NotesFAQContact Us
Search Tips
ERIC Number: ED447192
Record Type: Non-Journal
Publication Date: 2000-Apr
Pages: 18
Abstractor: N/A
Reference Count: N/A
An Empirical Study of the Effect of Pooling Effect Sizes on Hedge's Homogeneity Test.
Kim, Jong-Pil
The homogeneity test provided by L. Hedges (1982) in meta-analysis has been widely used, mainly to test if the effect sizes share the same variance. Ignoring the intercorrelations among effect sizes affects the Type I error rate. The main purpose of this research was to study the impact of pooling effect sizes on the homogeneity test in effect size analyses. Simulations were conducted to study the effects of pooling effect sizes on Type I error rate and power of the Q test for varying sample size (N), number of studies (k), and proportion of pooling effect sizes (p) in the k studies. Simulation results show that composite meta-analysis seems to have smaller Type I error than typical meta-analysis. The difference in Type I error between typical and composite meta-analysis is relatively big when sample size or number of studies is large. Composite meta-analysis always has greater Type II error and smaller power than typical meta-analysis. These results mean that more caution is necessary when pooling effect sizes. (Contains 2 tables, 3 figures, and 16 references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A