[An outstanding text, with extensive treatment of multivariate analysis of variance, including numerous research examples. For many years it was unique in its treatment of qualitative data (e.g. contingency tables) in an Anova-type format but unfortunately does not cover factor analysis. The mathematical level is moderately difficult, although it has been prepared for a graduate course in behavioral science.
"It assumes two previous statistics courses, the first covering such topics as probability, descriptive statistics, tests of significance, chi- square, simple correlation, and regression, and the second devoted to least-squares methods, including basic analysis of variance and multiple regression analysis. ... In addition to the statistical prerequisites, this text assumes some mathematical preparation, including at least an elementary knowledge of algebra, coordinate geometry, differential and integral calculus, and matrix algebra."
Although it is more difficult than some of the other texts, it is a good choice as a reference book. Chapter 2 contains a fine summary of matrix algebra.]
[An intermediate level text, somewhat less technical than Morrison's, with good discussions of the practical aspects of multivariate analysis, including interpretation and computer applications with SPSSX, SAS and BMDP. There is good coverage of methods related to the general linear model, but no treatment of MDS, cluster analysis, or log-linear models.]
[An intermediate level text with much broader coverage than is typical, including MDS, cluster analysis, log-linear models, and structural equations models (LISREL) in addition to the usual fare of linear models.]
[This book presents a detailed treatment of the general linear model approach to regression, ANOVA and ANCOVA. The treatment is coordinated with Finn's computer program, MULTIVARIANCE, one of the most comprehensive program for these analyses. MULTIVARIANCE is now part of SPSSX (as MANOVA). A number of sample research problems are considered in great detail throughout the text; an appendix displays the input to and output from MULTIVARIANCE for each problem, which is a real aid in learning how to use the program, but rather boring if you don't (and we don't).]
[An excellent text with a strong emphasis on the geometric properties of matrix operations and their application to multivariate statistics. The topics covered relate mainly to the general linear model and its extensions, including regression, ANOVA, structural equations models (path analysis, LISREL models), and models for categorical data. Unfortunately, the book is out of print, but see John's new book, below.]
[Georges Monette's comments: "I have never read a book on regression that reflects as broad and profound a grasp of the concepts of statistics as this book does. In every topic John Fox deals with--and he does not avoid the slippery ones--he shows a clarity and depth of understanding that goes beyond anything else I have seen in textbooks and that matches the works of the leading researchers within each field."]
[Though not in any sense a text for a standard course in multivariate analysis, this book is notable for its emphasis on graphical and exploratory techniques for multivariate data.]
A very readable and applied book. Each technique is accompanied by a published research paper illustrating it. The appendices include both SPSS and SAS code (also a few other languages) that could be used to conduct the analyses presented as examples
[This book presents a reasonably balanced, elementary coverage of multivariate methods. It is geared to students with a modest mathematical background, and concentrates on conceptual or heuristic explanations, rather than mathematical proof. Together with Bock's text, it was one of the first books to emphasize the role of multivariate techniques in research and the practical considerations in their use. Contains a good selection of small demonstration problems and a chapter on "canned" programs for multivariate analysis.]
[This has been used for some years as a somewhat more mathematical text in Psy614 for some years. Its coverage of multivariate topics is reasonably complete, with the greatest emphasis on multivariate linear models for ANOVA and regression. Its two chapters on principal components and factor analysis are also quite good. A major strength of this text is its focus on the geometry of multivariate methods. Another strength is a fairly large number of research examples and datasets. A major weakness is the rather poor typesetting of mathematical formulae, which makes the book more difficult to read than it really is.]
[The first edition was used as the text for the course quite a while ago, and the second edition was used in 86/87. It is intermediate between Bock & Tatsuoka in difficulty. It is a sound book, but somewhat dryly written, with little of the practical information often needed in research. Good coverage of MANOVA, with numerous examples. Chapters 1 and 2 provide an excellent review of elementary statistical concepts and matrix algebra, but contain no exercises.]
[Presents a reasonably balanced treatment of multivariate techniques, though the greatest emphasis is on factor analysis methodology (4 chapters) and classification (3-4 chapters). The coverage of MANOVA is somewhat skimpy. Rather than choosing simplified examples from a variety of disciplines, the authors use data from a long-term coordinated program of clinical/psychiatric research (which, incidentally, explains their emphasis on factor analysis and classification). Also noteworthy is the inclusion of an extensive series of FORTRAN programs for carrying out the analyses described in the text.]
[The emphasis in this book is expressed in the title: USING. There is very little in the way of mathematical explication (for which most students are grateful), but a great deal on understanding, at a verbal level, what various procedures are about and how to carry them out using SPSSX, BMDP, and (added to the second edition) SAS. Students who have used it have also been grateful for the way in which the authors demonstrate how to use the various techniques: after showing the computer output, the authors provide a sample Results section from a research paper, showing how one would interpret the output verbally.]
[A book of intermediate difficulty, this text assumes a previous statistics course at the level of Glass & Stanley or Hayes. The coverage is quite varied: multivariate tests of differences among means (e.g. T}, MANOVA, Covariance) are treated extensively, but simply; regression and factor analysis are not treated at all. When this book was last used as a text for 6140 it was necessary to supplement it with other readings.
The second edition is strengthened considerably by a major re-writing and the contributions of Paul Lohnes.]
[This is a relatively difficult book, at or slightly above the level of Bock. The excellent treatment of inferential procedures is extensive, includes many worked examples, and makes this an excellent choice as a reference book for the graduate of a 6140-level course. The primary emphasis is on the general linear model - regression, MANOVA and related topics. Timm's treatment of principal components is brief, but excellent.]
[By far the simplest book, in terms of mathematical difficulty, this is an excellent choice for someone wishing to gain a general, conceptual understanding of multivariate techniques. Chapter 9 provides a marvelous nontechnical overview of multivariate techniques, which exposes the strong interrelationships among the seemingly diverse methods. This chapter, together with Harris' Ch. 1 provide an excellent nonmathematical introduction to multivariate methods; they should be required reading for all graduate students. The book has a major weakness in that inferential aspects of multivariate methods (e.g. hypothesis testing; MANOVA) are not treated at all.]
[A comprehensive set of worked examples for regression analysis, ANOVA, analysis of covariance and multivariate analysis using SAS procedures GLM and ANOVA. An excellent treatment ofhow to do itwith SAS.]
[This book is somewhat similar to Tabachnick & Fiddell in its focus on concrete applications, with extensive discussion of the practical details of interpreting results and writing results sections. The examples cover principal components analysis, exploratory and confirmatory factor analysis, path analysis, and structural equation modeling.
[A nicely done, novice-level how-to book, with an emphasis setting up (simple) SAS programs, interpreting results, and writing results sections in APA style, using a nice collection of substantively-based (but fictitious) examples. The book covers data screening, correlation analysis, t-tests, ANOVA (one-way, two-way, simple repeated measures designs), MANOVA, multiple regression, principal component analysis, and scale reliability.]
© 1995 Michael FriendlyAuthor: Michael Friendly