<h2>Background</h2> Paul Slovic argues that people are more influenced by stories about individual victims of tragedy than statistical information about the masses of victims of the tragedy (e.g., <a href="https://www.sas.upenn.edu/~baron/journal/7303a/jdm7303a.htm">Slovic, 2007</a>). <h2>Hypotheses</h2> We predict that people’s intentions to adopt public health recommendations (e.g., hand washing, physical distancing, etc.) will be better predicted by information about individual victims than statistical information about masses of victims. <h2>Methods</h2> <h4>Design</h4> Participants will be randomly assigned to either statistical or individuating messaging about either COVID19 or the flu. (A 2x2 design). <h4>Participants</h4> GPower suggests that 235 will confer 99% power to detect a typical effect size of f = 0.204 (Gignac & Szodorai, 2016). To ensure that we have enough participants after eliminating participants who fail the attention check, we will oversample until we have about 250 online participants. If we find the expected effect, then we will attempt to replicate the finding in a larger experiment (about 750 participants) including measures of potentially moderating factors (philosophical stances, cognitive measures, etc.), totaling about 1000 participants. <h2>Analysis</h2> We will first test for an interaction and—depending on the result—investigate the effect of the individual vs. statistical manipulation. We will also employ regression to see how thedifference between these message formats explain variance in responses to 6 questions about people's adoption of public health recommendations/conclusions.