In the present study, a growth mixture modeling (GMM) analysis performed on 7117 low-to-moderate-risk justice-involved youth disclosed the presence of four latent classes of neutralization scores measured over four non-overlapping six-month periods. These four patterns were subsequently labeled low accelerating, moderate accelerating, high accelerating and decelerating. The four classes displayed differing levels of recidivism, and so a series of binary logistic regression analyses were computed in which each individual class was compared to every other class. The results of these binary logistic regression analyses revealed that recidivism occurred significantly more often in the high accelerating group than in the low and moderate accelerating groups and significantly less often in the decelerating group than in the low, moderate and high accelerating groups, even after Time 1 moral neutralization scores were controlled. These results corroborate prior research denoting that declining moral neutralization scores portend decreased odds of future recidivism.
Read full abstract