Authors: Angela Hawken, Ph.D., Jonathan Kulick, Ph.D., Kelly Smith, Jie Mei, M.P.P., Yiwen Zhang, M.P.P., Sara Jarman, Travis Yu, Chris Carson, Tifanie Vial
Hawaii’s Opportunity Probation with Enforcement program relies on a regimen of regular, random drug testing tied to swift and certain, but modest, sanctions to motivate probationer compliance.
In two 2007 studies in Hawaii, HOPE was demonstrated to improve compliance with terms of probation at a 12-month follow-up, with large reductions in drug use, recidivism, and overall incarceration for offenders assigned to the program.
Following the original evaluations, HOPE was substantially expanded. The program grew from 34 participants in 2004 to approximately 2,200 in 2014.
This study extends the original HOPE evaluations to an almost ten-year follow up, addressing whether the improvements in criminal-justice outcomes observed during the active HOPE intervention persist after the term of probation. The study also documents changes in HOPE practices and ongoing implementation fidelity to the model.
Principal findings include:
1. HOPE probationers performed better than those supervised under routine supervision. They were less likely to be revoked and returned to prison.
2. Probationers’ perception of risk of punishment given a violation was higher than probation officers’ estimates, which in turn were higher than researcher estimates of the true risk. HOPE appears to benefit from a reputation effect that exceeds the certainty delivered in practice.
3. Probation-officer surveys suggest that probation officers support HOPE. It makes them more effective at their job and their probationers are more likely to succeed on HOPE.
Results of this project should inform legislation, policies, and practices in community supervision, as they show that the better outcomes in HOPE versus probation as usual persist to a large degree in some measures, modest sanctions can be effective in a HOPE probation program, and fidelity of implementation can decline once implementation is routine.