More so than other classes of statistical multivariate methods, factor analysis has suffered a somewhat curious fate in the statistical literature. In spite of its popularity among research workers in virtually every scientific endeavor (e.g., see Francis, 1974), it has received little corresponding attention among mathematical statisticians, and continues to engender debate concerning its validity and appropriateness. An equivalent fate also seems to be shared by the wider class of procedures known as latent variables models. Thus although high-speed electronic computers, together with efficient numerical methods, have solved most difficulties associated with fitting and estimation, doubt at times persists about what is perceived to be an apparent subjectiveness and arbitrariness of the methods (see Chatfield and Collins, 1980, p. 88). In the words of a recent reviewer, "They have not converted me to thinking factor analysis is worth the time necessary to understand it and carry it out." (Hills, 1977.)