Explores whether flexible nonlinear models (including neural networks and genetic algorithms) can reveal otherwise unexpected patterns of relationship in typical school-productivity data. Applying three types of algorithms alongside regression modeling to school-level data in 183 elementary schools proves the hypothesis and reveals new directions…

Uses simulated data to assess the adequacy of two econometric and linear-programming techniques (data-envelopment analysis and corrected ordinary least squares) for measuring performance-based school reform. In complex data sets (simulated to contain measurement error and endogeneity), these methods are inadequate efficiency measures. (Contains 40…

In assessing sex-discrimination suit damages, debate rages over the type and number of variables included in a single-equation model of the salary-determination process. This article considers single- and multiple-equation models, providing 36 different damage calculations. For University of Minnesota data, equalization cost hinges on the…

Recent within-twin estimates of schooling returns are considerably higher than existing estimates. This paper shows that small ability differences among twins can yield more upward omitted-ability bias (and more upward bias overall) in the instrumental variables estimate correcting for measurement error than in the standard within-twin estimate.…

Applies comparative techniques originated by K. Wolpin and G. Psacharopulos to discriminate between the "weak" and "strong" screening hypotheses, employed to determine whether societal investment in upgrading or expanding education is justified. Controlling for sample selection, this study finds evidence for weak screening in…