The theory of planned behavior suggests that individuals' projections of their likelihood of committing a crime will often correspond with their actual criminal behavior; however, empirical research has addressed this proposition only indirectly. The current study tested the hypothesis that there is a positive correlation between individuals' projections of the likelihood they would offend in a hypothetical vignette and whether they actually cheated to obtain extra money in a randomized laboratory experiment. The sample consisted of 209 students enrolled in a large public university in the southwestern United States. The sample was 47 percent male with an average age of 21. The experiment afforded participants both an opportunity and an incentive to cheat. The minuscule probability of honestly earning a bonus formed the premise of the experiment; subjects who claimed a trivia bonus were assumed to have cheated to gain the bonus. Also during the session, respondents completed a survey based on how they would handle a hypothetical scenario regarding a decision to drive home from a bar with a blood alcohol level likely to exceed the legal limit or find another way home. The latter option would involve considerable inconvenience and time investment. The survey measured the following variables: certainty and severity of punishment, projected likelihood of offending, impulsivity, influence of extralegal sanctions, previous drunk driving, and additional control variables. The study found that participants who cheated during the experiment reported they would be substantially more likely to drive drunk than those who did not cheat. This pattern remained after accounting for a variety of multivariate controls. Also, respondents who cheated despite greater punishment risk projected they would be more likely to drive drunk than did respondents who cheated despite lesser punishment risk. These findings support the concept of heterotypic continuity, which refers to consistency in the tendency to commit various types of problem behaviors. This study effectively integrated two empirical methods to produce insights regarding the efficacy of projected offending measures for investigating the causes of crime. 1 table, 1 figure, 86 references, and appended trivia quiz