Current Research

If a Tree Falls in the Forest…? Conflict, Balance, and the Mass Media’s (Lack of) Coverage of Expert Agreement


A large gulf separates expert and public opinion on a wide range of questions of science and economics directly relevant to public policy. This may problematically put pressure on politicians to enact policies that are welfare reducing in the long-run.  Multiple theories have been advanced to explain why these gulfs exist. First, individual-oriented research emphasizes that citizens may be motivated to cognitively resist information from experts that go against their social identity-anchored priors. These theories implicitly assume expert information is readily available in the information environment. Second, media-centred approaches emphasize that news outlets often neglect thematic, context-rich content, which may limit coverage of expert consensus, and towards conflict and balance. They may elevate dissenting experts and political actors, and in so doing dilute the value of the expert information. However, little effort has made to systematically study these biases beyond the issue of climate change. This study will present an analysis of over 280,000 newspaper, newswire, broadcast, and cable American news stories on 10 issues where there are important elements of agreement among scientists and economists. It will combine automated methods, such as machine learning and dictionaries, with hand coding. This paper aims to demonstrate the extent to which the news media provide expert information through its coverage, including consensus cues, and whether the presentation of such information is limited by episodic coverage and contaminated by conflict and balancing biases.

Prepared for presentation at the CPSA annual conference in Regina, SK. Link here.

Anti-Elitism, Anti-Intellectualism, and the Motivated Resistance to Expert Consensus

Public opinion is far apart from experts on a wide range of issues. The dominant explanation of this is ideologically-driven motivated skepticism. However, this is not a sufficient explanation for less salient and politically charged questions. This paper argues that more attention needs to be placed on anti-intellectualism. Using a survey of 3,600 Americans, this paper identifies two primary dimensions of anti-intellectualism using principle components analysis: expert hostility and expert mistrust. Both of these dimensions are strongly associated with support for scientific positions on climate change, nuclear power, GMOs, and water fluoridation, above and beyond ideology. An embedded survey experiment also shows that anti-intellectualism moderates the acceptance of messages related to scientific agreement where such messages produce a backfire effect among those most mistrusting of experts. Finally, this paper explores a link between populism and anti-intellectualism. It shows that the former is strongly associated with both dimensions of the latter, and with an experiment it demonstrates that anti-elite rhetoric has the potential to activate anti-intellectual predispositions in the processing of expert messages. These findings suggest rising anti-elite rhetoric may make anti-intellectual predispositions more salient for information processing.

The Dynamics of Polarization: How the Media and Party Elites Facilitated the Growth of Climate Skepticism

(with Dominik Stecula)

Commentators are often perplexed about American polarization on issues where there is scientific consensus. One such issue is climate change. This article explores how the dynamics of climate change news media coverage contributed to polarizing the American public. We show that climate skepticism has been on the rise in America by constructing a “mood” measure going back to 1992 using the Roper Centre polling archive. We use automated and human content analysis of nearly 12,000 news articles published by the New York Times, Washington Post and the USA Today over a 34 year period (1980-2014) to show that journalistic “balance” in climate coverage has declined overtime, while politicization has steadily increased. We provide evidence using simple time series models that rising skepticism is primarily a result of party elite-driven persuasion facilitated by a proliferation of party cues in media coverage. Democratic cues, higher in volume and more consistent in direction, appear to drive public skepticism rather than being primarily influenced by Republican elites. The results conform to Berinsky’s elite cue theory and suggest we should see climate change polarization as a more typical case of public opinion formation.

Prepared for presentation at the APSA Annual Conference 2016 in Philadelphia. Link here.

Updated paper prepared for presentation at the PNWPSA Annual Conference 2016 in Portland. Link here.

Framing Climate Change: Economics, Ideology, and Uncertainty in American News Media Content

(with Domink Stecula)

The news media play a seminal role in shaping public attitudes on a wide range of issues – climate change included. As climate change has risen in salience, the average American is much more likely to be exposed to news coverage now than in the past. Yet, the content of these news stories has been under-explored in academic literature, despite likely playing an important part in fostering or inhibiting public support and engagement in climate action. In this paper we use a combination of automated and manual content analysis of the most influential media sources in the U.S., including the New York Times, Wall Street Journal, the Washington Post, and the Associated Press, to illustrate the prevalence of different frames in the news coverage of climate change and their dynamics over time. We focus on three types of frames, based on previous research: economic costs and benefits associated with climate mitigation, appeals to conservative and free market values and principles, and uncertainties and risk surrounding climate change. We find that many of the frames found to reduce people’s propensity to support and engage in climate action have been on the decline in the mainstream media, such as frames emphasizing potential economic harms of climate mitigation policy or uncertainty. At the same time, frames conducive to such engagement by the general public have been on the rise, such as those highlighting economic benefits of climate action. News content is also more likely now than in the past to use language emphasizing risk and danger, and to use the present tense. To the extent that citizens may not be informed of the gravity of the risk posed by uncontrolled greenhouse gas emissions, or discount threats that appear to be far in the future, these are welcome developments.

Learning from Divided Parties? Legislator Dissent as a Cue for Opinion Formation

That legislator dissent may entail costs to divided parties has received scant empirical attention. One possibility is that it undermines support for related policy. Dissent can do this two ways: by weakening party cues sent to co-partisans of the divided party; or by signaling the spatial location of the policy proposal indirectly with the relative ideological position of dissenting legislators. These possibilities are evaluated here using a survey experiment. Respondents were exposed to mock news articles about a bill that varied the presence of dissent on government benches and its ideological location. Dissent appears to reduce the policy support of opposing partisans, but only when it is centrist. It also primarily influences those with high levels of political knowledge. These results suggest legislator dissent can act as a cue, if a complex one, to help citizens form evaluations by signaling the spatial location of a given policy. The findings help illustrate the important, but conditional effect of legislator dissent on public opinion.

“Back to the Future”: Democratic Responsiveness and the Anticipation of Future Public Opinion

(with Andrew Owen)

Research on the responsiveness of policy to public opinion has failed to confront the possibility that re-election seeking politicians’ policy choices may reflect their expectations about future public opinion. This paper reports observational and experimental findings from a survey of senior Canadian policy makers. Results from vignette-based experiments, that manipulate the characteristics of current and future opinion, show that policy makers are responsive to the anticipated direction of election-proximate opinion, but this relationship is conditional on high expected salience. Survey results shed additional light on the role that future, or ‘latent’, opinion plays in policy making. Combined, these experimental and observational results suggest existing empirical work on policy responsiveness has incomplete and must account for the role electoral incentives play in democracy.

Prepared for presentation at the CPSA Annual Conference 2017 in Toronto. Link here.

Partisan Bias in Economic News Coverage: New Evidence

Claims that the mainstream media are biased in favor of the Democratic Party are commonplace. However, empirical research has yielded mixed results and neglected potential bias in the dynamics of media behavior and the exploration of observable implications of a biased media environment. This paper contributes to this literature by using time series analyses of the dynamics in media tone based on almost 400,000 stories on inflation and unemployment from top-circulating American print media. The results suggest there is bias in favor of Democratic presidents. Media tone in unemployment and inflation coverage is more favorable during Democratic presidencies after controlling for economic performance. It is also generally more responsive to negative, short-term changes in economic conditions during Republican presidencies. In other words, bias is stronger with worsening economic conditions. Extensions to this analysis provide suggestive evidence that partisan bias may be systemic in the media environment and influences public opinion.

Whose News? The Media and the Distribution of Economic Gains and Losses

(with Tim Hicks, Alan Jacobs, and Scott Matthews)

There is considerable evidence that citizens’ choices at the ballot box are shaped by their assessments of the state of the economy. We also know that economic perceptions are considerably influenced by the news, and that economic news responds differently to different kinds of economic developments. This paper advances our understanding of how the news responds to the economy by investigating _whose_ economic welfare drives economic reporting. In particular, how responsive is economic news to developments shaping the material fortunes of the rich, the broad middle, or the poor? In this paper, we analyze a massive dataset of economic news content over the last three decades in the United States to examine how the tone of economic news responds to real economic developments with differing distributional consequences. The analysis draws on all economic news stories from the New York Times and the Washington Post between 1977 and 2013, together with all economic news from 13 other newspapers among the top 50 in print or digital circulation between the years 1991 and 2013, yielding a dataset of almost 1 million observations. We use Lexicoder, an automated, dictionary-based coding tool with high validity for large samples, to measure the positivity or negativity of all economic stories in the selected newspapers. This allows us to generate annual and quarterly tone scores for the economic news being consumed by a large share of the American electorate over the last three decades (cf. Soroka, Stecula & Wlezien, 2014). We validate this index as a measure of the tone of the broad news environment by demonstrating that it acts as a strong predictor of consumer evaluations of the state of the economy and of the government’s performance in managing the economy (when controlling for actual economic conditions). We then undertake time-series analysis to examine how well the tone of economic news tracks changes in economic fundamentals with differing distributional implications, including median earnings, corporate profits, and capital incomes. The results will provide novel insight into the political economy of economic information, yielding a systematic and detailed portrait of the informational environment within which voters punish and reward their representatives. The findings will also help us make further sense of the mass politics of inequality (Bartels, 2008; Gilens, 2012). Whether low- and middle-income voters are equipped to defend their economic interests at election time depends critically on the distributional content of the informational environment in which they make economic and electoral judgments.

Prepared for presentation at the APSA Annual Conference 2016 in Philadelphia. Link here.

A Canadian “Perceptual Screen”? Partisan Bias in Information Processing

Partisan bias exists when supporters of different parties have systematically different views of reality. A rich literature has shown that Republicans and Democrats systematically view the world in very different ways largely because they resist the assimilation of information not congenial to their party’s interests and goals. Studies of partisan perceptual bias, however, have not typically been extended beyond the American context. This paper presents the preliminary results of a survey experiment conducted on a broadly left-leaning sample of 157 undergraduate political science students. Subjects were exposed to mock economic news that contradicted their reported priors. This news was imbued with partisan implications for either the Liberal or Conservative parties. It shows that students selectively updated their economic evaluations to the benefit of the Liberal Party and to the detriment of the Conservatives. This effect was found largely among those with less political knowledge, providing evidence on how party cues may backfire for less knowledgeable citizens.