This investigation was a meta-analysis of the effectiveness of group sensitivity training. Analysis of 63 studies revealed a moderate size, heterogeneous effect (weighted mean d = 0.62) on all outcome measures. Categorical model testing indicated that group sensitivity training had significantly larger effects on behavioral measures than on self-report measures (mean ds = 1.03 and 0.44, respectively). Moreover, effect sizes for behavioral measures were moderated by the size of treatment groups, the number of sessions, and the precision of measurement recording. That is, interventions involving larger groups and meeting for more sessions had larger effect sizes, as did studies involving more discrete outcome measures. As recommended elsewhere, future studies need to explore group processes and mechanisms of change in group sensitivity training.