THE DETERRENCE HYPOTHESIS PREDICTS THAT CRIME RATES WILL BE INVERSELY RELATED TO THE CERTAINTY THAT SANCTIONS WILL BE IMPOSED AND THE SEVERITY OF THOSE SANCTIONS. THE ANALYSIS USES UNPUBLISHED FBI DATA TO ESTIMATE BANK ROBBERY RATES AND AVERAGE TAKES; DISPOSITION DATA FROM FEDERALLY PROSECUTED BANK ROBBERY CASES TO MEASURE SANCTION VARIABLES; AND CENSUS DATA TO CALCULATE DEMOGRAPHIC VARIABLES. NO CONSISTENTLY NEGATIVE ASSOCIATIONS WERE FOUND BETWEEN ANY OF THE SANCTION VARIABLES AND THE BANK ROBBERY RATE WHICH COULD BE JUDGED AS INDICATING THE DETERRENT EFFICACY OF THAT SANCTION ALTHOUGH THE EVIDENCE DID NOT INDICATE THAT DETERRENCE WAS NOT OPERATING. THE AUTHOR CONCLUDED THAT THE PROSPECTS OF RESOLVING THESE AMBIGUOUS RESULTS BY CONSTRUCTING A MORE SPECIFIC MODEL WERE UNLIKELY. FOR A TEST OF THE DETERRENCE HYPOTHESIS FOR A NON-INDEX CRIME, SEE NCJ-37448. (AUTHOR ABSTRACT MODIFIED)