Content area
Full text
John E. Hunter and Frank L. Schmidt. Methods of Meta-Analysis: Correcting Error and Bias in Research Findings (2nd edition). Newbury Park, CA: Sage Publication, 2004, 582 pages, $105.00.
Ralf Schulze. Meta-analysis: A Comparison of Approaches. Cambridge, MA: Hogrefe & Huber, 2004, 256 pages, $39.95.
Meta-analysis is the process of cumulating research design effect sizes over a set of studies of a similar construct or condition. Because of its potential to summarize any number of unique studies into a small set of information, there is a great deal of professional concern over the judgment calls, limitations, underlying statistical models, and permissible interpretations associated with the meta-analytic approach. Regardless of one's opinion of meta-analysis, because of the growing volume of meta-analytic studies on any number of subjects, it would be dubious to claim to be a competent scientist-practitioner without understanding this method. To get right to the point, the new edition of Methods of Meta-Analysis by Hunter and Schmidt is the single most important source available for gaining a working knowledge of meta-analysis.
Previous editions of the Hunter and Schmidt book (Hunter, Schmidt, & Jackson, 1982; Hunter & Schmidt, 1990-which was not reviewed in this journal) had typos and sometimes seemed to be compilations less focused on didactics than on summarizing the extant state of the art in meta-analysis. (The 1990 edition also engaged in a little peer bashing.) They also did not exceed the exceptional and parsimonious intellectual contribution of the duo's classic paper (Schmidt & Hunter, 1977). At the same time, others were making substantial methodological contributions (Hedges & Olkin, 1985; Rosenthal, 1991) with at times complementary or divergent approaches. One could not help but wonder which approach or features were correct (e.g., Bangert-Drowns, 1986; Schmidt et al., 1985) and about the use of meta-analysis for the purposes of validity generalization.
In the 2004 edition, Hunter and Schmidt settle many of these issues, and address others. They present new developments in meta-analysis in some detail. Thus, the current edition can safely be said to be a landmark in the study of meta-analysis.
Three influences seem to have guided the improvements in the current edition. First, journals started publishing meta-analyses, which pressured Hunter and Schmidt to focus on teaching about meta-analysis. This revision is better edited...