Fourteen studies (more than 495 participants) were included in the review: seven controlled studies (276 participants), of which six were randomised; and seven before/after studies (more than 219 participants; one study did not report the total number).
Most of the controlled studies were medium to high quality. Most of the before/after studies were medium to low quality. Sample sizes ranged from seven to 157 participants.
Most of the included studies found an immediate beneficial effect of the cognitive-behavioural intervention. Effect sizes, where reported, ranged from 0.03 to 1.16 for anger (nine studies), 0.28 to 3.46 for aggression (11 studies), 0.4 to 0.59 for behaviour (nine studies), 0.33 to 0.62 for anger-related social skills (three studies) and 0.33 to 1.77 for self-control (three studies).
Six studies reported longer-term follow-up. The two studies with four- to six-week follow-up found no significant improvements; the four studies with three to nine months' follow-up found beneficial effects of the intervention. It was not possible to calculate an effect size in most of these studies.