Meta-analysis: Techniques, applications, and functions

Meta-analysis is an increasingly popular method for conducting a research review. Because of its quantitative basis, it forces reviewers to make explicit a range of decisions that might pass unnoticed in traditional reviews. In exchange, it makes possible a more precise characterization of a research domain. A typical meta-analytic package consists of techniques for (a) combining probabilities across studies, (b) estimating average effect size, (c) determining the stability of results, and (d) identifying factors that moderate the outcome of separate studies. The application of these techniques requires careful attention to a number of potential problems, including biased selection of studies, inadequacies in the studies comprising the data base for the review, and violations of the assumptions of meta-analytic statistical procedures. Notwithstanding these problems, a meta-analysis can advance both theory and application because of its descriptive, diagnostic, predictive, and generative functions. (74 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).