IN A PREVIOUS EXAMINATION USING STEPWISE REGRESSION TECHNIQUES, SOME QUESTIONS WERE RAISED ABOUT THE STATISTICAL PROCEDURES EMPLOYED. THE STUDY FAILED TO SHOW THAT THE CERTAINTY OR SEVERITY OF IMPRISONMENT OR THE CERTAINTY OF EXECUTION WERE SIGNIFICANTLY RELATED TO HOMICIDE RATES FOR 1950 AND 1960 (THE ONLY YEARS WHERE ADEQUATE IMPRISONMENT, EXECUTION, SOCIODEMOGRAPHIC, AND HOMICIDE DATA WERE AVAILABLE. TO EXAMINE THE RELIABILITY OF THE PREVIOUSLY REPORTED FINDINGS, THE DATA WERE REANALYZED USING LEAST-SQUARES REGRESSION PROCEDURES. WITH THE EXCEPTION OF THE 1950 FINDINGS FOR THE CERTAINTY OF IMPRISONMENT, BOTH ANALYSES LED TO THE SAME GENERAL CONCLUSIONS. NEITHER THE CERTAINTY OF EXECUTION NOR THE SEVERITY OF IMPRISONMENT WERE FOUND TO BE EFFECTIVE DETERRENTS TO HOMICIDE. INSTEAD, HOMICIDE RATES APPEARED TO BE MUCH MORE RESPONSIVE TO CERTAIN SOCIOECONOMIC AND DEMOGRAPHIC CONDITIONS WITH SOME EVIDENCE SUGGESTING A POTENTIAL DETERRENT EFFECT FOR THE CERTAINTY OF IMPRISONMENT. WHILE FUTURE RESEARCH COMPARISONS REMAIN UNKNOWN, THE PICTURE RESULTING FROM THIS STUDY AND PREVIOUS INVESTIGATIONS IS HIGHLY CONSISTENT. REFERENCES ARE GIVEN. (MJW)