Factor Analysis: Chisquare as Rotation Criterion.
Department of Statistics: Technical Reports, Nr. 40
The rotation problem in factor analysis consists in finding an orthogonal transformation of the initial factor loadings so that the rotated loadings have a simple structure that can be easily interpreted. The most popular orthogonal transformations are the quartimax and varimax procedure with Kaiser normalization. In this paper we propose the classical chisquare contingency measure as a rotation criterion. We think that this is a very natural and attractive criterion, not only for rotations but also for oblique transformations, that is not to be found in our popular statistical packages up to now.