Heterogeneity Test

CORE DEFINITION

A heterogeneity test in meta-analysis examines whether the effect sizes reported by separate studies vary more than chance would explain. The Q statistic tests for the presence of such variation, and I squared reports the proportion of total variability attributable to it rather than to sampling error. The core proposition is that pooling is only meaningful when the studies estimate a common effect, so heterogeneity must be assessed before any combined estimate is reported. The key qualification is that statistical homogeneity does not establish clinical homogeneity, since studies may measure the same number while differing in population, dose or outcome definition.