RTA Re-Arrests

● Nearly half of 16-year-olds arrested in the first year of RTA were re-arrested as of January 31, 2020 (48%), including more than a third (35%) who were re-arrested for a felony and more than a quarter (27%) who were re-arrested for a VFO. These rates of re-arrest are higher than the re-arrest rates for 16-year-olds arrested in the comparison period and higher than those for 17-year-olds in either period: 38% to 40% were re-arrested, 25% to 28% were re-arrested for a felony, and 16% to 19% were re-arrested for a VFO, as of January 31 of the year following the October to September arrest period.

Multivariate Analyses

● To identify factors associated with re-arrest, a series of Cox regressions were performed in which many possible explanatory variables were tested as potential predictors. This type of survival analysis was selected specifically because cases that vary in time at risk can be included. The analysis examines how factors affect the rate of “failure” (re-arrest) over time.

● Cox regression, rather than the more familiar logistic regression, was used in this research because this type of proportional hazards analysis can estimate the relative speed with which re-arrest occurred while logistic regression can only address whether a youth was re-arrested.

● Two sets of models were developed: One set assessed the length of time without re-arrest, felony re-arrest, or VFO re-arrest among all arrests regardless of the prosecution status and the second set of models examined the length of time to first, first felony and first VFO re-arrest for only those youths who were arraigned in the adult court. The “arraigned” models only include youths charged with felonies because in accordance with RTA no youths charged with misdemeanors are processed in the adult court.

● The models reached the highest levels of statistical significance based on the overall chi-square statistic (which tests the hypothesis that all regression coefficients for the variables in the model are identically zero), which was statistically significant at p≤.001 in all six models.

● The six models are very similar to one another with nearly the same variables predicting survival without re-arrest over time. The predictor variables include gender, race, severity of the arrest charge, having more than one arrest stemming from the arrest event, assault charge at arrest, and, in the ‘arraigned’ models, having any open cases.

● The Cox multivariate survival analysis indicates that the RTA age/year is a statistically significant predictor of re-arrest over time in both sets of models after accounting for the other included variables. Youths aged 16 arrested during the first year of RTA were at significantly greater risk of re-arrest, felony re-arrest and VFO re-arrest than were 17-year-olds arrested that year, as well as 16- or 17-year-olds arrested the previous year. In the ‘arraigned’ models, 16-year-olds arraigned in adult court for a felony-level offense in the first year of RTA were at a significantly greater risk for re-arrest, felony re-arrest and VFO re-arrest than were 17-year-olds arraigned on felony charges that year or the previous year but were not at a significantly greater risk for any re-arrest than were 16-year-olds arraigned on felony charges the previous year.