Ghost Light or Light House? Investigating Poll Reporting, Subjective Assessments, and Media Bias in Public Opinion Coverage in an Election Context

• In Event: Who Will Win? The Media’s Role in the Formation of Electoral Expectations

Abstract

Various studies suggest that media coverage is an important source for perceptions of public opinion and that it contributes to the formation of electoral expectations. However, much less is known about the prevalence of cues to public opinion within the media, mostly because there has been very little content-analytic research on the topic. Especially research on electoral expectations largely focuses on poll coverage as the main element of public opinion coverage. Consequently, media are regarded as rather objective transmitters of poll results. But the dramatic increase of poll reporting across western democracies does not mean that the media do not present other cues that might help citizens to form expectations regarding the outcome of an election. In fact, there are numerous experimental studies suggesting that climate-of-opinion perceptions can also be influenced by vox pops and exemplars, by the general slant of their coverage, or by purely subjective assessments of public opinion. And these elements may be more prone to bias. Therefore, this paper investigates how important different cues to public opinion are in campaign coverage, how they relate to each other, and if they are subject to political media bias.
We address these questions using a quantitative content analysis of six newspapers, three television evening newscasts, and one news website in the last two weeks before the German national election in 2013 (N=831 articles and 1095 public-opinion statements): First, we show that poll coverage is an important, but by far not the only means by which public opinion is communicated in the news media. Subjective statements about public opinion are another important one. Second, we will show that in spite of the fact that relying on polls is often regarded to enhance accuracy and impartiality of news coverage both – poll reporting and subjective statements – are still prone to political media bias. Finally, we will discuss the implications and generalizability of our findings as well as directions for future research.

Authors

• Angela Nienierza, Ludwig-Maximilians-Universitat Muenchen

• Carsten E. Reinemann, U of Munich

• Thomas Zerback, LMU Munich