How meta-analysis increases statistical power.

• DOI: 10.1037/1082-989x.8.3.243
• Publication Date: 2003-12-15T18:57:43Z
• Authors: Lawrence D. Cohn; Betsy J. Becker
• Keywords: Statistical power; Statistical Analysis

One of the most frequently cited reasons for conducting a meta-analysis is increase in statistical power that it affords reviewer. This article demonstrates fixed-effects increases by reducing standard error weighted average effect size (T.) and, so doing, shrinks confidence interval around T.. Small intervals make more likely reviewers to detect nonzero population effects, thereby increasing power. Smaller also represent increased precision estimated size. Computational examples are provided 3 effect-size indices: d (standardized mean difference), Pearson's r, and odds ratios. Random-effects meta-analyses may show smaller However, authors demonstrate number studies random-effects does not always