program participants over the control group. Overall, Lipsey (1995) found that programs that targeted behavioral change in a relatively structured and concrete manner had a greater effect on reducing delinquency than programs that targeted psychological change through traditional counseling or casework approaches. Other meta-analyses have similarly found that cognitive-behavioral, skill-oriented, and multimodal programs have the best effects (Gottschalk et al., 1987; Mayer et al., 1986). This pattern held for programs conducted under the auspices of the juvenile justice system and for those run by other institutions.
Of particular concern are programs that increased delinquency. Lipsey (1995:74) says about them:
Most notable are the deterrence approaches such as shock incarceration. Despite their popularity, the available studies indicate that they actually result in delinquency increases rather than decreases. Unfortunately, there are distressingly few studies in this category, making any conclusions provisional. The studies we do have, however, raise grave doubts about the effectiveness of these forms of treatment.
A systematic review of evaluations of deterrence programs, such as Scared Straight, that involve exposing youngsters who have come in contact with the juvenile justice system to prison life and adult inmates was undertaken by Petrosino and colleagues (2000). None of the nine evaluations that involved random assignment of youngsters to the treatment or control groups found any positive effect on future delinquency. Seven of the studies found that the effects of the program were harmful, that is, youngsters in treatment were more likely to commit additional delinquent acts than were those in the control group who received no treatment.
Lipsey (1995) also found that the length of the program and how well it was planned and delivered affected how well the program reduced delinquency. Programs that were monitored to ensure that they were delivered as planned had larger effects than programs that were not monitored. More of an otherwise effective program appears to be better than less. In general, Lipsey (1995) recommended that programs should have 100 hours or more of total contact with the juvenile, delivered at two or more contacts per week, over a period of 26 weeks or longer. Because the average length of stay for juveniles in residential placement is less than four months (Smith, 1998)—significantly shorter than 26 weeks—it may be difficult to provide programs over a sufficient length of time to make a difference for many youth in residential placement. Continuity of programming after release may be a way to increase effectiveness. It should be noted, however, that Lipsey and Wilson (1998) found that characteristics of effective programming both inside and outside institutions differed.