How Peer Networks Influence Male Sexual Aggression

Psychology of Violence © 2012 American Psychological Association 2013, Vol. 3, No. 2, 157–171 2152-0828/13/$12.00 DOI: 10.1037/a0029997 The Company They Keep: How Peer Networks Inﬂuence Male Sexual Aggression

Kevin M. Swartout Georgia State University

Objective: The goal of the present study was to add to knowledge concerning predictors of sexually aggressive behaviors by extending an existing model of sexual aggression to include attitudinal and structural variables of participants’ peer groups. Method: A battery of questionnaires was administered to 341 college-aged men via web-based survey. Participants were asked to report their previous sexual behavior, attitudes toward women and sexual aggression, the strength of relationships within their peer network, and their peers’ attitudes toward women and sexual aggression. Results: Findings suggest perceived peer rape-supportive attitudes significantly influence individual members’ hostile attitudes toward women. Peer network density negatively predicted hostile attitudes—individuals with tightly knit peer groups tend to have less hostile attitudes toward women; there was a significant interaction between peer group density and perceived peer rape-supportive attitudes in predicting individuals’ hostile attitudes toward women—individuals in high-density, low-hostility peer groups had the lowest average levels of hostility toward women. Conclusion: The present findings suggest perceived peer attitudes and structure of peer networks influence individuals’ attitudes concerning violence and hostility toward women, factors long known to predict both physical and sexual violence against women. These findings may be implemented through peer-focused bystander intervention programs aimed at reducing sexual aggression.

Keywords: sexual aggression, peer inﬂuence, social networks, attitudes, aggression

Sexual aggression is a social phenomenon 2006). And, in hindsight, almost exclusive reli- most people would rather ignore; but it is dif- ance on intrapersonal variables may not have ﬁcult to ﬁnd anyone who has not been either been the ideal strategy for the study of sexual directly or indirectly affected by this unfortu- aggression. Although these acts are most nate reality. Researchers have been relatively commonly perpetrated by an individual unsuccessful in identifying a particular type of within an isolated situation, interpersonal man who would or could perpetrate these acts, variables may help researchers better under- possibly because sexually aggressive men are stand sexual aggression. Polk may have best quite a heterogeneous group (e.g., Groth, 1977; stated the impetus for the current project: Knight & Prentky, 1990; Oxnam & Vess, “whether or not a male engages in sexually aggressive behavior may, in part, be due to the values and expectations of his male friends”

This document is copyrighted by the American Psychological Association or one of its allied publishers. (Polk et al., 1981, as cited in Ageton, 1983, p. This article was published Online First September 24, 2012. This article is intended solely for the personal use of the individual user and isI not tothank be disseminated broadly. Jacquelyn W. White, Janet J. Boseovski, Rose- 388). To this point, however, peer influences mery Nelson–Gray, John J. Seta, and Paul J. Silva for their have not been integrated into our empirical un- advice throughout this project. derstanding of sexual aggression; an omission This article details a portion of the research conducted by addressed by the current research. Kevin M. Swartout in partial fulfillment of his doctoral dissertation. A presentation based upon this article was awarded the Sexual aggression is commonly defined as Southeastern Psychological Association’s Most Outstanding compelling sexual activity where consent is not Paper Award during their 2012 annual meeting. obtained (Centers for Disease Control and Pre- Correspondence concerning this article should be ad- vention, 2009). A nationally representative sur- dressed to Kevin M. Swartout, Department of Psychology, Georgia State University, 140 Decatur Street, Atlanta, GA vey of sexual aggression and victimization 30303. E-mail: [email protected] found that over 50% of college-aged women

157 158 SWARTOUT

reported experiencing some form of sexual ag- Keseredy, 1997, 2000), but, to date, peer atti- gression (Koss, Gidycz, & Wisniewski, 1987), tudes have not been integrated into empirical and 25% of college-aged men reported engag- models of sexual aggression (i.e., Knight & ing in at least one instance of sexually aggres- Sims–Knight’s, 2004; Malamuth Sockloskie, sive behavior after age 14 (i.e., attempted or Koss, & Tanaka, 1991). Malamuth, Sockloskie, completed sexual contact—ranging from un- Koss, and Tanaka’s (1991) confluence model wanted contact to rape—without full female has been replicated and extended by several consent); almost 8% of the male sample re- research teams (e.g., Anderson & Anderson, ported engaging in behaviors that met legal 2008; Parkhill & Abbey, 2008) and will serve as definitions for rape or attempted rape. This pat- the framework for the current analyses, because tern of sexually aggressive behavior has been it is the only published model that (1) is empir- supported by subsequent findings of numerous ically testable in its entirety; (2) is strictly fo- research teams (e.g., White & Smith, 2004; cused on sexual aggression toward adult wom- Thompson, Swartout, & Koss, under review). en; and (3) was developed using data from a Relatively little research has explored associ- nonincarcerated population. ations between peer influence and individual sexually aggressive behavior; the research that The Confluence Model of Sexual has been conducted, however, indicates a strong Aggression association. There is, in fact, a relation between levels of peer and individual sexual aggression The confluence model hypothesizes two (Alder, 1985; Gwartney–Gibbs, Stockard, & pathways—impersonal sex and hostile mascu- Bohmer, 1989). Simply being a member of a linity—leading to sexually aggressive behavior male peer group increases the likelihood of sex- (Malamuth et al., 1991). In the initial iteration ually aggressive behavior (Berkowitz, Burkhart, of the confluence model, the impersonal sex & Bourg, 1994). Although these findings seem pathway was indicative of high levels of sexual clear cut, researchers continue to puzzle over activity with little emotional attachment to the nature of this relationship. College men liv- one’s partner on the part of the male, also ing in all-male dormitories are more likely than termed “sociosexuality” (Mouilso & Calhoun, other men to endorse rape myths (Schaeffer & 2012). The hostile masculinity pathway was Nelson, 1993). Members of peer groups that comprised of two attitudinal constructs: atti- objectify women tend to engage in more severe tudes supporting violence and a more specific sexual aggression, as compared with men who construct reflecting hostile masculine atti- do not associate with this type of peer group tudes. Both the impersonal sex and hostile (Koss & Dinero, 1989). Research conducted by masculinity pathways were significantly in- Schwartz and DeKeseredy (1997, 2000) found fluenced by adolescence delinquency; delin- that male sexually aggressive behavior, in part, quency, in turn, was influenced by negative is a function of peer support of intimate partner childhood experiences, such as child abuse or violence. After interviewing 341 male college witnessing domestic violence. students, Kanin (1967) found that sexually ag- The confluence model utilizes individual- gressive men—as compared with nonsexually level behavioral and attitudinal variables to- aggressive men—reported experiencing more gether to predict sexually aggressive behavior. peer pressure to engage in premarital sex. Shot- In this respect, the model provides insight into This document is copyrighted by the American Psychological Association or one of its allied publishers. land (1992) took a more cognitive stance with both the content and process involved with sex- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. the assertion that male peer groups “reinforce ual aggression at the individual level. Content [sexually aggressive] beliefs and help keep referring to the attitudes and behaviors found them accessible so that the rapist is cognitively predictive of sexual aggression and process re- ready to act” (p. 139). ferring to the manner in which these variables These findings, taken together, indicate a are organized to form a person-level model of strong and robust association between men’s sexual aggression. As noted by Malamuth, how- sexually aggressive behaviors and their peers’ ever, “such research may also benefit from more attitudes toward women and sex. This assump- general analyses of social influence” (Malamuth tion has been included in theoretical models of et al., 1991, p. 680). Although Malamuth wrote sexual aggression (e.g., Schwartz & De- this passage over 20 years ago, the confluence PEER INFLUENCE ON SEXUAL AGGRESSION 159

model has yet to be extended to address any account for the relationship between peer and form of social influence on sexual aggression; individual sexual aggression (Alder, 1985; replications and extensions continue to use only Gwartney–Gibbs, Stockard, & Bohmer, 1989) individual-level variables. and would generally endorse peer-support inter- pretations of sexual aggression (Schwartz & Social Networks and Social Influence DeKeseredy, 1997, 2000). Recent findings suggest bystanders weigh peer attitudes toward sex- Developed and utilized by social scientists of ual aggression heavily when making their deci- various disciplines over the past 50 years, the sion whether or not to intervene in a sexually social network perspective has been almost en- aggressive situation (Brown & Messman– tirely absent from research on violence against Moore, 2010); this further supports the theme women (Wasserman & Faust, 1994). Social net- that men look to their peers for information on work analysis offers an opportunity to explore the degree to which sexually aggressive behav- many of the same social phenomena that vio- ior is acceptable. lence against women researchers have studied for decades, but from a different theoretical, methodological, and analytic standpoint. When The Current Research attempting to explain or predict individual behavior, the social network perspective focuses Based on previous findings, attitudes of on an individual’s relationship dynamics as op- men’s close friends should predict their sexu- posed to intraindividual variables. Through this ally aggressive attitudes and behaviors; this re- lens, researchers can describe and analyze pat- lation, at least in part, should be a function of terns of relationships between individuals as the density of their peer group. Highly hostile well as the effects associated with these rela- and highly dense peer groups should influence tionships—an object of analysis often disre- individual members to also hold highly hostile garded by social scientists (Wasserman & attitudes toward women. By the same token, Faust, 1994). high-density, low-hostility peer groups should Structural variables within social networks influence members to have low levels of hostil- can affect how information is transferred be- ity. Thus, the following research questions were tween people and the extent to which people are developed: (1) What is the relation between influenced by this information (Collins, 1988). peer and individual attitudes toward women, The current research focuses on a specific type sex, and sexual aggression? and (2) Do peer of social network—the peer network—a group attitudes interact with peer network density to of similarly aged people who sustain personal predict individual attitudes? The hypothe- relationships across time. It is well established sized peer influence main effects model will that peer networks play a major role in the include the impersonal sex and hostile mas- development of aggressive behaviors, especially among children and adolescents (for a culinity pathways of the confluence model review see Espelage, Wasserman, & Fleisher, with the addition of constructs representing 2007). More specifically, a focus of the current perceived peer attitudes and peer network research is peer network density. For the current density, each predicting the two attitudinal purposes, peer network density is defined as the constructs within the hostile masculinity path- This document is copyrighted by the American Psychological Association or one of its allied publishers. strength of relationships among participants’ way. The hypothesized peer influence inter- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. close male friends. This will serve as an indi- action model will include the same constructs cator of how tightly knit participants’ peer with the addition of an interaction term be- groups are. tween perceived peer attitudes and peer net- Numerous published reports have linked so- work density, also predicting both constructs cial influences to attitude development and at- in the hostile masculinity pathway. In accor- titude strength (for a review, see Prislin & dance with recent published extensions (e.g., Wood, 2005). Male peer groups may be largely Anderson & Anderson, 2008; Parkhill & Ab- responsible for the development of attitudes bey, 2008), the current project will extend the predictive and supportive of sexual aggression. confluence model without the childhood risk Peer influences on key attitudes may partially factor or social isolation constructs. 160 SWARTOUT

Method surement of participant perceptions of peer attitudes rather than actual peer attitudes. The Participants literature concerning social influences on attitudes about sex suggests perceptions of others’ Participants were recruited from a participant sexual attitudes and behaviors, rather than pool at a medium-sized public university. The peers’ actual attitudes and behaviors, directly group of men who met inclusion criteria and influence individual attitudes and behaviors consented to participate in the study (N ϭ 341) (Chia & Lee, 2008; Cohen & Shotland, 1996). constituted a large sample size for structural equation modeling per Kline (2011). Partici- Constructs and Measures pants completed a series of web-based surveys in exchange for course credit. Of the men who Delinquency. Three indicators of delin- participated in this study, the average age was quency were used: the Self-reported Delin- 18.9 years and 60.9% were Caucasian, 20.6% quency Scale developed by Elliott, Huizinga, were African American, 7.5% were Asian or and Ageton (1985), and two separate questions Pacific Islander, 4.6% were Hispanic, 0.6% developed by Malamuth et al. (1991). The Self- were Native American or Alaskan Native, and reported Delinquency Scale is a 21-item mea- 5.8% were of another ethnicity. sure that asks respondents how many times since age 14 they have engaged in specific de- Procedure linquent behaviors such as damaging property, stealing, selling drugs, cheating, and fighting. Prospective participants signed-up for the Participants responded to each item on 5-point study through a web-based research participant scales ranging from “never”to“more than 10 management system; per departmental regula- times.” Composite scores were calculated by tion, the only information prospective partici- averaging participants’ responses within the pants received concerning the study prior to scale. The two questions developed by Mala- signing-up was the name of the person oversee- muth et al. (1991, p. 673) are “When you were ing the project, the location of the study (online, growing up, how many of your friends, if any, in this case), and the inclusion criteria: “Partic- got in trouble with the law for minor offenses ipants must be male and at least 18 years of (e.g., fighting or running away)?” and “How age.” Prospective participants were immedi- many times, if any, did you run away from ately given a link to the web-based survey after home for more than 24 hours?” Participants signing-up; upon following the link, they were responded to each of these items on 5-point immediately screened for gender and age—in scales ranging from “no”to“yes, more than 5 case they did not mind the inclusion criteria— times.” Raw scores from both of these questions males ages 18 years of age and older were in addition to composite scores from the Self- forwarded to the informed consent page to learn reported Delinquency Scale were used to indi- about the study. Men who consented to partic- cate the delinquency construct. ipate were forwarded to the measures detailed in Attitudes supporting violence. Three the next section. Twenty-five prospective par- scales developed by Burt (1980)—the Adver- ticipants were immediately disqualified because sarial Sexual Beliefs (ASB), Acceptance of In- they were not males and two because they were terpersonal Violence (AIV), and Rape Myth This document is copyrighted by the American Psychological Association or one of its allied publishers. under the age of 18. After reading the informed Acceptance (RMA) scales—were used to indi- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. consent materials, 10 individuals declined to cate individuals’ attitudes supporting violence participate in the study. against women. The ASB is a 9-item measure of Each measure was presented on a separate the extent to which people judge male-to- page and contained specific instructions for in- female relationships to be antagonistic. A sam- terpreting and responding to survey items. Items ple item is “A man’s got to show the woman pertaining to individual and peer attitudes were who’s boss right from the start or he’ll end up counterbalanced to account for order effects. No henpecked.” The AIV contains 6 items that significant effects were found related to order of measure the extent to which men support vio- measures. Peer attitudes were collected via par- lence within intimate relationships. A sample ticipant self-report; this allowed for the mea- item is “Sometimes the only way a man can get PEER INFLUENCE ON SEXUAL AGGRESSION 161

a cold woman turned on is to use force.” The Peer network density. A modified version RMA is a 13-item measure containing items of the procedure used by Green, Richardson, such as “A woman who goes to the home or and Lago (1996) was used to assess peer net- apartment of a man on their first date implies work density. Participants were asked to com- that she is willing to have sex.” Participants plete a measure of peer network density by responded to all items within these three mea- providing responses to the statement “Please list sures on 7-point scales ranging from “strongly the five (5) male peers with whom you most disagree”to“strongly agree.” Composite often associated during high school (either face- scores were calculated by averaging partici- to-face, over the phone, or through electronic pants’ responses within each scale. means such as text messages, email, and social Hostile masculinity. The Sexual Domi- networking sites).” Answers to the aforemen- nance Scale (SDO; Nelson, 1979), the Hostility tioned statement were forwarded into a subse- Toward Women Scale (HTW; Check, 1985), quent series of questions that asked participants and the Adversarial Sexual Beliefs Scale (ASB; to “Rate the relationship strength of each of the Burt, 1980) were used to operationalize the hos- following pairs of peers with 0 meaning they tile masculinity construct. It should be noted have never met and 100 meaning that they are that the ASB, described above, was also origi- extremely close friends.” This statement was nally used to indicate both the attitudes support- followed by all 10 possible pairs of the five ing violence and hostile masculinity constructs peers previously listed by participants. Peer net- of the confluence model (Malamuth et al., work density was calculated as the average re- 1991). The SDO is an 8-item subscale of the lationship strength of participants’ peers; this Sexual Functions Inventory (Nelson, 1979). observed variable represented peer network This subscale measures the extent to which sex- density in the modeling process. Perceived peer attitudes. ual activity is motivated by desire for power or Two variables measured by the Justification of Rape Scale control over one’s sexual partner. A sample (JRS; Burgess, 2007) and the attitudes section item is “I have sex because: I enjoy the feeling of the Date Rape Attitudes Survey (DRAS; La- of having someone in my grasp.” The HTW nier & Elliot, 1997) were used to indicate the scale is a 21-item attitudinal measure of anger perceived peer attitudes construct. These two specifically directed at women. A sample item measures were chosen in an effort to prevent is “I feel that many times women flirt with men projection or anchoring effects because they just to tease or hurt them.” Participants re- differ from the measures used to assess the sponded to all items within these three measures individual attitudes supporting violence and on 7-point scales ranging from “strongly agree” hostile masculinity constructs. The instructions to “strongly disagree.” Composite scores were of the JRS and the DRAS were modified to ask calculated by averaging participants’ responses about peer attitudes rather than personal atti- within each scale. tudes: “For the following statements, please an- Impersonal sex. Four items indicated im- swer according to what your close friends think, personal sex. The first two were questions: specifically [names of the participant’s five “How many sexual partners have you had in friends were forwarded to this point]. If these your lifetime?” and “What is the approximate friends were hanging out, honestly discussing number of dates that you expect to go on with a each statement without you there, what re- This document is copyrighted by the American Psychological Association or one of its allied publishers. woman before you engage in sexual inter- sponses would they give?” To ensure that par- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. course?” (reverse-coded for all inferential anal- ticipants continually applied these instructions yses). The third and fourth were responses to each item was preceded by the statement “An- the statements: “Sex without love is okay” and swer for your friends.” “You enjoy casual sex with different partners” The JRS is a 10-item measure found to (Hendrick & Hendrick, 1987; Simpson & Gan- strongly relate to sexually aggressive behavior gestad, 1991); participants responded to both of and proclivity (Burgess, 2007). A sample item these statements on 7-point scales ranging from is “Using aggression or physical restraint is a “strongly disagree”to“strongly agree.” Scores legitimate way to acquire sex from a certain from each of these four items were used as type of woman.” The DRAS was specifically indicators of the impersonal sex construct. developed to assess attitudes toward date rape 162 SWARTOUT

among college students. The DRAS contains 20 advantage when they were too drunk, high, or items such as “Women provoke rape by their out of it to stop what was happening”; this behavior.” Participants responded to statements sample pairing constitutes rape. Men were in- on both the JRS and DRAS on 7-point scales structed to report “the number of times each with responses ranging from “strongly agree” experience has happened to you since your 14th to “strongly disagree.” Composite scores were birthday” with response options ranging from calculated by averaging participants’ responses “zero”to“three or more times.” within each scale. Sexual aggression. Sexual aggression was Results assessed using the revised version of the Sexual Experiences Survey for perpetration (SES-R-P; Data Analysis Strategy Koss et al., 2007). The current study employed the entire SES-R-P short form plus the alcohol- After descriptive statistics and internal con- and drug-related items as well as the multiple sistency scores were calculated (see Table 1), perpetrator item from the long form, which to- variables were centered around their means and taled 56 items. The SES-R-P measured the fre- fit to structural equation models (SEM) using quency and severity of men’s sexual experi- Mplus version 5.1 (L. K. Muthén & B. O. ences. Based upon participants’ responses to the Muthén, 1998–2010) to assess the measurement SES-R-P, sexual aggression was modeled as a models of each latent factor and the structural latent factor. Four manifest variables were con- relationships between factors. SEM can be sim- structed based on men’s frequency of each form ply described as a combination of confirmatory of sexual aggression—unwanted sexual con- factor analysis: where variables are assigned to tact, verbal coercion, attempted rape, or a predetermined set of factors, and path analy- rape—as proposed by the SES Collaborative sis: where several regression paths are fit within (Koss et al., 2007). A sample act measured by the same model. SEM corrects for measurement this survey is “I had oral sex with a woman or error, allows structural relations between both had a woman perform oral sex on me without latent and observed variables, and provides sta- her consent by” and a sample tactic is “taking tistics relative to overall model fit (Kline, 2011).

Table 1 Means, Standard Deviations, Ranges, and Reliability Estimates for Each Indicator

Variable MSDRange ␣ Acceptance of interpersonal violence toward women 1.58 0.86 0–4 .61 Rape myth acceptance 1.33 0.66 0–3.16 .86 Adversarial sexual beliefs 2.49 0.96 0–5.67 .81 Hostility toward women 2.42 0.71 .10–4.30 .87 Sexual dominance orientation 2.69 1.24 0–5.63 .88 Peer date rape acceptance 2.44 .79 .29–4.52 .86 Peer justiﬁcation of rape 1.31 .99 0–4.60 .86

This document is copyrighted by the American Psychological Association or one of its allied publishers. Delinquency scale 0.66 0.58 0–3.14 .86 Delinquency (friends in trouble) 1.44 1.25 0–4 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Delinquency (ran away) 0.14 0.57 0–4 Number of sexual partners 3.91 4.86 0–20 Number of dates before sex 13.81 9.10 0–35 Sex without love 2.90 2.12 0–6 Enjoy casual sex 2.40 2.05 0–6 SES: Unwanted contact 0.99 2.78 0–24 .85 SES: Verbal aggression 0.45 1.80 0–18 .87 SES: Attempted rape 0.60 4.03 0–54 .97 SES: Rape 0.69 4.55 0–54 .98 Network density 57.69 20.95 7.6–100 .81 Note. SES ϭ Sexual Experiences Survey. PEER INFLUENCE ON SEXUAL AGGRESSION 163

A nonsignificant chi-square (p Ͼ .05), root attempted rape or rape as part of a group of two mean square error of approximation (RMSEA) or more people. Notably, frequencies of at- below .05, comparative fit index (CFI) and tempted and completed rape were highly corre- Tucker-Lewis index (TLI) above .95, and a lated (r ϭ .95); this was addressed during the standardized root-mean-square residual modeling process. (SRMR) of below .08 all suggest a model fits Although perceived peer attitudes were moder- data well. RMSEA below .08, CFI and TLI ately correlated with all five individual-level atti- above .90, SRMR below .10 all suggest a model tudes indicating the hostile masculinity pathway adequately fits data. It should be noted that the (see Table 2), two distinct factors emerged when chi-square statistic is too liberal when applied to all seven indicators were analyzed in an explor- large samples. atory factor analysis; individual attitudes indica- Although SEM has become quite common tors loaded most strongly onto one factor and across academic disciplines, methodologists con- perceived peer attitude indicators loaded most tinue to develop this approach. An example of this strongly onto the other. This suggests participants continual development is our relatively recent actually reported perceptions of their peers’ atti- ability to model latent interaction terms. This in- tudes when prompted, not their own attitudes. novation allows researchers to test interactions Peer network density was a significant, positive between latent variables, or between latent and correlate of responses to the impersonal sex indi- observed variables. Some of the models tested in cator “enjoy casual sex” and responses to delin- the current research include a latent interaction quency questions concerning running away from variable calculated through a modified version of home and friends getting into trouble during ado- the latent moderated structural equation method lescence; it negatively correlated with all indica- (Klein & Moosbrugger, 2000), known as quasi- tors of hostile masculinity, individual attitudes, maximum likelihood estimation (QML; Klein & and perceived peer attitudes, although none of Muthén, 2007). Of the methods available to model these relations reached conventional significance nonlinear relations within a latent variable frame- levels: p Ͻ .05. work, the QML method is thought to be the most accurate because it uses the expectation-maximi- Structural Equation Models zation algorithm to produce maximum likelihood estimates that take the inherent non-normality of As a final preliminary analysis, Malamuth et latent product terms into account (Klein & Moos- al.’s confluence model of sexual aggression brugger, 2000; Kline, 2011). Approximate fit in- (1991)—which included latent constructs indi- dices and model chi-square statistics are not avail- cating delinquency, impersonal sex, attitudes able for random-effects models containing latent supporting violence, hostile masculinity, and interaction terms using the QML method because sexual aggression—showed adequate fit (␹2[98, of non-normality (Klein & Moosbrugger, 2000; 341] ϭ 287.36, p Ͻ .05, RMSEA ϭ .076, 90% Muthén, 2012). Fit of a model containing a latent confidence interval [CI] ϭ .063–.084, CFI ϭ interaction term can be assumed adequate when .93, TLI ϭ .91, and SRMR ϭ .08. Adversarial (1) fit of the model without the interaction is Sexual Beliefs did not load strongly onto the adequate and (2) all structural paths associated hostile masculinity construct (r2 Ͼ .01); there- with the interaction are statistically significant fore, it was only used to indicate the individual (Muthén, 2004). attitudes construct moving forward. Also, the This document is copyrighted by the American Psychological Association or one of its allied publishers. errors associated with the rape and attempted This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Preliminary Analyses rape indicators of sexual aggression were allowed to correlate. These two modifications im- Approximately 25% of the men reported en- proved model fit (␹2[98, 341] ϭ 227.42, p Ͻ gaging in some form of sexual aggression with .05, RMSEA ϭ .063, 90% CI ϭ .052– 11.4% reporting behavior that meets legal def- .073, CFI ϭ .95 TLI ϭ .94, and SRMR ϭ .06). initions for either attempted rape or rape; these To address the first substantive research ques- rates correspond with previously published tion—What is the relation between peer and indi- findings (Koss et al., 1987; White & Smith, vidual attitudes toward women, sex, and sexual 2004). Over 5% of the men reported participat- aggression?—a peer influence main effects model ing in an act that met legal definitions of either was estimated by adding the latent variable repre- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 6 SWARTOUT 164

Table 2 Correlations Among Indicators

12345678910111213141516171819 Ϫ.04 05. 07. ءء16. ءء22. ءء21. 09. ءء19. 02. 07. 03. ءء16. ءء40. ءء42. ءء31. ءء44. ءء47. ءءAcceptance of interpersonal violence — .54 .1 Ϫ.05 11. 10. ءء16. ءء24. ءء23. 06. ءءϪ.06 .19 09. ءء14. ءء21. ءء57. ءء54. ءء39. ءء51. ءءRape myth acceptance — .53 .2 01. 03. 02. 0.09 ءء19. ءء34. ءء18. ءء21. ء12. ء12. ء11. ءء12. ءء32. ءء45. ءء42. ءءAdversarial sexual beliefs — .61 .3 Ϫ.09 10. 09. ءء15. ءء25. ءء31. ءء19. ءء23. ء14. ءء15. ءء20. ءء19. ءء35. ءء45. ءءHostility toward women — .42 .4 Ϫ.08 07. 05. ءء14. ءء23. ءء34. ءء28. ءء19. 10. ءء19. ءء12. ءء19. ءء28. ءءSexual dominance orientation — .37 .5 Ϫ.03 03. 03. ءء15. ءء21. ءء29. ءء23. ءء30. 05. ءء15. ءء19. ءء24. ءءPeer date rape acceptance — .74 .6 Ϫ.04 06. 06. ءء19. ءء22. ءء22. ءء17. ءء29. 02. ء13. ءء14. ءءPeer justiﬁcation of rape — .23 .7 05. 08. 10. ءء16. ءء18. ءء20. ءء30. ءء15. ءء19. ءء76. ءءDelinquency 1 — .65 .8 ء12. 04. 06. 0.05 0.17 ءء27. ءء34. ءء19. ءء40. ءءDelinquency 2 — .66 .9 ء12. 06. 05. 0.09 ء14. ءء21. ءء33. ءء18. ءءDelinquency 3 — .23 .10 09. 03. 06. ء11. ءء15. ءء43. ءء38. ءءNumber of sexual partners — .32 .11 01. ء14. 10. ءء17. ءء20. ءء46. ءءNumber of dates before sex — .45 .12 06. 01. 01. ء11. 0.09 ءءSex without love — .67 .13 ء11. 08. 07. ءء17. ءءEnjoy casual sex — .20 .14 01. ءء73. ءء72. ءءSES: Unwanted contact — .73 .15 01. ءء78. ءءSES: Verbal aggression — .76 .16 Ϫ.02 ءءSES: Attempted rape — .95 .17 18. SES: Rape — Ϫ.05 19. Network density — Note. SES ϭ Sexual Experiences Survey. .p Ͻ .01 ءء .p Ͻ .05 ء PEER INFLUENCE ON SEXUAL AGGRESSION 165

senting peer attitudes and manifest variable repre- final peer influence main effects model, they senting peer network density, both predicting in- were reintroduced to the peer influence interac- dividual attitudes and hostile masculinity. Per- tion model to specifically assess the interactive ceived peer attitudes positively predicted attitudes effects of the peer-level constructs. The latent supporting violence but did not significantly pre- interaction positively predicted individual hos- dict hostile masculinity. Peer network density neg- tile masculinity but did not significantly predict atively predicted hostile masculinity but did not individual attitudes; perceived peer attitudes significantly predict individual attitudes. The ini- continued to positively predict individual atti- tial peer influence main effects model fit the data tudes and peer network density continued to adequately well (␹2[143, 341] ϭ 347.89, p Ͻ .05, negatively predict hostile masculinity; all of the RMSEA ϭ .066, 90% CI ϭ .057–.074, CFI ϭ confluence model pathways remained signifi- .93, TLI ϭ .91, and SRMR ϭ .07). After trim- cant; and delinquency and perceived peer atti- ming nonsignificant paths, one at a time, overall fit tudes significantly covaried. of the final peer influence main effects model was Approximate fit indices are not available for almost identical to the initial model (␹2[145, 341] random-effects models containing latent inter- ϭ 348.59, p Ͻ .05, RMSEA ϭ .065, 90% CI ϭ action terms estimated using the QML method .056–.074, CFI ϭ .93, TLI ϭ .91, and SRMR ϭ (L. K. Muthén, February 07, 2012); however, fit .07). can be inferred through comparing the loglike- To address the second research question—Do lihood value of this model with that of the initial peer attitudes interact with peer network density peer influence main effects model, reported to predict individual attitudes?—a peer influ- above. Model fit is significantly improved by ence interaction model was estimated by adding adding the latent interaction (–2*loglikelihood the latent interaction between perceived peer difference ϭ 4.36, p Ͻ .05); this is also indi- attitudes and peer network density, predicting cated by the significant interaction effect. The individual attitudes and hostile masculinity. Al- nonsignificant paths from peer network density though the paths from perceived peer attitudes and the interaction term to individual attitudes to hostile masculinity and from peer network were removed from the final peer influence in- density to individual attitudes were cut from the teraction model (see Figure 1); but the path

Peer Network Perceived Density Peer Attitudes .27*** -.08* -.11 .52*** .08*

Attitudes 1.09*** Hostile Supporting Masculinity .21**

This document is copyrighted by the American Psychological Association or one of its allied publishers. Violence This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Sexual .20* Aggression .20*

.46*** Impersonal Delinquency Sex

Figure 1. Final peer inﬂuence interaction model with unstandardized estimates. Note: Black p Ͻ .001), standardized ءءء ,p Ͻ .01 ءء ,p Ͻ .05 ء) dot indicates latent interaction term estimates are not available for models containing a latent interaction term. 166 SWARTOUT

from peer attitudes to hostile masculinity, Perceived peer though nonsignificant, was left in the final attitudes model in order to correctly estimate the inter- High (+1 SD) active effect on hostile masculinity. Average Mediation and moderation within the final Low (-1 SD) peer network influence model. There was a 1 positive and significant indirect effect between 0.8 perceived peer attitudes and hostile masculinity ϭ 0.6 via attitudes supporting violence (b* .57, 0.4 SE ϭ p Ͻ standard error [ ] .12, .001); the direct 0.2 effect between perceived peer attitudes and hos- 0 tile masculinity is slightly negative, leaving a -0.2 positive and significant total effect (b* ϭ .46, -0.4 SE ϭ .07, p Ͻ .001). This indicates that essen- -0.6 tially all of the influence the perceived peer

Individual hostileIndividual masculinity -0.8 attitudes construct has on hostile masculinity is -1 mediated by attitudes supporting violence. Simple slopes were calculated to probe the Low High significant latent interaction. To accomplish Peer network density this, latent factor scores were saved and used in a multiple linear regression with scores of peer Figure 2. Simple slopes of peer network density predict- network density, perceived peer attitudes, and ing hostile masculinity as a function of perceived peer their interaction predicting individual hostile attitudes. masculinity. The resulting coefficients were used to test the effect of peer network density on individual hostile masculinity at Ϫ1 standard confluence model of sexual aggression (1991) deviation [SD], mean, and ϩ1 SD levels of to include perceived peer attitudes concerning perceived peer attitudes. The simple slopes (see sexual aggression and peer network density. In Figure 2) were negative and significant for both the final peer influence interaction model, per- average and low levels of perceived peer atti- ceived peer attitudes positively predicted atti- ء ء tudes (bϪ1SD ϭϪ.18, SE ϭ .03, p Ͻ .001; bM ϭ tudes supporting violence against women, peer Ϫ.08, SE ϭ .02, p Ͻ .01) but nonsignificant for group density negatively predicted hostile mas- ء high levels (bϩ1SD ϭ .03, SE ϭ .03, p ϭ .46). culinity, and perceived peer attitudes and peer Although this interaction could also be inter- network density interacted to positively predict preted as peer group density moderating the hostile masculinity. The positive relation between relation between perceived peer and individual perceived peer attitudes and hostile masculinity is attitudes, the current interpretation allows read- mediated by attitudes supporting violence. The ers to see the protective influence of peer net- significant interaction can be interpreted as the work density on individuals across levels of relation between peer network density and hos- peer group acceptance of sexual aggression. tile masculinity varying as a function of per- This interpretation is thought to have greater ceived peer attitudes. clinical and policy implications, to be discussed It is not surprising that peers share similar This document is copyrighted by the American Psychological Association or one of its allied publishers. shortly. ideas concerning sexual aggression and that This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. peer network density plays a role in this rela- Discussion tion. These results correspond with more general findings concerning peer influence on atti- The current study sought to answer two re- tudes among adolescents (Harton & Latané, search questions: (1) What is the relation be- 1997), and the relations between peer network tween peer and individual attitudes toward structure and individual attitudes (Espelage et women, sex, and sexual aggression? and (2) Do al., 2007). The negative main effect of peer peer attitudes interact with peer network density network density on individual hostile masculin- to predict individual attitudes? These questions ity suggests highly dense peer groups generally were addressed by extending Malamuth et al.’s have a positive influence on members, in that PEER INFLUENCE ON SEXUAL AGGRESSION 167

these members report less hostile masculinity. tudes and behaviors, correlates of this aggression, Upon further inquiry, it is individuals in lower and social network factors would allow for direct aggression peer groups who were increasingly tests of these remaining questions. affected by peer network density. This suggests Aggressive teens are generally found to have tightly knit peer groups that hold attitudes less strong peer relationships, often with other ag- accepting of sexual aggression protect against gressive teens, and increase their status among members developing high levels of hostile mas- their peers through aggressive acts (Cairns & culinity, making these individuals less likely to Cairns, 1994). In contrast, some findings sug- perpetrate acts of sexual aggression. Alterna- gest teens in low density peer networks engage tively, influence within peer groups highly sup- in significantly more direct aggression com- portive of sexual aggression may be so strong, pared with teens in high density peer networks. even at low levels of group density, that there is Although not a main emphasis of their paper, a ceiling effect making peer group density a Green, Richardson, and Lago (1996) found nonsignificant factor for these groups. male teens who were members of low-density peer groups tend to be more physically aggres- Research Implications sive compared with members of more dense groups. Based on both past and present findings, Admittedly, the tenor of this paper suggests there may be two different forces involved with that peer groups influence individual members peer influence on individuals’ sexually aggres- to adopt the prevailing attitudes of the group, sive attitudes and behavior: (1) a general pres- and new group members generally conform to sure to think and act like peers and (2) a general the status quo. This notion of assimilation is protective effect of peer network density on supported by large bodies of psychological and aggressive attitudes and behavior. More re- sociological literature that have consistently research is needed to confirm and distinguish ported this general trend (e.g., Deutsch & Ge- these two proposed social forces. rard, 1955). An equally sound explanation, however, is the tendency for like-minded people Clinical and Policy Implications to attract one another through a process of se- lection (Byrne, 1971). It is plausible that men Programs designed to weaken men’s attitudes who hold hostile attitudes toward women and supporting sexual aggression have yielded who are accepting of sexual aggression are at- mixed results. These programs have been found tracted to one another. If this is the case, what is to be effective for men with moderately aggres- the mechanism that brings these men together? sive attitudes, but largely ineffective for men When they first meet one another, it is unlikely with highly aggressive attitudes (Stephens & that these men converse about their hostility George, 2008). This may be due to the peer toward women, or how they view rape as ac- networks within which men reside. The current ceptable; if these men select one another, it is findings suggest men who hold highly hostile probably through a more nuanced mechanism. attitudes toward women tend to associate with They may send and receive subtle verbal or peers who share these attitudes. Individual par- nonverbal cues, especially when they talk about ticipation in a rape prevention program does not topics related to women, sex, and masculinity. It address peer influences; there will be sustained is also possible there is a third variable that pressure on these men to embody their peer This document is copyrighted by the American Psychological Association or one of its allied publishers. brings these men together, one that is related groups’ attitudinal and behavior norms. Men This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. to sexually aggressive attitudes and behav- who hold attitudes moderately supportive of iors, but is more public, readily observable, sexual aggression may not have such strong and and likely to bring like-minded men together; consistent pressure from their peers. This might two strong contenders are general aggression explain why rape prevention programs are far (Anderson & Anderson, 2008) and alcohol and less effective with men who hold attitudes drug use (Gallagher, Hudepohl, & Parrott, highly supportive of sexual aggression. 2010; Parkhill & Abbey, 2008; Swartout & There has been a recent resurgence in re- White, 2010). More research is needed to sort search on bystander intervention in preventing out these remaining questions. Longitudinally sexual aggression (e.g., DeKeseredy, Schwartz, measuring individuals’ sexually aggressive atti- & Alvi, 2000; Fabiano, Perkins, Berkowitz, 168 SWARTOUT

Linkenbach, & Stark, 2003; Katz, 1995). Al- their high school peers, but were not asked though sexual assaults most commonly occur in about college peers. Future research on this an isolated context, antecedent situations are topic should address the attitudes and network commonly social—such as parties or bars (Par- density of college students’ current peer groups rott et al., 2012). Taken together, it is not a and should seek to replicate these findings in surprise that recent studies suggest the utility of noncollege-student populations of men. Finally, bystander intervention programs in reducing sex- men who reported no history of sexual activity ual assault (e.g., Banyard, Moynihan, & Plante, were included in the analyses. This decision 2007; Foubert, Langhinrichsen–Rohling, Bras- was made from a prevention perspective: just field, & Hill, 2010; Gidycz, Orchowski, & because a man is not sexually active does not Berkowitz, 2011). Prospective bystanders, how- preclude him from engaging in future sexually ever, weigh perceived peer attitudes toward sexual aggressive behavior. Data concerning men’s ag- aggression heavily when making the decision to gressive and hostile attitudes, past delinquent intervening in a sexually aggressive situation, behaviors, and peer network density inform the even more heavily than they weigh their own current models, regardless of sexual activity or attitudes (Brown & Messman–Moore, 2010). lack thereof. In a post hoc model test, when fit Based upon the informational social influence lit- only to data from sexually active men, the peer erature, social norms campaigns—similar to those influence interaction model showed similar developed to reduce alcohol use on college cam- structural relations to those reported in Figure 1. puses—could be implemented to encourage atti- Some structural relations within this post hoc tude change at a broader level. Young men look to model did not reach conventional significance their peers for information, especially concerning levels, this is most likely due to the decreased women, dating, and sex (Sim & Koh, 2003). If statistical power to detect significant effects as a information concerning what constitutes accept- result of decreasing the sample size (n ϭ 257). able behavior within these domains were readily Future research seeking to replicate these find- available, harmful peer influence might be less- ings should collect a larger sample and conduct ened. This could be combined with bystander in- a multiple-groups analysis to detect differences tervention programs to institute a dual-pronged as a function of sexual activity status. approach to reduce sexual aggression. Conclusions Limitations The present study adds to existing knowledge The measures used in this study to assess rape concerning peer-level predictors of sexually myth acceptance, adversarial sexual beliefs, and aggressive behaviors and extends a popular attitudes toward interpersonal violence are over model of sexual aggression etiology. Findings 30 years old and have been argued to lack suggest perceived peer rape-supportive atti- construct validity (Lonsway & Fitzgerald, tudes influence individual hostile and violent 1995). Although more valid and up-to-date attitudes toward women, while peer network measures of these constructs were available, density negatively predicts individuals’ hostile Burt’s (1980) scales were selected for this study masculinity—suggesting individuals with in an effort to maintain consistency with the highly dense peer groups, in general, tend to original iteration of the confluence model (Mal- harbor less hostility toward women. Further- This document is copyrighted by the American Psychological Association or one of its allied publishers. muth et al., 1991) as well as other recent exten- more, perceived peer attitudes moderated the This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. sions (e.g., Anderson & Anderson, 2008; Parkh- relation between peer network density and in- ill & Abbey, 2009). Future research assessing dividual hostile masculinity; on average, the peer influence on individuals’ attitudes toward least hostile men are in low-aggression, high- violence, women, and sex will employ more density peer groups. These findings illustrate current and valid scales (e.g., Lonsway & how perceived peer attitudes and peer network Fitzgerald, 1995; Payne, Lonsway, & Fitzger- density function through individuals’ hostile at- ald, 1999). titudes to indirectly affect men’s sexually ag- Because they were mostly first-semester col- gressive behavior. These findings could be im- lege students, participants were asked to report plemented into bystander intervention programs on attitudes and relationship characteristics of and social norms campaigns aimed at reducing PEER INFLUENCE ON SEXUAL AGGRESSION 169

sexual aggression. Future research concerning media consumption on college students’ misper- peer influence on individual sexually aggressive ception of sex-related peer norms. International attitudes and behaviors should approach these Journal of Public Opinion Research, 20, 52–73. relationships longitudinally. doi:10.1093/ijpor/edn005 Cohen, L. L., & Shotland, R. L. (1996). Timing of first sexual intercourse in a relationship: Expecta- tions, experiences, and perceptions of others. Jour- References nal of Sex Research, 33, 291–299. doi:10.1080/ Ageton, S. S. (1983). Sexual assault among adoles- 00224499609551846 cents. Lexington, MA: Lexington Books. Collins, R. (1988). Theoretical sociology. Orlando, Alder, C. (1985). An exploration of self-reported sexually FL: Harcourt Brace Jovanovich. aggressive behavior. Crime and Delinquency, 31, DeKeseredy, W. S., Schwartz, M. D., & Alvi, S. 306–331. doi:10.1177/0011128785031002009 (2000). The role of profeminist men in dealing Anderson, C. A., & Anderson, K. B. (2008). Men with women on the Canadian college campus. Vi- who target women: Specificity of target, generality olence Against Women, 6, 918–935. doi:10.1177/ of aggressive behavior. Aggressive Behavior, 34, 10778010022182218 605–622. doi:10.1002/ab.20274 Deutsch, M., & Gerard, H. B. (1955). A study of Banyard, V. L., Moynihan, M. M., & Plante, E. G. normative and informational social influences (2007). Sexual violence prevention through by- upon individual judgment. The Journal of Abnor- stander education: An experimental evaluation. mal and Social Psychology, 51, 629–636. doi: Journal of Community Psychology, 35, 463–481. 10.1037/h0046408 doi:10.1002/jcop.20159 Elliott, D. S., Huizinga, D., & Ageton, S. S. (1985). Berkowitz, A. D., Burkhart, B. R., & Bourg, S. E. Explaining delinquency and drug use. Beverly (1994). Research on college men and rape. New Hills, CA: Sage. Directions for Student Services, 65, 3–19. doi: Espelage, D. L., Wasserman, S., & Fleisher, M. 10.1002/ss.37119946503 (2007). Social networks and violent behavior. In Brown, A. L., & Messman–Moore, T. L. (2010). D. J., Flannery, A. T., Vazsonyi, & I. D., Waldman Personal and perceived peer attitudes supporting (Eds.), The Cambridge handbook of violent behav- sexual aggression as predictors of male college ior and aggression (pp. 450–464). New York, students’ willingness to intervene against sexual NY: Cambridge University Press. doi:10.1017/ aggression. Journal of Interpersonal Violence, 25, CBO9780511816840.023 503–517. doi:10.1177/0886260509334400 Fabiano, P. M., Perkins, W., Berkowitz, A., Linken- Burgess, G. H. (2007). Assessment of rape- bach, J., & Stark, C. (2003). Engaging men as supportive attitudes and beliefs in college men: social justice allies in ending violence against Development, reliability, and validity of the Rape women: Evidence for a social norms approach. Attitudes and Beliefs Scale. Journal of Interper- Journal of American College Health, 52, 105–112. sonal Violence, 22, 973–993. doi:10.1177/ doi:10.1080/07448480309595732 0886260507302993 Foubert, J. D., Langhinrichsen–Rohling, J., Brasfield, Burt, M. R. (1980). Cultural myths and supports for H., & Hill, B. (2010). Effects of a rape awareness rape. Journal of Personality and Social Psychol- program on college women: Increasing bystander ogy, 38, 217–230. doi:10.1037/0022-3514.38.2 efficacy and willingness to intervene. Journal of .217 Community Psychology, 38, 813–827. doi: Byrne, D. (1971). The ubiquitous relationship: Atti- 10.1002/jcop.20397 tude similarity and attraction: A cross-cultural Gallagher, K. E., Hudepohl, A. D., & Parrott, D. J. study. Human Relations, 24, 201–207. doi: (2010). Power of being present: The role of mind- 10.1177/001872677102400302 fulness on the relation between men’s alcohol use This document is copyrighted by the American Psychological Association or one of its allied publishers. Cairns, R. B., & Cairns, B. D. (1994). Lifelines and and sexual aggression toward intimate partners. This article is intended solely for the personal use of the individual user and is notrisks: to be disseminated broadly. Pathways of youth in our time. New York, Aggressive Behavior, 36, 405–413. doi:10.1002/ab NY: Cambridge University Press. .20351 Centers for Disease Control and Prevention. (2009). Gidycz, C. A., Orchowski, L. M., & Berkowitz, A. D. Understanding sexual violence: Fact sheet [PDF (2011). Preventing sexual aggression among col- document]. Retrieved from http://www.cdc.gov/ lege men: An evaluation of a social norms and violenceprevention/ bystander intervention program. Violence Check, J. V. E. (1985). The hostility toward women Against Women, 17, 720–742. doi:10.1177/ scale. Unpublished doctoral dissertation. Winni- 1077801211409727 peg, Manitoba, Canada: University of Manitoba. Green, L. R., Richardson, D. R., & Lago, T. (1996). Chia, S. C., & Lee, W. (2008). Pluralistic ignorance How do friendship, indirect, and direct aggression about sex: The direct and the indirect effects of relate? Aggressive Behavior, 22, 81–86. doi: 170 SWARTOUT

10.1002/(SICI)1098-2337(1996)22:2Ͻ81::AID- nal of Consulting and Clinical Psychology, 57, AB1Ͼ3.0.CO;2-X 242–250. doi:10.1037/0022-006X.57.2.242 Groth, A. N. (1979). Men who rape. New York, NY: Koss, M. P., Gidycz, C. A., & Wisniewski, N. (1987). Plenum Press. The scope of rape: Incidence and prevalence of Gwartney–Gibbs, P. A., Stockard, J., & Bohmer, S. sexual aggression and victimization in a national (1987). Learning courtship aggression: The influ- sample of higher education students. Journal of ence of parents, peers, and personal experiences. Consulting and Clinical Psychology, 55, 162–170. Family Relations: An Interdisciplinary Journal of doi:10.1037/0022-006X.55.2.162 Applied Family Studies, 36, 276–282. doi: Lanier, C. A., & Elliot, M. N. (1997). A new instru- 10.2307/583540 ment for the evaluation of a date rape prevention Harton, H. C., & Latané, B. (1997). Social influence program. Journal of College Student Development, and adolescent lifestyle attitudes. Journal of Re- 38, 673–676. search on Adolescence, 7, 197–220. doi:10.1207/ Lonsway, K. A., & Fitzgerald, L. F. (1995). Attitu- s15327795jra0702_5 dinal antecedents of rape myth acceptance: A the- Hendrick, S., & Hendrick, C. (1987). Multidimensional- oretical and empirical reexamination. Journal of ity of sexual attitudes. Journal of Sex Research, 23, Personality and Social Psychology, 68, 704–711. 502–526. doi:10.1080/00224498709551387 doi:10.1037/0022-3514.68.4.704 Kanin, E. J. (1967). An examination of sexual ag- Malamuth, N. M., Linz, D., Heavey, C. L., Barnes, gression as a response to sexual frustration. Jour- G., & Acker, M. (1995). Using the confluence nal of Marriage and the Family, 29, 428–433. model of sexual aggression to predict men’s con- doi:10.2307/349577 flict with women: A 10-year follow-up study. Katz, J. (1995). Reconstructing masculinity in the Journal of Personality and Social Psychology, 69, locker room: The mentors in violence prevention 353–369. doi:10.1037/0022-3514.69.2.353 project. Harvard Educational Review, 65, 163–174. Malamuth, N. M., Sockloskie, R. J., Koss, M. P., & Klein, A. G., & Moosbrugger, H. (2000). Maximum Tanaka, J. S. (1991). Characteristics of aggressors against women: Testing a model using a national likelihood estimation of latent interaction effects sample of college students. Journal of Consulting with the LMS method. Psychometrika, 65, 457– and Clinical Psychology, 59, 670–681. doi: 474. doi:10.1007/BF02296338 10.1037/0022-006X.59.5.670 Klein, A. G., & Muthén, B. O. (2007). Quasi- Mouilso, E. R., & Calhoun, K. S. (2012). A media- maximum likelihood estimation of structural equation model of the role of sociosexuality in the tion models with multiple interaction and quadratic associations between narcissism, psychopathy, and effects. Multivariate Behavioral Research, 42, sexual aggression. Psychology of Violence, 2, 16– 647–673. doi:10.1080/00273170701710205 27. doi:10.1037/a0026217 Kline, R. B. (2011). Principles and practice of struc- Muthén, B. O. (2004, November 01). Re: Mplus tural equation modeling (3rd ed.). New York, NY: discussion Ͼ structural equation modeling Ͼ in- Guilford Press. teraction [online forum comment]. Retrieved from Knight, R. A., & Prentky, R. A. (1990). Classifying http://www.statmodel.com/discussion/messages/ sexual offenders: The development and corrobora- 11/499.html tion of taxonomic models. In W. L. Marshall, D. R. Muthén, L. K. (2012, February 07). Re: Mplus dis- Laws, & H. E. Barbaree, (Eds.), Handbook of sex- cussion Ͼ structural equation modeling Ͼ test for ual assault: Issues, theories, and treatment of the moderating variables [online forum comment]. offender (pp. 23–52). New York, NY: Plenum Press. Retrieved from http://www.statmodel.com/ Knight, R. A., & Sims–Knight, J. E. (2004). Testing discussion/messages/11/69.html an etiological model for male juvenile sexual of- Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus fending against females. Journal of Child Sexual user’s guide (2nd ed.). Los Angeles, CA: Muthén This document is copyrighted by the American Psychological Association or one of its allied publishers. Abuse: Research, Treatment, & Program Innova- & Muthén. This article is intended solely for the personal use of the individual user and is nottions to be disseminated broadly. for Victims, Survivors & Offenders, 13, 33– Nelson, P. A. (1979). Personality, sexual functions, 55. doi:10.1300/J070v13n03_03 and sexual behavior: An experiment in methodol- Koss, M. P., Abbey, A., Campbell, R., Cook, S., ogy. Unpublished doctoral dissertation, University Norris, J., Testa, M.,...White, J. (2007). Revising of Florida. the SES: A collaborative process to improve as- Oxnam, P., & Vess, J. (2006). A personality-based sessment of sexual aggression and victimization. typology of adolescent sexual offenders using the Psychology of Women Quarterly, 31, 357–370. Millon Adolescent Clinical Inventory. New Zea- doi:10.1111/j.1471-6402.2007.00385.x land Journal of Psychology, 35, 36–44. Koss, M. P., & Dinero, T. E. (1989). Discriminant Parkhill, M. R., & Abbey, A. (2008). Does alcohol analysis of risk factors for sexual victimization contribute to the confluence model of sexual as- among a national sample of college women. Jour- sault perpetration? Journal of Social and Clinical PEER INFLUENCE ON SEXUAL AGGRESSION 171

Psychology, 27, 529–554. doi:10.1521/jscp.2008 Sim, T. N., & Koh, S. F. (2003). Domain conceptual- .27.6.529 ization of adolescent susceptibility to peer pressure. Parrott, D. J., Tharp, A. L., Swartout, K. M., Miller, Journal of Research on Adolescence, 13, 58–80. C. A., Hall, G. N., & George. W. H. (2012). Validity Simpson, J. A., & Gangestad, S. W. (1991). Individ- for an integrated laboratory analogue of sexual ag- ual differences in sociosexuality: Evidence for gression and bystander intervention. Aggressive Be- convergent and discriminant validity. Journal of havior, 38, 309–321. doi:10.1002/ab.21429 Personality and Social Psychology, 60, 870–883. Payne, D. L., Lonsway, K. A., & Fitzgerald, L. F. doi:10.1037/0022-3514.60.6.870 (1999). Rape myth acceptance: Exploration of its Stephens, K. A., & George, W. H. (2008). Rape structure and its measurement using the Illinois Rape prevention with college men: Evaluating risk sta- Myth Acceptance Scale. Journal of Research in Per- tus. Journal of Interpersonal Violence, 24, 996– sonality, 33, 27–68. doi:10.1006/jrpe.1998.2238 1013. doi:10.1177/0886260508319366 Prislin, R., & Wood, W. (2005). Social inﬂuence in Swartout, K. M., & White, J. (2010). The relationship attitudes and attitude change. In D. Albarraci, B. T. between drug use and sexual aggression in men Johnson, & M. P. Zanna (Eds.), The handbook of across time. Journal of Interpersonal Violence, 25, attitudes (pp. 671–705). Mahwah, NJ: Lawrence 1716–1735. doi:10.1177/0886260509354586 Erlbaum Associates Publishers. Thompson, M., Swartout, K. M., & Koss, M. P. Schaeffer, A. M., & Nelson, E. S. (1993). Rape- (under review). Trajectories and predictors of sex- supportive attitudes: Effects of on-campus resi- ually aggressive behaviors during emerging adult- dence and education. Journal of College Student hood. Psychology of Violence. Development, 34, 175–179. Wasserman, S., & Faust, K. (1994). Social network Schwartz, M. D., & DeKeseredy, W. S. (1997). Sex- analysis: Methods and applications. New York, ual assault on the college campus: The role of NY: Cambridge University Press. doi:10.1017/ male peer support. Thousand Oaks, CA: Sage Pub- CBO9780511815478 lications, Inc. White, J. W., & Smith, P. H. (2004). Sexual assault Schwartz, M. D., & DeKeseredy, W. S. (2000). Ag- perpetration and reperpetration: From adolescence gregation bias and woman abuse: Variations by to young adulthood. Criminal Justice and Behav- male peer support, region, language and school ior, 31, 182–202. doi:10.1177/0093854803261342 type. Journal of Interpersonal Violence, 15, 555– 565. doi:10.1177/088626000015006001 Shotland, R. L. (1992). A theory of the causes of court- Received September 6, 2011 ship rape: II. Journal of Social Issues, 48, 127– Revision received May 3, 2012 143. doi:10.1111/j.1540-4560.1992.tb01161.x Accepted May 18, 2012 Ⅲ

Call for Papers: Special Issue on Issues in Sexual Minority Parenting

Professional Psychology: Research and Practice will feature a special issue on sexual minority parenting. We invite articles addressing issues and professional practice focused on parenting, grandparenting, or growing up in a family with a sexual minority parent or child. We seek manuscripts addressing issues and interventions for families that include an individual who identiﬁes as lesbian, gay, bisexual, or transgender, or with other diverse or minority sexual orientations or gender identities, including issues of bias, minority stress, and microaggressions

This document is copyrighted by the American Psychological Association or one of its allied publishers. relevant to these populations. We are particularly interested in receiving qualitative and quantitative

This article is intended solely for the personal use of the individual user and is not to be disseminatedempirical broadly. studies, comprehensive literature reviews, conceptual work, and program and therapy outcome studies, although descriptions of innovative programs or interventions may also be considered. Manuscripts can be submitted through the Journal’s electronic portal, under the Instructions to Authors at: http://www.apa.org/pubs/journals/pro/index.aspx

Please note in your cover letter that you are submitting for this special issue and send in attention to Kathi A. Borden, Ph.D., Associate Editor, Professional Psychology: Research and Practice.