Political Conflict & Media Sentiment in Shaping Leader Popularity during Crises

• In Event: Media and Political Attitudes

Abstract

Under ordinary circumstances, executive approval is politically consequential, and leaders follow it closely. Approval ratings provide leaders “a continuing monthly referendum” on their actions (Brace and Hinckley 1992, 19), and they shape government action. Popular leaders leverage their standing to enact their legislative agenda and to pass reforms that bolster their power (Corrales 2018). Unpopular leaders often fail to get reelected (Abramowitz 2016; Stegmaier et al. 2017) or are pushed out prematurely (Pérez-Liñán 2007).

Under extraordinary circumstances, public support for executives is arguably more crucial. During crises, policy stakes are set higher, and executive power is often at its apex. And while crises are periods of extraordinary politics and stress for a political system, they are not exceptionally rare (Carlin et al., 2023). Thus, understanding executive approval during crises is essential for understanding the dynamics of democratic government behavior and accountability. While there is growing literature on the effects of crises on the popularity of political leaders, particularly around foreign policy crises and the COVID-19 pandemic, much of it is theoretically and empirically structured around single crises and/or countries. Furthermore, this link between a crisis and approval is often theorized to involve intermediary factors, such as media coverage and political discourse (Mueller 1970; 1973), which are not measured or examined.

In this paper, part of a larger NSF-supported research project, we go beyond assuming a mechanism linking the effects of a crisis and approval. Instead, we develop a theoretical framework to understand the over-time dynamics of public support during crises and how media and information shape public responses to crises. Specifically, we argue that under certain conditions shaping salience, mass public evaluations of political leaders may begin as emotional, affective responses (i.e., rallies) that transition to an evaluative response as media sentiment and political conflict (expressed in the media) turn negative.

To test our arguments about the mediating role of the media and informational environment, we rely on measures of the popular approval of political leaders using a new weekly and monthly version of the Executive Approval Dataset (Carlin et al. 2023), which measures the approval ratings of political executives in over 23 countries. To capture political and media sentiment across numerous contexts, we introduce a new multi-country dataset of media sentiment, salience, and conflict for 13 of these countries. This new media coverage dataset utilizes a new transformer-based Large Language Model, mDeBERTa (He et al. 2021), fine-tuned to classify hundreds of thousands of news articles that mention the political executive during periods of crises, allowing for a first-of-its-kind daily comparable measures of media coverage across these countries and numerous crises. This analysis will advance our knowledge regarding how the media’s tone and emphasis of unity or conflict shapes the public’s views of the executive during different types of crisis events.

Authors

• Ryan Carlin, Georgia State University

• Jonathan Hartlyn, Univ North Carolina, Chapel Hill

• Timothy Hellwig, University at Buffalo SUNY

• Will Horne, Clemson University

• Gregory Love, University of Mississippi

• Cecilia Martinez-Gallardo, University of North Carolina, Chapel Hill

• Matthew Singer, University of Connecticut