Evaluation of Statistical Power in Random Effect Meta Analyses for Correlation Effect Size

Year 2022, Volume: 26 Issue: 3, 554 - 567, 30.06.2022

Burçin Öner

https://doi.org/10.16984/saufenbilder.1089793

Cited By: 1

Abstract

In meta-analysis, numerical index is used as an estimate of effect size to describe the results of each study and thereafter these estimates of across studies are combined to obtain summary of results.
It should be known that calculations of the power of statistical tests are important in planning research studies and for interpreting situations in which a result has not proven to be statistically significant. Although statistical power is often considered in the design of primary research studies, it is rarely considered in meta-analysis. Despite the importance of statistical power, few studies have been examined the performance of simulated power in meta-analysis. (In this study, calculations of statistical power for statistical tests that are used for unequal sample size on random effects model in meta-analysis using correlation coefficient as effect size were conducted.)
The power of the test for the overall effect size was calculated by using both analytical method and simulation method. Thus, it was investigated whether there was any difference between the simulation power and analytical power in random effects meta-analysis by using correlation coefficient as an effect size.

Keywords

meta-analysis, simulation power, correlation, analytical power

References

• [1] F. Mosteller, G.A. Colditz, “Understanding research synthesis (Meta–analysis)”, Annual review of public health, vol. 17, no. 1, pp. 1-23, 1996.
• [2] L. V. Hedges, T. D. Pigott, “The power of statistical tests in meta-Analysis”, The American Psychological Association, vol. 6, no.3, pp. 203-217, 2001.
• [3] A. P. Field, “Meta-analysis of correlation coefficients: A Monte Carlo comparison of fixed- and random-effects methods”, Psychological Methods, vol. 6, no. 2, pp. 161-180, 2001.
• [4] L. D. Cohn, B. J. Backer, “How meta-analysis increases statistical power”, The American Psychological Assosiation, vol. 8, no. 3, pp. 243-253, 2003.
• [5] L. V. Hedges, T. D. Pigott, “The power of statistical tests for moderators in meta-analysis”, The American Psychological Association, vol. 9, no. 4, pp. 426–445, 2004.
• [6] G. Cafri, J. D. Kromrey, M. T. Brannick, “A SAS macro for statistical power calculations in meta-analysis” Behavior Research Methods, vol. 41, no. 1, pp. 35-46, 2009.
• [7] J. C. Valentin, T. D. Pigott, H. R. Rothstein, “How many studies do you need? A primer on statistical power for meta-analysis”, Journal of Educational and Behavioral Statistics, vol. 35, no. 2, pp. 215-247, 2010.
• [8] J. Liu, “Statistical power in meta-analysis”, Ph.D. dissertation, Dept. Educational Studies, South Carolina University, Carolina, USA, 2015.
• [9] J. Liu, F. Pan, “A SAS Macro to investigate statistical power in meta-analysis”, in Paper presented at the Southeast SAS Users Group, Savannah, GA, pp. 109-114., 2015.
• [10] A. P. Field., “Meta-analysis of correlation coefficients: A Monte Carlo comparison of fixed- and random-effects methods”, Psychological Methods, vol. 6, no. 2, pp. 161-180, 2001.
• [11] H. Gamgam, Ö. Ünver, B. Atunkaynak, “Temel istatistik yöntemler”, Seçkin Yayıncılık, Ankara, 2011.
• [12] S. C. Chow, J. Shao, H. Wang, “Considerations Prior to Sample Size Calculation” in Sample size calculations in clinical research, Second ed., New York, USA, Boca Raton: Taylor & Francis, 2008, 25-49.
• [13] F. Schmidt, “What do data really mean? Research findings, meta-analysis, and cumulative knowledge in psychology”, American Psychologist, vol. 47, no. 10, pp. 1173-1181, 1992.
• [14] L. Hedges, I. Olkin, “Statistical methods for meta-analysis”, San Diego, CA, USA, Academic Press, 1985, pp. 224-237.
• [15] J. Cohen, “Statistical power analysis for the behavioral sciences”, Hillsdale, NJ, USA, Erlbaum, 1988, pp. 47-51.
• [16] A.P. Field, “The problems of using fixed-effects models of meta-analysis on real-world data”, Psychological Methods, vol. 6, no. 2, pp.161-18, 2003.
• [17] M. Borenstein, L. V. Hedges, J. P. T. Higgins, H. R. Rothstein, “Effect Sizes Based on Correlations” in Introduction to meta-analysis, First ed., United Kingdom, John Wiley & Sons, 2009, pp. 41-44.