This study assessed contextual risks for polydrug use in a triethnic sample (non-Hispanic white, African American, Hispanic) of young women with a low income. For the current analysis, a total of 712 young women aged 18 to 31 years who sought care in state-funded family planning clinics in southeast Texas from December 2001 to May 2003 participated in the survey. The main outcome of the study was the number of illicit drugs (including marijuana, MDMA [ecstasy], crack cocaine, and other hard drugs) used in the last 12 months. Of the 712 subjects, 198 (28%) reported using illicit drugs in the past 12 months. Neighborhood socioeconomic status was significantly associated with drug use in a bivariate model. The proportion of women living in the most advantaged neighborhoods who reported drug use was more than twice that of women living in the most disadvantaged neighborhoods. However, the significance of neighborhood socioeconomic status was eliminated after controlling for ethnicity or for personal network characteristics in a multivariate ordinal logistic regression model. In contrast, in multivariate models, personal network indicators, such as a larger number of monthly contacts with friends (odds ratio [OR] = 1.32, 95% confidence interval [CI] = 1.11, 1.56) and a larger number of friends who used illicit drugs (OR = 1.47, 95% CI = 1.33, 1.62) were associated with increased drug use. In addition, not being currently married (vs. being married) (OR = 2.73, 95% CI = 1.44, 5.16) was associated with a larger number of drugs used in the last 12 months. In conclusion, we found that neighborhood socioeconomic status was not directly associated with more drug use when controlling for ethnicity or for personal network characteristics. Personal networks may mediate the relationships between neighborhood and drug use. Strategies to reduce polydrug use should target personal networks where friends use illicit drugs.