BORROWING THE LANGUAGE OF ROLE THEORY, THE PAPER DISCUSSES A RESEARCH AGENDS WITH WHICH TO IDENTIFY WHAT IS KNOWN AND WHAT IS NOT KNOWN ABOUT POLICING. THE KNOWLEDGE IS CONCEPTUALIZED IN ROLE TERMS AS ROLE CONTENT, ROLE ENACTMENT, ROLE CAPACITY, AND ROLE OUTCOMES. IT IS ARGUED THAT MOST OF THE RESEARCH CONDUCTED ON POLICE IN THE UNITED STATES HAS FOCUSED ON THE CAPACITY TO PERFORM AND ON OUTCOMES. THE ISSUES OF WHAT THE POLICE OUGHT TO DO (THE CONTENT) AND HOW THEY OUGHT TO DO IT (THE PERFORMANCE) AND HOW THESE INTERACT WITH THE CAPACITY TO PERFORM REMAIN FOR THE MOST PART UNEXAMINED. AT BEST, UNDERSTANDING OF DEMAND IS STATIC AND INSUFFICIENT, AND INDICATORS OF QUALITY PERFORMANCE ARE HIGHLY SUSPECT. UPGRADING OF POLICE SERVICE DELIVERY NECESSITATES KNOWLEDGE OF THE RELATIONSHIP BETWEEN SELECTION AND TRAINING CRITERIA AND ROLE PERFORMANCE. HOWEVER, EFFORTS TO VALIDATE POLICE SELECTION AND TRAINING STANDARDS HAVE BEEN BASED PRIMARILY IN STUDIES OF ROLE CONTENT. INTERACTION BETWEEN DEMAND, QUALITY, CAPACITY, AND OUTCOME IS SO COMPLEX AND INTRICATE THAT ALL OF THEM MUST BE UNDERSTOOD. FUTURE RESEARCH SHOULD FOCUS ON THE RELATIONSHIP OF POLICE STRATEGY TO THE DEMAND STRUCTURE AND TO THE QUALITY OF POLICE PERFORMANCE. EXTENSIVE REFERENCES ARE PROVIDED.