Practice Profile: Cognitive–Behavioral Interventions for Justice-Involved Juveniles and Young Adults Adjudicated for an Offense in Europe

Using meta-analytic techniques, Koehler and colleagues (2013) analyzed the effect of treatment programs on recidivism for juveniles and young adults adjudicated for an offense in Europe, and conducted a moderator analysis to assess the differential effectiveness of the various treatment types including cognitive behavioral interventions. To identify studies, numerous online databases were used, including International Bibliography of the Social Sciences, PSYCINFO, PSYCHARTICLES, PUBMED, COCHRANE Library, EMBASE, ISI Web of Knowledge, Cambridge Scientific Abstracts, ERIC, and Science Direct. A variety of key search terms were used, such as juvenile, delinquent, therapy, rehabilitation, cognitive, mandated, outcome, recidivism, and reoffending.

To be eligible for inclusion in the meta-analysis, studies had to be evaluations of interventions that contained a circumscribed correctional program that aimed to reduce reoffending. The literature search focused on evaluations conducted in Europe that targeted juveniles and young adults adjudicated for an offense, defined as previously adjudicated youths up to age 25. Published and unpublished evaluations that appeared between the years 1980 and 2009 were considered, and any report in a commonly used European language was deemed eligible. To be eligible, the evaluation had to compare the effect of the intervention in a treatment group with the level of reoffending in a control group. The control group could be defined as no treatment, treatment as usual, or an alternative treatment, but there had to be a clear indication of equivalence between groups (this could be achieved through randomization, matching, or statistical comparisons of equivalence). Evaluations had to include a form of reoffending as an outcome, which included both formal legal measures (rearrest, reconviction, reincarceration, and revocation of probation or parole) and self-reported data pertaining to crime. Measures of lower level offenses that do not constitute as crime, such as antisocial behavior, were not included. Finally, for the study to be considered eligible, effects had to be reported as any common statistics or raw data that would allow for the calculation of effect sizes.

The sample of studies comprised 21 evaluations with 25 discrete comparisons between treatment and control groups. The sample consisted of a total n of 7,940 juveniles and young adults adjudicated for an offense, with 3,883 in the treatment groups, and 4,057 subjects in the control groups. The mean age of participants was 15 and younger for 8 studies, ages 16 to 20 for 14 studies, and ages 21–25 for 3 studies. The mean age of participants across all studies was 17.9 years. The sample was more than 90 percent male in 8 studies, less than 90 percent male in 11 studies, and not reported in 6 studies. Most evaluations were conducted in the United Kingdom (64 percent), but evaluations from the Netherlands (16 percent), Germany (8 percent), Norway (8 percent), and Sweden (4 percent) also were included. Of the 25 studies, 16 percent were random experiments. The remainder had either moderate or strong statistical control (52 percent and 32 percent respectively). Most interventions were community based (68 percent) and voluntary (56 percent). Interventions that were evaluated included behavioral/cognitive–behavioral interventions (11 studies), nonbehavioral (10 studies), and intensive supervision/deterrence (4 studies). The CrimeSolutions review focused on the 11 studies that evaluated behavioral or cognitive–behavioral interventions.

The effect sizes for the outcome variable (reoffending) were calculated using Cohen’s d, which were converted into an odds ratio. When necessary values were omitted from the study report, the effect size was calculated based on available p, t, or F values.