2006 年 33 巻 2 号 p. 149-177
In this article we study effectiveness of training school programs relative to probation on recidivism for Japanese juvenile delinquents with differing criminal experiences—early versus late involvement, and first-time as opposed to repeat. We measure effectiveness by the times elapsed from release to reincarceration in the Juvenile Classification Homes. We employ proportional hazards model to examine the relationship between the form of treatment and the times to reincarceration. The effect of treatment in training school is found to be complex. Institutionalization of the late-starting first-time offenders (N = 5,675) and of the early-starting repeat offenders (N = 158) is associated with significantly longer times to reincarceration. But for the late-starting repeat offenders (N = 946), we do not find any positive effect of training school.