Biden-Trump Election Polling Most Inaccurate in 40 Years: Research Group Says
By Charlie McCarthy | Thursday, 13 May 2021 12:44 PM
Pre-election polling of last year's race between former President Donald Trump and President Joe Biden was the most inaccurate in 40 years, according to the top public opinion research trade group.
Members of the American Association for Public Opinion Research's task force on the 2020 election disclosed their preliminary findings Wednesday at the trade group's annual conference, The Wall Street Journal reported.
Polls of the 2020 race overestimated Biden's national support by nearly four percentage points in the two-week period leading up to the election, according to the findings.
That marked the most substantial error in polling since 1980, when surveys overestimated support for former President Jimmy Carter by about six percentage points. Ronald Reagan took a landslide victory that year.
Last year's overestimation was a larger error than the 1.3-point overstatement in 2016 surveys for Hillary Clinton, who won the popular vote but lost the Electoral College.
The 2020 polls overstated Democrat support "in every type of contest we looked at: the national popular vote, the state-level presidential vote as well as senatorial and gubernatorial elections," said Joshua D. Clinton, a professor of political science at Vanderbilt University who led the review for the AAPOR.
"This was true no matter how respondents were polled. It didn’t matter whether you’re doing online or telephone polls. A similar level of error was pervasive."
Among the task force’s potential explanations for 2020 surveys were:
- Some voters simply refused to participate in polling.
- Trump might have been able to turn out new or infrequent voters who don’t identify with either major political party, making it difficult to project how the electorate would look.
- Trump often criticized public polling, potentially discouraging his supporters from participating.
While offering possible reasons for the inaccurate polling, experts have struggled to pinpoint a solution to the problem.
Trump's presence on the ballot might have had a significant effect, AAPOR President Dan Merkle said according to Journal. That idea is supported by the fact polling of the 2018 midterm elections were largely accurate.
The 2020 error in presidential polling was, other than in 1980, the highest since Republican Dwight Eisenhower's landslide victory in 1952.
Surveys understated Trump’s support in two consecutive elections, though in some years polling has overstated GOP support, as in 2012 when Democrat President Barack Obama beat Republican Mitt Romney.
The task force has not released its final 2020 polling report, which was based on an analysis of more than 2,000 polls taken ahead of the November election, The Hill reported.
The panel was formed after people wondered why the 2020 presidential race was much closer than many pre-election polls suggested. Many public surveys showed Biden carrying wide leads over Trump, including in states the former president carried.
The 2020 polling error wasn't limited to the presidential race. Surveys on average were too favorable to Democrats by six points in Senate and governors' races in the final two weeks of the campaign, compared with the ultimate election results.