Study 1
Weinrath and colleagues (2016) conducted a quasi-experimental designed study, to compare youth in the Spotlight Serious Offender Services program in the Canadian city of Winnipeg, Manitoba, with a comparison group of youth selected, using propensity score matching (PSM). The original samples consisted of 109 youths receiving program services and an initial comparison group of 217 youths. Youths were selected based on custodial release in 2006 and 2007. Reinvolvement in the criminal justice system was compiled up to a cut-off date of July 31, 2009. For the PSM comparison, there were 57 youths in the Spotlight (intervention) group and 85 youths in the comparison group. Comparison group youths were also high-risk gang-affiliated youths but did not receive Spotlight services for a variety of reasons (e.g., participation in another form of treatment, not referred by the probation officer, or were released from custody without a follow-up supervision requirement).
All participants were males ages 12 to 19. At admission, Spotlight intervention group participants had a mean age of 15.5, and comparison group participants a mean age of 15.3. In both the intervention and comparison groups, the great majority of participants were Indigenous (86.0 percent and 85.9 percent, respectively); the other 14 percent of participants were categorized as “other.” Treatment and comparison groups did not statistically significantly differ at baseline on their score on the Primary Risk Assessment (a version of the Youth Level of Service Inventory) or on any other baseline characteristics.
The outcome of interest was recidivism. The primary outcome was legal reinvolvement through court convictions. Crime severity also was assessed using the Manitoba three-category severity scale: high (e.g., homicide, aggravated assault), medium (e.g., assault, break and enter), and low (e.g., drug possession, theft). A fourth separate category was added that reflected breach of probation. Time to failure was coded as the length of time after release/assignment to the program before a new offense was charged.
Program effects on new convictions were evaluated using logistic regression analysis. Survival analysis was used to evaluate potential effects on length of time that reoffending was avoided. No subgroup analysis was conducted. A subset of youths in the intervention were interviewed, with responses indicating most youths felt they had a positive relationship with their mentor. Informal observations of street mentors indicated that the mentors had positive interactions with youth.