Twenty-six studies were included in the review; the number of participants was not stated.
There was not a significant amount of heterogeneity among the studies that could not be explained by the model.
Institutional adjustment: in 12 studies with 13 ESs, the weighted mean ES was 0.66. When one outlier with a large ES result was removed, the overall ES dropped to 0.43 (P<0.0001) for those inmates who received group treatment relative to the control groups. Hence, studies that used true control groups reported improvement for inmates relative to the control.
Anger: in 7 studies with 11 ESs, the weighted mean ES was 0.50. When one outlier with a large ES result was removed, the overall ES dropped to 0.45 (P<0.0001) for those inmates who received group treatment relative to control groups. Hence, inmates who participated in group psychotherapy reported less anger than those inmates who participated in control procedures.
Anxiety: in 3 studies with 7 ESs, the weighted mean ES was 0.85. When one outlier with a large ES result (due to the large number of males in the sample) was removed, the overall ES increased to 0.94 (P<0.0001) for those inmates who received group treatment relative to control groups. Hence, inmates who participated in group psychotherapy reported decreased anxiety compared with control group inmates.
Depression: in 3 studies with 4 ESs, the weighted mean ES was 0.57 (P<0.01) for those inmates who received group treatment relative to the control groups.
Interpersonal functioning: in 6 studies with 6 ESs, the weighted mean ES was 0.74. When one outlier with a large ES result was removed, the overall ES dropped to 0.36 (P<0.05) for those inmates who received group treatment relative to control groups. Hence, inmates who participated in group psychotherapy displayed improved interpersonal functioning compared with inmates in control groups.
Locus of control: in 5 studies with 5 ESs, the weighted mean ES was 0.79 (P<0.001). Publication bias was suspected in this result. Hence, group psychotherapy led to improvements in locus of control for participants receiving treatment relative to the control groups.
Self-esteem: in 4 studies with 7 ESs, the weighted mean ES was 0.52. When one outlier with a large ES result was removed, the overall ES dropped to 0.31 (P<0.05) for those inmates who received group treatment relative to the control groups. Hence, inmates who participated in group psychotherapy experienced increased levels of self-esteem when compared with inmates who participated in control groups.
Supplemental analyses: there were 23 observations available for supplemental analyses. The only secondary variable that was statistically significant was the weighted mean ES for homework exercises (ES=1.22) compared with no homework exercises (ES=0.53). Other secondary variables that failed to reach statistical significance were type of theory implemented, group format, and reason why inmates participated in therapy groups.