Were the 2020 election polls really so wretched?

The American Association for Public Opinion Research (AAPOR) recently confessed that opinion polling in the 2020 election campaign suffered from errors of “unusual magnitude.” AAPOR reported that pre-election polling understated President Trump’s support by an average of 3.3 percent and overstated the projected margins of Democratic candidates in senatorial and gubernatorial races over their Republican rivals.

As the November election approached, the opinion polls certainly gave Democrats hope for a wave election. But it did not work out that way. While Joe Biden scored an electoral college victory, there was no overwhelming rejection of Trumpism. Democrats did not win some Senate seats they were expecting to win, numerous House seats were lost and state gubernatorial and legislative results were disappointing. What happened?

AAPOR attributed the erroneous results to faulty methodology. While there may have been polling errors, the discrepancy between the poll results and the election outcome can largely be explained by factors not necessarily accounted for in the opinion polls.

Trump was by far the most important factor in driving voter turnout on both sides. No political figure in decades has been so revered by some and so reviled by others. His candidacy produced 22 million more votes in 2020 than in 2016, a 7 percent increase.

That being said, the outcome of the election was largely determined by the way it was structured. In deference to voters concerned about the coronavirus, many states made it easier to vote early. The Democrats strongly urged their supporters to vote early by mail or in person. Trump tried to discourage mail-in balloting, correctly thinking it would increase the Democratic vote. Forty-six percent of voters cast mail-in ballots and 65 percent of those votes were for Biden.

Others, perhaps wishing to avoid the crush and possibility of infection on Election Day, opted for early in-person voting. The early vote comprised 27 percent of the total, and Trump received 52 percent of that vote.

We all recall media reports of the massive turnout for early in-person voting and the tremendous number of requests for mail-in ballots. Television newscasts showed blocks-long lines of early voters waiting hours to cast their ballots, and suggested that most were supporting Joe Biden. These repeated stories gave the impression of a pending landslide for the Democrats, which was in keeping with the public opinion polling of “likely” voters. All of that would certainly have scared the living daylights out of Trump supporters.

The old Latin proverb praemonitus, praemunitus (“forewarned is forearmed”) seems to have came into play as Election Day neared. The stories about the impending Democratic tsunami may have put Trump devotees into high gear and caused millions who otherwise might have stayed home to show up at the polls to rescue their champion from electoral disaster. And while they were in the voting booth, most also supported down-ballot Republican candidates. Twenty-seven percent of the electorate voted in person on Election Day. Trump received 65 percent of that vote. 

Trump’s power to draw his fervent supporters to the polls is illustrated by white evangelical Christian voters, who turned out in massive numbers for Trump. White evangelicals constituted about one-third of Republican voters in 2020. Trump’s support from evangelicals increased from 77 percent in 2016 to 84 percent last year. They would obviously have been energized to turn out to support Trump when media reports indicated he was facing practically certain defeat.

While the record turnout has been remarked upon considerably by political commentators, little mention has been made of the role of the election structure in producing the disappointing results for Democrats. But the media hype about the impending Democratic tsunami probably frightened many Trump-loving voters, motivating them to vote on or before Election Day — essentially a panic-induced response.

The polling for this election was off the mark in several ways, but AAPOR need not place all of the blame for the mistaken projections on its own doorstep. The effect of the easier voting measures brought about by the pandemic and the media overplay of the process would not necessarily have shown up in polling until Trump voters streamed to voting booths on Nov. 3 to save their candidate.

Jim Jones is a Vietnam combat veteran who served eight years as Idaho attorney general (1983-1991) and twelve years on the Idaho Supreme Court (2005-2017).