| ||Literature Coverage Dates||Number of Studies||Number of Study Participants|
|Meta-Analysis 1||1993 - 2010||14||3226|
Henwood, Chou, and Browne (2015) analyzed the impact of cognitive behavioral therapy (CBT)–based anger management on adult male offenders sentenced to probation or incarceration. To be eligible for inclusion, the studies had to use random assignment or quasi-random assignment to the CBT program. Studies that did not use random assignment had to use a matched control, wait-list control, intent-to-treat group, or a comparison to attrition group. Eligible studies must have had a sample that comprised at least 50 percent adult offenders with a history of violence, or those who had been screened for, and found to have, dysfunctional anger. Studies had to test the effectiveness of CBT-based treatment for anger or violence delivered in prison or the community and include an outcome of general reoffending and/or violent reoffending. It is important to note that studies that focused solely on domestic violent offenders or offenders with a mental health diagnosis were excluded from the review.
To identify studies, keywords were used to search the following databases: the Cochrane Library, the Campbell Collaboration, Medline, PsychInfo, ASSIA, SCOPUS, and Web of Science. Additionally, reference lists of relevant systematic reviews were hand-searched, as well as governmental portals such as the Canadian Correctional Service, Australian Institute of Criminology, and the UK Ministry of Justice. The search yielded 3,362 references; however, after removing duplicates, book chapters, meta-analyses, opinion papers, or other irrelevant studies, 104 references remained. Of these 104 references, 14 studies met the eligibility criteria. These studies were derived from 13 articles (in other words, one article contained two studies using different samples, measures, and results, thus making it eligible for inclusion).
The included studies were published between 1993 and 2010 and were conducted in the United States (N=1), Canada (N=9), United Kingdom (N=1), and New Zealand (N=3). All included studies used quasi-experimental designs, and a mean age of participants that ranged from 23.5 to 35.6 years old. The offense history and demographics of participants varied widely; however, participants in all studies were male. Three of the 14 studies screened offenders for dysfunctional anger prior to treatment; however, all but one study focused on violent offenders. The one study that did not focus on violent offenders had a sample that was 40 percent nonviolent offenders (i.e., individuals who had no violent history). The risk level of the included participants also varied: some studies included only high-risk offenders and others measured both high- and low-risk offenders. Eleven of the studies took place in prison, and the other three were community-based programs. Although all the included studies used CBT components, the studies varied in their specific focus. For example, some programs combined anger management components with substance abuse treatment, while some also targeted other criminogenic needs (such as victim empathy). Furthermore, while the studies varied in duration, the mean length of the intervention across all studies was 190 hours. Finally, all included studies reported results on general and/or violent recidivism.
Given the level of diversity in programs included in the meta-analysis, an inverse-variance random effects model was used to account for the heterogeneity across studies. Effect sizes were calculated for each outcome and then averaged to create a mean effect size for each outcome, also known as a standardized mean difference.