Psychopathology and Boolean Factor Analysis: a mismatch

It has recently been claimed that Boolean Factor Analysis (BFA) is superior to more traditional approaches to empirical syndrome identification in psychopathology (Weber & Scharfetter, 1984). This paper examines the underlying assumptions and the empirical basis of this claim. A hypothetical data set, with a more complex structure than the examples used by Weber & Scharfetter to support their claim, is analysed by means of both BFA and Ordinary Factor Analysis (OFA). It is shown that, as data sets resemble empirical conditions more closely, BFA may break down completely while OFA continues to yield adequate results.