Challenges Associated with Media Coverage of Polls and Election Forecasts

Session Submission Type: Full Paper Panel

Session Description

Polls, surveys, and other quantitative estimates of public opinion remain among the most important sources of information in media coverage of elections, helping campaigns, journalists, and voters themselves track how candidates are faring and which are likely to prevail on election day. In 2018, the Associated Press published a new stylebook chapter offering journalists more detailed guidance concerning best practices for coverage of polls and surveys. It was the latest in a series of long-running attempts by pollsters and journalists to improve coverage of survey data and public opinion, calling for more “rigorous inspection of a poll’s methodology, provenance and results” and warning that “the mere existence of a poll is not enough to make it news.”

Concerns about the quality of news coverage of survey data have mounted over the past decade as vast amounts of public opinion data now circulate freely online. More and more news organizations, particularly in the US, have adopted increasingly sophisticated methods for aggregating available data and modeling likely winners and losers. This is thought to be an improvement over the tendency of news organizations to otherwise cherry pick individual poll results using outlier numbers (Searles, Ginn, and Nickens 2016), but such approaches to horse race journalism—what Butterworth (2014) refers to as the “statistical frame”—vary widely across news organizations and many struggle to cover statistical forecasts without misleading or confusing readers. The surprising 2016 election results have led some scholars to ask whether widely reported quantitative predictions may produce less accurate perceptions of the competitiveness of elections, potentially suppressing turnout among relevant segments of the public (Westwood, Messing, and Lelkes 2018).

This panel examines these and other related challenges associated with contemporary horse race journalism including issues surrounding disclosure of survey methodologies and interpretation of quantitative data. Using a range of methodological approaches, presentations address both how practitioners use and evaluate opinion data as well as how the public processes such information.

Sub Unit

• Related Groups / Political Forecasting Group

Cosponsor

• DIVISION 38: Political Communication

Individual Presentations

• Numeracy and Statistical Biases Among Journalists and their Public - Yanna Krupnikov, Stony Brook University; John Barry Ryan, Stony Brook University; Kathleen Searles, Louisiana State University

• Don’t Shoot the Messenger - Rachel Lynn Bitecofer, Christopher Newport University

• Polls, Forecasts, and Voters’ Perceptions of Uncertainty - Michael Decrescenzo, University of Wisconsin, Madison; Benjamin Toff, University of Minnesota; Zach Warner, Cardiff University

• Learning Who Will Win the Election - Andreas Erwin Murr, University of Warwick

Chair

• Benjamin Toff University of Minnesota

Discussants

• Natalie M. Jackson Public Religion Research Institute

• Solomon Messing Facebook