Partial and mixed support were found for two hypotheses, suggesting that deterrent effects varied by arrest ratios according to crime types. Mixed results were found for the hypothesis that lower perceptions of the certainty of arrest and subsequent sanctions would lead to increased involvement in crime. Support for this hypothesis was found for assault and motor vehicle theft; however, significant results in the opposite direction were found for burglary; i.e., the commission of more burglaries was linked with higher perceptions of the certainty of arrest and sanctions. The second hypothesis was that high arrest ratios would lead to perceived certainty of sanctions for crimes. Four out of the eight crimes demonstrated this effect. Beginning in July 1988, sampling was conducted 1 day a week through December 1989. The dependent variable was the perception of sanction certainty, and it was measured for each crime that the respondent reported committing in the previous 12 months. Independent variables were the total number of crimes committed and the arrest ratio. Control variables were the number of lifetime arrests; the respondent's age, sex, race, level of education, age at first arrest, and whether the respondent reported drinking alcoholic beverages frequently during periods when crimes were committed. 2 tables, 5 notes, and 60 references