Rethinking the Impact of "Bad Press" in Politics: Testing an Identity Model of Partisan Media Effects

Abstract

The rise in negativity in media coverage and campaign messaging in U.S. political elections is of pressing concern for scholars of political communication. Negative information has not only been shown to be more attention-grabbing than positive information, but there is also evidence that citizens respond to that information in diverse and often counter-intuitive ways. I argue our degree of attachment to political party groups, operationalized using the Social Identity Theory framework, is a useful tool to predict and understand these heterogeneous media effects. Building on existing literature in psychology, I suggest that elections in the U.S. two-party system are a “high group threat” context where biased processing and affective reactions are likely to be provoked for those who strongly identify with a political party group. Next, I review growing evidence in political science supporting the conceptualization of partisanship as psychological attachment to a group in contrast to competing ideology- or attitude-focused theories. Finally, I argue that although communication studies work frequently uses Social Identity Theory (SIT) as a tool for understanding the mechanisms driving various predominant media effects, little research has directly applied SIT to the study of partisanship as a predictor (and moderator) of outcomes. In sum, I connect theory and evidence across three disciplines to suggest an improved and more potent approach to exploring identity-based heterogeneous media effects in politics. I provide further substantiation of my approach using American National Election Studies data and tracking of media coverage tone over the course of five previous elections. Strength of partisanship proves to be an important moderator of the impact of media tone on voter attitudes. For evaluations of in-party candidates, increased negative tone of coverage increases in-party voter feelings of warmth towards those candidates, and the effect is amplified for the strongest partisans. Using these results as a springboard, I then present a novel measure of partisanship operationalized as degree of psychological attachment to the political group. Subsequently, I field two surveys during the 2020 presidential election and find that the new measure (called PSIM) outperforms the most common existing measure of partisanship (PID) in predicting the hostile media effect, selective exposure behavior and motivated reasoning among partisans. I also find evidence suggesting that PSIM is a useful tool for understanding heterogeneity in the impact of negative news on voter attitudes. In sum, political partisans respond to negative media messages in diverse and biased ways depending on (a) the group dynamic at play and (b) the degree to which the partisan’s identity is “attached” to the political group. I believe the new PSIM measure, and the results of my analyses, are useful to scholars of political communication as they may help them better (more acutely) navigate studying the tense communication climate in today’s political elections. Although bias is deeply ingrained in human processing, I hope that continued research into patterns of bias and conditions where it becomes salient will, at a minimum, increase the public’s awareness and, optimally, inspire improved techniques in overcoming it.