An exact confidence interval for a common effect size
Approximative confidence intervals for a common standardized mean difference from a series of independent experiments are reasonably accurate when effect sizes are less than 1.5 in absolute magnitude and the sample sizes in each group are at least 10. In this article, we derive an exact confidence interval for this effect size where the bounds have to be determined by solving nonlinear equations. As a by-product, we obtain a median unbiased estimator of the common effect size.
KeywordsMeta-analysis parameterized p value standardized difference of normal means homogeneity of effect sizes
AMS Subject Classification62F25
Unable to display preview. Download preview PDF.
- Hollands, G. J., I. Shemilt, T. M. Marteau, S. A. Jebb, H. B. Lewis, Y. Wei, J. P. T. Higgins, D. Ogilvie. 2015. Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Database of Systematic Reviews 14(9):CD011045. doi:10.1002/14651858. CD011045.pub2.Google Scholar