SLOPE-RANGE AND GENERALIZED LEAST SQUARES: ESTIMATION OF THE STRUCTURAL RELATIONSHIP WITH REPLICATED OBSERVATIONS, UNEQUAL MEANS AND VARIANCES

抄録

A model where the true variable has a different mean and distinct variance which can differ among groups is considered for a bivariate linear structural relationship with replicated observations. The asymptotic properties of two estimators, the slope-range estimator and the generalized least squares estimator, are compared under two situations. In the first, the number of replications tends to infinity and in the second, the number of incidental parameters goes to infinity.