Visualization aesthetics influence trust in science, news, and social media
Chujun Lin, Mark Thornton
Abstract
Scientists, policymakers, and the public rely on data visualizations to inform important decisions. Visualization aesthetics may vary independent of data quality. Across five-preregistered studies (3,405 participants, 1,219 graphs), we comprehensively investigated whether the beauty of a graph influences how much people trust it. Results indicate that graph beauty predicts increased trust in real-world graphs across three important domains: social media, news, and scientific publications (Studies 1-3). Moreover, beauty had consequential correlates, predicting increased citation and comment numbers. We next identified visual features that are associated with perceptions of graph beauty (Study 4). We used these insights to manipulate graph beauty, and observed that doing so causally increased trust (Study 5). This effect was fully mediated by people’s inferences of the general competence of the graph maker (Study 5b). These findings reveal a novel factor shaping trust in real-world information that is relevant to a wide range of domains.