THE THREE DESIGNS OFFERED ADDRESS A NUMBER OF GAPS OF KNOWLEDGE ABOUT HALFWAY HOUSE EFFECTIVENESS WHICH HAVE CONTRIBUTED TO THE PROBLEMS INHERENT IN MAILING POLICY DECISIONS CONCERNING PROGRAM THEORY AND OPERATIONS. ALL THREE UTILIZE A CONTINUOUS OUTCOME VARIABLE DESIGNED TO OVERCOME THE DIFFICULTIES OF POST-RESEARCH; ATTEMPT TO INSTITUTE SOME FORM OF QUASI-EXPERIMENTAL DESIGN THROUGH COMPARISON GROUPS OR STATISTICAL CONTROL; AND COMPARE POST-RELEASE BEHAVIOR BETWEEN ALTERNATE AFTERCARE MODELS. EVALUATION DESIGN NUMBER ONE PROVIDES AN ANALYSIS OF THE ACTIVITIES, ACCOMPLISHMENT OF OBJECTIVES, COST, AND EFFECT ON PROGRAM OUTCOME. EVALUATION DESIGN NUMBER TWO PROVIDES AN ANALYSIS OF COST AND OUTCOME OF RESIDENTIAL INMATE AFTERCARE PROGRAMS BY PROGRAM TYPE AND OFFENDER CHARACTERISTICS. EVALUATION DESIGN NUMBER THREE COMPARES THE POST-RELEASE ADJUSTMENT OF CLIENTS OF RESIDENTIAL AFTERCARE FACILITIES WITH THEIR EXPECTED POST-PAROLE ADJUSTMENT AS DETERMINED BY A PREDICTION TECHNIQUE SUCH AS BASE EXPECTANCY OR SALIENT FACTOR SCORES. PROPOSED BUDGETS FOR EACH DESIGN ARE ALSO INCLUDED. DESIGN NUMBER ONE IS RECOMMENDED SINCE IT IS THE ONLY ONE WHICH ADDRESS THE GAPS IN KNOWLEDGE IN THE AREA OF INTERMEDIATE OBJECTIVES. FOR RELATED PHASE ONE REPORTS, SEE NCJ-36379-83.