Content area
Full Text
Applied Multivariate Statistics for the Social Sciences, 4th ed., by James P. Stevens (2002). Hillsdale, NJ: Lawrence Erlbaum, 712 pages, $59.95 (paper).
This is the fourth edition of Stevens's text on applied multivariate statistics. As was the case with previous editions, this text is targeted toward those who wish to develop a strong conceptual understanding of how to interpret results from multivariate analyses using various software programs like SPSS and SAS. The text is well suited for people who are neither statisticians nor mathematicians but who are interested in understanding and applying multivariate techniques. Familiarity with multivariate statistics or matrix algebra is not assumed, but an appreciation for hypothesis-testing procedures and statistical concepts like (mu) (population mean), t tests, analysis of variance, correlation, and power is quite helpful. The text provides a review of these concepts, but due to space limitations the treatment is necessarily limited. Stevens also provides a review of matrix algebra in preparation for the more complex material to come; his treatment of matrix algebra is very clear and provides an excellent foundation for later material. The inclusion of syntax commands for SPSS MATRIX and SAS IML procedures for manipulating matrix data is helpful. I particularly enjoyed the geometric interpretation of the meaning of generalized variance on pages 68-70.
Formulae appear when necessary, but the author focuses primarily on developing an understanding of use of the various techniques rather than on the derivation of mathematical formulae or proofs. If you are a reader who is looking for a more technical, theoretical, or mathematical treatment of multivariate analyses along the lines of Johnson and Wichern (2002), Anderson (1984), or Martha, Kent, and Bibby (1979), then this is not the text for you. However, if you are looking for a text that coherently describes multivariate statistical analyses using clearly defined terms and with a minimal use of formulae, then this text fits your needs.
Topics include matrix algebra, multiple regression, multivariate analysis of variance, discriminant analysis, factorial analysis of variance, analysis of covariance, stepdown analysis, confirmatory and exploratory factor analysis, canonical correlation, repeated measures analysis, and log-linear models. Included in the text are statistical tables, exercises, and a CD-ROM containing all of the data sets used as examples. A wide variety of real-world examples...