Abstract: We use a large-scale news knowledge survey conducted just before the 2020 US presidential election, alongside monthly survey data, to explore how partisan differences in political news beliefs evolve. We exploit questions repeated in multiple surveys to identify changes in beliefs about the same news stories as the election approaches. Our findings indicate that partisan bias intensifies two to threefold during election periods. Within a framework of motivated beliefs, this change in partisan bias is predominantly driven by an amplification of the partisan identity effect, rather than differences in partisan recall. We also present findings from a counterfactual analysis that assesses the impact of a hypothetical targeted misinformation campaign during and outside of elections.

for information patrizia.pellizzari@unibocconi.it