题名
作者 Bruce Thompson.
版本
美国心理学会(APA)出品,195页,清晰PDF格式,可编辑,非扫描
简介 Intended for use by graduate students in a factor analysis or a multivariate statistics class, this unique book presents both exploratory and confirmatory methods within the single category of the general linear model (GLM).
目录
Preface
Introduction to Factor Analysis 3-7
Foundational Concepts 9-25
Exploratory Factor Analysis Decision Sequence 27-48
Components Versus Factors Controversies 49-56
Comparisons of Some Extraction, Rotation, and Score Methods 57-65
Oblique Rotations and Higher-Order Factors 67-81
Six Two-Mode Techniques 83-91
Confirmatory Rotation and Factor Interpretation Issues 93-98
Internal Replicability Analyses 99-108
Confirmatory Factor Analysis Decision Sequence 109-132
Some Confirmatory Factor Analysis Interpretation Principles 133-149
Testing Model Invariance 153-162
Appendix A: LibQUAL -super(TM) Data
Appendix B: SPSS Syntax to Execute a "Best-Fit" Factor Rotation
Notation
Glossary
References
Index
About the Author
书评
As the title suggests, this book provides a unified approach to exploratory and confirmatory factor
analysis. A combined treatment of both these methods in a single text is unusual, but has clear
advantages in drawing out the similarities in the two approaches as well as emphasizing their
differences.
After a very brief introductory chapter, Chapters 2–6 cover all the basic concepts and methods
of EFA. Explanations are clear throughout and the decision sequence for a factor analysis
(method, number of factors, rotation, calculating scores) is carefully explained. The discussion
of principal components vs. factor analysis proper (which gets its own chapter) is particularly
clear, and includes some interesting issues which are often not mentioned in more elementary
texts. There is also a clearer coverage than usual of sources of error in factor analysis, with
sampling as well as measurement error being discussed. The emphasis on the sample here and
elsewhere in the book is helpful as it gives the reader a better feel for what their sample size
should (ideally) be than can be gained from many discussions of factor analysis, and also
encourages thought about capitalization on chance features of a sample, a theme returned to in
the CFA section. Another good feature both here and in the CFA section is the presentation of
outputs from different methods using the same data (e.g. principal components vs. principal
axis methods). Chapter 7 concludes the EFA section by describing the six modes of factor
analysis listed by Cattell, with some discussion of practical aspects of using these. The book
continues (Chapters 8–12) with an equally clear and thorough coverage of CFA, bringing out
both similarities with and differences from EFA, and providing clear explanations of points
which often cause confusion to the beginner in this area, such as model identification, setting a
scale for latent variables and fit indices. Again, a good description of the CFA decision sequence
is provided. The final chapter in the CFA section covers testing model invariance via various
versions of multi-group modelling.
Each chapter concludes with a helpful section entitled ‘major concepts’ which emphasizes the
important issues covered rather than summarizing the chapter contents, and there is also a useful
glossary of technical terms at the end. All examples in the book are based on a dataset which is
supplied in an appendix. I was a bit surprised not to find either a CD or a web link for this; I
suspect that in not providing this material in a readily accessible format, the publisher is overestimating
the motivation of readers to get at the data by typing it in themselves. Similar comments
apply to the useful but lengthy SPSS syntax for targeted rotation provided in a second appendix.
The book’s style is clear and accessible, with technical material being both kept to the
necessary minimum and clearly explained. The only problematic feature, which might perhaps
make the book a little harder to dip into as opposed to reading through, is the use of some slightly
non-standard terminology, in particular the adoption of terms ‘score world’ (correlation
coefficients, etc.) and ‘area world’ (squares of these quantities). Similar comments apply to the use
of the non-standard term ‘pattern/structure coefficients’ instead of ‘loadings’, although this usage
is carefully justified. The book could clearly be used as a text for a postgraduate-level course on
factor analysis, but students coming to either EFA or CFA for the first time would require some
supplementary assistance from texts illustrating the use of specific packages, and probably also
need to do a bit of additional reading to amplify the information given in this very concise text. The
book would also be useful for anyone already familiar with factor analysis at undergraduate-course
level wishing to enhance their knowledge and to progress from EFA to CFA.
此书得来颇费周折,所以才收些辛苦费
此书在amazon的售价是$30.20