Covariance and regression slope models for studying validity generalization

Two new models, the covariance and regressionslope models, are proposed for assessing validity generalization.The new models are less restrictive in thatthey require only one hypothetical distribution (distributionof range restriction for the covariance modeland distribution of predictor reliability for the regressionslope model) for their implementation, in contrastto the correlation model which requires hypotheticaldistributions for criterion reliability, predictor reliability,and range restriction. The new models, however,are somewhat limited in their applicability since theyboth assume common metrics for predictors and criteriaacross validation studies. Several simulation(monte carlo) studies showed the new models to bequite accurate in estimating the mean and variance ofpopulation true covariances and regression slopes. Theresults also showed that the accuracy of the covariance,regression slope, and correlation models is affectedby the degree to which hypothetical distributionsof artifacts match their true distributions; theregression slope model appears to be slightly more robustthan the other two models.