A need exists for a controlled examination of the variables that may affect the results of a behavior-based safety program. These include the rate of risky behaviors and how they relate to a contingent bonus payout (Experiment 1), and the relative amounts of honesty in injury reporting and rule following (Experiment 2). A simple task consisting of moving a computer-generated image within a grid was used to simulate working behavior. Participants were awarded small monetary payments in an effort to further simulate reality. Results indicate that participants were more dishonest in reporting simulated injuries when: (1) a contingency for low rates of reporting of injuries was present and (2) in some cases, when not given proper training (or a rule) on how to work most safely. Considering the recent growth of behavioral safety programs in business and industry, the utility of an examination of the important and relevant variables within a controlled environment becomes apparent.