School voucher programs and other similar policies are increasingly prevalent in the United States. Evaluations of existing voucher programs have generally focused on student test score outcomes, with results indicating modest positive impacts in some cities and insignificant differences in others. A similar pattern has been demonstrated in analyses of charter school outcomes.

In recent years, researchers have begun to consider academic outcomes other than standardized test scores. Most notable are studies of student attainment, usually defined as graduation from high school and, or, enrollment in postsecondary institutions. In both voucher and charter school contexts, substantial attainment differences associated with choice have appeared even in areas where test score differences were small or unapparent. Such results would suggest that school choice may ultimately be most promising as a policy alternative that stresses long-term rather than short-term academic success.

In this paper we consider one measure of student attainment—enrollment in a four-year college—as it differs between students in a voucher program and those in traditional public schools. Using a rich set of survey and administrative data drawn from the city of Milwaukee, we extend previously published work showing modest but statistically significant average differences in attainment. We focus particularly on the interaction between student background, school choice, and college enrollment, stressing differences associated with parental education and the characteristics of the colleges students choose.

Our results broadly indicate that different rates of college enrollment are particularly strong for students whose parents had low levels of education: voucher students whose parents had high school or less than high school education only were more likely to attend college than similar public school parents, while public school students whose patterns had college degrees were more likely to attend college than voucher students with similar parental backgrounds. In our investigation of the colleges voucher and public students attend, we find important differences in the relationship between student background, voucher attendance, and the type of post-secondary institution.

Framed in the more general context of school choice policymaking, our results underscore the importance of focusing on student outcomes other than test scores, and suggest new ways for policymakers to target these programs as they expand in number and scope.