The goal of the ISP is to offer an alternative to incarceration that assures public safety and addresses offender needs in a cost-effective and less restrictive manner than incarceration. Quantitative data on the ISP were obtained from automated criminal justice data bases and a supplemental data collection form developed specifically to evaluate the program. Qualitative data were obtained primarily from interviews conducted with individuals involved in ISP operation and use. The evaluation effort also examined probation and parole files of a stratified sample of ISP offender terminations and an equivalent matched control group. Program evaluation findings demonstrated that offenders in the ISP group had higher risk and need scores than controls. The ISP fulfilled contact and supervision requirements, although ISP officers did not feel adequately equipped and/or trained to perform their jobs safely and effectively. The ISP appeared to be more effective in districts operating team programs than in districts operating with a single ISP officer. Offenders in the ISP group received higher levels of drug abuse treatment than controls. Methods of documenting offender treatment progress, however, were inadequate and poorly communicated to supervising ISP officers. The lack of a clearly defined ISP target population and a standard offender selection method led to a diverse program population. ISP officers did not have sufficient training and resources to appropriately manage diverse ISP caseloads. Almost one-third of ISP offenders were rearrested while on active ISP supervision, and nearly 11 percent of rearrests were for felony offenses against the person. ISP offenders were not more or less likely to be rearrested while on active supervision than controls. The ISP proved to be significantly less costly than incarceration and was only slightly more expensive than regular offender supervision. Recommendations are offered on how to improve the ISP. Appendixes contain the data collection and interview instruments and additional information on the study methodology. 30 references and 20 displays