Unsafe acts such as ignoring scheduled inspections can cause serious consequences. This study examines the effects of two reinforcing stimuli and four reinforcement schedules in maintaining sampling behavior in a virtual inspection task. Participants were asked to decide (yes or no) for each “product” whether it should be sampled for inspection. In Experiment 1, “yes” responses were reinforced with the message that defectives were found, once for every five times (FR), once for every one to nine (on average five) times (VR), only the first time (FTO), or never (None). The sampling behavior declined gradually in FR and VR and somewhat surprisingly more sharply in FTO than in None. In Experiment 2, the sampling behavior was effectively maintained when the participants were regularly provided with the evaluative feedback on their sampling behavior, although they were kept informed that defectives were not found. These results indicate the importance of utilizing reinforcing stimuli whose administration is independent of the outcome (e.g., defective or not) of the response (e.g., inspection) to be reinforced.