Interventions for Institutionalized Juveniles
Of the 200 studies analyzed, 83 dealt with programs for institutionalized youth; of these programs, 74 were in juvenile justice institutions and 9 were in residential facilities under private or mental health administration. Using the same method of regression analysis, researchers examined the same four clusters of variables as in the sample of studies with noninstitutionalized offenders. The clusters associated with the largest variation in method-adjusted effect size were, in decreasing order of magnitude:
This model was further reduced (using the same procedure described earlier) to weed out the weakest variables. This process indicated that the characteristics of institutionalized juveniles accounted for the smallest proportion of effect size variation. This was in contrast to noninstitutionalized juveniles, for whom juvenile characteristics were most important. This means that the conclusions yielded by this model need not be differentiated by juvenile characteristics such as age, gender, ethnic mix, or prior offense history.
Two variables emerged that were important in terms of the amount of treatment provided. First, monitoring to ensure that all juveniles received the intended treatment was essential. Second, the length of treatment was related to the size of treatment effect; that is, the longer the treatment (the average in this sample was 25 weeks), the larger the effects. The type of treatment also was important and is discussed in the next section.
General program characteristics (i.e., the way in which a program is organized, staffed, and administered) were more related to the size of recidivism effects than the type or amount of treatment. The largest treatment effects were found for well-established programs (2 years or older). However, the variable most strongly related to effect size was administration by mental health personnel, in contrast to juvenile justice personnel.
Type of Treatment and Effects on Recidivism
The different types of treatment for institutionalized juveniles were grouped according to the magnitude of mean effect sizes and the consistency of effect sizes. Again, it is important to note that the small number of studies forming the basis of these estimates limits the ability to draw strong conclusions. Two types of treatment showed relatively large, statistically significant mean effect sizes for institutionalized offenders across all estimation procedures: interpersonal skills programs (involving training in social skills and anger control) and teaching family homes (community-based, family-style group homes). Interpersonal skills training was also one of the treatments that had a stronger effect on noninstitutionalized juveniles. Strong, but less consistent, results appeared for multiple service programs, community residential programs (mostly other than juvenile justice programs), and other miscellaneous treatments.
Milieu therapy (in which the total environment, including peers, is structured to support the goals of treatment) showed consistent null results. Drug abstinence programs, wilderness/challenge programs, and employment-related programs did not show statistically significant or consistent mean effects.
The middle tier consisted of behavioral programs and individual, group, and guided group counseling (involving a facilitated group in which members develop norms, give feedback, and make decisions that regulate behavior). Some were statistically significant and some were consistent across the three estimation procedures, but none met all the criteria. In the case of behavioral programs, this may have been because only two studies were included. For the three varieties of counseling, the effect size estimates were inconsistent. Observed effects were confounded with other study characteristics, making it difficult to determine actual treatment effects.