STATISTICAL TESTING OF THE DETERRENCE HYPOTHESIS BEGAN IN EARNEST IN 1975, WITH ISAAC EHRLICH'S APPLICATION OF ECONOMETRIC TECHNIQUES TO TIME-SERIES DATA. EHRLICH'S TIME-SERIES STUDY, WHICH WAS LATER CONFIRMED IN A SEPARATE ANALYSIS OF CROSS-SECTIONAL DATA, FOUND A DETERRENT EFFECT FOR THE DEATH PENALTY AND SINCE HAS BEEN CRITICIZED ON GROUNDS OF DATA IMPERFECTIONS, METHODOLOGICAL PROBLEMS, AND QUESTIONABLE ASSUMPTIONS. THE CONTROVERSY OVER EHLICH'S WORK HAS STIMULATED THE DEVELOPMENT OF SEVERAL MODELS DESIGNED TO TEST THE CAUSAL RELATIONSHIP BETWEEN THE DEATH PENALTY AND HOMICIDES, USUALLY WITH A VIEW TO REFUTING EHRLICH'S CONCLUSIONS. IN THE PRESENT ATTEMPT TO ISOLATE, IDENTIFY, AND MEASURE ECONOMETRICALLY THE DETERRENT IMPACT OF CAPITAL PUNISHMENT ON HOMICIDE RATES, MANY OF THE CRITICISMS OF PRIOR DETERRENCE RESEARCH ARE TAKEN INTO ACCOUNT, AND HYPOTHESES ARE DESIGNED AND SPECIFIED TO PROVE OR DISPROVE WHETHER CAPITAL PUNISHMENT IS A DETERRENT TO CRIME. THE ANALYSIS IS PERFORMED AT STATE AND REGIONAL LEVELS AND EMPLOYS A STATISTICAL TECHNIQUE (ORDINARY LEAST SQUARES IN A POOLED CROSS-SECTIONAL TIME-SERIES FRAMEWORK) THAT COMBINES THE ADVANTAGES OF TIME-SERIES AND CROSS-SECTIONAL DATA. ONLY PROBABILITY OF APPREHENSION IS SHOWN TO HAVE A SIGNIFICANT DETERRENT EFFECT, AND THAT ONLY AT THE STATE LEVEL. NO DETERRENT EFFECT IS FOUND FOR THE DEATH PENALTY. THE ANALYSIS DOES NOT PROVE THAT EXECUTIONS DO NOT DETER, BUT IT DOES LEND SUPPORT TO THE THOUGHT THAT THE PRESENCE OR ABSENCE OF A DETERRENT EFFECT FOR THE DEATH PENALTY MAY NOT BE DEMONSTRABLE BY STATISTICAL MEANS, GIVEN AVAILABLE DATA AND PROBLEMS IN ESTIMATION. THUS FAR, THE ONLY EVIDENCE FAVORING A DETERRENT EFFECT IS EHRLICH'S. IT DOES NOT SEEM REASONABLE TO WEIGH LIVES AGAINST QUESTIONABLE STATISTICS AND FINDINGS IN FAVOR OF THE NUMBERS. DETAILS OF THE ANALYSIS AND SUPPORTING DATA ARE PROVIDED. (LKM)