gender differences and gender-related constructs in dating aggression

14
http://psp.sagepub.com/ Bulletin Personality and Social Psychology http://psp.sagepub.com/content/28/8/1106 The online version of this article can be found at: DOI: 10.1177/01461672022811009 2002 28: 1106 Pers Soc Psychol Bull Stephen S. Jenkins and Jennifer Aubé Gender Differences and Gender-Related Constructs in Dating Aggression Published by: http://www.sagepublications.com On behalf of: Society for Personality and Social Psychology can be found at: Personality and Social Psychology Bulletin Additional services and information for http://psp.sagepub.com/cgi/alerts Email Alerts: http://psp.sagepub.com/subscriptions Subscriptions: http://www.sagepub.com/journalsReprints.nav Reprints: http://www.sagepub.com/journalsPermissions.nav Permissions: http://psp.sagepub.com/content/28/8/1106.refs.html Citations: What is This? - Aug 1, 2002 Version of Record >> at WEST VIRGINIA UNIV LIBRARY on June 27, 2014 psp.sagepub.com Downloaded from at WEST VIRGINIA UNIV LIBRARY on June 27, 2014 psp.sagepub.com Downloaded from

Upload: j

Post on 20-Feb-2017

216 views

Category:

Documents


0 download

TRANSCRIPT

http://psp.sagepub.com/Bulletin

Personality and Social Psychology

http://psp.sagepub.com/content/28/8/1106The online version of this article can be found at:

 DOI: 10.1177/01461672022811009

2002 28: 1106Pers Soc Psychol BullStephen S. Jenkins and Jennifer Aubé

Gender Differences and Gender-Related Constructs in Dating Aggression  

Published by:

http://www.sagepublications.com

On behalf of: 

  Society for Personality and Social Psychology

can be found at:Personality and Social Psychology BulletinAdditional services and information for    

  http://psp.sagepub.com/cgi/alertsEmail Alerts:

 

http://psp.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

http://psp.sagepub.com/content/28/8/1106.refs.htmlCitations:  

What is This? 

- Aug 1, 2002Version of Record >>

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

PERSONALITY AND SOCIAL PSYCHOLOGY BULLETINJenkins, Aubé / GENDER AND AGGRESSION

Gender Differences and Gender-RelatedConstructs in Dating Aggression

Stephen S. JenkinsJennifer AubéUniversity of Rochester

This study examined frequency and severity of physical, sym-bolic, and psychological aggression between college men andwomen in 85 heterosexual dating relationships and the extentto which gender role constructs predicted reports of aggression.Although there were no differences on self-reports of perpetration,men reported higher victimization levels than women and higherphysical and psychological victimization levels than perpetra-tion levels, whereas women reported higher symbolic perpetrationlevels than victimization levels. As a result, averaging reportsfrom both partners suggested that women in existing college dat-ing relationships are more aggressive than men. For both gen-ders, stereotypically negative masculine (i.e., instrumental)characteristics were the best predictors of aggressive acts. Perpe-trators’ positive masculinity and femininity predicted self-reports of decreased aggression that were not confirmed by theirpartners. Whereas men’s traditional attitudes about the male rolepredicted greater male aggression, women’s less traditional atti-tudes predicted increased severity of female physical aggression.

As recent meta-analyses (Archer, 1999, 2000a) havedescribed, there is considerable theoretical disagree-ment about the origins, composition, and appropriatesampling method of aggression between romantic part-ners, each of which affects study findings of gender dif-ferences. Whereas clinical and law enforcement dataderived from female victims have shown predominantuse of violence by men (Dobash, Dobash, Wilson, & Daly,1992), studies (e.g., Arias & Johnson, 1989) based on theConflict Tactics Scales (CTS; Straus, 1979) and usingcommunity samples have typically found similar levelsof aggression between men and women. These latterfindings have, on one hand, raised concerns about thegender sensitivity of aggression research (Cascardi &Vivian, 1995) and, on the other hand, elevated interestin variables that might predict aggressive behavior betterthan gender alone (e.g., Thompson, 1991). Specifically,

when examining gender differences in courtship aggres-sion, research has often overlooked or poorly definedgender-role processes. Other criticisms of past researchinclude its (a) predominant focus on frequency ofaggression, neglecting severity of aggression (e.g.,Cantos, Neidig, O’Leary, 1994); (b) use of the CTS with-out additional measures of aggression (i.e., sexual, psy-chological) (e.g., Smith, 1994); and (c) sampling indi-vidual men and women for the purpose of aggregate-gender data rather than within-couple data (e.g.,Jouriles & O’Leary, 1985). This study sought to addressthese concerns.

Studies of Physical Aggression

Aggression between partners in dating relationshipsoften has been defined operationally by sections of theCTS (Straus, 1979) that assess frequency of self-reportedphysically aggressive acts, including behaviors such as“kicking,” “slapping,” and “using a knife or gun.”

The literature on physical aggression between hetero-sexual dating partners is not consistent with research onsame-sex aggression that has shown that men are gener-ally more aggressive than women (e.g., Buss & Perry,1992). Rather, many studies have found similar propor-tions of men and women inflicting some form of court-ship aggression (Arias & Johnson, 1989; Carlson, 1987;Cate, Henton, Koval, Christopher, & Lloyd, 1982;Sigelman, Berry, & Wiles, 1984). Illustratively, White andKoss’s (1991) large national college survey found that

Authors’ Note: This work is dedicated to the memory of Emory L.Cowen (1926-2000), who was instrumental in the preparation and en-couragement of this study. Correspondence concerning this articleshould be addressed to Stephen S. Jenkins, Department of Clinical andSocial Sciences in Psychology, Meliora Hall, University of Rochester,Rochester, NY 14627; e-mail: [email protected].

PSPB, Vol. 28 No. 8, August 2002 1106-1118© 2002 by the Society for Personality and Social Psychology, Inc.

1106

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

37% of men and 35% of women reported inflicting someform of physical aggression on a partner. In a study ofmarried, cohabiting, and dating couples, Stets andStraus (1989) found that the most frequent pattern incouples was for both partners to be aggressive, followedby female-only, with male-only aggression the least fre-quent. Indeed, several studies have found more womenthan men reporting aggression toward dating partners(O’Leary et al., 1989; Riggs, O’Leary, & Breslin, 1990;Sugarman & Hotaling, 1989). Taking these together,Archer’s (2000a) meta-analysis of aggression studiesdemonstrated an overall small effect size of d = –.12 inthe direction of women when using self-reports of perpe-trating physical aggression.

Another approach for comparing male and femaledating aggression has been to measure self-reports of vic-timization. Some studies have found that women reportvictim status in a conflict more often than men(Makepeace, 1986; Sugarman & Hotaling, 1989), andStets and Pirog-Good (1987) found that women reportmore aggressive victimization from their partners, con-sistent with men’s report of perpetrating more aggres-sion than women. However, other studies have foundthat more college men than women report they were vic-tims of dating aggression (Arias, Samios, & O’Leary,1987; White & Koss, 1991). Given the inconsistent find-ings to date, it is not surprising that Archer (2000a)found the average effect size across studies to be d = –.016for differences in victimization rates, indicating a com-parable level of aggression received by—and thereforepresumably perpetrated by—each gender within hetero-sexual relationships.

Nonphysical Aggression: Psychological and Symbolic

In the context of dating relationships, psychologicalaggression is defined as any nonphysical act intended tohurt one’s partner and may include direct or indirect,verbal or nonverbal elements. Although psychologicalaggression has been identified as a precursor to physicalaggression (Hydén, 1995; Kasian & Painter, 1992;Murphy & O’Leary, 1989), and in many cases its effectsmay be more severe and longer felt than those of physi-cal abuse alone (Walker, 1984), few studies have mea-sured both physical and psychological aggressiontogether (Archer, 2000b).

In most cases where nonphysical aggression isassessed, these studies use a second CTS factor of “sym-bolic aggression,” measuring “verbal and nonverbal actswhich symbolically hurt the other or the use of threatswhich hurt the other” (Straus, 1992b, p. 32), includingitems such as “insults,” “smashing items,” and threats toharm. In the largest study of dating aggression using thismeasure, women reported inflicting significantly moreaggression than men, although both genders reported

experiencing similar levels as victims (White & Koss,1991).

Another method of assessing nonphysical aggressionis the Psychological Maltreatment Inventory (PMI),developed by Kasian and Painter (1992) to measure neg-ative relationship strategies, including isolation andemotional control, attacks on self-esteem, jealousy, ver-bal abuse, and withdrawal. Using a dating sample, Kasianand Painter (1992) found that men reported that theirpartner used fewer positive behaviors (e.g., respect) andmore negative behaviors compared to women’s reportsof male partners (Kasian & Painter, 1992).

Although there is some overlap between the items ofthe CTS symbolic subscale and PMI, namely around ver-bal abuse, there are several differences that warrant theuse of both measures of nonphysical aggression.Whereas the CTS alone includes threats of physicalharm and aggression against items that are substitutes,or “symbols,” of the partner, the PMI targets a wider vari-ety of manipulative and denigrating behaviors. In thecurrent study, both measures are used to assess non-phys-ical aggression. For clarity, the term symbolic aggression isused to refer specifically to CTS nonphysical aggressionitems and the term psychological aggression is used to referto the PMI items.

Severity of Aggression

Researchers are challenged to explain the unexpect-edly high levels of reported female aggression relative tomen in nonclinical populations (Riggs & O’Leary,1989). On the surface, these findings contradict reportsthat women are more likely than men to experience vio-lence from partners, be hospitalized, or require batteredspouse services (Dobash et al., 1992; Walker, 1984,1989). One given explanation is that greater malestrength and size results in greater physical injury forfemale than male victims (O’Leary et al., 1989; Straus,1992a; Walker, 1989), but because the CTS did not mea-sure injury or severity, researchers can mistakenly equatetwo very different acts of aggression. For example, wiveshave reported more negative consequences than hus-bands even when aggression frequency was similar(Cantos et al., 1994; Vivian & Langhinrichsen-Rohling,1994). Langhinrichsen-Rohling, Neidig, and Thorn(1995) found that although 83% of married couples intreatment were mutually aggressive, men were morelikely than women to use extreme forms of violence andless likely either to incur injury or to report fear frompartner aggression. Makepeace (1986) found that mod-erate to severe injuries were 4 times higher for womenthan men. Finally, Archer’s (2000a) meta-analysis foundthat in contrast to low frequency differences betweengenders, injuries were inflicted more by men thanwomen; the values were d = .54 for samples treated for

Jenkins, Aubé / GENDER AND AGGRESSION 1107

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

marriage problems, d = .13 for students, and d = .11 forthe general community. These findings suggest thatalthough gender differences may not be detected in fre-quency of acts, men may inflict more serious injuriesthan women through such acts.

Although Straus, Hamby, Boney-McCoy, andSugarman (1996) have recently published a revised setof scales (CTS2) with provision for measuring injuries,the present study chose to measure perceptions of theseverity of acts of aggression rather than outright inju-ries. This step allows for comparing severity as a variableacross various forms of aggression, including psychologi-cal acts, which are not expected to result in observablewounds or hospitalization. Furthermore, given the lowfrequency of physical injury in samples age 14 to 22(Archer, 2000a), which encompasses this study’s sample,emphasis on perceived consequences appears moreappropriate.

Surveying Couples

Whereas marriage studies frequently involve bothpartners in surveys, dating aggression studies tend toreport lone respondents of current or former relation-ships (Archer, 2000a). This method is problematicbecause individuals are less likely to report themselves asperpetrators than victims of aggression (Archer, 1999;Makepeace, 1981; Moffitt et al., 1997), even when ques-tionnaires are anonymous (Riggs, Murphy, & O’Leary,1989). Thus, rather than using a single partner’sresponses, some researchers have suggested surveyingboth partners (e.g., Margolin, 1987; Moffitt et al., 1997;Szinovacz, 1983) and averaging response rates (Barling,O’Leary, Jouriles, Vivian, & MacEwen, 1987). A recentmeta-analysis of aggression studies (Archer, 1999) hasshown that studies with couples exhibit less systematicunderreporting, particularly among men, than thosewith individuals, indicating that the responses of cou-ples—who know their partners are being assessed in aseparate room—are more frank and accurate than thoseof lone respondents. In the current study, self-report ofperpetrated aggression, self-report of victimization, andcouple-averaged scores using the mean between onepartner’s perpetration score and the other partner’s vic-timization score, are each presented as methods ofassessing aggression between dating partners.

Gender-Related Constructs and Aggression

Gender-related constructs, such as characteristics andideology traditionally attached to one or the other gen-der, may be important in understanding aggression inmen and women (Burke, Stets, & Pirog-Good, 1989;Spence, Losoff, & Robbins, 1991), perhaps more impor-tant than gender itself (Thompson, 1991). These con-structs are relatively stable and a strong source of motiva-

tion leading people to behave in ways consistent withtheir gender identities (Burke et al., 1989). Priorresearch with gender-related constructs has been guidedby social role theory (Eagly & Steffen, 1986), particularlythe suggestion that aggression is dictated by culturalnorms surrounding gender relations. Eagly and Steffen(1986) contended that whereas “the male gender roleincludes norms encouraging many forms of aggres-sion . . . the traditional female gender role places littleemphasis on aggressiveness” (p. 310). Thus, to theextent that individuals adhere to traditional masculinity,they are more likely to engage in aggressive behavior.Research on the relationship between gender-relatedconstructs and dating aggression has, however, pro-duced few consistent findings. One source of disparatefindings about the role of gender-related constructs inaggressive behavior is inconsistency in conceptualizingand measuring these constructs.

The gender literature has had two primary foci: gen-der-related characteristics or personality and gender-related attitudes or ideology. Gender-related character-istics describe stable dispositions believed to be charac-teristic of one or the other gender. Gender-related atti-tudes are beliefs about the way men and women act andshould act.

Gender-related characteristics. Aggressors typicallyendorse stable, pervasive masculine characteristics.Using the Bem Sex Role Inventory (BSRI; Bem, 1974),Bernard, Bernard, and Bernard (1985) found that menwho are aggressive toward partners have stronger mascu-line gender typing than nonaggressive men. Masculinecharacteristics relate to minor and severe physicalaggression, regardless of respondents’ gender (Thomp-son, 1991). Using the Extended Personal AttributesQuestionnaire (EPAQ; Spence, Helmreich, & Holahan,1979), Spence et al. (1991) found that a masculine char-acteristic of negative instrumentality predicts sexuallyaggressive behaviors in dating men. Finally, Malamuth,Sockloskic, Koss, and Tanaka (1991) reported thatnonsexual aggression is primarily a function of hostilemasculinity and sexual aggression a function of thefusion of hostile masculinity and promiscuity.

Other researchers have suggested that feminine char-acteristics are important in determining aggression.Feminine characteristics were found to temper aggres-sion in Caucasian husbands (Boye-Beaman, Leonard, &Senchak, 1993), and Lisak and Ivan (1995) found thatsexually aggressive men rated themselves less feminineand empathic, but not more masculine, thannonaggressive ones.

Gender-related attitudes. Traditional masculinity ideol-ogy involves beliefs about male aggression and domina-tion over females (Thompson & Pleck, 1986), and men

1108 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

with traditional gender-related attitudes are more likelyto be aggressive to partners. Pleck, Sonenstein, and Ku’s(1993) study of adolescent men’s traditional gender ide-ology shows that a traditional attitude toward the malegender role relates to a belief that relationships betweenmen and women are adversarial. Men who are high insexual stereotyping are more likely to accept aggressionas a means of resolving conflict (Burt, 1980), and thosewith more traditional gender-related attitudes towardwomen report engaging in more aggressive behavior indating relationships (Sigelman et al., 1984).

Because traditional gender expectations portraywomen as passive and men as aggressive (Currie, 1998),it follows that aggressive behavior is more likely inwomen whose attitudes are least traditional and leastlikely to accept the traditional gender dichotomizationregarding aggression. Women with nontraditional gen-der expectations may be more willing to initiate aggres-sion, retaliate, or actively select as partners nontradi-tional men, who are less focused on toughness and statusdominance, allowing women a greater opportunity for amutual or dominant aggressive position. Prior researchhas found that women who identify themselves as femi-nists, compared to nonfeminists, are more likely toresolve conflicts with confrontation than submission(Wilson, 1982), and those with less traditional gender-role attitudes are more likely to define themselves as vic-tims (DeGregoria, 1987) as well as offenders (McKinney,1986) in aggressive conflicts. Although far from conclu-sive, these studies suggest that both men and women’sgender attitudes are important in predicting aggressivebehavior and should be studied more closely in thatcontext.

Although gender-related characteristics and attitudeshave both been viewed as components of gender identity(Ashmore, 1990), they are not highly intercorrelated(Spence et al., 1991) and they relate differently to out-come variables (Aubé, Norcliffe, Craig, & Koestner,1995). This separation is noted in studies of aggression.Spence et al. (1991), for example, found that after gen-der characteristics are accounted for in women, attitudesmake no significant further contribution in predictingexperiences either as victims or perpetrators of sexualaggression. In contrast, male sexual aggression, whichoccurred more frequently than female aggression, waspredicted by traditional gender attitudes after control-ling for negative instrumentality. Based on these find-ings, researchers have increasingly recognized the valueof assessing both gender-related characteristics and atti-tudes in studying gender role processes. Traditional gen-der attitudes seem particularly important for menbecause of findings showing a consistent relationshipbetween attitudes and aggressive behavior.

Hypotheses

This study tested three sets of issues relevant to aggres-sion among dating couples: (a) differing methods of cal-culating aggression, (b) gender differences in datingaggression, and (c) gender-related constructs that pre-dict reports of aggression. First, because respondents aremore likely to report themselves victims than perpetra-tors of socially undesirable acts (e.g., Makepeace, 1981),victimization scores were expected to be higher thanperpetration scores for both genders. Second, consis-tent with reports of greater injury for women than men(e.g., Morse, 1995), gender differences were expected,with men indicated as responsible for physically, symbol-ically, and psychologically aggressive acts rated moresevere (as perceived by both perpetrator and victim)than those performed by women. However, no genderdifferences were predicted for contrasting men andwomen’s frequency of acts. Third, gender-related char-acteristics of negative instrumentality were expected touniquely predict physical, symbolic, and psychologicalaggression in both men and women, with perpetrator’sgender-related traits predicting both self-reported per-petration and his or her partner’s victimization reports.In addition, higher levels of physical, symbolic, and psy-chological aggression were expected to be found in menwith more traditional attitudes; higher aggression inwomen was expected in those with less traditionalattitudes.

METHOD

Participants

Participants in this study were 170 heterosexual, non-married, university students (85 couples) recruited for astudy on “relationships and personality,” with induce-ments of course credit or entry into a lottery for cashprizes, as they preferred. The minimum relationshipperiod for entering the study was 3 months. Once pres-ent, all participants consented to the study after beinginformed that it included questionnaires about relation-ship aggression. Descriptive characteristics of the sampleare given in Table 1.

Measures and Procedures

Partners were led to separate rooms prior to beinginformed of the purpose of the study and giving consentto reduce the opportunity for collusion on measures ofaggression. Same-gender research assistants distributedmeasures to optimize participant comfort, honesty, andcompletion of surveys. Aggression measures were giventwice for frequency and severity of aggression toward,and received from, the current partner. Participants

Jenkins, Aubé / GENDER AND AGGRESSION 1109

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

were assured of confidentiality at the outset andinformed at the end of the study about services availablefor persons seeking more information about physicaland psychological abuse. The following measures wereused:

CTS (Straus, 1979). The CTS, Form R, is a 19-item self-report inventory that assesses frequency of behaviorsengaged in during interpersonal conflict. Only the 7-item Symbolic Aggression and 9-item Physical Aggres-sion scales were used (Straus, 1992a, 1992b). The modi-fied CTS used here added one additional symbolicaggression item (“Threatened to break up”) used in aprincipal components analysis by Caulfield and Riggs(1992) and one physical aggression item (“Physicallyforced to have sex”) suggested by Cantos et al. (1994)and dropped one symbolic aggression item (“Cried”)that Straus (1992b) excluded from scoring. The revisedCTS then included 17 aggression items, 7 symbolic (αs =.81/.84) and 10 physical (αs = .58/.60). Reliability valuesfor this study are for perpetration and victimization,respectively.

Frequency of CTS items was scored on a scale from 0(never) to 6 (more than 20 times) based on the past 6months of experience with the respondent’s currentpartner. Each item was followed by a severity rating scale

that asked the respondent to indicate the severity of the“clearest instance” of that act of aggression on a scaleincluding 0 (did not occur or not severe), 1 (mildly severe), 2(moderately severe), and 3 (very severe).

For each form of aggression surveyed, three mean fre-quency scores were calculated for each respondent:“perpetration” (i.e., mean of aggression a respondentreported using), “victimization” (i.e., mean of aggres-sion a respondent reported receiving from his or herpartner), and “couple-averaged” (i.e., average valuebetween the respondent’s self-reported perpetrationscore and his or her partner’s self-reported victimizationscore). Three mean severity scores were calculated in thesame way.

Partner concordance values for the CTS are shown inTable 2. These are correlations between one partner’sperpetration scores and the other partner’s victimiza-tion scores, using interval-level data, to assess consistencyof reports about the aggression committed by each gen-der. These values are similar to those presented byArcher (1999), whose meta-analysis yielded weightedmean r = .55 and .53 for men and women, respectively. Inanother study, Jouriles and O’Leary (1985) reportedCTS interspousal reliability ratings of couple agreementabout the occurrence or nonoccurrence of partners’aggression. These kappa were .43 and .40 for husbandsand .40 and .41 for wives in therapy and community sam-ples, respectively. Moffitt et al. (1997) reported values of.46 for men and .34 for women with couples in mostlydating and cohabiting relationships.

PMI (Kasian & Painter, 1992). This 37-item version ofthe PMI assesses psychologically negative aspects of dat-ing relationships. It has five factors: “Isolation and emo-tional control” (e.g., “My partner blamed me for his or

1110 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

TABLE 1: Description of the Sample

As individual partners (N = 170)Average age 19.9 years (SD = 1.19)Primary ethnicity

Caucasian 75.9%Latino 8.2%African American 4.7%East Asian 4.2%Middle Eastern 4.1%Native American 0.6%

ReligionRoman Catholic 32.4%a

Protestant 21.2%Non-religious 20.0%b

Jewish 13.0%Buddhist, Hindu, or “other” 13.6%

As dating couples (N = 85)Length of relationship at time of studyc

3-6 months 27.1%6-12 months 36.5%1-2 years 19.4%2-4 years 11.2%More than 4 years 5.9%

Couples including persons of differingEthnicity 31.7%Religion 69.4%

a. Men: 25.9%, women: 38.8%.b. Men: 25.9%, women: 14.1%.c. Length of relationship had no significant impact on perpetration orvictimization rates of either gender for any form of aggression.

TABLE 2: Interpartner Correlations of Aggression by Gender(N = 85)

Correlation Values

Male Aggressiona Female Aggressionb

CTS-PhysicalFrequency .27* .67**Severity .44** .76**

CTS-SymbolicFrequency .64** .56**Severity .31** .40**

PMI-PsychologicalFrequency .49** .49**Severity .38** .30**

NOTE: CTS = Conflict Tactics Scales, PMI = Psychological Maltreat-ment Inventory.a. Self-reports of male perpetration and female victimization.b. Self-reports of female perpetration and male victimization.*p ≤ .05. **p ≤ .01.

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

her problems”) (αs = .78/.86), “Attacks on self-esteem”(e.g., “My partner treated me like I was stupid”) (αs =.84/.82), “Jealousy” (“My partner accused me of seeinganother man/woman”) (αs = .84/.82), “Verbal abuse”(e.g., “My partner called me names”) (αs = .81/.82), and“Withdrawal” (e.g., “My partner sulked and refused totalk about a problem”) (αs = .80/.78). Reliability valuesfrom this study are given above for perpetration and vic-timization scores, respectively. In this study, the PMI cor-related with the CTS-symbolic scale in a range from r =.71 to .92 across methods of scoring.

Frequency and severity of PMI items were rated in thesame manner as for CTS items. Perpetration, victimiza-tion, and couple-averaged scores were calculated forboth frequency and severity. Partner concordance valuesfor the PMI are reported in Table 2.

EPAQ (Spence et al., 1979). This 28-item measureassesses both socially desirable and undesirable charac-teristics judged to be more typical of, and more likely tobe endorsed by, one gender than the other (Spenceet al., 1979). Items comprise two scales, each with sociallydesirable and undesirable subscales. The Masculinescale (M) contains eight desirable, self-assertive, instru-mental (I+) characteristics, such as independent anddecisive (α = .75), and eight undesirable, self-centered,instrumental (I–) characteristics, such as egotistical andgreedy (α = .75). The Feminine scale (F) contains eightdesirable, expressive (E+) characteristics, such as gentleand helpful to others (α = .75), and four undesirable,expressive (Ec–) characteristics, such as servile or gull-ible (α = .62). Items are rated on a 5-point scale (1 = verytrue of me to 5 = not at all true of me). In this study, I– is corre-lated negatively (r = –.38) with E+ and E– is correlatednegatively (r = –.52) with I+.

Male Role Norms Scale (MRNS; Thompson & Pleck,1986). The 26-item MRNS assesses attitudes towardappropriate behaviors by men. Each item is rated on a 7-point Likert-type scale (1 = strongly agree to 7 = strongly dis-agree). The MRNS has three subscales of “status norms”(e.g., “A man should never back down in the face of trou-ble”; α = .82), “Toughness norms” (e.g., “A real manenjoys a bit of trouble now and then”; α = .78), and“Antifemininity norms” (e.g., “It is a bit embarrassing fora man to have a job that is usually filled by a woman”; α =.78). Intercorrelates of any of the three subscales rangefrom r = .51 to 64. In this study, the measure as a wholehas an internal reliability of α = .90. MRNS scores oftraditionality correlate positively with EPAQ scores of I+(r = .18) and I– (r = .42) but negatively with E+ (r = –.22),suggesting some relationship between traditional atti-tudes about the male role and gender-related traits, pre-dominantly negative instrumentality.

RESULTS

Frequency and Severity

Table 3 presents, for each form of aggression (i.e.,physical, symbolic, psychological), the relationshipbetween the frequency and severity scores when at leastone act of aggression is reported. Listed separately bygender, these values depict a positive relationshipbetween frequency and severity scores in both self-reports of perpetration and victimization for symbolicand psychological aggression but not physical forms.Notably, the number of cases in which at least one exam-ple of aggression was reported and rated for perceivedseverity was much lower for physical than other form ofaggression, making the pattern of correlations for physi-cal aggression difficult to interpret.

Using only cases with an accompanying frequencyabove zero (i.e., where at least one act occurred), themean severity scores for perpetration, averaged acrossthe three forms of aggression, were 1.65 and 1.56 forwomen and men, respectively; for reports of victimiza-tion, 1.67 and 1.77 for women and men, respectively.Severity scores did not vary greatly by form of aggression.Presented for descriptive purposes, these values placeaverage perceptions of severity by respondents in thisstudy in the mild-to-moderate range.

Differences in Scoring Method

Table 4 presents paired comparisons of scoring meth-ods for calculating aggression for each gender.Nonsignificant values indicate comparable estimates ofaggression from a respondent’s self-report of perpetra-tion and his or her partner’s self-report of victimization.For male aggression, male perpetration scores werehigher than female victimization for frequency of sym-bolic aggression. There were no other significant differ-

Jenkins, Aubé / GENDER AND AGGRESSION 1111

TABLE 3: Correlations Between Frequency and Severity Scores inCases Where at Least One Act Was Reported by Gender

Female Male

CTS-PhysicalPerpetration scores .27 (N = 22) –.14 (N = 28)Victim scores .19 (N = 18) –.12 (N = 34)

CTS-SymbolicPerpetration scores .39** (N = 77) .24* (N = 74)Victim scores .51** (N = 68) .27* (N = 72)

PMI-PsychologicalPerpetration scores .50** (N = 83) .49** (N = 83)Victim scores .56** (N = 83) .55** (N = 84)

NOTE: CTS = Conflict Tactics Scales, PMI = Psychological Maltreat-ment Inventory. N = number of cases where at least one act was re-ported.*p ≤ .05. **p ≤ .01.

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

ences between male perpetration and female victimiza-tion scores. For female aggression, male victimizationwas higher than female perpetration in two instances:frequency of physical and frequency of psychologicalaggression. Effect sizes for significant discrepanciesranged from .32 to .34, using Cohen’s d statistic (Cohen,1988).

Gender Differences

Table 5 presents means and standard deviations ofperpetration, victimization, and couple-average scoresfor frequency and severity of aggression separately formen and women and results of paired t tests comparingthe genders in these measures. No gender differenceswere found either for frequency or severity in self-reportof perpetration for any form of aggression surveyed.Men, however, reported higher victimization receivedfrom female partners than females reported receivingfrom male partners, for all forms of aggression exceptseverity of psychological aggression. It should be notedthat there were no gender differences in comparisons onthe lone item “Forced to have sex.” Each of the findingswith perpetration and victimization scores was con-firmed by post hoc analyses in which aggression itemswere standardized, which weights each response by itsfrequency in the sample, and average Z scores comparedby gender. On couple-averaged scores, using the averageof one partner’s perpetration and the other’s victimiza-tion scores, women had higher frequency scores thanmen on all forms of aggression. Effect sizes for gender

differences ranged from .29 to .39, using Cohen’s d statis-tic (Cohen, 1988).

Paired comparison t tests for perpetration and victim-ization scores by gender are presented in Table 6.Whereas men reported higher victimization than perpe-tration scores for frequency of physical and frequencyand severity of psychological aggression, womenreported higher perpetration than victimization scores

1112 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

TABLE 4: Comparisons of Means of Perpetrator and Victim Scoresby Gender (N = 85)

Female MaleVictim Perpetrator

M SD M SD d t

Male aggressionPhysical-frequency .09 .24 .10 .22 .04 –.47Physical-severity .03 .10 .03 .10 .00 .10Symbolic-frequency .75 .90 1.02 .95 .34 –3.20**Symbolic-severity .27 .48 .30 .40 .16 –.64Psychological-frequency .93 .87 .94 .82 .12 –.04Psychological-severity .30 .42 .26 .31 .10 –.89

Male FemaleVictim Perpetrator

Female aggressionPhysical-frequency .21 .35 .12 .30 .33 3.03**Physical-severity .07 .16 .06 .18 .08 .62Symbolic-frequency 1.12 1.05 1.00 .99 .13 1.20Symbolic-severity .35 .41 .32 .51 .06 .63Psychological-frequency 1.18 .84 .92 .77 .32 2.93**Psychological-severity .33 .31 .27 .38 .15 1.46

**p ≤ .01.

TABLE 5: Comparison of Means of Female and Male Perpetration,Victimization, and Couple-Averaged Scores of Aggression(N = 85)

Female Male

M SD M SD d t

Perpetration scorePhysical-frequency .12 .30 .10 .22 .06 .55Physical-severity .06 .18 .03 .10 .15 1.16Symbolic-frequency 1.00 .99 1.02 .95 .02 –0.25Symbolic-severity .32 .51 .30 .40 .04 .31Psychological-frequency .92 .82 .94 .87 .02 –0.18Psychological-severity .27 .38 .26 .31 .02 .20

Victimization scorePhysical-frequency .09 .24 .21 .35 .34 –3.24**Physical-severity .03 .10 .07 .16 .27 –2.08*Symbolic-frequency .75 .90 1.12 1.05 .39 –3.63***Symbolic-severity .27 .48 .35 .41 .15 –1.44Psychological-frequency .93 .87 1.17 .84 .29 –2.77**Psychological-severity .30 .42 .33 .31 .07 –0.77

Couple-averaged scorePhysical-frequency .17 .30 .10 .18 .29 2.73**Physical-severity .06 .16 .03 .08 .20 1.78Symbolic-frequency 1.06 .90 .88 .83 .37 3.25**Symbolic-severity .33 .39 .28 .37 .20 1.83Psychological-frequency 1.05 .70 .93 .73 .35 3.04**Psychological-severity .30 .28 .28 .31 .12 1.19

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

TABLE 6: Comparisons of Means of Each Gender’s Victimizationand Perpetration Scores (N = 85)

Victimization Perpetration

M SD M SD d t

MalePhysical-frequency .21 .35 .10 .22 .38 3.43***Physical-severity .07 .16 .03 .10 .23 1.83Symbolic-frequency 1.12 1.05 1.02 .95 .16 1.43Symbolic-severity .35 .41 .30 .40 .18 1.72Psychological-frequency 1.17 .84 .94 .82 .48 4.65***Psychological-severity .33 .31 .26 .31 .36 3.50***

FemalePhysical-frequency .09 .24 .12 .30 .12 –1.25Physical-severity .03 .10 .06 .18 .19 –1.43Symbolic-frequency .75 .90 1.00 .99 .48 –4.37***Symbolic-severity .27 .48 .32 .51 .14 –1.32Psychological-frequency .93 .87 .92 .77 .02 .29Psychological-severity .30 .42 .27 .38 .12 1.10

***p ≤ .001.

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

for frequency of symbolic aggression. Effect sizes for dif-ferences between perpetration and victimization scoresranged from .36 to .48, using Cohen’s d statistic (Cohen,1988).

Post hoc comparisons were calculated using percent-age of each gender reporting at least one act of aggressionin the past 6 months. Only one significant gender differ-ence arose: Men (41%) were more likely than women(21%) to report being the victim of at least one act ofphysical aggression. Perpetration of physical aggressionranged without significant difference between 27.1%(women) and 34.1% (men). Symbolic aggressionranged from 88.2% to 90.6% for perpetration, 81.2% to85.9% for victimization, and 92.9% to 95.3% whenreported by either partner. Psychological aggression wasequal for the genders: 97.6% for perpetration, 98.8% forvictimization, and 100% when reported by either part-ner. Given these percentages, there appears to be a ceil-ing effect for psychological aggression, with all couplesreporting at least one act.

Gender-Related Constructs

Because of nonindependence of samples, all regres-sion analyses were done separately by gender. In the firstset of analyses, presented in Table 7, perpetrators’ gen-der-related construct scores were regressed onto perpe-tration scores of aggression. The contribution of the setof five gender-related constructs (i.e., EPAQ characteris-tics and MRNS attitudes) to male self-reports of perpe-trated aggression ranged between 16.9% of the variancein the case of severity of physical aggression, ∆R2 = .169,F(5, 79) = 3.22, p ≤ .05, and 38.1% of the variance in the

case of frequency of symbolic aggression, ∆R2 = .381, F(5,79) = 9.73, p ≤ .001. For female-perpetrated aggression,women’s scores on gender-related measures contributeda range between 16.3% of the variance in the case ofseverity of symbolic aggression, ∆R2 = .163, F(5, 79) =3.07, p ≤ .05, and 27.4% of the variance in the case of fre-quency of psychological aggression, ∆R2 = .274, F(5, 79) =5.97, p ≤ .001.

Both self-reported gender-related characteristics andattitudes about the male gender role were uniquely asso-ciated with self-reports of perpetrated aggression. Inmen, negative instrumentality uniquely predictedincreases in the frequency of all forms of aggression butdid not predict increases in severity. In contrast, men’spositive instrumentality predicted lower perpetrationscores for all forms of aggression except severity of physi-cal aggression. More traditional attitudes about the malegender role predicted increases in the severity of men’ssymbolic aggression and the frequency and severity ofpsychological aggression. For women, negative instru-mentality predicted increases in frequency of self-reported symbolic and frequency and severity of psycho-logical aggression. Frequency of female physical aggres-sion increased for women who reported more negativeexpressiveness but decreased for women who reportedmore positive expressiveness. Less traditional attitudesabout the male gender role predicted increases inwomen’s severity of physical aggression.

In the second set of analyses, presented in Table 8,perpetrators’ gender-related construct scores wereregressed onto partners’ self-reported victimizationscores of aggression. Men’s scores on gender-related

Jenkins, Aubé / GENDER AND AGGRESSION 1113

TABLE 7: Regression of Perpetrator’s Gender-Related Constructs on Perpetration Scores of Male and Female Aggression

R2 I-Pos sr2/ I-Neg sr2/ E-Pos sr2/ E-Neg sr2/ MRNS sr2/

Phys-FreqMale .204** .058/–.273* .048/.239* .010/–.110 .012/–.119 .030/.182Female .230*** .004/–.081 .010/.084 .046/–.236* .067/.338* .023/–.175

Phys-SevMale .169* .023/–.174 .029/.185 .039/–.214 .011/–.114 .012/.116Female .175** .003/–.067 .040/.248 .006/–.084 .031/.229 .061/–.285*

Symb-FreqMale .381*** .040/–.227* .073/.301** .009/–.104 .000/.023 .127/.306***Female .188** .030/–.225 .044/.261* .000/–.002 .012/.143 .000/–.013

Symb-SevMale .229*** .088/–.338** .025/.173 .001/.018 .004/.074 .053/.244*Female .163* .008/–.117 .035/.234 .000/–.023 .031/.231 .021/–.166

Psych-FreqMale .287*** .059/–.275* .084/.316** .002/–.050 .001/.041 .050/.236*Female .274*** .034/–.239 .104/.403*** .000/.017 .007/.107 .000/.008

Psych-SevMale .310*** .140/–.421*** .039/.200 .001/.038 .006/.088 .057/.253*Female .221*** .016/–.163 .075/.341** .003/–.057 .016/.164 .003/–.065

NOTE: I = instrumental, E = expressive, MRNS = Male Role Norms Scale, phys = physical aggression, symb = symbolic aggression, psych = psychologi-cal aggression, freq = frequency, sev = severity.*p ≤ .05. **p ≤ .01. ***p ≤ .001.

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

measures contributed to women’s scores of victimizationin all forms of aggression, in a range from 13.6% of thevariance in the case of severity of physical aggression,∆R2 = .136, F(5, 79) = 2.48, p ≤ .05, and 21.0% of the vari-ance in the case of frequency of psychological aggres-sion, ∆R2 = .210, F(5, 79) = 4.20, p ≤ .01. Women’s scoreson gender-related measures contributed to men’s scoresof victimization in four of six analyses, in a range from14.6% of the variance in the case of frequency of physicalaggression, ∆R2 = .274, F(5, 79) = 2.70, p ≤ .05, and 18.4%of the variance in the case of severity of symbolic aggres-sion, ∆R2 = .184, F(5, 79) = 3.58, p ≤ .01.

Men’s self-reported negative instrumentalityaccounted uniquely for variance in the frequency ofphysical acts and frequency and severity of psychologicalacts reported received by female partners. Men’s moretraditional attitudes about the male role contributed towomen’s scores of victimization for frequency of sym-bolic and severity of psychological aggression. Women’sself-reported negative instrumentality accounteduniquely for variance in the frequency of symbolic andpsychological acts and severity of physical and psycho-logical acts reported received by male partners.Women’s less traditional attitudes about the male rolecontributed to men’s scores of victimization for severityof physical aggression. For both men and women, perpe-trators’ positive characteristics were unrelated to victimreports.

DISCUSSION

Scoring Method and Gender Differences

The present findings suggest that research with (a)unrelated individuals rather than couples, (b) perpetra-tors or victims alone rather than together, or (c) eithergender without the other may provide inconsistent snap-shots of aggression in heterosexual dating relationships.The results of these various scoring methods highlightthe need for careful survey of both partners to reflect thecomplex nature of relationship aggression.

In particular, the choice of method for calculatingaggression has a powerful impact when examining gen-der differences. Comparing couple-averaged scoresreveals small but consistent differences, with womenreported to be greater than men in frequency of all threeforms of aggression. Notably, these averages are a prod-uct of two other forms of calculation: perpetration scoresthat show no gender differences, such that men andwomen in college dating relationships report compara-ble frequency rates of perpetration, and victimizationscores in which men report more victimization thanwomen. These patterns contrast with Archer’s (2000a)meta-analysis, where comparing partner (i.e., victim)reports produced effect sizes near to zero and self-report(i.e., perpetration) produced effect sizes in the femaledirection, but they are consistent with his findings of anoverall effect indicating greater female than maleaggression.

1114 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

TABLE 8: Regression of Perpetrator’s Gender-Related Constructs on Victim’s Scores of Received Aggression

Perpetrator’s Gender-Related Constructs

∆R2 I-Pos sr2/β I-Neg sr2/β E-Pos sr2/β E-Neg sr2/β MRNS sr2/β

Victim scorePhys-Freq

Male .146* .003/–.071 .031/.219 .008/–.096 .028/.217 .031/–.204Female .165* .007/–.093 .052/.250* .004/–.069 .002/.051 .037/.202

Phys-SevMale .181** .000/–.010 .125/.441*** .001/.032 .006/.099 .088/–.342**Female .136* .009/–.107 .032/.194 .013/–.126 .015/–.136 .028/.165

Symb-FreqMale .125 .000/.004 .075/.342* .004/–.074 .002/.058 .009/–.107Female .140* .005/–.084 .033/.198 .004/–.066 .001/.033 .045/.225*

Symb-SevMale .184** .011/.134 .113/.419*** .007/–.091 .000/.011 .023/–.175Female .142* .001/–.033 .024/.169 .025/–.172 .002/–.055 .032/.188

Psych-FreqMale .159* .002/.058 .077/.345** .002/–.051 .004/.083 .000/.020Female .210** .010/–.112 .109/.360** .000/–.016 .000/.025 .027/.174

Psych-SevMale .078 .001/.041 .028/.209 .003/–.059 .001/–.045 .003/.065Female .193** .004/–.069 .067/.281* .004/–.072 .001/.042 .044/.222*

NOTE: I = instrumental, E = expressive, MRNS = Male Role Norms Scale, phys = physical aggression, symb = symbolic aggression, psych = psychologi-cal aggression, freq = frequency, sev = severity.*p ≤ .05. **p ≤ .01. ***p ≤ .001.

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

Further examining gender differences by proportionof each gender reporting at least one act of aggressionreveals findings similar to the approach with intervaldata. No gender differences are revealed in perpetrationof aggression, but nearly twice as many men report beingthe victim of at least one act of physical aggression. Onelimitation of this approach is a ceiling effect for psycho-logical aggression, meaning that nearly every couplereported at least one act of psychological aggression,suggesting that it is more productive to study the magni-tude of aggression in a sample than the number ofoffenders and victims. Overall, taking both proportionaland interval data into account suggests that in collegedating environments, aggression is widespread (i.e.,most couples have experienced aggression) but, withincouples, generally low in frequency.

Findings were expected to confirm reports that maleaggression was more severe than female (Archer, 2000a;Morse, 1995). However, perceived severity in cases whereat least one act occurred fell in the mild-to-moderaterange for both men and women, and across all surveyeditems, the severity index tended to yield findings of com-parable male and female aggression. The one exceptionwas higher male than female victimization reports forseverity of physical aggression. Although these findingscontrast with Archer’s (2000a) overall conclusions, com-parable reports of perceived severity may be consistentwith his report of no difference in injury by gender in theage 14 to 22 category, which embraces this current sam-ple. Taken collectively, these findings suggest thatwomen of college age and education are at least as likelyas males—perhaps more so—to engage in aggressive actsof all forms in the context of dating relationships andthat these female acts are perceived as no less severe thanmale acts. However, it is notable that these findingsappear to rely as much on the absence of gender differ-ences in perpetration as on the surprising willingness ofmen in the sample to report higher victimization levelsthan their female counterparts, a finding inconsistentwith the notion that men are especially predisposed bygender role socialization to tolerate and hide their pain(e.g., Stets & Straus, 1992).

There appears to be no systematic underreporting ofperpetration by men in this study. Comparisons of maleperpetration and female victimization scores suggestthat couples generally agreed on the overall absolutelevel of aggression by men; this level of agreement wasmet despite a procedure of separating the partners priorto informing them that the study involved questionsabout aggression, therefore reducing opportunities forcollusion. The sole exception in this sample is that menreported more frequent perpetration of symbolicaggression than women reported victimization. Curi-

ously, the direction and size of discrepancy in eachpartner’s own reports of perpetration versus his or hervictimization show that both men and women thoughtthat male partners were the greater victim, but for differ-ent forms of aggression. Specifically, whereas menreported their female partners to be more physically andpsychologically aggressive toward them than they aretoward their partner, women reported that they aremore aggressive symbolically toward their male partnersthan their partners are toward them. Thus, althoughboth the reports of men and women tended to viewwomen as the more aggressive partner, the genders didnot agree on the forms of aggression that womendominate.

An alternate interpretation of these findings ofgreater aggression by women is provided by Currie(1998), who suggests that both men and womenunderreport male aggression because it is normal andexpectable, whereas they exaggerate female aggressionbecause it is abnormal and remarkable. However, it canbe argued that gender expectations also tend to framemale aggression toward women as unacceptable whiletrivializing female aggression as harmless (Archer,2000a; Flynn, 1990), leading men (but not women) toinhibit dating aggression (Arias et al., 1987). Whetherbecause of perceptual bias or awareness of actual behav-ior, a majority of college students appear to accept asaccurate findings that men in heterosexual relationshipsare recipients of aggression as frequently as women(Fiebert, 1996). Currie’s (1998) suggestion, however,may account for the notably lower concordance rates forpartners’ reports on men’s physical aggression com-pared to women’s. Although these correlations do notindicate disagreement about the absolute level of maleaggression, they do indicate relatively less consistency inpartner’s joint recall in frequency of specific acts by menthan by women. Because there are marginal culturalexpectations about aggression in women versus that inmen, it may be easier to recall women’s specific forms ofphysical aggression than that of men’s in relationships.

Greater degree of female than male aggression alsomay be accounted for by the relatively low degree ofaggression that exists in the sample. Whereas highlyaggressive environments tend to be unilaterally male-to-female (Vivian & Langhinrichsen-Rohling, 1994), a lessaggressive environment, such as the one studied, mayallow women the perception of greater control and lessfear of retaliation (Archer, 2000a; Fiebert & Gonzalez,1997), resulting in bilateral or greater female aggres-sion. That aggression was low in this study does notnegate the importance of the finding that when overallaggression is low, men are not necessarily the sole oreven primary aggressors.

Jenkins, Aubé / GENDER AND AGGRESSION 1115

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

Gender-Role Constructs

The findings of this study suggest that gender-roleconstructs play a significant role in predicting theaggressive behavior of men and women in college datingrelationships. Specifically, for both genders, hostile, con-trolling, and self-absorbed “masculinity” best predictscommitting aggressive acts against one’s partner. Thisfinding suggests, as with prior findings (e.g., Thompson,1991), that socially undesirable characteristics tradition-ally consistent only with men are not in fact restricted tomen and that both men and women with these charac-teristics are more likely to be aggressive. Indeed, it iswomen’s self-report of these stereotypically masculinecharacteristics that most strongly predicts a partner’sreport of victimization.

Trait endorsement also was related to a number of dis-crepancies in partners’ reports of aggression. For exam-ple, men who self-reported more socially desirableinstrumentality also reported lower rates of perpetra-tion, but these traits were not related to lower femalereports of victimization. Although no reporting bias wasfound for men as a whole, it appeared that in men butnot women, a “positivity bias” (Aubé et al., 1995) resultedin a tendency for masculine men to view themselves asless aggressive than their partners viewed them. Simi-larly, a small feminine “positivity bias” in women was evi-denced by a decline in positive feminine women’s self-reported physical aggression not supported by a declinein male victimization. Although it has been suggestedthat feminine characteristics limit male aggression(Lisak & Ivan, 1995), the above findings suggest ratherthat men and women are most likely to perceive them-selves as least aggressive when they possess characteris-tics both socially desirable and stereotypically consistentwith their gender.

For women but not men, expressive traits associatedwith self-reports of being gullible, subordinate, and ser-vile predict self-reports of more frequent acts of physicalaggression. These findings suggest that women who seethemselves in socially undesirable, stereotypically femi-nine terms also may be most likely to feel resentment andaggression toward their partners to whom they feel sub-ordinate. However, this finding also may be a result of a“negativity bias” in women (i.e., an unrealistically highperception of self as aggressive), a suggestion supportedby the finding that male victimization scores are not pre-dicted by women’s negative expressiveness.

The results of this study also suggest that men’sreports of symbolic and psychological aggression arerelated to traditional attitudes about men (i.e., need forstatus, antifeminine behavior, and toughness). Aggres-sive behavior, therefore, is related not simply to being aman but to one’s view about how men ought to behave.In contrast, women’s less traditional views about male

behavior predicted heightened severity of physicalaggression in both female perpetration reports and malevictim reports. This finding modestly suggests that het-erosexual women who oppose men’s traditional behav-iors of dominance and toughness may be more willing orable to appropriate some of these behaviors forthemselves.

Limitations

One limitation of this study is that findings from a stu-dent sample may not generalize to all dating couples.First, university environments tend to provide extensivesocial resources, such as violence prevention programsand relatively positive support for female autonomy,compared to the private home life of married or cohabit-ing couples. Second, these findings extend only to rela-tionships that are established, not initial dating experi-ences or chance sexual encounters where aggression ismost intense (Makepeace, 1989) and may have involvedsuch an intolerable level of aggression that a further rela-tionship was precluded. Studies that can assess couplesin the earliest stages of relationships may have a betteropportunity to assess the fullest range of dating aggres-sion, particularly acts of sexual violence that are predom-inantly male-to-female (Makepeace, 1986).

A second theme that requires further investigationinvolves context of aggression, such as whether acts areinitiated or self-defensive. This is a complex matter forsurvey research because judging an act as retaliatory ver-sus unprovoked may be based in perception or rational-ization by a partner as often as in objective fact. Further-more, certain forms of aggression, such as physical acts,may be used in response to psychological aggression bythe other partner, and vice versa, making it difficult tojudge them as first attacks or responses, and some actsmay be carefully planned and initiated many days later,appearing unprovoked but actually are in response toprior acts. These situations challenge our notions ofwhat is properly considered self-defense. As oneresponse, future research may choose to use either inter-view or diary methodologies to take into account the top-ographical and chronological context of relationshipaggression.

REFERENCES

Archer, J. (1999). Assessment of the reliability of the conflict tacticsscales: A meta-analytic review. Journal of Interpersonal Violence, 14,1263-1289.

Archer, J. (2000a). Sex differences in aggression between heterosex-ual partners: A meta-analytic review. Psychological Bulletin, 126,651-680.

Archer, J. (2000b). Sex differences in physical aggression to partners:A reply to Frieze (2000), O’Leary (2000), and White, Smith, Koss,and Figueredo (2000). Psychological Bulletin, 126, 697-702.

Arias, I., & Johnson, P. (1989). Evaluations of physical aggressionamong intimate dyads. Journal of Interpersonal Violence, 4, 298-307.

1116 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

Arias, I., Samios, M., & O’Leary, K. D. (1987). Prevalence and corre-lates of physical aggression during courtship. Journal of Interper-sonal Violence, 2, 82-90.

Ashmore, R. D. (1990) Handbook of personality: Theory and research. NewYork: Guilford.

Aubé, J., Norcliffe, H., Craig, J., & Koestner, R. (1995). Gender char-acteristics and adjustment related outcomes: Questioning themasculinity model. Personality and Social Psychology Bulletin, 21,284-295.

Barling, J., O’Leary, K. D., Jouriles, E. N., Vivian, D., & MacEwen, K. E.(1987). Factor similarity of the Conflict Tactics Scales across sam-ples, spouses, and sites: Issues and implications. Journal of FamilyViolence, 2, 37-54.

Bem, S. L. (1974). The measurement of psychological androgyny.Journal of Consulting and Clinical Psychology, 42, 155-162.

Bernard, J. L., Bernard, S. L., & Bernard, M. L. (1985). Courtship vio-lence and sex-typing. Family Relations, 34, 573-576.

Boye-Beaman, J., Leonard, K., & Senchak, M. (1993). Male premaritalaggression and gender identity among Black and White newlywedcouples. Journal of Marriage and the Family, 55, 303-313.

Breslin, F. C., Riggs, D. S., O’Leary, K. D., & Arias, I. (1990). Familyprecursors: Expected and actual consequences of dating aggres-sion. Journal of Interpersonal Violence, 5, 247-258.

Burke, P. J., Stets, J. E., & Pirog-Good, M. A. (1989). Gender identity,self-esteem, and physical and sexual abuse in dating relationships.In M. A. Pirog-Good & J. E. Stets (Eds.), Violence in dating relation-ships (pp. 72-93). New York: Praeger.

Burt, M. R. (1980). Cultural myths and support for rape. Journal of Per-sonality and Social Psychology, 38, 217-230.

Buss, A. H., & Perry, M. (1992). The aggression questionnaire. Journalof Personality and Social Psychology, 63, 452-459.

Cantos, A., Neidig, P. H., & O’Leary, K. D. (1994). Injuries of womenand men in a treatment program for domestic violence. Journal ofFamily Violence, 9, 113-124.

Carlson, B. (1987). Dating violence: A research review and compari-son with spouse abuse. Social Casework, 68, 16-23.

Cascardi, M., & Vivian, D. (1995). Context for specific episodes ofmarital violence: Gender and severity of violence differences. Jour-nal of Family Violence, 10, 265-293.

Cate, R., Henton, J., Koval, J., Christopher, F., & Lloyd, S. (1982). Pre-marital abuse: A social psychological perspective. Journal of FamilyIssues, 3, 79-80.

Caulfield, M. B., & Riggs, D. S. (1992). The assessment of datingaggression. Journal of Interpersonal Violence, 7, 549-558.

Cohen, J. (1988). Statistical power analyses for the behavioral sciences (2nded.). New York: Academic Press.

Currie, D. H. (1998). Violent men or violent women? In R. K. Bergen(Ed.), Issues in intimate violence (pp. 97-111). London: Sage Ltd.

DeGregoria, B. (1987). Sex role attitude and perception of psycho-logical abuse. Sex Roles, 16, 227-235.

Dobash, R. P., Dobash, R. E., Wilson, M., & Daly, M. (1992). The mythof sexual symmetry in marital aggression. Social Problems, 39, 71-91.

Eagly, A. H., & Steffen, V. J. (1986). Gender and aggressive behavior:A meta-analytic review of the social psychological literature. Psy-chological Bulletin, 100, 309-330.

Fiebert, M. S. (1996). College students’ perceptions of men as victimsof women’s assaultive behavior, Perceptual and Motor Skills, 82, 49-50.

Fiebert, M. S., & Gonzalez, D. M. (1997). College women who initiateassaults on their male partners and the reasons offered for suchbehavior. Psychological Reports, 80, 583-590.

Flynn, C. P. (1990). Relationship violence by women: Issues and impli-cations. Family Relations, 36, 295-299.

Hydén, M. (1995). Verbal aggression as prehistory of woman batter-ing. Journal of Family Violence, 10, 55-71.

Jouriles, E. N., & O’Leary, K. D. (1985). Interspousal reliability ofreports of marital violence. Journal of Consulting and Clinical Psy-chology, 53, 419-421.

Kasian, M., & Painter, S. L. (1992). Frequency and severity of psycho-logical abuse in a dating population. Journal of Interpersonal Vio-lence, 7, 350-364.

Langhinrichsen-Rohling, J., Neidig, P., & Thorn, G. (1995). Violentmarriages: Gender differences in levels of current violence andpast abuse. Journal of Family Violence, 10, 159-176.

Lisak, D., & Ivan, C. (1995). Deficits in intimacy and empathy in sexu-ally aggressive men. Journal of Interpersonal Violence, 10, 296-308.

Makepeace, J. (1981). Courtship violence among college students.Family Relations, 30, 97-102.

Makepeace, J. (1986). Gender differences in courtship violence vic-timization. Family Relations, 35, 383-388.

Makepeace, J. (1989). Dating, living together, and courtship vio-lence. In M. A. Pirog-Good & J. E. Stets (Eds.), Violence in datingrelationships (pp. 94-107). New York: Praeger.

Malamuth, N. M., Sockloskie, R. J., Koss, M. P., & Tanaka, J. S. (1991).Characteristics of aggressors against women: Testing a modelusing a national sample of college students. Journal of Consultingand Clinical Psychology, 5, 670-681.

Margolin, G. (1987). The multiple forms of aggression between mari-tal partners: How can we identify them? Journal of Marital and Fam-ily Therapy, 13, 77-84.

McKinney, K. (1986). Perceptions of courtship violence: Gender dif-ference and involvement. Free Inquiry into Creative Sociology, 14, 61-66.

Moffitt, T. E., Caspi, A., Krueger, R. F., Magdol, L., Margolin, G., Silva,P. A., et al. (1997). Do partners agree about abuse in their relation-ship? A psychometric evaluation of interpartner agreement. Psy-chological Assessment, 9, 47-56.

Morse, B. J. (1995). Beyond the Conflict Tactics Scale: Assessing gen-der differences in partner violence. Violence and Victims, 10, 251-272.

Murphy, C., & O’Leary, K. D. (1989). Psychological aggression pre-dicts physical aggression in early marriage. Journal of Consultingand Clinical Psychology, 57, 579-582.

O’Leary, K. D., Barling, J., Arias, I., Rosenbaum, A., Malone, J., &Tyree, A. (1989). Prevalence and stability of physical aggressionbetween spouses. Journal of Consulting and Clinical Psychology, 57,263-268.

Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993). Masculine ideologyand its correlates. In S. Oskomp & M. Costanzo (Eds.), Genderissues in contemporary society (pp. 85-110). Newbury Park, CA: Sage.

Riggs, D. S., Murphy, C. M., & O’Leary, K. D. (1989). Intentional falsi-fication in reports of interpartner aggression. Journal of Interper-sonal Violence, 4, 220-232.

Riggs, D. S., & O’Leary, K. D. (1989). A theoretical model of courtshipaggression. In M. A. Pirog-Good & J. E. Stets (Eds.), Violence in dat-ing relationships (pp. 53-71). New York: Praeger.

Riggs, D. S., O’Leary, K. D., & Breslin, F. C. (1990). Multiple correlatesof physical aggression in dating couples. Journal of Interpersonal Vio-lence, 5, 61-73.

Sigelman, C. K., Berry, C. J., & Wiles, K. A. (1984). Violence in collegestudents’ dating relationships. Journal of Applied Social Psychology,14, 530-548.

Smith, M. D. (1994). Enhancing the quality of survey data on violenceagainst women: A feminist approach. Gender and Society, 8, 109-127.

Spence, J., Helmreich, R., & Holahan, C. (1979). Negative and posi-tive components of psychological masculinity and femininity andtheir relationships to self-reports and neurotic and acting outbehaviors. Journal of Personality and Social Psychology, 37, 1673-1682.

Spence, J. T., Losoff, M., & Robbins, A. S. (1991). Sexually aggressivetactics in dating relationships: Personality and attitudinal corre-lates. Journal of Social and Clinical Psychology, 10, 289-304.

Stets, J. E., & Pirog-Good, M. A. (1987). Violence in dating relation-ships. Social Psychology Quarterly, 50, 237-246.

Stets, J. E., & Straus, M. A. (1989). The marriage license and a hittinglicense: A comparison of assaults in dating, cohabiting, and mar-ried couples. In M. A. Pirog-Good & J. E. Stets (Eds.), Violence indating relationships (pp. 34-51). New York: Praeger.

Stets, J. E., & Straus, M. A. (1992). Gender differences in reportingmarital violence and its medical and psychological consequences.In M. A. Straus & R. J. Gelles (Eds.), Physical violence in American

Jenkins, Aubé / GENDER AND AGGRESSION 1117

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from

families: Risk factors and adaptations in 8,145 American families (pp.151-166). New Brunswick, NJ: Transaction Books.

Straus, M. A. (1979). Measuring intrafamily conflict and violence:The Conflict Tactics (CT) Scales. Journal of Marriage and the Family,41, 75-88.

Straus, M. A. (1992a). The Conflict Tactics Scales and its critics: Anevaluation and new data on validity and reliability. In M. A. Straus& R. J. Gelles (Eds.), Physical violence in American families: Risk fac-tors and adaptations in 8,145 American families (pp. 49-73). NewBrunswick, NJ: Transaction Books.

Straus, M. A. (1992b). Measuring intrafamily conflict and violence:The Conflict Tactics (CT) Scales. In M. A. Straus & R. J. Gelles(Eds.), Physical violence in American families: Risk factors and adapta-tions in 8,145 American families (pp. 29-47). New Brunswick, NJ:Transaction Books.

Strauss, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B.(1996). The revised Conflict Tactics Scales (CTS2). Journal of Fam-ily Issues, 17, 283-316.

Sugarman, D. B., & Hotaling, G. T. (1989). Dating violence: Preva-lence, context, and risk markers. In M. A. Pirog-Good & J. E. Stets(Eds.), Violence in dating relationships (pp. 3-31). New York: Praeger.

Szinovacz, M. E. (1983). Using couple data as a methodological tool:The case of marital violence. Journal of Marriage and the Family, 45,633-644.

Thompson, E. H., Jr. (1991). The maleness of violence in dating rela-tionships: An appraisal of stereotypes. Sex Roles, 24, 261-279.

Thompson, E. H., Jr., & Pleck, J. H. (1986). The structure of male rolenorms. American Behavioral Scientist, 29, 531-543.

Vivian, D., & Langhinrichsen-Rohling, J. (1994). Are bi-directionallyviolent couples mutually victimized? A gender-sensitive compari-son. Violence and Victims, 9, 107-124.

Walker, L. E. (1984). The battered woman syndrome. New York: Springer.Walker, L. E. (1989). Psychology and violence against women. Ameri-

can Psychologist, 44, 695-702.White, J. W., & Koss, M. P. (1991). Courtship violence: Incidence in a

national sample of higher education students. Violence and Victims,6, 247-256.

Wilson, G. D. (1982). Feminism and marital satisfaction. Personalityand Individual Differences, 3, 345-347.

Received July 26, 2001Revision accepted January 4, 2002

1118 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

at WEST VIRGINIA UNIV LIBRARY on June 27, 2014psp.sagepub.comDownloaded from