BOTH PROGRAM MANAGERS AND POLICYMAKERS NEED PROGRAM EVALUATIONS THAT APPLY QUANTITATIVE DATA AND TECHNIQUES NORMALLY ASSOCIATED WITH IMPACT ANALYSIS TO PROCESS QUESTIONS ABOUT ASPECTS OF THE PROGRAM WORK. QUANTITATIVE ANALYSIS CAN IMPROVE DESIGN BY LINKING TOGETHER PROGRAM ELEMENTS AND POSTPROGRAM IMPACTS. IT CAN COMBINE PROCESS VARIABLES WITH IMPACT DATA TO ANALYZE THE USEFULNESS OF PERFORMANCE INDICATORS FOR FEEDBACK TO PROGRAM OPERATORS AND CAN TEST OPERATOR ASSERTION OF PROGRAM EFFECTIVENESS. LACK OF DATA, INSUFFICIENT SAMPLE SIZE, AND SELECTION BIAS INHIBIT SUCH LINKED ANALYSES. IN EVALUATION OF THE SUPPORTED WORK PROGRAM--A TRANSITIONAL WORK EXPERIENCE PROGRAM FOR GROUPS WITH GREAT DIFFICULTY IN GETTING OR KEEPING REGULAR JOBS, RESULTS SUGGEST THAT TYPE OF TERMINATION AND LENGTH OF STAY MAY BE SIGNIFICANTLY RELATED TO POSTPROGRAM PERFORMANCE AND CAN PROVIDE POLICY HINTS ON THE LENGTH OF TIME PARTICIPANTS SHOULD STAY IN THE PROGRAM. EVEN MORE IMPORTANT, A PRELIMINARY ANALYSIS OF TREATMENT AND POSTPROGRAM OUTCOME VARIATIONS SUGGESTS HOW FURTHER INQUIRY ABOUT PROGRAM PROCESS MIGHT BE INITIATED. IT COULD INDICATE NEEDED IMPROVEMENTS IN PROGRAM DESIGN OR FORESTALL INVESTMENT IN INEFFECTIVE CHANGES. EMPIRICAL REQUIREMENTS FOR SUFFICIENTLY LARGE SAMPLES (IN THIS CASE, MORE THAN 1,000), AN INTERNAL DATA SYSTEM, FAST FEEDBACK, AND OPERATOR SUPPORT, AS WELL AS INTERACTION BETWEEN QUANTITATIVE AND QUALITATIVE PROCESS ANALYSTS AND PROGRAM MANAGERS, ARE NEEDED FOR A SUCCESSFUL SEQUENTIAL SEARCH TO IMPROVE PROGRAM DESIGN AND IMPLEMENTATION. TABULAR DATA AND NOTES ARE PROVIDED. (AOP)