<p>Informing people of an impending hazard can lead them to adopt behavior to mitigate the harm. In this study we examine whether giving more information, and giving it earlier, leads to a greater behavioral response. Our results, which are contextually dependent, show that providing more information has no effect on behavior, and that longer lead times lead to less behavioral change. These results conflict with those from previous studies. These previous studies differed from ours in terms of the research methods: while past studies examined people's anticipated responses to hypothetical warnings, we conducted a field experiment to observe people's responses to actual warnings of real hazards. Theory from cognitive science suggests that this difference matters. In situations of high stress people may make decisions using a faster decision pathway that is rather emotion-driven, while in less stressful situations they are more likely to base their decisions on information. The difference between actual and hypothetical warnings would capture this mismatch in stress levels, and account for the divergent findings. At the same time, the cognitive theory has been hard to test in the field, because of the ethical challenge of submitting people to actually dangerous conditions. Therefore, our results are not only relevant for the design of warning information, but also provide important empirical support for the theory of different decision-making pathways.</p>