Bridging Disciplinary Boundaries (January 11 - 14, 2007)

The paper reports on an evaluation of a youth justice team in Scotland with the aim of diverting young people away from crime. A multi-disciplinary team of social workers and psychologists provides group work programs as well as individual counseling to help young people involved in crime to make positive changes in their lives. A secondary aim of the study was to integrate evaluation procedures into the daily practice of the youth justice team to ensure that evaluation findings informed practice.

Method

The evaluation method consisted of single system designs with each young person, using two outcome measures—the risk assessment ASSET and the number of offences. Offences were tracked in a baseline period, during the intervention and a follow-up phase, creating ABA designs for each young person. The data was then converted into a pretest posttest design comparing the baseline with the intervention phase with all the young people. A comparison group design was also used between those young people involved in crime who received the youth justice team intervention and those that did not. Finally, binary logistic regression models were used to investigate the circumstances in which the intervention was more or less likely to be effective

Findings

The ASSET risk assessment was used repeatedly with 69 young people in a three-year period, and it was found that 79% improved on the risk scores. Offence data was available for 93 young people, and it was found that 87% reduced the number of offences during the intervention phase. However, the comparison with those not receiving the youth justice team's services indicated no significant differences in the rate of reduction of offences between the two groups, mainly because the two groups were not matched as those who were persistent offenders were more likely to be referred to the youth justice team. The binary logistic regression model indicated that those who had an offending peer group were 25 times more likely to offend during the intervention then those that did not, and also that those with the worst baseline risk scores were 1.1 times more likely to offend then those with better risk scores.

Implications for practice

The findings indicate that the intervention should be targeted at those with the worst ASSET scores and that it should be developed to deal with negative peer influences. The study also indicates that outcomes, interventions and client circumstances should be systematically tracked and analyzed regularly to inform future practice, and that when evaluation is integrated into practice the findings can help to improve the effectiveness of the interventions. Binary logistic regression models can be a useful tool for practice in investigating the circumstances in which the interventions are more likely to be effective.