Fig. 2

Proportion of “hazard-present” responses as a function of median hazardousness ratings for each movie (represented by each circle). Across all experiments, the probability of a “hazard-present” response increased as videos became more dangerous. Consistent with the LPE, there were fewer “hazard-present” responses under low hazard prevalence (green) compared to high hazard prevalence (purple) across all hazardousness levels. Interestingly, this pattern was observed even in the no feedback experiment, although the magnitude of the LPE is much smaller than when feedback was present. These results are consistent with previous work that found that low hazard prevalence induces a more conservative criterion, such that participants needed the video to be more hazardous before they are willing to respond “hazard-present” (e.g. Wolfe et al., 2007)