Violations are a problem for safety-critical organizations, which need to trust in their workers' compliance with rules and regulations. In two studies we investigate the impact of the framing of production outcome goals. The first study was designed in accordance with the classic framing experiments by Kahneman and Tversky (1983) as a paper-based case with two groups (N = 72) reading a story in which they are to envisage themselves in the position of the production supervisor, who decides to comply with or violate a safety-relevant rule in order to meet the production goal. The two groups differ in the framing of the production outcome goals in terms of losses or gains. In contrast to our assumptions and to previous findings, we were unable to replicate the framing effect for a production setting, but 35-45 % of participants in both groups were willing to violate the normative rule, independently of the framing. In the second experiment (N = 39), the experimental groups acted out the role of the supervisor and were trained to use two different start-up strategies for the plant. In the final trial, participants were told that their salary would depend on their performance in this last trial. As a result, in the loss framing conditions, significantly more participants violated the normative rule than in the gain framing condition. Finally, we discuss the implications of framing effects for production settings and their impact on the probability of choosing to violate normative rules.