Fraternity Membership, Traditional Masculinity Ideologies, and Impersonal Sex: Selection and Socialization Effects

Abstract

Fraternity culture perpetuates traditional masculinity ideologies, but little research has considered the process by which men internalize these ideologies. Men may select into fraternities based on preexisting ideologies, or fraternities may have a socializing effect on ideologies. We used two longitudinal datasets to explore selection and socialization effects of fraternity membership on masculinity ideologies (gendered beliefs, gendered traits, and sexual double standard beliefs) and impersonal sex (sexual motives and multiple sex partners) among ethnically and racially diverse college men. Using dataset one (n = 166, M = 18.0 years old fall of first year), we explored the selection and socialization effects of fraternity membership on male role norms, masculine traits, and endorsement of the sexual double standard. Men who more strongly endorsed male role norms about status and the sexual double standard were more likely to join fraternities than other men, indicating selection effects. Using dataset two (n = 256, M = 18.5 years old fall of first year), we explored selection and socialization effects of fraternity membership on sex motives and multiple sex partners. We did not find much evidence for selection or socialization effects on sex motives and multiple sex partners. Our findings may inform intervention efforts for men before and during college.

Fraternity culture perpetuates traditional masculinity ideologies, or socially constructed beliefs about men’s roles (Thompson & Bennett, 2017). However, the process of how men internalize these ideologies is not well understood (Harway & Steel, 2015). These ideologies are particularly concerning because of their association with sexual aggression, including rape, attempted rape, forced touching, forced kissing, verbal coercion, and threat (Bagwell-Gray, Messing, & Baldwin-White, 2015; Krebs, Lindquist, Warner, Fisher, & Martin, 2009; Mouilso, Fischer, & Calhoun, 2012). Fraternity members are more likely to be sexually aggressive than other college men (Adams-Curtis & Forbes, 2004; Tharp et al., 2012). Fraternity membership may be associated with traditional masculinity ideologies via both selection effects (men with traditional masculinity ideologies select into fraternities; Koss & Cleveland, 1996) and socialization effects (the fraternity environment influences traditional masculinity ideologies; Godenzi, Schwartz, & DeKeseredy, 2001). The purpose of the current paper was to use longitudinal data to explore both selection and socialization effects of fraternities on college men’s traditional masculinity (Lemaster, Strough, Stoiko, & DiDonato, 2015; Murnen, 2015) and impersonal sex (a distinct indicator that describes one way in which men enact masculinity ideologies; Lesmester et al., 2015; Murnen, 2015). We aim to contribute to theory on fraternity membership and masculinity ideologies, and to inform the prevention of sexual aggression among fraternity members.

Fraternity Culture, Traditional Masculinity Ideologies, and Impersonal Sex

Experts have often described fraternity culture as adhering to traditional masculinity ideologies. An interdisciplinary body of research describes fraternities as highly masculine environments where men not only “have” traditionally masculine beliefs, but also “do” gender by engaging in traditionally masculine activities (Lesmester et al., 2015, Murnen & Kohlman, 2007; Sanday, 1996). Traditional masculinity ideologies may include endorsement of male role norms about status (beliefs that men should strive for high social status), endorsement of male role norms about toughness (beliefs that men should be physically and emotionally tough), endorsement of antifemininity male role norms (beliefs that men should avoid behaving in feminine ways), masculine traits (traits traditionally considered to be masculine, such as dominance and forcefulness), and endorsement of the sexual double standard (beliefs that men and women should be held to different standards for sexual behavior). These traditional masculinity ideologies are of particular interest to researchers and practitioners because of their associations with intimate partner violence and sexual aggression (Locke & Mahalik, 2005; Murnen, Wright, & Kaluzny, 2002), and because of their prominence in “rape-prone” societies (Sanday, 1996). Conformity to fraternity culture, then, may contribute to sexual aggression. Indeed, conformity to masculine norms mediates the association between fraternity membership and acceptance of violence (Seabrook, Ward, & Giaccardi, 2018).

One way of “doing” gender (Lester et al., 2015) may be engaging in impersonal sex (Sanday, 1996; Murnen, 2015; Murnen & Kohlman, 2007), or casual and unemotional sex that is often characterized by sex with many partners (Malamuth, Heavey, & Linz, 1996; Malamuth, Sockloskie, Koss, & Tanaka, 1991). This enactment of gender is common in fraternity culture (Murnen & Kohlman, 2007; Sanday, 1996) and is particularly concerning because impersonal sex is a theoretical pathway to sexual aggression, according to the Confluence Model of sexual aggression (Malamuth et al., 1996; Malamuth et al.1991). Some indicators of impersonal sex may be multiple sex partners (e.g., two or more partners in a three-month period; Fromme, Corbin, & Kruse, 2008; Malamuth et al., 1991) and impersonal sexual motives, or choosing to have sex for reasons that are not relationship or intimacy focused (Malamuth et al., 1991; Malamuth et al., 1996; Patrick & Lee, 2010). Sexually aggressive men may be less motivated to have sex for relationship-focused reasons (e.g., intimacy motives; Malamuth et al., 1991; Malamuth et al., 1996; Patrick & Lee, 2010), more motivated to have sex because of peer pressure (peer pressure motives; Abbey, Parkhill, BeShears, Clinton-Sherrod, & Zawacki, 2006; Tharp et al., 2012), and more motivated to have sex to cope with negative emotions (coping motives; Cortoni & Marshall, 2001; Ward & Beech, 2006).

Selection and Socialization Effects of Fraternity Membership

One possible explanation for the association between fraternity membership and traditional masculinity ideologies is that men who already have these ideologies select into fraternities. That is, men with more traditional masculinity ideologies men may be more likely to join fraternities than men with less traditional ideologies. Developmental Systems Theory suggests that individuals select into environments that are consistent with their pre-existing beliefs and behaviors (e.g., Gottlieb, 1996). Past longitudinal research has found evidence of selection into fraternities based on binge drinking and rape myth acceptance (Borsari, Hustad, & Capone, 2009; McMahon, 2010; McCabe et al., 2005; Park, Sher, & Krull, 2008), providing support for Developmental Systems Theory. For instance, individuals who binge drink more frequently before college are more likely to join a fraternity or sorority than individuals who binge drink less frequently before college, indicating that individuals who binge drink select into the Greek environment, where such binge drinking is common (Borsari et al., 2009; McCabe et al., 2005; Park et al., 2008; Seabrook, McMahon, & O’Conner, 2018).

Another possible explanation for the association between fraternity membership and traditional masculinity ideologies is that fraternities have a socializing effect on members. According to the Male Peer Support Model of sexual aggression, men with all-male social networks may adopt particular masculinity ideologies, including ideologies that are hostile to women such as the acceptance of sexual aggression. These ideologies are associated with sexually aggressive behavior (Godenzi et al., 2001; Schwartz et al., 2001). Men who are more attached to a particular group are more likely to adopt the group’s ideologies than men who are less attached (Godenzi et al., 2001; Schwartz et al., 2001). Past research on fraternity culture described how fraternity brothers adopt ideologies, supporting the Male Peer Support Model. For instance, fraternity members tend to have particularly strong attachment to their fraternity, stronger than men’s attachment to other university clubs and organizations (Godenzi et al., 2001). Fraternity members’ ‘brothers’ are often their closest social network, in part because they live together, take classes together, and organize social events together (Murnen & Kohlman, 2007). The intense process of joining a fraternity, called pledging, secures these attachments. In addition, fraternities have a secretive nature that emphasizes loyalty to the group (Adam-Curtis & Forbes, 2004).

The Current Paper

Although Developmental Systems Theory and the Male Peer Support Model provide explanations for both selection and socialization effects of fraternity membership on traditional masculinity ideologies and impersonal sex (Godenzi et al., 2001; Gottlieb, 1996; Schwartz, DeKeseredy, Tait, & Alvi 2001), there is little empirical evidence to indicate whether sexually aggressive college men select into fraternities or if fraternity membership has a socializing effect on college men (Adam-Curtis & Forbes, 2004; Godenzi et al., 2001; Koss & Cleveland, 1996; Seabrook et al., 2018). Previous scholars have called for longitudinal research to explore selection and socialization effects of fraternity membership and masculinity ideologies (Adam-Curtis & Forbes, 2004; Koss & Cleveland, 1996; Murnen & Kohlman, 2007; Seabrook et al., 2018). To separate selection and socialization effects of fraternity membership, longitudinal data are critical. Using longitudinal data, one can examine selection effects by measuring the association of traditional masculinity ideologies and impersonal sex with college men’s subsequent fraternity membership. In addition, one can examine socialization effects by measuring how fraternity membership is associated with subsequent changes in traditional masculinity ideologies and impersonal sex.

The current paper is one of the first papers to examine both selection and socialization effects of fraternity membership on traditional masculinity ideologies and impersonal sex (Malamuth et al., 1991; Malamuth et al., 1996; Thompson & Bennett, 2017) using longitudinal data. Using secondary data from two separate longitudinal studies, we explored five indicators of traditional masculinity ideologies in dataset one (male role norms about status, toughness, and antifemininity, masculine traits, and endorsement of the sexual double standard) and four indicators of impersonal sex in dataset two (sex motives for intimacy, peer pressure, and coping, and multiple sex partners). Developmental Systems Theory and the Male Peer Support Model work together to provide rationale for both selection and socialization effects. Thus, we hypothesize that results will indicate both selection and socialization effects, in that traditional masculinity ideologies and impersonal sex at the start of college will be positively associated with subsequent fraternity membership, and that fraternity membership will be positively associated with subsequent traditional masculinity ideologies and impersonal sex. The current research contributes to theory on selection and socialization in this area. In addition, our findings have practical implications for the timing of interventions to change the fraternity culture which scholars believe contributes to sexual aggression (Sanday, 1996).

Method

Dataset One: Traditional Masculinity Ideologies

Participants

Participants were new, full-time, traditionally-aged first-year college students recruited via the university Registrar in the fall of their first year to enroll in a study on sexuality and gender (Lefkowitz, Shearer, Gillen, & Espinosa-Hernandez, 2014). The university was a large state university in the Northeast, with an undergraduate population of over 40,000 undergraduate students (approximately 47% female), and a strong fraternity presence on campus (over 10% of the students and over 40 fraternities). The culture of the fraternities on campus was consistent with the culture described in the introduction. The sample draws from a larger study of gender and HIV risk among male and female college students that took about 45 minutes to complete at each wave. Participants received initial emails about the study. To obtain a diverse sample, we invited all African American and Latino American first-year students to participate, as well as nine percent of European American first-year students (by random sample). The response rate was 52%, yielding a Semester 1 sample of 434 students.

Participants responded to surveys for four semesters beginning in the Fall of their first year. The current paper used data from the first three consecutive semesters, Semesters 1 (S1), 2 (S2), and 3 (S3) due to the availability of relevant constructs. We excluded women in the current paper (n = 226). In addition, we excluded men who reported being in a fraternity at S1 (n = 12), because we were unable to measure their beliefs and behaviors before fraternity membership. Additionally, we excluded participants for whom we could not determine their fraternity membership status because of missing data (i.e., they dropped out of the study; n = 29), and one participant missing all masculinity ideologies indicators at S3. Thus, the analytic sample was 166 college men, aged 17–19 at S1 (M = 18.0 years, SD = 0.4).

Participants could identify as more than one race/ethnicity; thus, the sample was 49.4% White/European American, 31.9% Black/African American, 28.3% Hispanic/Latino American, and 1.2% Asian/Asian American/Hawaiian/Pacific Islander. We conducted five t-tests and three Chi-squares to determine whether men in the analytic sample (n = 166) differed from men not in the analytic sample (n = 42) on S1 variables. Participants in the analytic sample were significantly older (t(206) = 2.17, p < .05) than participants not in the analytic sample. Groups did not differ on race/ethnicity, male role norms, masculine traits, or endorsement of the sexual double standard.

Procedures

Participants signed consent forms before completing the questionnaire at S1. They completed paper questionnaires each semester in centrally located classrooms on campus. Participants received an email and call with the study time and location. They received $25 at S1, $30 at S2, and $35 at S3. Participants were assigned a random ID number and the master list of ID numbers and corresponding contact information was stored separately from the data. The study was approved by the university’s Intuitional Review Board (IRB).

Measures

Fraternity membership by S2

At S1 and S2, participants responded to the question, “Have you joined any clubs, organizations, teams, or fraternities/sororities on campus?” and responded yes or no. Participants then responded to the question, “If yes, please list all of them.” Undergraduate coders recorded if participants listed a social fraternity at each semester. Coders consulted a list of all university-affiliated organizations on the university’s website to determine whether an organization was a social fraternity/sorority or an academic/professional organization. We coded participants who listed a social fraternity in S2 as in a social fraternity (= 1) and participants who did not list a social fraternity as not in a social fraternity (= 0).

Male role norms about status, toughness, and antifemininity

At S1 and S3, participants rated their agreement with a series of statements about male role norms on a scale from Strongly disagree (= 1) to Strongly agree (= 7; Male Roles Attitudes Scale; Thompson & Pleck, 1986). The male role norms about status scale included 11 statements such as “It is essential for a man to always have the respect and admiration of everyone who knows him” and “I always like a man who’s totally sure of himself.” The male role norms about toughness scale included eight statements such as “A real man enjoys a bit of danger now and then” and “Fists are sometimes the only way to get out a bad situation.” The male role norms about antifemininity scale included seven statements such as “It bothers me when a man does something that I consider ‘feminine’” and “It is a bit embarrassing for a man to have a job that is usually filled by a woman.” Reliability (Cronbach’s α) at S1/S3 was .85/.87 for status, 75/.79 for toughness, and .82/.84 for antifemininity. For each sub-scale, we created a mean of all items after reverse coding appropriate items. Previous research using this scale demonstrates its validity given its association with attitudes toward gender roles, relationships, homosexuality, and with traditional male procreative attitudes (Lefkowitz et al., 2014; Pleck, Sonenstein, & Ku, 1994).

Masculine traits

At S1 and S3, participants responded to the prompt, “Please indicate how well these adjectives apply to your personality by circling the appropriate number” on a scale from Never or almost never true (= 1) to Always or almost always true (= 7; Bem Sex Roles Inventory; Bem, 1974). The scale was the mean of responses on ten adjectives that described traditionally male personality traits such as “Dominant” and “Forceful.” Reliability at S1/S3 was .81/.80. Regarding validity, this scale is associated with several other similar measures of masculinity (Bem, 1974).

Endorsement of sexual double standard

At S1 and S3, participants responded to a shortened version (Lefkowitz et al., 2014) of the Sexual Double Standard Scale (Muehlenhard & Quackenbush, 1996), which measures endorsement of a double standard allowing men more sexual freedom than women. Participants responded to 17 items such as “It’s worse for a woman to sleep around than it is for a man” and “A man should be more sexually experienced than his wife” on a scale from strongly disagree (= 1) to strongly agree (= 4). Reliability at S1/S3 was.71/.72. We calculated the final sum based on original scoring guidelines, which included reverse coding some items and subtracting others. Regarding validity, this scale is associated with other gender role attitudes and with gendered-stereotyped sexual beliefs (Lefkowitz et al., 2014).

Dataset Two: Impersonal Sex

Participants

We used data from the University Life Study, a longitudinal study of college students at the same university (Patrick, Maggs, & Lefkowitz, 2015). The sample is from a larger study of sexual behavior and alcohol use among male and female college students that took about 35 minutes to complete at each wave. Participants were first-time, traditionally-aged college students who responded to online surveys for seven consecutive semesters beginning in Fall of their first year. They initially received emails and mailings with information about the study. Eligible students were U.S. citizens or permanent residents under 21 years of age who lived within 25 miles of campus. We used a stratified sampling procedure at S1 to achieve a diverse sample of first-year students with respect to gender and four major race/ethnicity groups. The response rate was 65.5%, yielding a S1 sample of 744 participants.

In the current paper, we used data from S1, S2, and S3 (due to the availability of relevant constructs). We excluded women (n = 378) and men who reported fraternity membership at S1 (n = 38). Additionally, we excluded participants for whom we could not determine fraternity membership status because of missing data (i.e., they dropped out of the study; n = 72). Thus, the analytic sample was 256 college men, aged 17.6 to 19.8 at S1 (M = 18.5 years, SD = 0.4).

Participants could identify as more than one race/ethnicity; thus, the sample was 47.3% White/European American, 32.4% Asian/Asian American/Hawaiian/Pacific Islander, 25.0% Hispanic/Latino American, and 15.2% Black/African American. We conducted three t-tests and four Chi-squares to determine whether participants in the analytic sample (n = 256) differed from men not in the analytic sample (n = 110) on S1 variables. Participants in the analytic sample were younger (t(364) = −2.56, p < .05) than men not in the analytic sample. Groups did not differ on sex motives, multiple sex partners, or race/ethnicity.

Procedures

Students received a mailed informational letter with a $5 pre-incentive and a pen. Subsequently they received an email message with a link to the first Web-based baseline survey. If students did not respond to the email, we also contacted students by phone. Students earned $20-$30 for completing each of the S1, S2, and S3 online baseline surveys. Participants were assigned a random ID number and the master list of ID numbers and corresponding contact information was stored separately from the data. The study received IRB approval and a Certificate of Confidentiality from the National Institutes of Health.

Measures.

Fraternity membership by S2

At S1 and S2, participants responded to the question, “What extracurricular activities do you participate in? Check all that apply.” There were 17 options including “Fraternity/Sorority (social).” We coded participants who selected social fraternity in S2 as in a social fraternity (= 1) and participants who did not select social fraternity as not in a social fraternity (= 0).

Sex motives

At S1 and S3, participants responded to a series of questions about their motives for sex on a scale from Almost never/never (= 1) to Almost always/always (= 5). In the current study, we used the means of three of these sex motive sub-scales, each with five questions: intimacy (e.g., “How often do you have sex to express love for your partner?” and “How often do you have sex to make an emotional connection with your partner?”); peer pressure (e.g., “How often do you have sex because you worry that people will talk about you if you don’t have sex?” and “How often do you have sex because others will kid you if you don’t?”); and coping (e.g., “How often do you have sex to cope with upset feelings?” and “How often do you have sex to cheer yourself up?”). Prior research demonstrates validity of the scales (Cooper, Shapiro, & Powers, 1998). Reliability at S1/S3 was .95/.95 for intimacy, .89/.94 for peer pressure, and .94/.95 for coping. Previous research using this scale demonstrates its validity given its association with relevant constructs such as emotional responses to sex and unrestricted orientation toward sex (Cooper et al., 1998).

Multiple sex partners

At S1 and S3, participants responded to the question, “In the past 12 weeks, have you had vaginal and/or anal sex with a partner?” yes (= 1) or no (= 0). Participants who responded yes were then asked, “In the past 12 weeks, how many different female partners have you had vaginal and/or anal sex with?” and, “In the past 12 weeks, how many different male partners have you had vaginal and/or anal sex with?” We summed these two responses to determine how many sex partners participants had in the past 12 weeks. We coded participants who had not had vaginal and/or anal sex in the past 12 weeks as having zero partners. We chose past 12 weeks as has been recommended in previous research (Stone, 1999). We dichotomized this variable into two or more sex partners (= 1) and zero or one sex partners (= 0; Fromme et al., 2008).

Analysis Plan

To test for selection effects, we performed eight logistic regression analyses. Using dataset one, we performed five logistic regressions predicting S2 fraternity membership from the five indicators of traditional masculinity ideologies at S1 (three male role norms subscales, masculine traits, and endorsement of the sexual double standard). Using dataset two, we performed four logistic regressions predicting S2 fraternity membership from the four indicators of impersonal sex at S1 (three sex motive scales and multiple sex partners).

To test for socialization effects, we performed seven multiple regression analyses and one logistic regression. We assessed change in each indicator of masculinity ideologies or impersonal sex by controlling for the indicator at S1. Using dataset one, we performed separate multiple regression analyses predicting change in each indicator of traditional masculinity ideologies from fraternity membership at S2. Using dataset two, we predicted change in each indicator of impersonal sex from fraternity membership with multiple regression analyses for each of the three sex motives, and a logistic regression for the dichotomous measure of multiple sex partners.

We controlled for S1 age and race/ethnicity in all analyses. In the analyses using dataset one, we controlled for Black/African American and Hispanic/Latino American. We could not control for Asian/Asian American because there were so few Asian/Asian American participants in this dataset. Thus, the reference group was non-Hispanic, non-multiracial White/European and Asian/Asian American. In the analyses using dataset two, we controlled for Asian/Asian American, Black/African American, and Hispanic/Latino American with non-Hispanic, non-multiracial White/European American as the reference group. In both datasets, we chose the largest group as the reference group. We did not control for fraternity membership at S1 because participants who reported fraternity membership at S1 were removed from analysis.

Results

Dataset One: Traditional Masculinity Ideologies

We present bivariate correlations between indicators of traditional masculinity ideologies in Table 1, and descriptive statistics and bivariate tests of masculinity ideologies and fraternity membership for dataset one in Table 2. Approximately 9.6% of college men (n = 16) in dataset one joined a fraternity by S2. According to bivariate tests, men who joined a fraternity by S2 had endorsed the sexual double standard more at S1 then men who had not joined a fraternity. Men who joined a fraternity by S2 subsequently had more traditional male role norms about status, more traditional beliefs in the sexual double standard, and had more masculine traits at S3 than men who did not join a fraternity.

Table 1.

Dataset One: Bivariate correlations between indicators of traditional masculinity ideologies (ns = 164 to 166)

	S1 MRN status	S1 MRN toughness	S1 MRN anti-femininity	S1 Masculine traits	S1 Sexual double standard	S3 MRN status	S3 MRN toughness	S3 MRN anti-femininity	S3 Masculine traits
S1 MRN toughness	.36***
S1 MRN anti-femininity	.30***	.48***
S1 Masculine traits	.13+	.25**	.15*
S1 Sexual double standard	.37***	.33***	.431***	.12
S3 MRN status	.74***	.32***	.30***	.19*	.40***
S3 MRN toughness	.33***	.68***	.44***	.25**	.36***	.49***
S3 MRN anti-femininity	.30***	.43***	.73***	.18*	.40***	.44***	.58***
S3 Masculine traits	.16*	.19*	.13	.73***	.11	.24**	.32***	.18*
S3 Sexual double standard	.29**	.25**	.37***	.17*	.61***	.40***	.30***	.49***	.14+

Note. MRN = Male role norms.

⁺p < .10

^*p < .05.

^**p < .01.

^***p <.001.

Table 2.

Descriptive statistics

	Semester 1 Beliefs			Semester 3 Beliefs
	S2 Fraternity members	S2 Not fraternity members		S2 Fraternity members	S2 Not fraternity members
Dataset One (n = 166)	M (SD)	M (SD)	ta	M (SD)	M (SD)	ta
Male role norms: Status	4.61 (1.11)	4.18 (1.07)	0.13	4.79 (1.14)	4.19 (1.07)	2.12*
Male role norms: Toughness	3.80 (0.95)	3.94 (0.97)	0.57	4.22 (1.03)	4.22 (0.96)	0.03
Male role norms: Antifemininity	3.80 (1.06)	3.59 (1.12)	0.72	3.88 (0.99)	3.66 (1.18)	0.73
Masculine traits	5.41 (0.78)	5.08 (0.75)	1.63	5.39 (0.76)	5.03 (0.68)	2.01*
Sexual double standard	16.78 (4.81)	13.70 (4.91)	2.39*	16.62 (6.35)	13.67 (5.07)	2.15*
Dataset Two (n = 256)
Sex motives: Intimacy	2.38 (0.93)	2.10 (1.40)	1.12	2.05 (1.04)	2.02 (1.33)	0.12
Sex motives: Peer pressure	0.83 (0.87)	0.61 (0.83)	1.42	1.10 (1.03)	0.78 (1.08)	1.16
Sex motives: Coping	0.61 (0.84)	0.37 (0.72)	1.79	0.82 (0.98)	0.45 (0.82)	2.42*
	%	%	χ2b	%	%	χ2b
Two or more sexual partners	6.76	14.17	2.56	11.26	20.59	.01

Note.

^aIndependent samples t-test examining the the difference between fraternity and non-fraternity members.

^bChi-square test examining the difference between fraternity and non-fraternity members. None of the variables were skewed beyond a skewness statistic of +/− 2 (George & Mallery, 2010).

^*p < .05.

In step one of the logistic regression model testing selection effects for fraternity membership (Table 3), we entered all control variables. As indicated by the Cox & Snell R² (Cox & Snell, 1989), this step was not significant, meaning that the control variables did not explain a significant amount of variance in fraternity membership. The variance explained in step two of this model, in which we entered all masculinity ideologies indicators, was significant. Traditional masculinity ideologies at S1 were associated with subsequent fraternity membership. Specifically, men who endorsed male role norms about status more strongly at S1 were more likely to join a fraternity by their second semester than other men. Each one-unit increase (on a seven-point scale) in male role norms about status was associated with a 105% greater, or more than double, likelihood of joining a fraternity. Men who endorsed male role norms about toughness more strongly were less likely to join a fraternity by their second semester than other men. Each one-unit increase (on a seven-point scale) in male role norms about toughness was associated with a 68% lower likelihood of joining a fraternity. Men who endorsed the sexual double standard more strongly in their first semester of college were more likely to join a fraternity by their second semester than other men. Each one-unit increase (on a 51-point scale) in endorsement of the sexual double standard was associated with a 20% greater likelihood of joining a fraternity. In addition, the term for masculine traits approached significance (p < .10). More masculine men were more likely to join a fraternity than less masculine men. Each one-unit increase (on a seven-point scale) in masculine traits was associated with a 128% greater likelihood of joining a fraternity.

Table 3.

Dataset One: Logistic regression analysis predicting fraternity membership by S2 (selection effects; n = 166)

	Step 1		Step 2
	OR	95% C.I.	OR	95% C.I.
Age	0.73	(0.18, 2.93)	0.75	(0.17, 3.29)
Black/African American	0.61	(0.19, 1.91)	0.26+	(0.06, 1.08)
Latino/Hispanic American	0.13+	(0.02, 1.06)	0.04*	(0.00, 0.43)
S1 Male role norms: Status			2.05*	(1.05, 4.00)
S1 Male role norms: Toughness			0.32**	(0.14, 0.74)
S1 Male role norms: Antifemininity			1.07	(.57, 1.99)
S1 Masculine traits			2.28+	(0.98, 5.34)
S1 Sexual double standard			1.20*	(1.04, 1.39)
Step Chi-square			18.59**
Model Chi-square	5.80		24.40**
Cox & Snell R2	.04		.14**

Note.

⁺p < .10

^*p < .05.

^**p < .01.

^***p < .001.

In step one of the regression models testing socialization effects of fraternity membership (Table 4), we entered all control variables. Across models, this step was significant, meaning that the control variables explained a significant amount of variance in masculinity ideologies. The change in R² of step two, in which we entered fraternity membership, was not significant in any of the models. Thus, we did not find evidence of socialization effects of fraternities on indicators of traditional masculinity ideologies.

Table 4.

Dataset One: Multiple regression analyses predicting S3 traditional masculinity ideologies (socialization effects; ns = 163 to 165)

	Male role norms: Status		Male role norms: Toughness		Male role norms: Antifemininity		Masculine traits		Sexual double standard
	Step 1	Step 2	Step 1	Step 2	Step 1	Step 2	Step 1	Step 2	Step 1	Step 2
	β	β	β	β	β	β	β	β	β	β
Age	−.04	−.04	.03	.03	−.01	−.01	.09+	.09+	−.02	−.02
Black/African American	.06	.07	.03	.04	.04	.04	.09	.09	.10	.10
Latino/Hispanic American	−.02	−.01	−.02	−.01	−.03	−.02	−.03	−.01	.07	.08
S1 Indicator of masculinity ideologies	.73***	.71***	.67***	.67***	.73***	.73***	.73***	.72***	.61***	.60***
Fraternity membership by S2		.08		.03		.01		.06		.07
R2	.55***	.56***	.46***	.46***	.53***	.53***	.56***	.56***	.37***	.37***
Δ R2 from Step 1 to Step 2		.01		.00		.00		.00		.00

Note.

⁺p < .10

^***p < .001.

Dataset Two: Impersonal Sex

We present bivariate correlations between indicators of impersonal sex in Table 5, and descriptive statistics and bivariate tests of impersonal sex and fraternity membership for dataset two in Table 2. Approximately 13.3% (n = 34) of college men in dataset two joined a fraternity by S2. Men who joined a fraternity by S2 were more motivated to use sex to cope at S3 than men who did not join fraternities.

Table 5.

Dataset Two: Bivariate correlations between indicators of impersonal sex (ns = 251 to 256)

	S1 SM1 for intimacy	S1 SM for peer pressure	S1 SM for coping	S1 Multiple sex partners	S3 SM for intimacy	S3 SM for peer pressure	S3 SM for coping
S1 SM for peer pressure	.39***
S1 SM for coping	.23***	.64***
S1 Multiple sex partners	−.01	.12	.06
S3 SM for intimacy	.53***	.22***	.08	−.02
S3 SM for peer pressure	.31***	.53***	.39***	.01	.37***
S3 SM for coping	.16*	.35***	.49***	.02	.21**	.65***
S3 Multiple sex partners	−.03	.20**	.04	.13	−.10	.10	−.03

Note.

¹SM = Sexual motives.

^*p < .05.

^**p < .01.

^***p <.001.

Step one of the logistic regression model testing selection effects for fraternity membership (Table 6), in which we entered all control variables, was not significant. That is, the control variables did not explain a significant amount of variance in fraternity membership. The variance explained by step two of the models, in which we entered all indicators of impersonal sex, was not significant. That is, impersonal sex at S1 was not associated with subsequent fraternity membership. Thus, we did not find evidence of selection effects for impersonal sex.

Table 6.

Dataset Two: Logistic regression analysis predicting fraternity membership by S2 (selection effects; n = 256)

	Step 1		Step 2
	OR	95% C.I. (lower, upper)	OR	95% C.I. (lower, upper)
Age	1.02	(0.43, 2.46)	0.99	(0.41, 2.42)
Asian/Asian American	1.19	(0.47, 3.01)	1.22	(0.47, 3.19)
Black/African American	1.02	(0.29, 3.55)	0.97	(0.27, 3.53)
Hispanic/Latino American	2.55*	(1.05, 6.17)	2.35+	(0.94, 5.85)
S1 Sex motives: Intimacy			1.11	(0.81, 1.54)
S1 Sex motives: Peer pressure			1.00	(0.56, 1.79)
S1 Sex motives: Coping			1.33	(0.75, 2.36)
S1 Multiple sex partners			2.19	(0.70, 6.86)
Step Chi-square			4.41
Model Chi-square	5.07		9.49
Cox & Snell R2	.02		.04

In step one of the regression models and logistic regression model testing socialization effects of fraternity membership (Table 7), we entered all control variables. Across models, this step was significant, meaning that the control variables explained a significant amount of variance in impersonal sex. The variance explained in step two, in which we entered fraternity membership, approached significance (p < .10) in the model that examined sex motives to cope. Specifically, men who joined fraternities were more motivated to use sex to cope than men who did not join fraternities. Men who joined fraternities did not differ from other men on sexual motives for intimacy, sexual motives for peer pressure, or multiple sex partners.

Table 7.

Dataset Two: Multiple regression and logistic regression analyses predicting S3 indicators of impersonal sex (socialization effects; n = 256)

	Sex motives: Intimacy		Sex motives: Peer pressure		Sex motives: Coping
	Step 1	Step 2	Step 1	Step 2	Step 1	Step 2
	β	β	β	Β	β	β
Age	−.04	−.04	−.06	−.06	−.05	−.05
Asian/Asian American	−.06	−.06	.03	.03	.12+	.12+
Black/African American	−.02	−.02	.01	.01	.06	.06
Hispanic/Latino American	−.05	−.04	.06	.05	.00	−.01
S1 Indicator of impersonal sex	.53***	.53***	.52***	.51***	.48***	.47***
Fraternity membership by S2		−.03		.05		.11+
R2	.29***	.29***	.28***	.29***	.25***	.26***
Δ R2 from Step 1 to Step 2		.00		.00		.01*
Multiple sex partners
	Step 1		Step 2
	OR	95% C.I. (lower, upper)	OR	95% C.I. (lower, upper)
Age	0.48	(0.16, 1.43)	0.48	(0.16, 1.43)
Asian/Asian American	3.07+	(0.97, 9.72)	3.07+	(0.97, 9.80)
Black/African American	4.28*	(1.32, 13.86)	4.28*	(1.32, 13.87)
Hispanic/Latino American	1.71	(0.59, 4.97)	1.72	(0.59, 5.01)
S1 Multiple sex partners	1.99	(0.68, 5.85)	1.99	(0.68, 5.86)
Fraternity membership by S2			1.31	(0.36, 2.68)
Step Chi-square			0.01
Model Chi-square	12.16*		12.16+
Cox & Snell R2	.08		.08

Note.

⁺p < .10

^*p < .05.

^**p < .01.

^***p < .001.

Discussion

In the current paper, we used longitudinal data to test both selection and socialization effects of fraternity membership on masculinity ideologies and impersonal sex. The current paper responds to calls for a more detailed understanding of the social processes by which fraternity membership contributes to the traditional masculinity ideologies that are associated with sexual aggression (Adam-Curtis & Forbes, 2004; Harway & Steel, 2015; Koss & Cleveland, 1996; Murnen & Kohlman, 2007; Sanday, 1996). In our multivariate analyses, we found that men with more traditional beliefs about men’s roles were more likely to join fraternities than other men, suggesting that men with these beliefs tend to select into fraternities. We found much less evidence for a socialization effect of fraternities in multivariate analyses. Thus, although both selection and socialization effects of fraternity membership may co-exist, we found stronger evidence for selection effects. The current paper both contributes to theory on selection and socialization and provides implications for prevention.

Traditional Masculinity Ideologies

Men who at the start of college more strongly endorsed male role norms about status, or the belief that men should strive for high social status, were more likely to join fraternities. This finding is consistent with previous literature describing the high social status of fraternity men on campus. Although a minority of college men are in fraternities, fraternities enjoy a high degree of visibility in the college social scene through class privilege, ability to hold parties, and ability to provide alcohol to underage students (Armstrong & Hamilton, 2013; Jozkowski & Wiersma-Mosley, 2017). Men who endorse the belief that men should strive for high social status may be more attracted to this visibility and power than other men. Because men who select into fraternities hold these social beliefs, the ways in which fraternities appear to control social scripts about women’s behavior, dress, and alcohol consumption may appeal to these men; for example, women attending parties are expected to dress in scant clothing, be very friendly to men, and drink heavily (Harris & Schmalz, 2016). Scholars attribute higher rates of sexual aggression among fraternity men in part to these social scripts (Armstrong, Hamilton, & Sweeney, 2006; Harris & Schmalz, 2016; Jozkowski & Wiersma-Mosley, 2017; Sweeney, 2014). Men who select into fraternities to achieve high status on campus are unlikely to disrupt the male-controlled social scripts that contribute to sexual aggression.

Contrary to previous research describing hypermasculine values among fraternity men (Godenzi et al., 2001; Schwartz et al., 2001), in multivariate analyses we found that men who joined fraternities endorsed male role norms about toughness less strongly than men who didn’t join fraternities. However, this finding should be interpreted with caution. At the start of college, the means of male role norms about toughness for fraternity members and nonmembers were similar. It is possible that, because male role norms about toughness are associated with the other indicators of masculinity ideologies, this finding reflects a statistical suppression effect. Suppression occurs when including particular independent variables in an analysis improves the amount of variance explained by another independent variable. This effect is due to correlations between the independent variables in the analysis (Tzelgov & Henik, 1991). That is, after accounting for the shared variance with the other three indicators of traditional masculinity ideologies, the unique variance of male role norms about toughness became a significant predictor. Future work should attempt to replicate this finding before interpretation.

We also found that men who more strongly endorsed the sexual double standard during the first semester of college were more likely to join a fraternity than men who less strongly endorsed the sexual double standard. The sexual double standard is the belief that men and women should be held to different standards for sexual behavior (Kreager, Staff, Gauthier, Lefkowitz, & Feinberg, 2016). According to several qualitative studies of fraternity and sorority members, students in Greek life frequently describe the sexual double standard as part of Greek culture. Women involved in Greek culture are expected to exhibit sexual attraction for men, such as dressing provocatively for parties, but are stigmatized for overt sexual behavior (Armstrong et al., 2006; Armstrong & Hamilton, 2013; Jozkowski & Wiersma-Mosley, 2017; Sanday, 1996; Sweeney, 2014). Meanwhile, within Greek culture men are viewed as always wanting sex (Armstrong et al., 2006; Armstrong & Hamilton, 2013; Jozkowski & Wiersma-Mosley, 2017; Sweeney, 2014). These gendered beliefs create an environment in which men’s sexual aggression is normalized and women are blamed for being victims of sexual aggression (Armstrong et al., 2006). Our findings, taken together with previous qualitative studies, suggest that men select into the fraternity context because this context is consistent with their pre-existing beliefs, and may contribute to the continuation of these beliefs in their fraternity over time. These findings are congruent with Developmental Systems Theory, which proposes that individuals select into environments consistent with their pre-existing beliefs and behaviors (Gottlieb, 1996).

Impersonal Sex

We found that men who joined fraternities became more motivated to use sex to cope with negative emotional states than men who did not join fraternities. This finding, which approached significance, suggests that fraternities may have a socialization effect on men’s sex motives to cope. The effect for this finding was small. However, in light of previous research on using sex to cope, future research should attempt to replicate this finding. Previous research has found that sexually aggressive men often use sex to alleviate negative emotions (Cortoni & Marshall, 2001), which may indicate an underlying difficulty in handling emotional distress, such as a lack of other means to improve mood (Ward & Beech, 2006). That is, men who lack the skills to handle negative emotions may use sex and sexual aggression to meet emotional needs (Ward & Beech, 2006). The transition to college can be difficult for all students, who during this transition commonly experience negative emotions such as loneliness (Cleary, Walter, & Jackson, 2011). Because the college transition is stressful, coping skills become particularly salient during this time. Fraternity culture, which emphasizes impersonal sex (Harris & Schmalz, 2016; Murnen & Kohlman, 2007), may push men toward coping with stress by having sex instead of engaging in healthier coping methods. This finding is consistent with the Male Peer Support Model (Godenzi et al., 2001; Schwartz et al., 2001), which states that men with all-male social networks may adopt traditional masculinity ideologies. In turn, using sex to cope may be a factor in the higher rates of sexual aggression among fraternity membership, compared to other college men (Ward & Beech, 2006).

Socialization Effects

Although the multivariate analyses provided stronger evidence for selection effects than socialization effects, the bivariate analyses indicated that men who joined fraternities subsequently differed from men who did not join in their traditional masculinity ideologies and impersonal sex. The multivariate socialization analyses controlled for indicators of previous traditional masculinity ideologies and impersonal sex, whereas the multivariate selection analyses did not. Given stability of measures, it may have been easier to detect selection than socialization in multivariate models.

Limitations and Future Research Directions

The current paper has several limitations that indicate areas for future research. First, we used two different, pre-existing datasets. An advantage of this strategy is that we could test for selection and socialization effects using longitudinal data. In addition, we could explore indicators of both traditional masculinity ideologies and impersonal sex. However, a disadvantage of this strategy was that fraternity membership was measured differently in the two datasets. In dataset one, participants were asked to list their campus activities, and coders determined whether participants listed a social fraternity. In dataset two, participants chose from a list of campus activities that included social fraternities. Dataset two had a higher rate of fraternity membership than dataset one and these different measurement strategies may have contributed to the different rates of fraternity membership. It is possible that some participants in dataset one forgot to list fraternity participation or did not consider their fraternity an on-campus activity. Because dataset one had a lower rate of fraternity membership (only 17 of 166 men joined a fraternity by S2), we may have been underpowered to find significant effects. We believe the measurement strategy used in dataset two was stronger and suggest that future research use that strategy when measuring fraternity participation. Second, power may also have impacted the fact that in dataset two, men with multiple sex partners were more than twice as likely to join fraternities than other men, but this difference did not reach statistical significance. Third, because this data was collected in participants’ second semester at the university, it is unlikely that they lived in fraternity houses yet. Socialization effects may be stronger after participants have been living in fraternity houses. Fourth, our samples differed from each other in terms of race/ethnicity. Specifically, we oversampled for Asian Americans in dataset two but not in dataset one. Future research could replicate our results with different populations to strengthen the literature on selection and socialization. Finally, we explored traditional masculinity ideologies and impersonal sex. Future researchers could explore selection and socialization effects of fraternities on sexually aggressive behavior.

Practice Implications

Despite these limitations, findings have potential implications for practice. Overall, we found that men with more traditional masculinity ideologies select into a fraternity environment consistent with their pre-existing beliefs. Because fraternity members are more likely to perpetrate sexual aggression than other college men (Adams-Curtis & Forbes, 2004; Tharp et al., 2012), fraternity members are often targeted for psychosocial interventions that aim to change beliefs and behaviors associated with sexual aggression (Vladutiu, Martin, & Macy, 2011). In addition, men intending to join a fraternity are likely at risk (Seabrook et al., 2018). Universities could target these men for prevention before they join, either in person at first-year orientations or using online programs.

We found less evidence that fraternity membership had a socialization effect on traditional masculinity ideologies. According to our results, it is possible that men develop these beliefs before college, and that these beliefs remain stable over time despite fraternity membership. There is limited discussion in the field about addressing college sexual assault through early prevention (Schewe, 2007). In addition to interventions for college students, early interventions that aim to help boys develop healthier beliefs about gender and sex during middle and high school may also be effective strategies for preventing college sexual aggression (DeGue et al., 2014). In the current study, beliefs remained stable over time. In contrast, at younger ages, boys are still developing these beliefs, and beliefs may be more malleable than college students’ beliefs. Thus, programs aimed at younger boys may be more developmentally appropriate. The purpose of developmentally appropriate prevention, or prevention that is properly timed in the course of development, is to change individuals’ pattern of beliefs and behaviors that lead to undesired outcomes (Nation et al., 2003). Early prevention is a promising strategy. For example, program effects of the Safe Dates program to prevent dating violence can last for at least four years (Foshee et al., 2004). In addition, some early intervention programs target gender role beliefs. For example, the Coaching Boys into Men program aims to promote gender equity and reduce dating violence (Miller et al., 2012). We suggest that future research complete long-term evaluation of early prevention programs on sexual aggression through the college years. These early interventions may be integrated into comprehensive sexuality education and may occur through schools, parents, or tech-based sources like websites or applications.

Intervening at the individual level may not be sufficient to produce change. It is important not to ignore broader societal issues that contribute to fraternity members’ individual development of their beliefs about gender and sexuality (Tharp et al., 2012; Ward, 2003). Our results suggest that gendered beliefs are established prior to college, reflecting a lifetime process of socialization about men and women’s roles. This process occurs through influence from family, friends, media, and other socialization forces (e.g., Epstein & Ward, 2008; Ward, 2003). Preventing rape-supportive attitudes, therefore, requires cultural shifts in the messages young people receive about gender.

Conclusions

The current paper explored selection and socialization effects of fraternity membership on indicators of traditional masculinity ideologies. In addition, the current paper explored one way that men may enact masculinity ideologies, impersonal sex. Although theory provides explanations for both selection and socialization effects of fraternity membership, there is little empirical evidence to indicate whether college men select into fraternities or if fraternity membership has a socializing effect on college men (Adam-Curtis & Forbes, 2004; Godenzi et al., 2001; Gottlieb, 1996; Koss & Cleveland, 1996). We found evidence that men with more traditional masculinity ideologies tend to select into fraternities. We suggest that pre-college interventions facilitate boys’ healthy and egalitarian beliefs about gender and sex.

Public Significance Statement.

Findings indicate that men who more strongly endorsed male role norms about status and the sexual double standard may self-select into fraternities. These findings indicate a need for early intervention on traditional masculinity ideologies. Intervention work with younger boys may be better suited to encourage healthier beliefs about gender and sex during middle and high school development and provide an effective strategy for preventing college sexual aggression.