Three additional factors should be considered in relation to demand for school choice. One is that the first and third components of the decision – eligibility and use – may be confounding. Campbell et al. (2005) find that minority families are more likely to apply for CSFadministered private vouchers, but conditional on eligibility, they are less likely to use them. (Data from the Chicago Public Schools lotteries shows both that (a) black students are less likely to participate and (b) lottery winners from high-achieving schools are more likely to enroll in a new school than winners from low-achieving schools, Cullen et al., 2005, Tables 1 and 2). It is unclear how school choice preferences generate this result. But it suggests that in actuality voucher programs may be less equitable than their supporters claim based on a simple reading of the program design (Kemerer and Vitteriti, 2002). The second factor is that a non-trivialproportion of students who are eligible for and receive vouchers are already in private school. In the DC Scholarship program, 28% of the winners were already in private schools when they applied. The third factor is that, regardless of school quality, changing school induces an academic penalty for children (as they learn the standards, rules, and expectations at their new school; for empirical estimates of the magnitude of this effect, see Hanushek et al., 2004). Parents might therefore be wary of changing school and would be unlikely to repeatedly select a different school for their children.