In the experimental condition, the male and female inmates were given a pre-test before the AIDS education program and a post-test about 1 month after the program but before they were released into the general population. A randomly selected post-test only comparison group was also surveyed to determine whether observed changes in experimental group scores from pre-test to post-test might be attributable to testing effects. The comparison group had already participated in the AIDS education program but had not been pre-tested. T-tests for paired samples were employed to determine whether any significant changes occurred within groups (male and female), and t-tests for independent samples were used between groups to determine whether males or females experienced the greatest magnitude of change. Based on data obtained using the AIDS Knowledge and Attitudes Questionnaire, multiple regression analyses explored the relationships between selected independent variables, post-test perceptions, and magnitude of change. Findings revealed the men's levels of perceived risk declined significantly from pre-test to post-test, whereas the women's levels of perceived risk increased, although not significantly. Regression analyses indicated change in perceptions of risk was related to several variables outside of the prison's control. Implications of the findings are discussed, and suggestions are offered for modifying existing prison-based AIDS education programs. The authors conclude that AIDS education should be sensitive to culture and sex and should focus on preventing participation in risky behaviors in elementary schools and that street-based initiatives should be developed to reach individuals who are already engaged in risky behaviors but who have not yet had contact with the criminal justice system. 35 references and 6 tables