Content area
Full text
Elazar J. Pedhazur. Multiple Regression in Behavioral Research: Explanation and Prediction (3rd edition). Fort Worth, TX: Harcourt Brace College Publishers, 1997, 1,058 pages. Reviewed by Chockalingam Viswesvaran, Assistant Professor, Florida International University, Miami, FL.
A course in statistics is approached by many students (even graduates) with trepidation. Having taught methodology classes, I am aware of the dislike that most students have for such classes. The problem lies partly with a mindset that gets cultivated in introductory statistics classes where concepts of regression are first introduced-a mindset that considers anything associated with statistics as difficult to understand. Even the brave souls who enjoy mental gymnastics are not able to see the link between the methods and the substantive questions that are answered by their application. Pedhazur's Multiple Regression in Behavioral Research is an excellent antidote to this problem.
The book is arranged in four parts. The first part (Chapters 1-8) begins with simple linear regression and correlation, moves on to elements of multiple regression, and concludes with a preview of the issues to be considered in prediction. Pedhazur begins his discussion of multiple regression with the simple case of two independent variables, and uses it to introduce all essential concepts of multiple regression. Once the concepts are introduced he generalizes them to the case of many variables using matrix notation. Also included in Part I is a discussion of regression diagnostics, where influence analysis and detection of outliers are presented. Finally, Part I also describes the format in which computer input and output statements (used throughout the text) are presented. Although Pedhazur relies primarily on SPSS, he also provides information for SAS, BMDP, and MINITAB software packages.
Part II (Chapters 9-17) focuses on the use of multiple regression for explanation and theory building. Coding for categorical independent variables, curvilinear regression, multilevel analysis, and logistic regression are presented. Part III (Chapters 18-19) presents structural equations models, whereas Part IV (Chapters 20-21) presents multivariate analysis....