A Practical Guide to Estimate Statistical Power in Meta-Analysis

Abstract

Statistical power is important in meta-analysis as in primary studies to detect the existing treatment effects in target populations. The mathematical computation of the approximated power (i.e., analytical power) in meta-analysis has been developed. Alternatively, Monte Carlo simulation can be used to estimate statistical power. The current study was conducted to identify the discrepancy between analytical power and simulated statistical power for the hypothesis test of two group means. Results indicated noticeable discrepancies between analytical and statistical power under certain conditions. To conclude, recommendations for the average sample size and number of studies were provided to practical researchers who seek to meet the desired statistical power in meta-analysis. We provided R code for applied researchers to obtain power estimates through simulation.