How do emotions shape the way people process information? We study this question with a pre-registered survey experiment given to 9,000 US respondents who are randomly exposed to (i) an emotionally charged news video about violent crimes committed by immigrants (treatment) or a neutral cartoon (control), (ii) statistical information about immigration, and (iii) summaries of academic research on the effectiveness of border enforcement. The videos carry no informational signal about the elicited posterior beliefs, isolating the role of emotions. We show that emotions bias information processing along three margins. First, how people feel influences what they learn, but only when the information is incongruent with the emotions triggered by the treatment. Second, emotions distort the understanding of statistics even after sharing the correct information. Third, emotions impede research evidence from guiding policy conclusions. These results are consistent with the notion that, by reallocating attention toward internal cues and the associated impulses, emotions can eclipse facts.
Research on Europe’s historical academic market earns the European Economic Association’s Hicks–Tinbergen Award.
