We conducted a meta-analysis using 20 randomised controlled studies of cognitive behavioural therapy (CBT) for anxiety disorders in children and adolescents. The mean pre-post effect size was d = 0.94, which was maintained at follow-up. The mean effect size when comparing the CBT and control group was d = 0.61. Within the CBT group, the mean effect size of university clinics (d = 0.77) was larger than that of other clinics (d = 0.37). The difference in effect sizes was hardly noticeable when comparing CBT with family or parents and CBT with child only (d = 0.03). Further studies are required to examine the effectiveness of family CBT versus child CBT.