Yearly Count of Recidivism Events by County and Condition (Adjusted for Length of Follow-Up and Time Incarcerated).
Note. SB416 = Oregon Senate Bill (SB) 416 Program. PAU = Probation as Usual. The figure displays predicted yearly count of recidivism by county and intervention condition. The estimates are based on a negative binomial regression model (log link) with a natural log transformed exposure term to adjust for each participant’s length of follow-up and time incarcerated. The incarceration outcomes reflect the count of convictions associated with the incarceration. The Incarceration (All) outcome is based on all incarceration dates that followed randomization, regardless of the associated crime date. The Incarceration (New Crime) outcome is based on incarcerations for crime dates that followed randomization.
*p < .05. **p < .01. ***p < .001.