Public Opinion in the News: Examining Portrayals and Viewpoint Heterogeneity

• In Event: The Story of Tonight: Frames, Indexing, and Attitudes

Abstract

A large corpus of literature has shown that perceptions of public opinion influence people’s own judgments and opinions (e.g. Mutz, 1992; Noelle‐Neumann, 1974). But how do the news media portray public opinion? Disparate studies have examined various portrayals of public opinion: (1) opinion polls; (2) vox pops, or man-on-the-street interviews; (3) inferences to public opinion; (4) protests; and (5) social media references. Some of these portrayals, such as coverage of polls, clearly communicate to citizens how majority opinion stands on an issue, while others, like protests, reveal the existence of different perspectives.

This study builds on the work on media portrayals of public opinion in two ways. First, we undertake a broad, systematic investigation of how the five types co-occur. Research shows, for instance, that public opinion inferences seem to serve as context to other public opinion displays (Lewis et al., 2005), and individual tweets appear to be combined regularly with generalizing statements about public opinion (Beckers & Harder, 2016). In addition, a news item about the popularity of politicians might contain not only poll results, but also vox pops, to illustrate the poll data (Zillmann & Brosius, 2000). Recognizing that public-opinion displays do not occur in a vacuum, we study the extent to which they co-occur.

Studying public opinion displays in an integrated manner will also provide a better understanding of the diversity of viewpoints covered in the news. Although a balanced representation of points of view is a central journalistic convention around the world, it is not known whether this principle of balance is mirrored in public opinion representations (Schultz, 1998). We will analyze the direction of the viewpoints reflected in various public-opinion displays and whether these public opinion portrayals are consonant or not. We expect that the public opinion displays will mostly be consonant, as journalists are usually found to create narratives of public opinion based on their own interpretations (Entman & Paletz, 1981; King & Schudson, 1995).

Our second goal is to examine differences between routine and election news. Given how elections are the most direct and active expression of public opinion and campaigns are rife with pre-election polls, it is not surprising that public-opinion portrayals appear more frequently during elections. In addition, more explicit guidelines exist about the diversity of opinions journalists must present during election campaigns (VRT, 2019). While the very salient nature of election campaigns generates much research about news coverage during such periods, much less is known about news coverage during more routine political periods. Thus, we undertake a comparison of election and routine news portrayals of public opinion. Given the many points of controversy raised during an election campaign, we expect the representation of viewpoints to be more diverse when references to public opinion appear in election news.

Methods. To obtain a robust picture of how public opinion is represented in the news during election and politically routine periods, we conducted a manual content analysis of four constructed weeks of Flemish (Belgian) print and television media. Manual coding was necessary, as not all displays of public opinion are explicit and therefore would not be captured by automated coding.

The sample comprised three constructed weeks of news from the first half of 2018, a non-election (routine) news period. In 2019, one week was constructed from the month before the federal, regional and European elections of May 26. Characterized by a ‘democratic corporatist’ model with rules about representation of political viewpoints (Hallin & Mancini, 2004), the media in Flanders include a strong public service broadcaster (Eén) and only one commercial broadcaster (VTM) with a primetime news bulletin. Newspapers in Flanders are non-partisan, and for print news, we selected the main popular newspaper (Het Laatste Nieuws) and elite newspaper (De Standaard). In addition, the evening broadcasts of the two main TV broadcasters were analyzed. The population of public opinion portrayals during these four weeks included 2947 news items during the routine news period and 947 news items during the election period.

Data collection has been completed and the data are expected to be cleaned and recoded by the end of January 2019. In all, this paper will shed light on how public opinion is represented in the news and the diversity of opinions (or lack thereof) in such portrayals. This is important as journalists often find it difficult to ‘read’ public opinion (Herbst, 1998), and therefore intentionally or not, project erroneously public opinion about important issues. Understanding how public opinion is portrayed can help scholars better grapple with the process by which such portrayals can shape public opinion itself.

Authors

• Kathleen Beckers, University of Antwerp

• Patricia Moy, University of Washington