Factor analysis of multiple batteries by maximum likelihood

McDonald's (1970) generalization of Tucker's (1958) inter‐battery factor analysis model to multiple batteries is considered. The identification of parameters is examined and an iterative algorithm for obtaining maximum‐likelihood estimates is provided. Consideration of the relationship between inter‐battery factor analysis and canonical correlation analysis in the case of two batteries suggests a generalization of canonical coefficients to the situation where there are several batteries. Examples of the application of the procedure are given.