The population analyzed consisted of all 823 offenders who were placed on furlough during calendar years 1992 and 1993 in the State and who had at least one reclassification. In addition to the reclassification instrument, data came from two other sources, one of which was a special monthly report submitted by the field offices to the central office. The other data source was a needs assessment accomplished by caseworkers on offenders who stayed an appreciable length of time. There were 12 items scored, but this study is limited to five items that were of chief concern to caseworkers and policymakers: sexual deviance, violence proneness, alcohol dependency, drug dependency, and employment history. Findings show that generally the person most likely to have a furlough revoked was a younger offender with a misdemeanor assault or property crime as a current offense. In addition, this offender was more likely to have problems with authority and have a history of rule infractions. Thus, the typical candidate for intensive supervision is often the type of offender who is most likely to fail in the community. Moreover, the current culture of community supervision only exacerbates the problem by encouraging officers to bring violations. The author suggests ways to reduce the prison crowding caused by revocations. 5 tables and 22 references