Applied Multivariate Research
Design and Interpretation
- Lawrence S. Meyers - California State University, Sacramento, USA
- Glenn Gamst - University of La Verne, USA
- A.J. Guarino - Massachusetts General Hospital Institute of Health
Multivariate concepts presented in an extremely applied manner, strongly emphasizing written results and offering SPSS examples for greater comprehension.
Today, through the sophistication of statistical software packages such as SPSS, virtually all graduate students across the social and behavioral sciences are exposed to the complex multivariate statistical techniques such as correlation and multiple regression, exploratory factor analysis, MANOVA, path analysis and structural equation modeling. This book is designed to provide full coverage of the wide range of multivariate topics in a conceptual, non-mathematical, approach. It is geared toward the needs, level of sophistication, and interest in multivariate methodology of students in applied programs in the social and behavioral sciencesáthat need to focus on design and interpretation rather than the intricacies of specific computations.
Very nice book. Especially our students with a linguistic background who have never been thaught statistics like it very much.
This book is very discriptive and appropriate to be considered as a supplemental. Also I have no access to teh instructor materials.
College Review Board did nto approve it.
Despite retaining the general structure of providing pairs of chapters for each topic, this second edition represents a considerable revision of the first edition. Every chapter was extensively reviewed and, in most places, substantially rewritten to make the narrative more readable, complete, and/or current. We have also added a good deal of new material to cover topics not included in the earlier edition. The changes we have made include the following:
· We have revised somewhat the ordering and organizational schema of the chapters in an effort to more smoothly build from conceptually simpler to conceptually more complex procedures. In very global terms, we begin with basics, then cover comparisons of means, followed by prediction, analyses of structure, and finally model fitting.
· Our treatment of missing values with the IBM SPSS Missing Value Analysis module has been updated and expanded.
· We have collapsed the three pairs of MANOVA chapters down to a single pair.
· To supplement the basic statistical regression procedures, we added a pair of multiple regression chapters covering some advanced topics such as polynomial regression, mediation, interaction effects, and dummy and effect coding.
· We added a pair of chapters on multilevel modeling.
· Both the logistic regression and the discriminant analysis chapters now include three group designs as well as the two-group designs we presented in the first edition.
· We have now included within binary logistic regression the topic of ROC curves and ROC analysis applied to the classification decision criterion.
· We have expanded on the strategy to perform exploratory principal components/factor analysis, displaying the results from several extraction techniques. We have also added to that material the procedure for performing a reliability analysis based on the factor analysis results, how to compute subscales based on the reliability analysis, and how to use the computed scales in a subsequent (albeit simple demonstration) analysis.
· We have added a new pair of chapters on multidimensional scaling.
· We have added a new pair of chapters on cluster analysis, including both hierarchical clustering and k-means clustering.
· We separated the full information structural modeling material dealing with path analysis and structural equation modeling into separate pairs of chapters.
· We have updated the chapters on model invariance.
· We have included an Appendix covering many of the frequently used IBM SPSS Amos commands.
· Our sample reports of results now includes exact probability values and confidence intervals consistent with the new APA publication guidelines.