In an article otherwise lamenting political scientists' inability to influence public policy (“Politics, Political Science, and Specialization,” June 2002, pp. 187–190), Robert Jervis takes issue with recent research on school vouchers, to which we have contributed (2002). Jervis argues that voucher experiments “are largely irrelevant, or at best a first step.” Comparisons of students in private and public schools are of little value as long as most students remain in public school. Jervis therefore directs analysts to look for “large-scale improvements [that] can only come through the impact of the possibilities of choice on the bulk of the public school system.”