The 3R program combined competency-based, individualized instruction with the "open-entry/open-exit" feature provided by computer software. The program planners intended to take advantage of this feature by providing for the program's continuation in the community. Although CAI proved to be an effective method for delivering education -- as measured by GED retesting pass rates for inmates in the program, the percentage of GED certificates earned, the large number of participants who continued 3R in the community, and a lower recidivism rate than those who did not participate -- the program ended August 31, 1991, after more than 19 months of operation. The termination resulted from complex interagency interactions and lack of a cohesive planning structure; these failings led to the placement of too few and the wrong type of offenders in the program. Program administrators did not establish procedures during the diagnostic/classification process for early identification of eligible inmates. The major lesson to be learned from 3R's failure is the importance of designing effective multiagency planning, coordination, and implementation mechanisms for correctional treatment programs. 12 charts, 9 notes, and 16 references