The study site was a poor, Black community on the northern city boundary of Champaign, Illinois. The interviews focused on both the behavior of the respondent and the respondent's perceptions of the behaviors of others. Interviews were conducted with Champaign police and local residents as well as 50 acknowledged female gang members. A multi-stage sampling design captured both a numerically balanced sample of active and inactive women in the north end's main gangs, as well as minor gangs and gang women's friendship networks. The social network analysis showed that women were more likely to choose a friend from within their own gang rather than another gang. Gang affiliation had a friend-choice effect, but friend choice was first determined by social and physical distance, as well as complex historical patterns of the community. Descriptive analysis showed that approximately 50 percent of active and inactive gang women's ego-gang networks were composed of gang women who belonged to gangs different from their own. The policy and program recommendations derived from the study findings are linked to gang women's friendship networks. The social network analysis suggests that the most effective social units of intervention should be the most prestigious and influential gang women and cohorts of strongly linked gang women and their children. Middle-teen years were indicated to be the optimum time of prevention and intervention, and school (before girls drop out) would be the optimum location. The selection of the best intervention target by gender, age, and long-term outcome is critical. This research suggests that teenage gang women with children are a better long-term intervention target than teenage gang males. Providing social, emotional, and employment training, as well as mental health counseling to adolescent and young-adult gang women, along with other mother-focused support services, should increase material resource levels within women's personal networks and eventually strengthen community solidarity and reduce crime in ego-gang networks. 103 tables, 8 figures, and 41 references