Dear colleague's, What is the appropriate way to incorporate error associated with replicate studies in meta-analysis when the CV of effect sizes is used instead of individual effect sizes themselves? That is, we are calculating CV of multiple effect sizes for a single study that measured the same response variable on different occasions (e.g., in different years). We would like to weight each study by the variance associated with each study. In typical meta-analysis, a pooled variance is calculated for each study which incorporates sample sizes and standard deviations for each treatment (in some cases at least). However, it is not clear how to weight each study when the CV is used as the response variable. Thanks for your help!
Scott Chamberlain EEB, Rice Univ.
