Using student-level data from Durham, North Carolina, we examine the potential impact of school choice programs on the peer environments of students who remain in their geographically assigned schools. We examine whether the likelihood of opting out of one's geographically assigned school differs across groups and compare the actual peer composition in neighborhood schools to what the peer composition in those schools would be under a counterfactual scenario in which all students attend their geographically assigned schools. We find that many advantaged students have used school choice programs in Durham to opt out of assigned schools with concentrations of disadvantaged students and to attend schools with higher achieving students. Comparisons of actual peer compositions with the counterfactual scenario indicate only small differences in peer composition for nonchoosers on average. More substantial differences in peer environment emerge, however, for students in schools with concentrations of disadvantaged students and schools located near choice schools attractive to high achievers. The results suggest that expansions of parental choice may have significant adverse effects on the peer environments of a particularly vulnerable group of students. (Contains 6 tables, 1 figure and 12 footnotes.)