Date of Award


Document Type


Degree Name

Master of Science (MS)


Department of Mathematics

First Advisor

Pei Geng


A meta-analysis is a tool commonly used to try and gain an understanding of a given topic by using multiple studies conducted on the topic. A key element of properly interpreting the results of a meta-analysis is the test to check for heterogeneity within the studies included. This is currently done using Cochrane’s Q-statistic to test a null hypothesis that the studies included share a common effect size. However, this method has been scrutinized for some of its downsides such as its low power in cases with small sample sizes. This can often create issues because meta-analysis is commonly used in scenarios in which there are few studies. Therefore, in this paper we decide to propose some alternative methods for testing heterogeneity among the studies of a meta-analysis. These methods include using U-statistics because of a few helpful characteristics that may make their interpretation with regard to meta-analysis easier. It is important to note that meta-analysis is a widely used tool in research and is not limited to use in statistical fields. Therefore, the interpretability of the statistic used to test the heterogeneity is crucial in some cases and we aim to make this easier by recommending an alternative to Cochrane’s Q. This proposed method, while simple, will hopefully lay the groundwork for easily interpretable and more accurate tests for heterogeneity using U-statistics.


Imported from Smith_ilstu_0092N_12000.pdf


Page Count