VIOLENT CRIME AND PROPERTY CRIME DATA WERE DRAWN FROM THE UNIFORM CRIME REPORTS. POLICE EMPLOYMENT RATES WERE DEFINED AS THE NUMBER OF POLICE EMPLOYEES (INCLUDING NONUNIFORMED PERSONNEL) PER 100,000 POPULATION. POLICE EMPLOYMENT DATA WERE ALSO TAKEN FROM THE UNIFORM CRIME REPORTS. THE PARAMETERS OF THE EQUATIONS USED WERE ESTIMATED USING ORDINARY LEAST-SQUARES ESTIMATORS, WITH THE EXCEPTION OF THE EQUATIONS FOR CRIME AND POLICE EMPLOYMENT RATES, WHERE TWO-STAGE LEAST-SQUARES TECHNIQUES WERE USED IN ESTIMATING THE COEFFICIENTS FOR CRIME AND EMPLOYMENT RATES. POLICE EMPLOYMENT WAS FOUND TO BE RESPONSIVE TO CRIME RATES, REGARDLESS OF POPULATION SIZE, DENSITY, RACIAL, AGE, OR SOCIOECONOMIC COMPOSITION. THIS SUGGESTS THAT COMMUNITIES RESPOND TO HIGH CRIME RATES BY ALLOCATING MORE RESOURCES FOR POLICE PROTECTION. POLICE EMPLOYMENT ALSO EXERTED A POSITIVE EFFECT ON CRIME RATES, WHICH WOULD APPEAR TO SUPPORT THE LABELING RATHER THAN THE DETERRENCE PERSPECTIVE. GIVEN THE AGGREGATE NATURE OF THE DATA, HOWEVER, THE DETERRENCE IMPACT OF INCREASED POLICE RESOURCES CANNOT BE DISCOUNTED. VIOLENT CRIME RATES MORE CLEARLY INFLUENCE COMMUNITIES TO INCREASE POLICE RESOURCES, COMPARED TO PROPERTY CRIME RATES. DATA ALSO SUGGEST THAT THE PERCENTAGE OF LOW-INCOME FAMILIES AND BLACKS IN AN AREA INFLUENCE DECISIONS TO INCREASE POLICE RESOURCES. FINDINGS ARE RELEVANT ONLY TO SUBURBAN AREAS AND NOT NECESSARILY TO OTHER ECOLOGICAL UNITS. GRAPHIC AND TABULAR DATA AND REFERENCES ARE PROVIDED. (RCB)