The roles that peers play in the decision to go to college are not well understood. Logistic regression is used to explore the role that peers play in the college-going decisions of a sample of low-income urban minority public high school graduates drawn from the National Education Longitudinal Study (NELS:88) database. Friends' plans are found to be the single best predictor of 4-year college enrollment for these low-income urban minority students, even when controlling for variables traditionally assumed to affect college going. These peer variables are stronger predictors of 4-year collegegoing behavior for this group than they are for a comparison sample of all U.S. high school graduates. These data are important for the formation of policies to improve all levels of college access programming.A college education is, now more than ever, critical for economic selfsufficiency and social mobility. The lifetime economic benefits afforded by college graduation are well documented, and growing. Between 1979 and 1993, the entry-level wage deferential between high school and college graduates more than doubled. The median 2004 income for men age 25 years and older who completed only high school was U.S. $31,624, whereas the median income for men with a bachelor's degree was 64% higher, $51,876. For women, high school graduates made $20,928 in 2004, whereas those with bachelor's degrees made $35,663, or 70% more (U.S. Census Bureau, 2004). This represents an increase over the 1997 discrepancies for men and women, and in our service and/or information economy we can expect this gap to continue to increase, even as the supply of highly educated workers surges (Donahue