This article examines the extent to which service-learning courses affect students' attitudes and opinions. Elsewhere, we used a pre/postsurvey field experiment to demonstrate that volunteering with a homeless person tends to erode the stereotypes held by the domiciled--a confirmation of the venerable contact hypothesis. Here we use the same research design to assess whether students in service-learning courses exhibit a similar type of opinion change after spending a day with a homeless person. We find that even with limited contact a significant number of service-learning students came away from their time with homeless individuals holding fewer stereotypes and with a more nuanced perspective on the causes and consequences of homelessness. Nevertheless, working with a homeless person did have a negative effect on some students and contact generally failed to change students' views on public policy.