level of the partici-pants used in this experiment may have influencedthe results. Previous research indicates that the aver-age juror has an eighth-grade education (Cooper etal., 1996). Because we used only college students, whohave a much higher education level than the averagejuror, the results may not generalize to the largerpopulation. In addition, because most college stu-dents have a background in science and may have

studied DNA, they may have more confidence in thereliability of DNA testing than the typical juror.Another limitation with the sample was that all par-ticipants were between the ages of 18 and 23. Thisage group is the least likely to actually serve on a realjury. Although one study (Kassin & Barndollar, 1992)found that students and nonstudents (adults) do notsignificantly differ in their sensitivity to factors thataffect eyewitness evidence, the present study shouldbe conducted again using both college students andadults to verify that the findings are robust across agegroups.

Another limitation of the present study was thatparticipants did not engage in the jury deliberationprocess. Participants rendered a verdict and com-pleted the questionnaire individually without knowl-edge of other participants’ responses. Results mayhave varied if the participants had discussed theirverdicts with other mock jurors.

To increase knowledge as to how jurors perceivevarious types of evidence, several future studies couldbe conducted. One possible study could address theissue of how DNA evidence is presented. The use ofvisual aids could affect persuasiveness of the evidence.It would be interesting to investigate how jurors’understanding of DNA evidence is affected by the waythe DNA is presented. In addition, understandingcould be influenced by jurors’ previous knowledgeof mathematics and science. Jurors who have a vastunderstanding of mathematics and science may bemore prone to trust DNA and believe it is a reliableform of evidence than those jurors who are not asexperienced with math and science. Further researchconducted in the psychology laboratory can provideadditional insight concerning the variables thatinfluence juror decision making.

References

Brigham, J. C., & Bothwell, R. K. (1983). The ability of prospec-tive jurors to estimate the accuracy of eyewitness identifications.Law and Human Behavior, 7, 19–30.

Cooper, J., Bennett, E. A., & Sukel, H. L. (1996). Complex scientifictestimony: How do jurors make decisions? Law and HumanBehavior, 20, 379–394.

Cutler, B. L., Penrod, S. D., & Martens, T. K. (1987). The reliabil-ity of eyewitness identification: The role of system and estima-tor variables. Law and Human Behavior, 11, 233–258.

Goodman, J. (1992). Juror’s comprehension and assessment ofprobabilistic evidence. American Journal of Trial Advocacy, 16,361–389.

Joyce, C. (1990, July 21). High profile: DNA in court again. NewScientist, 21, 24–25.

Kassin, S. M., & Barndollar, K. A. (1992). The psychology of eye-witness testimony: A comparison of experts and prospectivejurors. Journal of Applied Social Psychology, 22, 1241–1249.

Kennedy, T. D., & Haygood, R. C. (1992). The discrediting effectin eyewitness testimony. Journal of Applied Social Psychology, 22,70–82.



Loftus, E. F. (1980). Psychological aspects of courtroom testimony.In F. Wright, C. Bahn, & R. W. Rieber (Eds.), Forensic Psychologyand Psychiatry (pp. 27–37). New York: New York Academy ofSciences.

McCloskey, M., & Egeth, H. E. (1983). Eyewitness identification:What can a psychologist tell a jury? American Psychologist, 38,550–563.

Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Clas-sic and contemporary approaches. Dubuque, IA: Wm. C. Brown.

Smith, B. C., Penrod, S. D., Otto, A. L., & Park, R. C. (1996).

Juror’s use of probabilistic evidence. Law and Human Behavior,20, 49–82.

Wells, G. L. (1992). Naked statistical evidence of liability: Is sub-jective probability enough? Journal of Personality and Social Psy-chology, 62, 739–752.

Wright, E. F., MacEachern, L., Stoffer, E., & MacDonald, N. (1996).Factors affecting the use of naked statistical evidence of liabil-ity. The Journal of Social Psychology, 136, 677–688.

Wrightsman, L. S. (1991). Psychology and the legal system (2nd ed.).Pacific Grove, CA: Brooks/Cole.


Part-Set Cuing is Due to Strong,Not Weak, List Cues

Research in memory on the part-set cuing paradigm has shown that the presenta-tion of study list items as cues at retrieval impairs recall for the remaining items ofthe study list (e.g., Rundus, 1973). Subsequent research by Peynircioglu (1987)has indicated that the part-set cuing effect may also occur in nonmemory tasks. InPeynircioglu’s experiment, participants generated words from a larger source word.At retrieval, participants were provided with 0 or 8 exemplars as cues to helpgenerate other words. The results indicated that fewer words were generated in thecued condition than in the noncued condition, suggesting that the mechanismproducing the effect was similar to that found in traditional memory tasks (e.g.,blocking, strategy disruption). The procedure in the present study was similar toPeynircioglu’s except participants generated exemplars from common conceptualcategories (e.g., fruits). The number and popularity of cue words was manipu-lated. The results indicated that a part-set cuing effect occurred only when popu-lar cues were provided. This finding suggests that the part-set cuing effect is due tothe strength of the cue words and thus supports predictions made by blockingtheory, which suggests that study list items with higher retrieval strength blockaccess to study list items with weaker retrieval strength.

MONICA L. BOICE

GARY J. GARGANO*

Saint Joseph’s University

IN STUDIES OF CATEGORICAL RECALL, RECALL FOR

items related to a common conceptual category,findings indicate that the presentation of category

labels enhances recall (Tulving & Pearlstone, 1966).However, other research has shown the presentationof study list members has the opposite effect (Brown,1968). For example, Brown (1968) asked participantsto recall as many of the 50 states as possible. Brownfound when participants were provided with 25 states,as cues to help them recall the remaining 25, perfor-mance on the recall test decreased. Similar researchby Rundus (1973) and Roediger (1974), in which par-ticipants recalled studied items in the presence ofcues, has shown that there is an inverse relation be-tween the number of cues presented and the num-ber of category members recalled. This paradoxicalfinding in memory research is referred to as the part-set cuing effect.

Several theories have been proposed to explainthe part-set cuing effect, including strategy disrup-tion, blocking, and cue overload. Brown (1968) at-tributed the part-set cuing effect to strategy disrup-tion. According to strategy disruption, the specific re-trieval strategy of each participant is disrupted whencues are presented at retrieval. Brown believed eachparticipant has a unique retrieval strategy, and inter-

fering with that strategy, by providing cues, inhibitsrecall. For example, when participants recalled asmany states as they could, they may have done so sys-tematically by using alphabetical order or by usingproximity from east to west. The cues disrupt the re-trieval strategy because they are not presented in thesystematic strategy used by the participant.

Rundus (1973) attributed the part-set cuing ef-fect to blocking theory. According to blocking theory,items compete for access at retrieval. Items that havehigh retrieval strength inhibit, or block, access foritems with weaker retrieval strength. The retrievalstrength for an item becomes stronger and more ac-cessible when the item is studied. The accessibilityfor an item also increases each time the item is re-trieved (Bjork & Bjork, 1988). Thus, when some ofthe study list items are presented as cues, their re-trieval strength increases. According to blockingtheory, items with high retrieval strength will block

Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 118–122 / ISSN 1089 - 4136). *Faculty Supervisor

Author note. Monica L. Boice, Saint Joseph’s University; Gary J.Gargano, Saint Joseph’s University.

This research was presented at the annual meeting of the East-ern Psychological Association in March 2000.

Correspondence concerning this article should be addressedto Dr. Gary J. Gargano, Department of Psychology, Saint Joseph’sUniversity, 5600 City Avenue, Philadelphia, PA 19131, or via elec-tronic mail at [email protected].


PART-SET CUING � Boice and Gargano

Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 118–122 / ISSN 1089 -4136).

access to items with weaker retrieval strength, mak-ing it more difficult to recall noncued items.

Rundus (1973) used category members as cuesto determine if conceptually related items would bebetter cues than category names. Rundus found adecrease in the number of exemplars recalled as thenumber of exemplars presented as cues increased.Rundus attributed the effect to blocking, because cuewords intruded during recall, blocking access to otherless accessible words.

According to cue-overload theory (Mueller &Watkins, 1977), the number of retrieval cues has aninhibitory effect on recall. Each occurrence of anitem, whether it is presented on the study list or as acue, increases the number of instances related to thecategory. As the number of instances increases, thecue becomes overloaded and is less effective at aid-ing recall. Thus, when participants are presented withstudy list items as cues, recall will be lower than whenzero instances are provided as cues. Because bothintralist and extralist cues overload the cue, both willdecrease recall.

Mueller and Watkins (1977) manipulated thetype of retrieval cue to support cue-overload theory.After studying a list of categorized words, participantswere presented with either intralist or extralist cues.Intralist retrieval cues consisted of items that wereincluded in the original study list. Extralist retrievalcues consisted of items that were not included in theoriginal study list but were related to the list items.The results supported predictions of cue-overloadtheory because a part-set cuing effect was found whenboth intralist and extralist cues were used.

Peynircioglu (1987) has indicated that the part-set cuing effect may not be limited to memory tasks.Peynircioglu provided participants with a nonmemorytask of generating words from the letters of a largersource word. Participants were provided with zero oreight words as cues to help generate other words.Results indicated that fewer words were generated inthe cued condition compared to the noncued condi-tion. Peynircioglu concluded that a similar mecha-nism found in traditional memory experiments (e.g.,blocking, strategy disruption, cue overload) may beresponsible for the effect.

Although there is some evidence for each theorythat offers explanations for part-set cuing, the mecha-nism producing the effect is still not fully understood.The purpose of the present study is twofold. Researchhas indicated some support for each theoreticalexplanation of the part-set cuing effect, but fewstudies have contrasted the predictions of these theo-ries. The present study tested predictions made byblocking, strategy disruption, and cue-overload

theories. Additionally, the present study furtherinvestigated whether part-set cuing can be general-ized to nonmemory tasks.

In the present study, participants generated cat-egory members when provided with a category nameand, in some cases, category members as cues. Wemanipulated the number and popularity of cues pro-vided on the recall test. Participants were providedwith either zero or eight popular, unpopular, or ran-dom category exemplars as cues. The dependentmeasure was the number of noncue items generatedfrom each category. A part-set cuing effect will befound if recall is lower in the cued conditions than inthe uncued conditions.

Blocking, strategy disruption, and cue-overloadtheories make contrasting predictions based on dif-ferent theoretical explanations of the part-set cuingeffect. According to strategy disruption, cues shouldbe disruptive regardless of their popularity, becauseeach instance of a cue will disrupt the participants’unique retrieval methods (Brown, 1968). Accordingto blocking theory, popular cues should be more dis-ruptive than unpopular cues, because stronger, moreaccessible items will block access to less accessibleitems (Rundus, 1973). Cue-overload theory predictslower recall as the number, not type, of exemplarspresented as cues increases (Watkins, 1975). Thus,the popularity of the cue should not affect recall.

MethodParticipants

One hundred eighty undergraduate students,ranging in age from 19 to 22 years, at Saint Joseph’sUniversity volunteered to participate in the study asan option for fulfilling a course requirement. All par-ticipants were debriefed and given credits for theirparticipation.

DesignThe design was a 3 (popularity of cue) × 2 (num-

ber of cues) between-subjects factorial design. Onevariable was the popularity of exemplars provided(popular, unpopular, random). The other variablewas the number of cues provided (zero or eight). Thedependent measure was the number of words cor-rectly recalled. Participants were randomly assignedto one of the six groups.

MaterialsA separate norming study was conducted to de-

termine the most frequently generated responses foreach category. Fifteen conceptual categories wereselected from the Battig and Montague (1969) cat-egory norms (see Appendix A). A similar group of


120 participants were presented with each categoryname and asked to generate as many exemplars aspossible. Based on the results of the norming study,we selected the eight most frequent responses foreach category as the popular cues; we used the eightleast frequent responses as the unpopular cues (seeAppendix B). The eight random cues were items thatwere randomly chosen from all of the possible re-sponses.

The test form consisted of 15 pages, each pagehaving one of the category names at the top of thepage for the zero-cue conditions. In the cued condi-tions, depending on the condition for popularity,eight cues were listed below the category name. Forexample, in the popular-cue condition on each pageof the test form, the category name was followed bythe eight most popular responses selected as cues (seeAppendix C). We presented all the categories andcues in the same random order to all the participants.

ProcedureThe participants were randomly placed into one

of the six experimental groups receiving either zeroor eight popular, unpopular, or random cues. We toldthe participants that on each page of the test bookletthey would see the name of a conceptual category.Their task was to generate as many instances of thecategory as possible. In the cued conditions, partici-pants were further instructed that they were providedwith cues, in order to help them generate other cat-egory members. The participants were allowed 11⁄ 2min to recall words from each category. Participantswere debriefed following the experiment.

ResultsA 3 × 2 factorial analysis of variance revealed a

significant interaction between the type of cue andnumber of cues provided at recall, F(2, 174) = 42.84,MSE = 23217.92, p < .05. To determine if a part-setcuing effect occurred, a simple main effects analysiswas conducted for each cue condition (popular, un-

popular, and random) by the number of cues pro-vided at recall (0 or 8). A part-set cuing effect wasfound in the popular-cue condition, F(1, 174) =197.94, MSE = 107272.82, p < .05, and the random-cue condition, F(1, 174) = 18.33, MSE = 9933.07, p <.05, but not in the unpopular-cue condition (p > .05).A comparison of the mean scores indicated higherrecall when no cues were provided (M = 157.57) thanwhen cues were provided (M = 73.00) in the popular-cue condition. Similar findings were found in therandom-cue condition, with higher recall when nocues were provided (M = 144.30) than when cues wereprovided (M = 118.57). Means and standard devia-tions for each condition are displayed in Table 1.

To determine the source of the part-set cuingeffect, a second simple main effects analysis was con-ducted for the number of cues provided at recall(0, 8) for each cue condition (popular, unpopular,random). The results indicated that an equal num-ber of words were recalled for the zero-cue condi-tion (p > .10) but not for the eight-cue condition,F(2,174) = 68.95, MSE = 37365.74, p < .05. A Tukeypost hoc test indicated a linear trend with recall be-ing significantly higher (p < .05) for the unpopular-cue condition (M = 142.47) than for the random-cuecondition (M = 118.57) and the popular-cue condi-tion (M = 73.00). The decreased recall in the popu-lar-cue condition is due to the fact that participantscannot recall the more accessible popular words, be-cause they are provided as cues. These items, in turn,then block access to less accessible items. This findingis supported by a comparison of the frequency of thetypes of words recalled. Participants recall an aver-age of 78.28 popular words across all conditions,whereas they recall an average of only 3.51 unpopu-lar words across all conditions.

DiscussionThe results of the Cue Type × Number of Cues

interaction indicated that a part-set cuing effect wasfound in the popular- and random-cue conditions,



TABLE 1

Means and Standard Deviations for Each Condition

Popular Unpopular Random

Number of cues M SD M SD M SD

0 157.57 20.35 152.37 24.81 144.30 27.41

8 73.00 36.32 142.47 24.91 118.57 21.52

Note. N = 30 for each condition.



Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 118–122 / ISSN 1089 -4136).

but not in the unpopular-cue condition. This findingdoes not support the predictions of strategy disrup-tion and cue overload, because these theories pre-dict the popularity of the cues would have no affecton recall.

Strategy disruption predicted that recall woulddecrease when cue words were provided, regardlessof their popularity, because any type of cue disruptsthe participants’ unique retrieval methods. Strategydisruption could not account for the effect because apart-set cuing effect was found in the popular- andrandom-cue conditions, but not in the unpopular-cuecondition.

Likewise, cue-overload theory could not accountfor the findings. Cue-overload theory predicted thatrecall would decrease as the number of items associ-ated with the category increased. According to thistheory, a part-set cuing effect should be observed inall cue conditions. Cue-overload could not explainthe effect because a part-set cuing effect was onlyfound in the popular and random conditions, not inthe unpopular condition.

Blocking theory, however, can account for thefindings. Blocking theory predicted that popular cueswould be more disruptive than unpopular cues, be-cause more accessible items block access to less ac-cessible items. The results indicated that when themore accessible popular words were presented ascues, retrieval for the less accessible words decreased.Also, because some cue words in the random-cuecondition were popular, a smaller part-set cuing ef-fect occurred.

The results suggest that when participants wereunable to generate responses with high retrievalstrength in the popular-cue condition, because theywere provided as cues, recall was suppressed. Thesestrong cue items then blocked access to weaker, lessaccessible items. However, in the unpopular-cue con-

dition, the weaker, less accessible items did not blockaccess to stronger, more accessible items. When com-paring the frequency of the type of words recalledacross all conditions, participants recalled more popu-lar words than unpopular words. It appears as thoughparticipants retrieve the most popular responses first.Then, participants either stop recalling additionalwords, or the recall of popular cues blocks access toless accessible, or unpopular responses.

The findings of the present study support thoseof Peynircioglu (1987), that part-set cuing can befound in nonmemory tasks. In addition, a memorymechanism seems to be responsible for the effect.However, because Peynircioglu did not state whethershe used strong or weak cues, it is hard to generalizeher findings to a memory mechanism as she claims.

ReferencesBattig, W. F., & Montague, W. E. (1969). Category norms for

verbal items in 56 categories: A replication and extension ofthe Connecticut category norms. Journal of Experimental Psy-chology Monograph, 80(3, Pt. 2).

Bjork, E. L., & Bjork, R. A. (1988). On the adaptive aspects ofretrieval failure in autobiographical memory. In M. M. Grune-berg, P. E. Morris, & R. N. Sykes (Eds.), Practical aspects ofmemory: Current research and issues, Vol. 1: Memory in everyday life(pp. 283–288). London: Academic Press.

Brown, J. (1968). Reciprocal facilitation and impairment of freerecall. Psychonomic Science, 10, 41–42.

Mueller, C. W., & Watkins, M. J. (1977). Inhibition from part-setcuing: A cue-overload interpretation. Journal of Verbal Learningand Verbal Behavior, 16, 699–709.

Peynircioglu, Z. F. (1987). On the generality of the part-set cuingeffect: Evidence from nonmemory tasks. Journal of Experimen-tal Psychology: Learning, Memory, and Cognition, 13, 437–442.

Roediger, H. L. (1974). Inhibiting effects of recall. Memory andCognition, 2, 261–269.

Rundus, D. (1973). Negative effects of using list items as recallcues. Journal of Verbal Learning and Verbal Behavior, 12, 43–50.

Tulving, E., & Pearlstone, Z. (1966). Availability versus accessibil-ity of information in memory for words. Journal of Verbal Learn-ing and Verbal Behavior, 5, 381–391.

Watkins, M. J. (1975). Inhibition in recall with extralist “cues.”Journal of Verbal Learning and Verbal Behavior, 14, 294–303.



APPENDIX A

List of All Category Names Used

Furniture Flowers VegetablesFabrics Fruits Weather PhenomenonBirds Trees Carpenters’ ToolsProfessions Sciences FishSports Insects Musical Instruments

APPENDIX B

Example of Popular, Unpopular, and Random Cues for the Tree Category

Popular cues Unpopular cues Random cues

Pine Cottonwood BeechMaple Balsam HemlockRedwood Gingko SpruceOak Holly LocustDogwood Mahogany WalnutWillow Mimosa OliveBirch Mulberry SycamorePalm Poplar Sequoia

APPENDIX C

Example of Test Form Used in the Popular-Cue Condition

Instructions:Please write down as many members of the category listed below.

TREES

OakPineMapleWillowRedwoodBirchDogwoodPalm


Human Sex Differences in AggressionWithin an Evolutionary Model

This investigation used evolutionary theory and social learning theory to predictsex differences in human aggression. One hundred fifteen students estimated theirfrequency of aggression in the following contexts: (a) compete to attract members ofthe opposite sex, (b) co-opt resources from others, and (c) negotiate status hierar-chies. We assessed aggression type in 3 categories: direct physical, direct verbal,and indirect. We determined sex role by using the Bem Sex-Role Inventory (Bem,1978). The social model predicts that men will engage more frequently than womenin all types of aggression. The evolutionary model predicts that men and womenwill engage in similar frequencies of total aggression but differ in the type they use;our data supported this prediction. These findings were more consistent with theevolutionary model, indicating that this theory provides a sound basis for predict-ing and understanding sex differences in aggression.

BARBARA J. HAGENAH

CHRISTOPHER HEAPS*

EUGENE GILDEN

MICHAEL ROBERTS

Linfield College

AGGRESSION IS A UNIVERSAL HUMAN BEHAVIOR

that psychologists define as “any action under-taken with the apparent intent of causing

physical or psychological harm” (Burbank, 1987, p.72); it is generally used to gain resources or otheradvantages. The primary model used to explainaggression is social learning theory. According tothis theory, individuals are taught to be aggressivethrough the influence of surrounding social andenvironmental conditions (Feldman, 2001). In addi-tion, sex differences in aggression are a result of thedifferent social roles that each sex acquires, whichare regulated by social norms (Eagly, 1987). Eagly(1987) explains that the male gender role typically“encourages varying forms of aggression” (p. 72)whereas the female gender role has “little emphasison aggressiveness” (p. 72). A social learning modelargues that these preferences dictate aggressivepatterns in individuals.

Evolutionary theory provides another model forunderstanding aggression. It uses biological, socio-logical, and historical factors to explain differencesin aggression between the sexes and across cultures.According to evolutionary theory, aggression is notsimply a learned phenomenon, nor is it an inevitableinstinctual drive. Rather, aggression often has adap-

tive benefits and is sensitive to social contexts thathave the potential to evoke aggressive behavior (Buss& Shackelford, 1997). Though behavior is affectedby both learned social and underlying biological at-tributes, it is only in recent years that researchers haveconsidered both attributes simultaneously. We feelthat exploring aggression using evolutionary theorywill add to our understanding of it. For this reason,the majority of the introduction will be devoted tothis perspective.

Tooby and Cosmides (1992) discuss guidelinesfor approaching psychology from an evolutionary per-spective. They argue that humans have specializedrather than generalized information-processingmechanisms, and these mechanisms are context-specific. To explain a human behavior, the researchermust look at how it is related to an adaptive problem

Copyright 2001 by Psi Chi, The National Honor Society in Psychology (Vol. 6, No. 3, 123–131 / ISSN 1089 - 4136). *Faculty Supervisor

Author note. This article is based on the undergraduate researchwork of Barbara J. Hagenah, Department of Psychology, LinfieldCollege. Christopher Heaps and Eugene Gilden, Department ofPsychology, and Michael Roberts, Department of Biology, LinfieldCollege, formed the honors thesis committee and supported thedevelopment of this study. A poster of this study was presented atthe annual meeting of the Western Psychological AssociationConvention on April, 15, 2000, in Portland, Oregon.

Correspondence concerning this article should be addressedto Barbara J. Hagenah, c/o Christopher Heaps, Department ofPsychology, Linfield College, McMinnville, OR 97128.


HUMAN SEX DIFFERENCES IN AGGRESSION � Hagenah, Heaps, Gilden, and Roberts

and consider the ancestral environment in whichnatural and sexual selective pressures influencedhuman behavior (Tooby & Cosmides, 1992). Naturalselection is a process that controls variation of physi-cal, physiological, or behavioral traits within a popu-lation, and influences life, death, and reproduction.Traits having a genetic basis spread if they are morefavorable to survival and reproduction in the face ofselection pressures. Sexual selection is the specificprocess that selects traits influencing reproductivesuccess. The two main tasks involved in reproductivesuccess are finding and attracting mates and provid-ing care to offspring (Geary, 1998). The operation ofthese forces on human behavior means that the vari-ables we observe in cultures and individuals are “theproduct of a common, underlying evolved psychol-ogy, operating under different circumstances” (Tooby& Cosmides, 1992, p. 45). Moreover, these selectivepressures make testable predictions about sex differ-ences in aggression (Buss & Shackelford, 1997).

The evolutionary model argues that aggressionhas been an adaptive behavior for both sexes incertain situations. The model proposes that menand women experienced the pressures of sexualselection differently in our ancestral past (Barkow,Cosmides, & Tooby, 1992; Buss, 1995; Geary, 1998)and that these differences account for consistentsex differences in type of aggression (Björkqvist,Österman, & Kaukiainen, 1992; Burbank, 1987;Campbell, Sapochnik, & Muncer, 1997; Daly & Wil-son, 1988; Eagly & Steffen, 1986). Men and womendiffer according to the amount of effort and the strat-egies used in approaching mating and parenting.Specifically, the amount of parental energy that isrequired to successfully reproduce and rear offspringdiffers greatly between the sexes. Primarily due togestation and postnatal care, women must invest moreresources in each offspring than men. Men do nothave the same gestation requirements, and althoughmale investment of resources after birth typically in-creases survival of offspring, the amount of effort andresources expended by men is less than the amountexpended by women. These factors influence behav-ioral differences between the sexes (Geary, 1998). AsBuss (1995) notes, “Strong sex differences occur re-liably in domains closely linked with sex and mating,precisely as predicted by psychological theories basedon sexual selection” (p. 166).

Björkqvist, Lagerspetz, and Österman (1992) pro-vide an inclusive model for aggression that includesassessment of three types of aggression: direct physi-cal (e.g., hit, kick), direct verbal (e.g., yell, insult),and indirect (e.g., gossip, ostracize). Variation in thetype of aggression men and women use would be re-

lated to sex differences in amount of parental energyneeded for rearing successful offspring. The repro-ductive efforts required by women will discouragethem from using direct physical aggression becauserisk of injury would jeopardize biological resourcescrucial to success in bearing and rearing offspring(Daly & Wilson, 1988). Therefore, women who uti-lize indirect aggression could benefit from aggressingwith less associated risk of physical injury. Men, hav-ing a lower degree of parental investment, would nothave the same adaptive pressure to avoid physicalaggression (Daly & Wilson, 1988). Consequently, menwill employ direct physical aggression relatively fre-quently. Verbal aggression would be used by bothsexes with similar frequency because general verbalability has similar adaptive advantages for men andwomen, and a selective pressure does not likely existfor one sex to utilize direct verbal aggression morethan the other (Geary, 1998).

Another necessary consideration for understand-ing aggression is that aggressive behaviors can beunderstood as “context-sensitive solutions to parti-cular adaptive problems of social living” (Buss &Shackelford, 1997, p. 605). Aggressive behavior maybe present when such contexts emerge. Buss andShackelford (1997) describe seven contexts or situa-tions in which aggression elicits adaptive benefits.These contexts are: intrasexual (same-sex) competi-tion, co-opting the resources of others, negotiatingstatus and dominance hierarchies, deterring long-term mates from infidelity, aggressing toward unre-lated children, defense against attack, and deferringrivals from future aggression. These contexts can bebetter understood when divided into three broadercategories based on function.

The first category includes intrasexual competi-tion, co-opting resources from others, and negotiat-ing status and dominance hierarchies. We were in-terested in these categories for two reasons. First, bothmen and women gain important benefits from hav-ing greater resources, status, or better mates. For thisreason, differences in the type of aggression employedto reach these goals should relate to the differing strat-egies associated with reproduction and not to differ-ences in relative importance of the contexts. Second,the contexts are important social situations under-lying mate choice and retention prior to forminglong-term attachments or bearing children. Thus,these situations are particularly appropriate for mea-surement in a college-aged population.

A second category includes the contexts of de-terring long-term mates from infidelity and aggres-sion toward unrelated children. We did not includethese contexts in this study because they are related


to situations that occur after mating and, therefore,are not generally appropriate for measurement in acollege-aged population. A third category includesdefense against attack and deterring rivals from futureaggression. These contexts deal with reactions eitherto other’s aggression or with an individual’s aggressionin the past. These contexts are removed from the ini-tial aggressive intent, and for this reason we did notinclude them in this study.

Predictions about sex differences in aggressionwill differ depending on whether a social learningmodel or an evolutionary model is used to make thepredictions. In the contexts of intrasexual competi-tion, co-opting resources, and negotiating status anddominance hierarchies, a social learning model pre-dicts that overall aggression will be higher for menthan women. This prediction is made because themodel proposes that each sex has acquired a role inhow each should aggress, and these patterns shouldexist in each context. Direct physical aggressionshould show the greatest difference, men engagingmost frequently, but men should also show moreindirect and direct verbal aggression than women(Eagly, 1987).

An evolutionary model predicts that overallaggression in each of these contexts may have simi-lar benefits for both men and women. For this rea-son, in contrast to the social model, this model pre-dicts that the amount of aggression for both sexeswill be similar in each context and across all contexts.However, in light of sex differences in selective pres-sures associated with reproduction and parentalinvestment, this model proposes that differences willoccur in the type of aggression men and women usein these contexts. It predicts that men will consistentlyengage in more direct physical aggression thanwomen in each context, and utilize lower levels ofindirect aggression. It predicts that women will bemore indirectly aggressive than men in each context.This model predicts that men and women will havesimilar levels of high direct verbal aggression in allcontexts.

Utilizing evolutionary theory to explain sex dif-ferences implies that the differences are not arbitrarilysocialized. A sex role inventory that assesses mascu-linity, femininity, and androgyny will provide insightto participants’ perceptions of their gender identityand attempts to account for biological sex as a pre-dictor of aggression type independently of learnedgender roles. Social learning theory, an approach thatexplains sex differences in aggression by differinglearned gender roles, suggests that masculine andfeminine gender roles will have an influence on bothmen’s and women’s aggressive patterns as predicted

by the social model. However, an evolutionary modelwould predict that regardless of sex role, men andwomen would behave as predicted by the evolution-ary model, because sex role does not change thereproductive differences in men and women.

To test these predictions we administered a ques-tionnaire to college-aged men and women assessingreported aggressive frequency in three types of aggres-sion: direct physical, direct verbal, and indirect, inthe three contexts of intrasexual competition, co-opting resources, negotiating status and dominancehierarchies. We used the Bem Sex-Role Inventory(Bem, 1978) to assess participants’ perceived sex role.

MethodParticipants

One hundred and fifteen college students (49men, 66 women: mean age = 19.46, SD = 1.24) volun-teered to participate. Participants responded to pub-licity posted in the psychology department andcampuswide.

MaterialsWe adapted the Direct and Indirect Aggression

Scale (DIAS) questionnaire from Björkqvist, Lager-spetz, et al. (1992) to assess frequency in aggression.Adaptations required the modification of pronounsto be expressed in the first person. Aggression wasdefined in three categories: direct physical (e.g., hit,kick), direct verbal (e.g., yell, insult), and indirect(e.g., gossip, ostracize). There were 24 questions (7physical, 5 verbal, 12 indirect), and participants ratedfrequency of aggressive behaviors on a scale rangingfrom 0 (never) to 4 (very often). Appendix A shows themodified version of the DIAS. We included a descrip-tion of three contexts in which aggression could con-fer adaptive benefits (Buss & Shackelford, 1997), andreported frequency of aggression was considered ineach context. The contexts included intrasexual com-petition, co-opting resources, and negotiating status;these contexts were counterbalanced. Appendix Bshows the context descriptions that were used for thisstudy.

We used the Bem Sex-Role Inventory (Bem, 1978)to assess masculine, feminine, androgynous, and un-differentiated gender types. Sixty gender-descriptivewords are listed of which there were 20 masculine(e.g., forceful, willing to take risks), 20 feminine (e.g.,affectionate, sensitive to needs of others), and 20neutral (e.g., moody, reliable) items. Participants self-assessed each item on a scale ranging from 1 (never oralmost never true) to 7 (always or almost always true).Masculine sex roles result from a high masculine anda low feminine score. Feminine sex roles result from




a low masculine and high feminine score. Androgy-nous sex roles result from both high masculine andfeminine scores. Undifferentiated is defined as hav-ing a low masculine and feminine score such that agender type cannot be assigned.

ProcedureWe conducted testing in 4-hr blocks for 3 days;

participants came at their convenience. Each partici-pant responded to the three context descriptions and

self-reported the likelihood of their aggression foreach question in the last 4 years. Following the ag-gression questionnaire, participants filled out theBem Sex-Role Inventory.

ResultsA 2 (sex) × 3 (evolutionary context) × 3 (aggres-

sion type) analysis of variance (ANOVA) revealed thatmen and women aggress with equal frequency acrosscontexts. Table 1 shows the ANOVA. A significant Sex



TABLE 1

Analysis of Variance for a Three-Factor Study: Sex �� Context �� Aggression Type

Source of variation SS df MS F P

Sex (S) 1.536 1 1.536 0.990 0.322

Context (C ) 9.810 2 4.905 20.769 0.001

Aggression type (A) 133.523 2 66.761 198.051 0.001

SC interactions 0.323 2 0.161 0.683 0.506

SA interactions 4.201 2 2.101 6.232 0.002

AC interactions 10.875 4 2.719 24.945 0.001

SCA interactions 0.507 4 0.127 1.163 0.326

Error 173.810 112 1.522

Total 445.024 1 445.024 286.764 0.000

TABLE 2

Means and Standard Deviations of Aggression Frequency Scores forMen and Women According to Contexts and Aggression Types

Aggression type

Direct physical Direct verbal Indirect Context total Overall total

Context Men Women Men Women Men Women Men Women

Intrasexual M a.20a a.02a 1.09 .78 1.02 1.13 .80 .73 .75competition SD .46 .15 .69 .62 .63 .75 .52 .47 .49

Co-opting M b.23b b.13b .87 .78 .55 .62 .52 .51 .52resources SD .33 .23 .72 .71 .53 .57 .44 .44 .44

Negotiating M c.24c c.08c 1.21 1.05 .92 1.06 .78 .77 .78status SD .45 .17 .76 .77 .65 .77 .51 .54 .52

Aggression M d.23d d.08d 1.06 .87 .83 .93 .70 .67type total SD .38 .13 .60 .59 .51 .60 .44 .42

Overall M .14 .95 .89total SD .27 .60 .56

Note: All superscripts indicate p < .01. Comparisons a, b, c, and d are illustrated in Figures 1–4, respectively.




FIGURE 1

Men’s and women’s mean frequency scores fordirect physical aggression in the context

of intrasexual competition.

Men WomenSex

Mea

n F

req

uen

cy o

fP

hys

ical

Ag

gre

ssio

n

Intrasexual Selection

0.25

0.2

0.15

0.1

0.05

0

FIGURE 2

Men’s and women’s mean frequency scores fordirect physical aggression in the context of

co-opting resources from others.

Men WomenSex

Mea

n F

req

uen

cy o

fP

hys

ical

Ag

gre

ssio

n

Co-opting Resources

0.25

0.2

0.15

0.1

0.05

0

FIGURE 3

Men’s and women’s mean frequency scores fordirect physical aggression in the context of

negotiating status and dominance hierarchies.

Men WomenSex

Mea

n F

req

uen

cy o

fP

hys

ical

Ag

gre

ssio

n

Negotiating Status and Dominance

0.3

0.25

0.2

0.15

0.1

0.05

0

FIGURE 4

Men’s and women’s mean frequencyscores for direct physical aggression

in all contexts.

Men WomenSex

Mea

n F

req

uen

cy o

fP

hys

ical

Ag

gre

ssio

n

All Contexts

0.25

0.2

0.15

0.1

0.05

0

× Aggression Type interaction was found in whichdirect physical aggression was more frequent amongmen than women in the context of intrasexual com-petition, t(113) = –2.9, p < .004, co-opting resources,t(113) = –2.04, p < .043, negotiating dominance andpower hierarchies, t(113) = –2.73, p < .007, and overall contexts, t(113) = –2.99, p < .003. Figures 1, 2, 3,and 4 show the mean frequency scores of direct physi-cal aggression in each context and over all contextsfor men and women. Men and women were similarin their level of indirect aggression. No sex differencesin frequency of verbal aggression were found. Table

2 shows the mean and standard deviation scores foraggression types in each context, as well as scores fortotal aggression type over all contexts, for men andwomen.

There was also a significant effect for aggressiontype and context in which direct verbal aggressionwas more common in the contexts of achieving sta-tus/dominance than it was in the contexts onintrasexual competition, t(114) = –.204, p < .001, andco-opting resources, t(114) = –.299, p < .001. Indirectaggression was also more common in the contexts ofintrasexual competition, t(115) = .48, p < .001, and


status/dominance, t(115) = –.413, p < .001, than inthe context of co-opting resources.

A one-way ANOVA and Tukey’s test showed therewere no significant interactions in which masculin-ity, femininity, or androgyny was a significant predic-tor of how men or women of each gender type wouldbehave aggressively. Undifferentiated was not consid-ered in the analysis because this group does not re-veal sex role.

DiscussionOur results indicate that men and women report

aggressing with similar frequency across the contextsof intrasexual competition, co-opting resources, andnegotiating status and dominance hierarchies. Thispattern of results is consistent with the evolutionarymodel’s prediction and not with that of the sociallearning model. Men did report engaging in directphysical aggression significantly more than women,as predicted by both models. Men and women re-ported engaging with similar frequency in indirectaggression. This result was not predicted by eithermodel. Men and women reported similar frequen-cies of direct verbal aggression as predicted by theevolutionary model and not by the social learningmodel. Because no predictions were made from eithermodel regarding significant interactions betweencontext and aggression type, these findings, althoughinteresting, do not add clarification to sex differencesin aggression. In addition, the interaction betweencontext and aggression type does not directly lendsupport to one model or the other. For these reasons,these results will not be a part of the discussion.

As suggested by the evolutionary model, ourresults indicate that sex role identity is not a signifi-cant predictor of how men and women will reportfrequency of aggression or type of aggression. Regard-less of whether men or women were masculine orfeminine, these gender types did not influence howeach sex would report aggression. By comparing thepredictions each model makes regarding men’s andwomen’s aggressive behaviors, we suggest that an evo-lutionary model better explains our data than thesocial learning model. For this reason, the remain-der of the discussion will examine an understandingof the data from an evolutionary perspective and itsimplications for understanding human behaviorwithin this framework.

It is important to note again that frequency ofaggression in this study was determined by self-report,and this should be taken into account in interpret-ing the results. Also, the mean frequencies of aggres-sion, calculated on a scale from 1 to 4, did not goabove 2, resulting in a population that did not report



high frequencies of aggression. Overall aggression inour population was low, but the emphasis of this studywas to examine sex differences in aggressive behav-ior, and these differences were significant.

Because the evolutionary model is based on bothbiological and social attributes of humans, cross-cultural comparisons are important parts of theevolutionary approach. Using evolutionary theoryto interpret our data in conjunction with other stud-ies yields important implications for understandingaggression. First, both men and women reportedengaging in a similar amount of aggression. Thisresult suggests that aggression is beneficial to bothsexes in many social situations, and that the contextsof intrasexual selection, co-opting resources, andnegotiating status and power hierarchies are situa-tions in which aggression may be adaptive for bothmen and women.

Men and women differed in the frequency withwhich they reported to engage in the specific typesof aggression, and the differences were consistentacross the three contexts. In agreement with ourresults, other studies indicate that in women, physi-cal aggression is less frequent than other types ofaggression (Burbank, 1987; Buss & Shackelford, 1997;Eagly & Steffen, 1986). This pattern suggests a ten-dency for women to avoid direct physical aggression,which is consistent with sexual selection theory inthat high amounts of maternal care given to offspringrequire that women minimize their risk of physicalinjury (Buss & Shackelford, 1997). In a review ofpaternal investment, Geary (1998) demonstrated thatcross-culturally and historically the mortality ofmothers greatly increased the likelihood of the child’smortality. Thus, the constraints of the female repro-ductive strategy may have led to a selective pressurefor women to engage in low amounts of physicalaggression.

In this study, indirect aggression was the type mostfrequently reported by women, suggesting that theremay be a selective advantage for women to use indi-rect aggression. Other studies have shown that womenuse significantly more indirect aggression than men(Björkqvist, Österman, et al., 1992; Campbell et al.,1997). High levels of parental care that characterizethe female reproductive strategy seem to be stronglyconnected to utilizing indirect methods of aggression.Indirect aggression allows women to gain benefitssuch as mates, resources, or status, while simulta-neously decreasing their risk of physical injury.

Women also report high frequencies of directverbal aggression. Eagly and Steffen (1986) found thatwomen frequently use a form of aggression that causespsychological harm (i.e., aggressive acts that are not



physical). Burbank (1987) found that across cultureswomen frequently use verbal aggression (in whichdirect vs. indirect verbal aggression was not distin-guished). To more precisely understand direct ver-bal aggression and its relation with the female repro-ductive strategy of avoiding physical aggression, as-sessment of women’s beliefs about the risk of physi-cal injury as a result of verbal aggression are needed.

Our results suggest that men report using directphysical aggression significantly more than women.The predominance of direct physical aggression inmen has also been consistently reported in other stud-ies (Björkqvist, Österman, et al., 1992; Burbank, 1987;Daly & Wilson, 1988; Eagly & Steffen, 1986). Becauseof the lower degree of parental investment requiredof men, there has not been a selective pressure formen to avoid physical aggression to the same extentas women. Whereas men use physical aggression morethan women, direct verbal and indirect aggressionwere the types of aggression most frequently reportedby men in this study. Consideration of cultural differ-ences in paternal investment can provide understand-ing of the relation between the amount of male physi-cal aggression relative to the amount of direct verbaland indirect aggression. In contrast to the consistentamount of maternal investment given to offspring,the amount of paternal investment differs betweencultures (Geary, 1998). We suggest the differences inamount of paternal investment relate to type ofaggression men most frequently use.

In societies with high paternal investment, weexpect to see lower levels of direct physical aggres-sion and higher levels of direct verbal and indirectaggression. In contrast, in societies with low paternalinvestment, we expect higher levels of male directphysical aggression and lower levels of direct verbaland indirect aggression. Male investment is higherin societies that are industrial or stratified, wheremonogamous relationships are predominant (Geary,1998). Examples of such societies include the UnitedStates (in the present study), Finland (Björkqvist,Österman, et al., 1992), and England (Campbell etal., 1997). These studies showed that men use physi-cal aggression less than they use direct verbal andindirect aggression. In cultures characterized by po-lygamous systems with high incidence of warfare, raid-ing, and male social displays in which physical aggres-sion could lead to higher status or numerous mates,direct paternal investment needed to ensure offspringsurvival is low (Geary, 1998). Chagnon and Bugos’s(1979) observations of the Yanomamö indicate thistype of culture, and men in this society show very highamounts of physical aggression and low levels ofdirect verbal and indirect aggression. Although more

work is needed to understand aggressive behavior asa function of parental investment, the studies dis-cussed indicate that this type of relation may applyacross cultures.

We view this research as an initial attempt toencourage the value of using an evolutionary frame-work for understanding human behavior in generaland aggression in particular. Further work thatapproaches aggression in relation to differences inmale and female reproductive strategies is a worth-while consideration. However, to make strong claimsabout human aggression, direct testing of actualbehaviors is necessary. In order to better understandthis relation, we suggest that the continuation ofcross-cultural research in conjunction with definingaggression as a potentially adaptive behavior thatcan manifest in either direct or indirect form, willfurther our understanding of human aggression.

This research has barely touched on the idea oftesting predictions made by the evolutionary modelwith predictions made by a social model. To com-prehensively test one model against the other, manyfactors such as participants’ beliefs about their be-havior, other sex role inventories, and more diversepopulations are necessary to thoroughly examine thetwo models. It is apparent that social and culturalfactors will shape both individual and societal mani-festations of behavior. However, our work suggests thatthese manifestations cannot be understood withoutreference to evolutionary theory, and that evolution-ary theory provides a powerful basis for understand-ing general patterns of male and female aggression.Recognizing the trends that exist based on evolution-ary selective pressures will aid scientists in under-standing a complete picture of the nature of humanaggression.

ReferenceBarkow, J. H., Cosmides, L., & Tooby, J. (Eds.). (1992). The adapted

mind: Evolutionary psychology and the generation of culture. NewYork: Oxford University Press.

Bem, S. L. (1978). Bem Sex-Role Inventory. Redwood City, CA: Con-sulting Psychologists Press.

Björkqvist, K., Lagerspetz, K. M. J., & Österman, K. (1992). Directand indirect aggression scales (DIAS). In K. Björkqvist & K.Österman (Eds.), Pro Facultate: No. 4. Scales for research on inter-personal relations. Vasa, Finland: Faculty of Social and CaringSciences, Åbo Akademi University.

Björkqvist, K., Österman, K., & Kaukiainen, A. (1992). The devel-opment of direct and indirect aggressive strategies in malesand females. In K. Björkqvist & P. Niemlä (Eds.), Of mice andwomen: Aspects of female aggression (pp. 51–64). San Diego, CA:Academic Press.

Burbank, V. K. (1987). Female aggression in cross-cultural per-spective. Behavior Science Research, 21, 70–100.

Buss, D. M. (1995). Psychological sex differences: Origins throughsexual selection. American Psychologist, 50, 164–168.

Buss, D. M., & Shackelford, T. K. (1997). Human aggression inevolutionary psychological perspective. Clinical Psychology


APPENDIX A

Modified DIAS

Answer the questions by circling the number which seems to tell about your behavior in the closest way.

0 = never, 1 = seldom, 2 = sometimes, 3 = quite often, 4 = very often

1. Hit the person?

2. Shut the person out of the group?

3. Yell at or argue with the person?

4. Become friends with another as a kind of revenge?

5. Kick the person?

6. Ignore the person?

7. Insult the person?

8. Gossip about the person you are angry with?

9. Trip the person?

10. Tell bad or false stories about the person?

11. Say that you are going to hurt the person?

12. Plan secretly to bother the person?

13. Shove the person?

14. Say bad things behind the person’s back?

15. Call the person names?

16. Say to others, “Let’s not be with him/her!”?

17. Take things from the person?

18. Tell the person’s secrets to a third person?

19. Tease the person?

20. Write small notes where the person in criticized?

21. Push the person down on the ground?

22. Criticize the person’s hair or clothing?

23. Pull at the person?

24. Try to get others to dislike the person you are angry with?

Note: Physical aggression is indicated in numbers 1, 5, 9, 13, 17, 21, and 23.Verbal aggression is indicated in numbers 3, 7, 11, 15, and 19.Indirect aggression is indicated in numbers 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, and 24.



Review, 17, 605–619.Campbell, A., Sapochnik, M., & Muncer, S. (1997). Sex differ-

ences in aggression: Does social representation mediate formof aggression? British Journal of Social Psychology, 36, 161–171.

Chagnon, N. A., & Bugos, P. E., Jr. (1979). Kin selection andconflict: An analysis of a Yanomamö ax fight. In N. A. Chagnon& W. Irons (Eds.), Evolutionary biology and human social behav-ior: An anthropological perspective (pp. 213–238). North Scituate,MA: Duxberry Press.

Daly, M., & Wilson, M. (1988). Homicide. Hawthorne, NY: Aldinede Gruyter.

Eagly, A. H. (1987). Sex differences in social behavior: A social-role

interpretation. Hillsdale, NJ: Erlbaum.Eagly, A. H., & Steffen, V. J. (1986). Gender and aggressive behav-

ior: A meta-analytical review of the social psychological litera-ture. Psychological Bulletin, 100, 309–330.

Feldman, R. S. (2001). Social psychology (3rd ed.). Upper SaddleRiver, NJ: Prentice-Hall.

Geary, D. C. (1998). Male, female: The evolution of human sex differ-ences. Washington, DC: American Psychological Association.

Tooby, J., & Cosmides, L. (1992). The psychological foundationsof culture. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.),The adapted mind: Evolutionary psychology and the generation ofculture (pp. 19–136). New York: Oxford University Press.



APPENDIX B

Context Descriptions

1. Context description of intrasexual competition:

We are interested in the things people do to make other members of their own sex seem undesirable

to members of the opposite sex. For each of the following items, estimate how often (in approximately the

last four years) that you have used each of the methods to make other members of your own sex undesirable

to members of the opposite sex.

2. Context description of co-opting resources from others:

We are interested in the things that people do to get “resources” from other people. These resources

could be money, favors, objects—anything you might want that someone else has. For each of the following

items, estimate how often (in approximately the last four years) that you have used each of these methods to

get something you wanted from another person or group of people.

3. Context description of negotiating status or dominance hierarchies:

We are interested in the things that people do to increase their status or dominance—how they

make themselves “look better” or “seem better,” how they may try to appear tougher, smarter, in more control

of people or situations, etc. For each of the following items, estimate how often (in approximately the last

four years) that you have used each of these methods to increase your status or dominance.


Sincere appreciation is expressed for the hard workon the part of the following individuals who servedas reviewers for this issue. Without the assistance ofsuch dedicated professionals, the Psi Chi Journalsimply would not be able to function!

—EDITOR

Reviewers

John Broida, University of Southern Maine

Charles I. Brooks, King’s College

Sheila Brownlow, Catawba College

Lisa N. Coates-Shrider, McMurry University

Susan Conradsen, Berry College

David Cox, East Tennessee State University

F. Richard Ferraro, University of North Dakota

Robert W. Flint, Jr., The College of Saint Rose

James N. Forbes, Angelo State University

Gary Gargano, Saint Joseph’s University

John A. Juve, University of Missouri–Columbia

Allen Keniston, University of Wisconsin–Eau Claire

Defne Koraman, Weehawken, NJ

Jason P. Kring, University of Central Florida

John Kulig, Plymouth State College


Psi Chi Research Awardsand Grants

PSI CHI ANNUALLY SPONSORS NATIONAL UNDER-graduate and graduate research award com-petitions, as well as research awards for mem-

bers submitting the best research for the regionaland national paper/poster sessions. Members areencouraged to begin research papers early to submitfor presentation at local, state, regional, or nationalconventions. Chapters are encouraged to provide anopportunity for members to rehearse their papersbefore an audience prior to presenting them at aconvention.

In addition, Psi Chi also sponsors programs tofund student and faculty research. Descriptions of theaward/grant competitions follow. Further informa-tion and submission forms may be obtained from PsiChi’s national website (www.psichi.org ) or from thePsi Chi National Office, P.O. Box 709, Chattanooga,TN 37401-0709; telephone: (423) 756-2044; e-mail:psichi@ psichi.org.

Guilford AwardsAll Psi Chi undergraduate members are eligible to

submit their research for the Psi Chi/J. P. GuilfordUndergraduate Research Awards. Cash awards are$1,000 for first place, $650 for second place, and $350for third place. In addition, all winners and their fac-ulty research advisors receive award certificates. Theabstracts of the winning papers, as well as photographsand brief biographies of the top three winners, are

published in Eye on Psi Chi. The deadline for this awardis May 1 (postmark).

Allyn & Bacon AwardsThe Psi Chi/Allyn & Bacon Psychology Awards,

sponsored by Allyn & Bacon Publishers, are open toall undergraduate Psi Chi members and are awardedto those who submit the best overall empirical re-search papers. The awards are $500 for first place,$300 for second place, and $200 for third place. Inaddition, all winners and their faculty research advi-sors receive award certificates. The abstracts of thewinning papers, as well as photographs and briefbiographies of the top three winners, are publishedin Eye on Psi Chi. The deadline for this award isApril 1 (postmark).

Erlbaum AwardsThe new Psi Chi/Erlbaum Awards in Cognitive

Science, sponsored by publisher Lawrence ErlbaumAssociates, Inc., are open to all Psi Chi under-graduate and graduate Psi Chi members and areawarded to those who submit the best overallempirical studies in the area of cognitive science.The awards are $500 for the first-place graduatestudent and $500 for the first-place undergraduatestudent. In addition, the winners and their facultyresearch advisors receive award certificates. Theabstracts of the winning papers, as well as photo-


graphs and brief biographies of the top two winners,are to be published in Eye on Psi Chi. The deadlinefor this award is April 1 (postmark).

Newman Graduate AwardAll psychology graduate students are eligible to

submit their research for the Psi Chi/APA Edwin B.Newman Graduate Research Award. The winner re-ceives: (1) travel expenses to attend the APA/Psi ChiNational Convention to receive the award, (2) a three-year subscription to an APA journal of the winner’schoice, and (3) two engraved plaques, one for thewinner and one for the winner’s psychology depart-ment, as a permanent honor to the winner. In addi-tion, the abstract of the winning paper, as well as aphotograph and brief biography of the winner, ispublished in Eye on Psi Chi. This award is the onlystudent research award presented during the presti-gious APA/APF Awards ceremony at the APA/Psi ChiNational Convention in August. The deadline for thisaward is February 1 (postmark).

Regional Research AwardsAll Psi Chi members (undergraduate and gradu-

ate) are eligible to submit their research for the Re-gional Research Awards. Cash awards of $300 eachare presented to students submitting the best researchpapers to Psi Chi sessions at regional conventions.The number of awards in each region vary with thesize of the regions; 78 awards of $300 each are avail-able for the 2001–2002 year. Award monies are dis-tributed at the conventions following the presenta-tions. The Psi Chi regional vice-presidents each senda Call for Papers and a letter to the Psi Chi chaptersin their respective regions during the fall. These let-ters include information about the Regional ResearchAwards, the regional conventions, and submissiondeadlines for Psi Chi programs. Deadlines for sub-missions vary according to region and sometimesfrom year to year; check your fall regional mailing orPsi Chi’s national website (www.psichi.org) for details.

National ConventionResearch Awards

All Psi Chi members (undergraduate and gradu-ate) are eligible to submit their research for the Na-tional Convention Research Awards. Cash awards of

$300 each are presented to students submitting thebest research for Psi Chi sessions at the APA and APSnational conventions. Up to eight awards are given:four for the APA Convention and four for the APSConvention. Award monies are distributed at the con-ventions following the presentations. A Call for Pro-posals is mailed to all chapters in the fall and is alsoavailable from the Psi Chi National Office or the PsiChi national website (www.psichi.org). The deadlinefor submissions to the Psi Chi student sessions at boththe APA and APS conventions is December 1(postmark).

Hunt Research AwardsAll Psi Chi student and faculty members are eli-

gible to apply for a Thelma Hunt Research Award.Up to three awards of $3,000 each will be presentedannually to enable members to complete empiricalresearch that addresses a question directly related toPsi Chi, as posed by either (1) the Psi Chi NationalCouncil or (2) the researcher submitting a proposal.Unlike other national Psi Chi award/grant programs,the Hunt Awards focus on research directly relatedto the mission of Psi Chi. The deadline for this awardprogram is October 1 (postmark).

Undergraduate Research GrantsAll undergraduate Psi Chi members are eligible to

apply for these undergraduate research grants. Thepurpose of this program is to provide funds for mem-bers to defray the cost of conducting a researchproject. Applicants may request up to $1,500 for eachproject. A total of $45,000 has been allotted for thisstudent grant program. The deadline for this grantprogram is October 1 (postmark).

Faculty Advisor Research GrantsAll current faculty advisors and coadvisors who have

served an active Psi Chi chapter for at least one yearare eligible to apply for these new faculty advisor re-search grants. The purpose of this program is to pro-vide funds for advisors to defray the direct costs ofconducting a research project (no stipends included).Two grants will be awarded annually within each ofPsi Chi’s six regions. The maximum amount of eachgrant will be $2,000. The deadline for this grant pro-gram is June 1 (postmark).

REV. 7-00

*

Psi Chi Journal ofUndergraduate Research

Subscriptions and Back IssuesThe Psi Chi Journal of Undergraduate Research is a national, fully reviewed, quarterly journal dedicatedto the publication of undergraduate student research. All active Psi Chi chapters receive one compli-mentary subscription to the journal. We encourage each chapter to see that an additional subscriptionis obtained for the school library and that other organizations and interested individuals are madeaware of its availability. Every effort has been made to provide a high-quality publication and yet offerthe journal at affordable subscription rates to ensure its availability to all interested students, facultymembers, and institutions. Back issues and bulk orders for classroom use are also available.

Journal Subscription and Back Issue Order Form

Mail this form to: Psi Chi National Office (Subscriptions), P.O. Box 709, Chattanooga, TN 37401-0709

SUBSCRIPTION RATESIndividuals: $20/year (4 issues)Institutions: $40/year (4 issues)Subscription orders must be prepaid and are based on thecalendar year. Make checks payable to Psi Chi.

BACK ISSUE RATESEntire Volumes (Individuals): $20Entire Volumes (Institutions): $40Single issues (Individuals): $6 per copySingle issues (Institutions): $12 per copy

CLASSROOM USEBulk orders for classroom use are available at a discount.Faculty members should contact the Psi Chi National Office( [email protected]) for availability, rates, and shipping costs.

INTERNATIONAL (OUTSIDE U.S.)Canada: Add $10 for annual subscriptions; add $2.50 for

single issues.Other International: Contact Psi Chi National Office for rates.

SHIPPING ADDRESS (please type or print):

Name

Address

City/State /Zip+4

SUBSCRIPTION ORDER

Check one: � New Subscription � Renewal

Check all that apply:� 2001 � 2002 � 2003 � 2004

@ $20 ( individual) =

@ $40 ( institution) =

BACK ISSUE ORDER(see rates at left)

Enter number you are ordering beside all that apply:

2000 __Entire volume __Spring __Summer __Fall __Winter




1996 __Entire volume __Spring/Summer __Fall/Winter

Subscription total _________________= $

Back Issue total ___________________ = $

International total (see rates at left) __= $

Total amount enclosed _____________ = $

Journal of Psychological Inquiry• Founded 1996 by the Great Plains Behavioral

Research Association• Authors must be undergraduate students.• Publishes empirical studies, literature

reviews, and historical articles concerningany topical area in the psychological sciences.

• Submissions must (a) come from students atinstitutions that sponsor the Great PlainsStudents’ Psychology Convention and theJournal of Psychological Inquiry or (b) havebeen accepted for or presented at themeeting of the Great Plains Students’Psychology Convention, the Association forPsychological and Educational Research inKansas, the Nebraska Psychological Society,or the Arkansas Symposium.

• Contact: Mark E. Ware, Managing EditorDepartment of PsychologyCreighton University2500 California PlazaOmaha, NE 68178-0001Telephone: (402) 280-3193E-mail: [email protected] URL: http://puffin.creighton.edu/psy/

journal/JPIhome.html

Journal of Psychology and the BehavioralSciences• Founded 1966• One issue per year• Authors may be undergraduate or graduate

students with faculty mentor.• Contact: Dr. Daniel J. Calcagnetti

Faculty EditorDepartment of PsychologyFairleigh Dickinson University285 Madison AvenueMadison, NJ 07940-1099Telephone: (973) 443-8974E-mail: [email protected] URL: http://alpha.fdu.edu/psychweb

Modern Psychological Studies• Founded 1992• Two issues per year: September and March• Primary author must be an undergraduate

student.• Preferred submission deadlines: April and

October• Publishes experimental research, but will

also consider theoretical papers, literaturereviews, and book reviews.

• Contact: Editor, MPSDepartment of PsychologyUniversity of Tennessee at Chattanooga615 McCallie AvenueChattanooga, TN 37403-2598Telephone: (423) 785-2238, 755-4262E-mail: [email protected]

Journal for Undergraduate Research inPsychology• Founded 1997• Published quarterly (March, June,

September, December)• Online journal dedicated to undergraduate

research presented at local, regional, ornational conferences.

• Abstracts accepted from any conference.• Award winners may have full text of paper

published online.• Contact: Dr. Chris Koch

Director of Undergraduate Studies inPsychology

Department of PsychologyGeorge Fox UniversityNewberg, OR 97132Web URL: http://www.georgefox.edu/jurp/

Other Journals of Student Research in Psychology

The journals listed above all solicit and publish research in psychology conducted and written by students.Journals published internally (i.e., which only accept submissions from students within one institution ordepartment) are not listed. If you know of other journals that meet these criteria, please inform the Psi ChiNational Office, P.O. Box 709, Chattanooga, TN 37401-0709; Telephone: (423) 756-2044; Fax (toll-free):1-877-265-1529; E-mail: [email protected].

psi chi journal of undergraduate research - fall 2001

Documents