A Generalized Regression Estimate Formula for Factor Scores: Its Application to Singular Data

Summary

A generalized formula for determining regression estimate factor scores is developed for use with either singular or nonsingular data with either orthogonal or oblique factor solutions. Kaiser's component scores are shown to be a specialized case of the generalized formula. With use of the formula, estimated factor scores and component scores are developed for an oblique solution of the singular data of the Thurstone box problem and then compared to the true factor scores.