Date of Award

12-14-2015

Document Type

Open Access Dissertation

Department

Educational Studies

First Advisor

Xiaofeng Liu

Abstract

Statistical power is important in a meta-analysis study, although few studies have examined the performance of simulated power in meta-analysis. The purpose of this study is to inform researchers about statistical power estimation on two sample mean difference test under different situations: (1) the discrepancy between the analytical power and the actual power and (2) the influence of unequal sample size and unbalanced design on the power. Results indicated that there are noticeable discrepancies between the estimated power and actual power under certain conditions. In general, unbalanced design decreases the statistical power in the meta-analysis. Recommendations are provided for researchers who are interested in power of meta-analysis.

Rights

© 2015, Jin Liu

Share

COinS