A field experiment, lasting 26 weeks, was conducted to determine whether ROP increased the likelihood of arrest of targeted repeat offenders relative to the likelihood of their arrest in the ROP's absence. ROP officers' arrests, dispositions, and arrestees' criminal histories were compared with those of a random sample of officers in various other assignments. A total of 212 offenders were paired, and by a coin toss one was ROP assigned and the other was used as a control. The controlled experiment shows that ROP substantially increased the likelihood of arrest for those persons it targeted. ROP arrestees had longer and more serious criminal histories than did the comparison group arrestees. The ROP arrestees also were more likely to be prosecuted and convicted on felony charges and were more likely to be incarcerated. One cost of the program was that ROP officers made only half as many arrests while in ROP than they had before ROP. This cost appears to have been offset by the greater seriousness of the current and prior offenses of ROP arrestees. Thus, by virtually all measures selected to assess the program, the unit appears to have succeeded in selecting, arresting, and contributing to the imprisonment of repeat offenders. On the basis of the results, it is suggested that other large police departments seriously consider creating proactive, repeat offender programs. Tables, graphs, and 50 references are provided. (Author abstract modified)