LOGISTIC VERSUS HAZARDS REGRESSION-ANALYSES IN EVALUATION RESEARCH - AN EXPOSITION AND APPLICATION TO THE NORTH-CAROLINA COURT COUNSELORS INTENSIVE PROTECTIVE SUPERVISION PROJECT

This article examines and compares the advantages of employing logisti c and hazards regression techniques in assessing both the overall impa ct of a treatment program and the extent to which the impact varies am ong different client subgroups. Data gathered from the North Carolina Court Counselors' Intensive Protective Supervision Project provides a context for the application of these statistical techniques to assess this program's effectiveness. This experimental project was designed t o provide intensive supervision for status offenders who are placed un der the protective supervision of the juvenile courts in North Carolin a. It is found that, although intensive supervision had the desired im pact of reducing the probability of delinquent offenses during the ear ly (first year and a half) phase of the project, this effect deteriora ted over the full three and a half years of the experiment. In additio n, the hazards regression analysis shows that youths under intensive s upervision failed (recidivated with status or delinquent offenses) at a more rapid rate than those under regular supervision. Implications a re drawn for the management and effectiveness of intensive supervision programs.