Obsessive–compulsive disorder is ranked by the WHO as among the 10 most debilitating disorders and tends to be chronic without adequate treatment. The only psychological treatment that has been found effective is cognitive behavior therapy (CBT). This meta-analysis includes all RCTs (N=37) of CBT for OCD using the interview-based Yale–Brown Obsessive Compulsive Scale, published 1993 to 2014. The effect sizes for comparisons of CBT with waiting-list (1.31), and placebo conditions (1.33) were very large, whereas those for comparisons between individual and group treatment (0.17), and exposure and response prevention vs. cognitive therapy (0.07) were small and non–significant. CBT was significantly better than antidepressant medication (0.55), but the combination of CBT and medication was not significantly better than CBT plus placebo (0.25). The RCTs have a number of methodological problems and recommendations for improving the methodological rigor are discussed as well as clinical implications of the findings. •CBT yielded very large effect sizes compared to wait list and placebo.•CBT was significantly better than antidepressants.•The addition of antidepressants did not potentiate the effect of CBT.•There was no significant difference between ERP and cognitive therapy.•There was no significant difference between individual and group treatment.