抄録
In factor analysis both the maximum likelihood estimator and the generalized least squares estimator for the structural parameters (i.e. factor loadings and unique variances) are shown to be consistent weakly or strongly under Anderson and Rubin's sufficient condition for weak identifiability.
