A PAPER PRESENTED TO THE AMERICAN POLITICAL SCIENCE ASSOCIATION DISCUSSES THE PROBLEM OF DEVELOPING INDICATORS BY WHICH PRODUCTIVITY AND PERFORMANCE POLICIES MAY BE EVALUATED. THE SCALE OF PATROL SERVICE TO RESIDENTS IS ONE SUCH INDICATOR. THE AREA IN WHICH A PATROL OFFICER WORKS FOR A YEAR IS DESIGNATED THE PRIMARY ASSIGNMENT AREA (PAA). THE SPECIFICATION OF THIS MEASURE AS A DIMENSION OF DECENTRALIZATION IS DISCUSSED. METHODS OF MEASURING THE POPULATION SIZE OF THE PRIMARY ASSIGNMENT AREA IS OUTLINED AS A MULTISTAGED PROCESS INCLUDING DETERMINING THE BOUNDARIES, THE AREA TO WHICH AN OFFICER IS ROUTINELY ASSIGNED, AND THE DEGREE TO WHICH OFFICERS REMAIN WITHIN THE BOUNDARIES WHILE ON PATROL. WHEN THE BOUNDARIES INCLUDE TERRITORIES, THAT ACCOUNT FOR 70 PERCENT OF THE OFFICER'S TIME, THIS IS THE ADJUSTED PAA. THE POPULATION OF THE ADJUSTED PAA IS THEN DETERMINED. THE DETERMINATION OF PAA BOUNDARIES AS STUDIED IN THIS REPORT RELIED ON INTERVIEWS WITH POLICE ADMINISTRATORS AND PATROL PERSONNEL, DEPARTMENT RECORDS AND INDEPENDENT OBSERVATION BY TRAINED PROJECT RESEARCHERS FROM THE POLICE SERVICES STUDY IN ROCHESTER, N.Y., ST. LOUIS, MO., AND TAMPA-ST. PETERSBURG, FLA. DETAILS OF NEIGHBORHOOD CHARACTERISTICS IN THE THREE AREAS ARE GIVEN, AND ACTUAL DETERMINATION OF BOUNDARIES FOLLOWED GUIDELINES IN THE MODEL. SOME DIFFICULTIES ENCOUNTERED INCLUDED POLICE PERSONNEL ROSTER CHANGES BECAUSE OF PERSONNEL SHORTAGES, LACK OF IDEAL DATA FOR MONITORING ORGANIZATIONAL FEATURES, AND THE COST OF SUCH DATA COLLECTION ON A ROUTINE BASIS. THE APPLICATION OF THE PAA POPULATION IN POLICY ANALYSIS IS SHOWN, WHERE THE POPULATION SIZE OF THE PAA IS THE POLICY VARIABLE AND OBSERVED OFFICER ACTIVITY WITHIN THE STUDY AREA IS THE DEPENDENT VARIABLE. TABULAR DATA AND FOOTNOTES ARE INCLUDED.