This paper examines the impact of vouchers on student test scores in Dayton, Ohio, New York, New York, and Washington, DC, highlighting New York City parental assessment of private and public schools to investigate why vouchers seem to have differential effects depending on the students' ethnic background. Researchers collected baseline test scores and family data prior to the voucher lotteries, administered the lotteries, and collected follow-up information 1 and 2 years later. At pretest and posttest, students entering grades 1-8 took the Iowa Test of Basic Skills in reading and mathematics. Parents completed surveys regarding satisfaction with their children's schools, involvement in students' education, and demographics. Results varied systematically by ethnicity. There were no significant differences between the test scores of non-black students who moved to private schools and of the control-group students. Vouchers made a substantial, positive difference for black students but did not significantly impact Hispanic American students' test scores. All control group parents were satisfied with their children's education. Hispanic and black parents were very satisfied with private school placements, though they differed in perceptions. Private school impact on black students was not due to school size, class size, school disruptions, school communications, desegregation, dress rules, hallway monitoring, school resources, homework, parent involvement, or student involvement with the school. (Contains 10 tables and 21 endnotes.) (SM)