This repository includes data from approximately 11,000 unique users who interacted with the COVID-19 Event Risk Tool at https://covid19risk.biosci.gatech.edu/. Data collection took place between 9/22/21 and 1/22/22. Files also include code necessary to reproduce all results and figures reported in the preprint "Counteracting COVID-19 Risk Misestimation with an Interactive Website: An Online Informational Intervention." All data files are described below. The file **versions_dependencies.rtf** describes all R session information and package versions. If you have questions, please contact Allie Sinclair (ahs50@duke.edu). **map_willingness_data_OSF_1-22-22.csv:** Users viewed a county-level map of COVID-19 exposure risk levels across the USA for a particular event size (e.g., the risk of one or more individuals infected with COVID-19 at an event with 50 people). Below the map, users reported "change in willingness," whether they were more or less likely to participate in an event of that size after viewing the risk map. **risk_quiz_data_clean_OSF_1-22-22.csv:** Users completed a risk quiz where they guessed the exposure risk probability (from 0% to 100%) for four event sizes (20, 50, 100, and 1000 people) in a particular location. After guessing, users received veridical feedback about actual risk levels. This dataset includes users' guesses (g_eventsize), our prevalence-based risk estimates (pred_eventsize), and "change in willingness" ratings optionally provided after viewing the accuracy feedback. This dataset has been cleaned to exclude risk quiz submissions with all-default responses and repeat submissions from the same unique user. **risk_quiz_no_exclusions_OSF_1-22-22.csv:** An alternate version of the risk quiz dataset with no exclusions applied. **Qualtrics_merged_data_OSF_1-22-22.csv:** An external demographics survey completed by a subset of a users. When possible, users' responses to the survey have been merged with data collected via the Event Risk website. Note that the demographics survey also included an open-ended comment box, but this field has been omitted from the public dataset for privacy reasons.