Abstract
This study uses the National Educational Longitudinal Survey of 1988 (NELS) to replicate both the analysis in The Bell Curve and that of several of its previous replications. We examine the relative importance of test scores and family background in predicting dropping out of high school, starting college, arrests, and out-of-wedlock fertility. Our results relax several arbitrary assumptions made in The Bell Curve. We strongly reject The Bell Curve’s conclusion that family backgrou nd is almost always less important than test scores in predicting outcomes. In addition, our analysis casts doubt on some of The Bell Curve’s claims concerning reverse discrimination in education