摘要：A semiparametric estimator for evaluating the parameters of data gener-ated under a sample selection process is developed. This estimator is basedon the generalized maximum entropy estimator and performs well for smalland ill-posed samples. Theoretical and sampling comparisons with parametricand semiparametric estimators are given. This method and standard ones areapplied to three small-sample empirical applications of the wage-participationmo del for female teenage heads of households, immigrants, and Native Ameri-cans.
