Customer Reviews

Modern Factor Analysis

5 star:		(1)
4 star:		(0)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

This is an all-time classic on the subject. Want to know about principal axes and components, minimum residuals, alpha factor analysis, canonical factors, maximum liklihood, and the multiple group and centroid methods? Then read this book. How about factor rotation? Whether you prefer your factors served up on an orthogonal or oblique reference basis, you will find it here: the book discusses the notion of simple structure and clearly describes and compares many of the rotational methods by which it is sought--quartimax, varimax, oblimax, oblimin, Harris-Kaiser orthoblique, even procrustes. Need to measure your factors? Consult Chapter 16 for an in-depth analysis and comparison of different techniques for estimating factor scores. The only complaint I have is that some of the less popular oblique rotations such as PPFP and Oblisim were not covered. While Varimax is unquestionably king when it comes to orthogonal rotation, the popular oblique methods (like oblimax and promax, which appear in many stats packages) do not perform very well; currently, PPFP (Primary Product Functionplane), Oblisim, and possibly TRASID, are the best solutions to the oblique simple-structure rotation problem. I am sure that if there was a 4th edition, such methods would have been included. Nevertheless, I would not hesitate to recommend this book to anyone interested in factor analysis or needing to perform one.

Jeffrey Owen Katz, Ph.D.

Author (with Donna McCormick):

"The Encyclopedia of Trading Strategies" (McGraw Hill, 2000)

"Advanced Option Pricing Models" (McGraw Hill, 2005)

This was truly one of the classic books on exploratory factor analysis. However, the big problem is that the literature has moved on while (for obvious reasons) Harman has been frozen in place since its publication, which means that it is nearly forty years out of date. The literature made a substantial turn towards confirmatory factor analysis right as this was coming out. While EFA wasn't a hot topic anymore, there are developments of importance (e.g., geomin rotation, more focus on standard errors, the comprehensive review done by Cudeck and Browne in the 1990s, semi-confirmatory/semi-exploratory factor analysis, multiway/multimode models, models for non-normal data, etc.) that came after this book was published. Unless you want the best the 1960s had to offer and nothing else, a book like Stanley Mulaik's Foundations of Factor Analysis, Second Edition (2009, CRC Press), is likely to be more suitable.