“According to the Polls…” Opinion Poll Coverage in Network Evening News during the 2020 U.S. Election Campaign

Abstract

After news media and pollsters were unsuccessful in predicting recent political outcomes such as the 2016 U.S. election, opinion polls came under scrutiny. Journalists were accused of not providing audiences the tools to correctly interpret poll information. Using a content analysis of all evening news items from CBS, ABC, and NBC from the final two months before the 2020 U.S. general election, we analyzed the quality of poll coverage. We find that half of the references to election polls are “diffuse,” in which journalists refer to “the polls” in a general manner. When news items do cover specific poll results, media often disclose the absolute minimal essential information (e.g., population and error margin), but fail to provide additional methodological details.

Should we include margins of error in public opinion polls?

“…Merely mentioning that the projected winning margins of leading parties are within the margin of error can affect voters' responses to poll results. Accordingly, pollsters and public opinion experts should intensify their efforts to communicate the methodological details of opinion polls as transparently as possible (Beckers, 2021). In this regard, reporting margins of error can also be an efficient way of reducing horserace coverage and media attention to outlier polls.…”

“…Although public opinion pundits frequently demand margins of error to be transparently reported (BBC, 2019; ESOMAR/WAPOR, 2014; Pew Research Center, 2016), the media often fails to provide this essential information (Beckers, 2021; Bhatti & Pedersen, 2016; Welch, 2002). Brettschneider (2008) found that only 2 per cent of the polling results reported in the German media between 1980 and 2002 included the corresponding margins of error.…”

“…For more recent elections, the coverage of margins of error has increased. Beckers (2021) analysed television news during the U.S. 2020 presidential campaign and found that 76 per cent of the reported results of ‘in‐house’ polls included mentions of the corresponding margins, whereas the share was 39 per cent for ‘outside’ polls. Still, this information was only shown in small letters, and journalists explicitly referred to the margins of error in only 3 per cent of the cases.…”