Content area
Full Text
Winfred Arthur, Jr., Winston Bennett, Jr., and Allen Huffcutt. Conducting Meta-Analysis Using SAS. Mahwah, NJ: Lawrence Erlbaum Associates, 2001, 188 pages, $19.95 softcover.
Meta-analytic techniques are increasingly used in literature reviews and theory building in social sciences. This is especially true in 1-0 psychology, and a book that explains the basics to facilitate application by researchers is a welcome addition. Of the different meta-analytic frameworks that exist out there, the methods developed by Hunter and Schmidt are widely used in 1-0 psychology. A book that focuses on the Hunter-Schmidt methods, presents software to apply the HunterSchmidt methods, and walks readers through examples to illustrate the steps, will certainly be valued by readers of Personnel Psychology. In addition, given that SAS is a popular statistical software package available to many researchers in both the mainframe as well as in various PC platforms, this book will be well appreciated by researchers, students, instructors, and practitioners interested in conducting a meta-analysis.
The book comprises five chapters and four appendices. The first chapter lays out the theory of meta-analysis. An explanation of sampling error and the law of small numbers is used to introduce what is attempted in a meta-analysis. The authors state that there are 11 steps in the general process of implementing a meta-analysis, the 11 being: (a) topic selection-defining the research domain, (b) specifying the inclusion criteria, (c) searching and locating relevant studies, (d) selecting the final set of studies, (e) extracting data and coding study characteristics, (f) deciding on independence and nonindependence of data points, (g) testing for and detecting outliers, (h) data analysis, (i) deciding to search for moderators, (0) testing moderators, and (k) interpreting results. Although the authors identify all of the steps involved in a meta-analysis, this book focuses heavily on the data analysis phase. The chapters focus on computing the mean and variability of the observed and corrected distribution...