Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research

Many of us who have used factor analysis had only a vague notion of what we were doing, namely trying to reduce a large number of items into a smaller, less unwieldy, more readily interpretable set of variables. With user-friendly software such as SPSS, the mechanics -- entering items, extracting factors, rotation of factors, saving factor scores if needed, and calculating reliability coefficients -- are sufficiently obvious to permit rough and ready, sometimes quite useful factor solutions that provide insights that otherwise would not have been available.Without studying factor analysis as such, however, such quick and dirty applications often yield misleading results, something that anonymous reviewers of submitted manuscripts will be only too happy to acerbically explain. In my own work, I stared off routinely using principal components analysis, the SPSS default option for factor analysis, but had no notion that principal components analysis and factor analysis in its various forms are mathematically distinct. Principal components uses all three sources of variance -- shared, random, and error variance in formulating components, while factor analysis uses only shared variance. One common outcome is that principal components will typically yield a misleadingly clear-cut solution, while factor analysis rightly yields a solution that requires more interpretative effort.Furthermore, when trying to reduce a comparatively large number of items to a small set of themes or variables, we get our most informative results when the analysis is limited to shared variance. Thus, while principal components has its uses, one of the many forms of factor analysis, say alpha factoring, is usually better suited to the task at hand.Before I read Making Sense of Factor Analysis, I had a vague notion of how factor extraction was accomplished. Having read the book, I can now understand the procedures pretty clearly. I find the regression-based approaches intuitively appealing, and this book explains them very well.Rotation of factors was something I previously did ritualistically, but I had no idea what was really going on. Having read this book, however, a good deal of the mystery has been dispelled. I must admit, however, that while I can now see how rotation of factors is accomplished, and I understand that some factors yield orthogonal solutions while others yield factors that are associated, the reason why rotated factors are generally more readily interpretable still eludes me.Furthermore, I have done factor analyses in which the non-rotated factors were more readily interpretable than the rotated ones. The book is silent as to whether or not rotation is a requirement in all instances.Finally, when looking for composite variables to insert as an independent variable in a regression equation, I have sought and found single-factor solutions. These can be very illuminating, yielding powerful explanatory variables or predictors. Given the scale-construction nature of this book, however, these useful procedures are not discussed. (By the way, single factors cannot be rotated.)Factor analysis is a complex statistical tool that comes in explanatory and confirmatory forms. No one text can do justice to the entire topic, but Making Sense of Factor Analysis gives most of us most of what we need. If you're put off by the seemingly extraneous material at the beginning of the book, stay with it. It turns out to be really quite pertinent and definitely worth reading,This is a very good textbook, and the health-related questionnaire that the authors produce is exemplary, a tribute to their patient and methodical application of this complex statistical technique. Nevertheless, if you come away from the book with the feeling that factor analysis is not foreign to guess-work and fishing, and that some crucial issues remain the subject of heated debate, you're right.

I am a PH student at Universdad Nacional de Colombia (Engineering Faculty). The Pett et al.`s book presents a detail explanation (focuses on "HOW") of the procedure involved to make factor analysis (FA). The authors made a great effort in writing a book understandable to those inexperienced researchers with no strong formation in statistics. The only thing that I criticize is that this wonderful book was written emphasizing on the "how" with less focus on the "why" and very little treatment on the explanation and foundations of the descriptive and statistical approach of FA.

This is the best book I've found on factor analysis - I am not mathematically inclined, but I do want to understand "the sense" of factor analysis. This book is detailed, clear, with concise summaries as well as detailed explanations of the matrix algebra and math of factor analyses. It contains clear and useful diagrams. I have checked out several other books on factor analysis - this is the one that finally explained the subject to me.

My professor says it's a little outdated, but I had only a vague understanding of factor analysis before reading the book and after, I was able to correctly conduct factor analysis for work. Excellent text. Too bad it's a little outdated.

This is a great book for those who are new to factor analysis. Although written from the perspective of a health care researcher, it's easily accessible to folks in other disciplines (I'm in educational leadership). I especially appreciated the step-by-step process for doing factor analysis from beginning to end, along with thorough explanations of various choices (and why you'd make them) along the way.

I haven't used this in class yet, but I will next year. I really like the approach and level of the presentation. Hope students like it!

This book SAVED my life... well, my job but it felt like my life.

Good deal

Excellent and easy to read while also not cutting back on the theoretical background - I would definitely recommend to anyone interested or is required to conduct factor analyses (novices and experts alike!)