Participants were drawn from a population of male offenders sentenced to 2 years or more, which is the criteria for Federal custody in Canada. All offenders spend several weeks at the commencement of their sentence being assessed across a variety of areas, such as health, education, substance abuse, and psychology (for violent offenses), and for penitentiary placement. The participants were 207 offenders who completed the measure of alcohol abuse as part of a psychological assessment required during intake for violent offenders. These offenders were drawn from consecutive admissions over an 18-month period from 1991 to 1992. Average Alcohol Dependence Scale (ADS) scores, age, and incident types were calculated for each severity level and for the total sample. Findings show that results support, in part, the original hypothesis. ADS total score was not related directly to the outcome measures of institutional misconduct; however, specific items from the ADS were salient to the prediction of misconduct. Specifically, offenders with substantial alcohol dependence, which has been shown to be indicative of polysubstance abuse, were involved in more serious incidents than those with other levels of dependence; however, there was no direct correlation between the ADS total score and the outcome variables. Implications for institutional classification and prediction of misconduct are discussed. 4 tables, 1 figure, and 10 references