Fourteen studies (number of participants unclear) were included in the review.
Self-help versus control condition post-treatment (16 comparisons, n=490): there was a statistically significant benefit associated with self-help (ES 0.84, 95% confidence interval, CI: 0.65, 1.02). The fail-safe N calculation suggested that 50.8 null studies would be required to reduce the ES to an insignificant value.
Self-help versus contrast treatment post-treatment (16 comparisons, n=543): there was no significant overall difference between groups (ES -0.03, 95% CI: -0.20, 0.14).
Self-help versus control condition follow-up (5 comparisons, n=130): there was a statistically significant benefit associated with self-help (ES 0.76, 95% CI: 0.09, 1.42).
Self-help versus contrast treatment follow-up (9 comparisons, n=236): there was no significant overall difference between groups (ES -0.07, 95% CI: -0.33, 0.19).
There was considerable heterogeneity within the group of self-help versus control condition follow-up comparisons. All other pooled comparisons were statistically homogeneous.