Meta-Analysis

Meta-analysis entails a set of analytical techniques designed to synthesize findings from studies investigating similar research questions. While meta-analysis includes narrative integration of results, the current chapter will focus only on quantitative meta-analysis. Meta-analysis permits summary of studies’ results and is designed for scenarios in which the primary studies’ raw data are not available. The meta-analytic process involves summarizing the results of each study using an effect size (ES), calculating an overall average across studies of the resulting ESs, and exploring study-and sample-related sources of possible heterogeneity in the ESs. The overall average ES provides a single best estimate of the overall effect of interest to the meta-analyst. Meta-analysis can be used to explore possible differences in ESs as a function of study or sample characteristics. In the seminal article in which the term meta-analysis was coined, Smith and Glass used meta-analysis to summarize results from studies that had assessed the effectiveness of psychotherapy (1977). Thus, treatment effectiveness results provided the first type of ES to be synthesized using meta-analysis. Since the 1970s, the field of meta-analysis has grown to include methods for conducting the synthesis of other types of ESs including correlations, transformations of odds-ratios, validity coefficients, reliability coefficients, and so forth.