AN EXAMINATION OF UNIFORM CRIME REPORTS OVER THE PAST 14 YEARS (1964-1977) REVEALS LONG-RANGE CRIME RATE TRENDS WHICH GROW EXPONENTIALLY WITH TIME. THESE DATA, COMBINED WITH THE KNOWN EXPONENTIAL GROWTH IN POPULATION, RESULT IN A LINEAR RELATIONSHIP BETWEEN THE NUMBER OF INDEX CRIMES AND POPULATION. THE FEDERAL BUREAU OF INVESTIGATION HAS PROVIDED A MEANS OF COMPARING THE CRIME RATES IN TWO JURISDICTIONS OR TWO DIFFERENT VALUES OF CRIME RATE OVER TIME WITH A BAYSIAN APPROACH TO INTERVAL ESTIMATION. AN ALTERNATIVE MEANS OF COMPARISON IS PROPOSED. MATHEMATICAL EQUATIONS SHOW THAT THE INCIDENCE OF CRIME FROM ONE YEAR TO THE NEXT FROM JURISDICTION TO JURISDICTION CAN BE COMPARED BY USING REGRESSION ANALYSIS AND COMPUTING CONFIDENCE INTERVALS, THEN COMPARING RESULTS. BY SUCH COMPUTATIONS IT IS POSSIBLE TO DETERMINE WHETHER JURISDICTIONS HAVE A HIGHER OR LOWER INCIDENCE OF CRIME THAN IS TO BE EXPECTED AS A RESULT OF POPULATION CHANGES. TEST RESULTS SHOW THAT CRIME SLOPE IS A MORE CONSISTENT AND USEFUL INDICATOR OF CRIME TRENDS THAN IS CRIME RATE. CRIME RATE, ALTHOUGH EASIER TO COMPUTE, DOES NOT ALLOW COMPARISONS TO BE MADE IN A CLEAR, STATISTICALLY USEFUL MANNER. CRIME SLOPE ANALYSIS REQUIRES MORE LENGTHY COMPUTATIONS, BUT COMPUTERS MINIMIZE THIS DRAWBACK. TABULAR AND GRAPHIC DATA AND MATHEMATICAL EQUATIONS ARE PROVIDED. (RCB)