The study found no significant difference in the likelihood of reconviction between juveniles who received a custodial penalty and those given a noncustodial sentence. These results are inconsistent with the two previous Australian studies on specific deterrence (Kraus, 1974, and Cain, 1996). Both of these studies found evidence that juveniles given custodial penalties were more likely to be reconvicted. The difference between the findings of the previous studies and the current study is probably due to the current study's more effective control for prior criminal record. The current study's findings, along with previous studies that have shown the adverse effects of imprisonment on employment outcomes, suggest that custodial penalties ought to be used sparingly. Data for the current study were obtained from a longitudinal cohort study of juvenile offenders. A total of 152 juvenile offenders were given a detention sentence, and 243 juveniles were given a noncustodial sentence. All juvenile offenders were interviewed at length about their family life, school performance, association with delinquent peers, and substance abuse. Each juvenile was monitored for 6 months after the last juvenile was interviewed in order to determine whether they reoffended. Factors associated with the type of sentence given were controlled, since they might influence the risk for reconviction. Factors included in the multivariate analysis were gender, race, socioeconomic status, age, age at first contact with police, prior criminal record, number of prior commitments to custody, principal offense, number of concurrent offenses, and whether the juvenile was a victim of abuse. Other variables controlled included family structure, parenting style, and level of parental supervision. 4 tables, 20 references, and appended detailed list of factors examined for potential inclusion in the multivariate analysis and their relationship with time to reconviction.