RESEARCH ON DETERRENCE FAIRLY CONSISTENTLY REPORTS FINDING A NEGATIVE STATISTICAL ASSOCIATION BETWEEN CRIME RATES (CR) AND THE CERTAINTY OF PUNISHMENT (PR/CR). THE MONTE CARLO ANALYSIS TESTED THE SENSITIVITY OF THE ASSOCIATION BETWEEN THE TWO RATIO VARIABLES TO VARYING AMOUNT OF CONSTANT AND RANDOM ERROR IN MEASURING THEIR COMPONENTS, THAT IS CR (CRIMES PER CAPITAL) AND PR (PUNISHMENTS PER CAPITA). RESULTS INDICATED THAT THE OBSERVED ASSOCIATION BETWEEN CRIME RATES AND THE CERTAINTY OF PUNISHMENT ARE NOT IN FACT NEGATIVE. THEY DID INDICATE THAT AN ALTERNATIVE EXPLANATION--ERRORS IN MEASUREMENT--COULD HAVE ACCOUNTED FOR THE STATISTICAL ASSOCIATIONS OBSERVED. THE FINDINGS SUGGEST THAT, UNTIL MORE ACCURATE CRIME STATISTICS BECOME AVAILABLE, CAUTION SHOULD BE EXERCISED IN INTERPRETING RESULTS FROM DETERRENCE WORK OF THIS TYPE. WHENEVER POSSIBLE, EFFORTS TO MEASURE THE PROBABLE EXTENT AND CHARACTER OF ERROR IN THE DATA SHOULD BE UNDERTAKEN. AT MINIMUM, ALTERNATIVE CALCULATIONS ALLOWING FOR VARYING LEVELS OF ERROR IN MEASUREMENT SHOULD BE MADE TO TEST THE SENSITIVITY OF THE FINDINGS TO POSSIBLE ERROR SOURCES AND TO ESTABLISH REASONABLE BOUNDS ON DETERRENCE ESTIMATES. TABLES AND REFERENCES ARE GIVEN. (AUTHOR ABSTRACT MODIFIED--MJW)