EVALUATIONS OF INTERVENTION PROGRAMS CONCENTRATE ON THE ADEQUACY OF RESEARCH DESIGN AND THE SPECIFICATION OF OUTCOME, WHILE TENDING TO IGNORE THE INTEGRITY OF THE PROGRAMS BEING EVALUATED. ASSESSMENT OF PROGRAM INTEGRITY INVOLVES INFORMATION ON THE ADEQUACY OF THE CONCEPTUALIZATION OF TREATMENT; THE DURATION AND INTENSITY OF THE PROGRAM; THE QUALITY AND NUMBER OF PERSONNEL; AND THE MATCH OF TREATMENT, TREATER, AND TREATED. A LARGE-SCALE STUDY OF THE EFFICACY OF GROUP COUNSELING, CONDUCTED IN 1971, EMPLOYED EXEMPLARY RESEARCH DESIGN AND PROVIDED CONSIDERABLE INFORMATION ON PROGRAM INTEGRITY. DETAILED ANALYSIS OF THE STUDY DEMONSTRATES HOW THE LACK OF PROGRAM INTEGRITY MADE IT IMPOSSIBLE TO DRAW CONCLUSIONS ABOUT THE EFFICACY OF THE INTERVENTION. IT IS CONCLUDED THAT CONTINUED LACK OF ATTENTION TO PROGRAM INTEGRITY WILL LEAD TO CONTINUED, GENERALLY UNPRODUCTIVE DEBATE ABOUT 'WHAT WORKS.' IT IS RECOMMENDED THAT TECHNIQUES BE DEVELOPED FOR THE OBJECTIVE ASSESSMENT OF THE ELEMENTS OF PROGRAM INTEGRITY FOR COMMONLY USED TREATMENT STRATEGIES IN CORRECTIONS. (AUTHOR ABSTRACT MODIFIED).