What’s Sex (Composition) Got to Do with It? The Importance of Sex Composition of Gangs for Female and Male Members’ Offending and Victimization

Abstract

Sex composition of groups has been theorized in organizational sociology and found in prior work to structure female and male members’ behaviors and experiences. Peer group and gang literature similarly finds that the sex gap in offending varies across groups of differing sex ratios. Drawing on this and other research linking gang membership, offending, and victimization, we examine whether sex composition of gangs is linked to sex differences in offending in this sample, further assess whether sex composition similarly structures females’ and males’ victimization experiences, and if so, why. Self-report data from gang members in a multi-site, longitudinal study of 3,820 youths are employed. Results support previous findings about variations in member delinquency by both sex and sex composition of the gang and also indicate parallel variations in members’ victimization. These results are further considered within the context of facilitating effects such as gender dynamics, gang characteristics, and normative orientation.

Keywords:

• gangs
• sex composition
• sex ratio
• gender dynamics
• offending
• violence
• delinquency
• victimization
• victim–offender overlap

Acknowledgements

A prior version of this paper was presented at the 2012 American Society of Criminology meeting, Chicago, IL. We thank Cheryl Maxson, Finn Esbensen, and the anonymous reviewers for their insightful feedback on earlier drafts of this paper, and of course, accept all responsibility for any errors or omissions.

Notes

¹ Lauritsen and Laub (2007, p. 66) also assert that the study of victimization is often de-coupled from the study of offending because examining the victim–offender overlap contradicts conventional notions of victims as “good” and offenders as “bad”; this is likely, we argue, to be especially pertinent to the study of gang members as victims.

² An important parallel body of research also shows that composition (including sex ratio) and structure of adolescent friendship groups/networks are important for females’ and males’ deviance (Giordano, 1978; Giordano, Cernkovich, & Pugh, 1986; Haynie, Steffensmeier, & Bell, 2007; Kreager, Rulison, & Moody, 2011; McCarthy, Felmlee, & Hagan, 2004; Schwartz, Conover-Williams, & Clemons, 2015; Steffensmeier, 1983; Warr, 2002; Weerman & Bijleveld, 2007).

³ Some debate has been introduced by a few studies concluding that there is no unique contribution of gang membership to victimization (e.g. Gibson, Miller, Jennings, Swatt, & Gover, 2009; Katz, Webb, Fox, & Shaffer, 2011; Spano, Freilich, & Bolland, 2008). Although questions have been raised about this work (see Ozer and Engel’s 2011 critique of Gibson et al., 2009), we acknowledge this debate and note that our goal in this literature review and current study is not to assert a unique effect of gang membership on victimization, but to assert that gang members’ patterns of involvement in risky and delinquent behavior may be related to differential patterns of victimization (by sex and gang composition). That is, we do not compare gang to non-gang members, but rather gang members in different gang types to each other.

⁴ In a unique contribution to gang composition studies, Panfil’s (2017) research on gay gang-involved men revealed some differences in members’ experiences and crime involvement depending on the sexual orientation composition of men’s gangs, i.e. the ratio of gay to straight members in the gang.

⁵ Peterson et al. (2001) and Weerman (2012) found males in majority-male gangs to be most delinquent. In Weerman’s study, females and males in majority-female gangs were combined with those in sex-balanced gangs, perhaps reducing offending averages in sex-balanced gangs; results are therefore not directly comparable.

⁶ Importantly, these processes may not be mirrored in non-gang peer groups, highlighting the amplification of societal gender norms within gangs (see Peterson & Carson, 2012).

⁷ Miller (1998, 2001) references variations in females’ victimization and also in males’ victimization (Miller & Brunson, 2000), but this was not a specific focus of the work.

⁸ Miller (2002) cautions that females in majority-male gangs who are described as “one of the guys” (by themselves and by males in the gang) are not necessarily identifying their gender as male; rather, they may adopt “masculine” attitudes and behaviors (such as violence) to situationally enact masculinity, while maintaining female identity.

⁹ Although these same data were used in a previous study (Peterson & Carson, 2012), that study utilized a different operationalization of “gang member” (the Eurogang definition, as the study was presented at a Eurogang Research Workshop and published in a Eurogang-generated book) and included just one wave of data (Wave 4); our current study, by contrast, uses the self-nomination measure of “gang member” and incorporates data from Waves 3–6.

¹⁰ We acknowledge potential limitations inherent in using individual-level data to describe experiences that we argue stem in part from group-level processes (the “level of measurement” problem described by, among others, Short, 1998). We assert that hypotheses about how group dynamics affect individual-level behavior can be proposed based on theoretical and empirical work from sociology of organizations, peer group, and qualitative gang research and that findings from individual-level data can be interpreted within this larger framework; recent examples include Hennigan and Spanovic (2012) and Pyrooz et al. (2014) (see too arguments in Decker et al., 2013).

¹¹ Wave 1 and 2 gang members were excluded from the sample due to the use in this study of emotional victimization and offending measures not included in the survey instrument until Wave 3.

¹² Combining these categories is also consistent with prior work (e.g. Peterson & Carson, 2012; Peterson et al., 2001), and t-test of means comparisons ensured that combining the two groups was not masking important differences between the two; although females in majority-female gangs had lower violent offending and higher violent and property victimization means than those in all-female gangs, there were no statistically significant differences in mean levels of any offending or victimization types.

¹³ An alternative, but equally untestable, explanation is that the respondents may have been sex-assigned female at birth but were male-identified in terms of gender identity (or, assigned male at birth, but female-identified).

¹⁴ To streamline discussion, we include all-/majority-female gangs in our references to “same-sex” gangs, throughout the paper.

¹⁵ Due to the unequal sample sizes across gang sex composition, we also checked significant differences using Tukey’s Kramer.

¹⁶ Because the survey instrument did not ask respondents to report the name of their gang, we have no way of knowing whether females and males within the same gang types are reporting on the same gang.

¹⁷ The reader is reminded, though, that Hispanic/Latino youth comprised the majority of the overall G.R.E.A.T. evaluation sample.

¹⁸ There were no significant differences across the seven cities in gang members’ sex, the sex composition of their gangs, or their victimization levels. With regard to offending, gang youth in Philadelphia reported fewer violent and property crimes than youth in Greeley and the Dallas-Fort Worth area city, with no other significant differences. Given these collective findings, we believe there are no confounding site differences with which to be concerned. We thank an anonymous reviewer for inquiring about this possibility.

¹⁹ This differs from Peterson et al. (2001) and Weerman (2012), who found males in majority-male gangs to be most delinquent; the current findings, however, fall more in line with predictions based on institutional theories and are consistent with Peterson and Carson (2012), despite using a different operationalization of “gang member” and different data waves. Further, in Weerman’s study, delinquency rates of boys in sex-balanced gangs were combined with majority-female gangs, the latter of which may have pulled down average offending rates.

²⁰ We thank the anonymous reviewer who encouraged us to further explore this finding.

²¹ To explore this issue, we looked at the subset of males who reported membership in sex-balanced gangs at two time points (N = 55), predicting their time 2 offending with time 1 victimization and vice versa. While these results suggest the relationship is stronger when victimization predicts offending than when offending predicts victimization, tests for equality of coefficients (Clogg, Petkova, & Haritou, 1995) indicate statistically significant differences only for property (not violent or emotional) victimization and offending.

²² An alternative, or addition, to this perspective is the fact that some females just “like to fight”: e.g. female gang members in Hagedorn and Devitt’s (1999) study described their affinity for fighting, mostly over respect and gang-related representation. Being in majority-male gangs may offer young women space to escape societal gendered constraints and expectations and engage in what would be societally-viewed as non-normative behaviors.

²³ Analyses of the distribution of demographic characteristics, offending, and victimization were re-run with this smaller sample and the same general pattern of findings from the larger sample (n = 287) was upheld, providing confidence that findings from the subsample might be applicable also to the larger gang member sample.

²⁴ Although these gangs may be largely social in nature, the protective function of the group is evident: 100% of females in same-sex gangs reported that their gangs provide protection for each other (see Table 5).

²⁵ Haynie, Doogan, and Soller (2014), for example, suggest that violent or delinquent females may be motivated, given rejection by peers who scorn their non-normative deviance, to select peers who engage in similar behaviors.

²⁶ In their research, Miller and Brunson (2000) also note that it is unclear whether youth in their sample self-selected into gangs, or if they adopted certain values and behaviors after they joined; furthermore, the type of group youths were able to join may have been constricted by which gangs exist in their immediate surroundings.

²⁷ For example, those that are more organized and perhaps able to be more selective, which may help explain findings for females in majority-male gangs.

Additional information

Notes on contributors

Dana Peterson

Dana Peterson, Associate Professor, School of Criminal Justice, University at Albany, researches intersectional identities, youth gangs, and violence. She co-authored Youth Violence: Sex and Race Differences in Offending, Victimization, and Gang Membership (2010, Temple) with long-time friends and colleagues Finn-Aage Esbensen, Terrance J. Taylor, and Adrienne Freng, and co-edited (with Vanessa R. Panfil) the Handbook of LGBT Communities, Crime, and Justice (2014, Springer).

Dena C. Carson

Dena C. Carson is an Assistant Professor in the School of Public and Environmental Affairs at Indiana University Purdue University Indianapolis. Her research interests include youth violence, victimization, gangs and delinquent peer groups. Her recent publications have appeared in Criminal Justice Review, Journal of Crime & Justice, and Youth Violence and Juvenile Justice.

Eric Fowler

Eric Fowler is the crime analyst for the Burlington (VT) Police Department and a PhD student in criminal justice at the University at Albany. He is an advocate of and a contributor to transparency in police and public safety data. His research interests include juvenile crime and justice, youth street gangs, and policing.