Reparole is the procedure whereby a parolee or a mandatory releasee violates community supervision standards, is returned to prison, and is again considered for parole. The study findings indicate that the National Council on Crime and Delinquency's standard definition of parole performance categories is not a good indicator of implied or articulated reparole policy. A secondary review of the data was done, using the U.S. Parole Commission's offense behavior severity typology to test conditional release revocation/reparole decisions. Data analysis showed that the introduction of violation decision guidelines generally structured discretion well for minor offense violations; however, there was a tendency to make decisions below the recommended policy guidelines in the more serious reported offense behavior. An alternative reparole policy is recommended that provides for the conditional release violator to be reassessed in terms of 'reparole risk' rather than to be automatically placed in the 'poor risk' category as is customary. The study also examined 'relative adjustment' as it relates to the efficacy of parole supervision, and it is recommended that testing and comparison of the U.S. Parole Commission's Cohort Justice Model data be analyzed in terms of relative adjustment. The study also examines the possibility of agency goal displacement, which involves preoccupation with the maintenance of bureaucratic apparatus at the expense of specified goals. Materials related to the study are appended, and a bibliography of 55 listings is provided. (Author abstract modified)