A common critique of school choice is that it will have the practical effect of further increasing racial and ethnic segregation in American education. Although most survey evidence indicates that the highest priority of parents in choosing schools is academic quality, with the race and ethnic composition of the student body being much less important, many critics doubt that private preferences are being publicly revealed. In this paper, we use an experimental design embedded in a survey to obtain an alternative measure of educational quality and racial diversity as considerations for household school choice. While both academic quality and race/ethnic diversity had an effect on preferences, academic quality was a more important predictor. We then examined the relationship between preference and actual choice outcomes. Race-related opinions were nonpredictive of outcomes, but a stress on high test scores by parents predicted school choice among students who are not “at risk.”