Iterative factor clustering of categorical data reconsidered