Attacking the credibility of polls (and pollsters) has become a kind of bank shot for those aiming to destroy the credibility of the institution that polls simulate: elections. Among MAGA adherents, if the polls are off, it’s not because polling has become more difficult. It’s because everything is corrupt and nothing is true.
The political science department of Vanderbilt University has said that pre-election polls in 2020 exhibited the “largest errors in 40 years.” That mistake hovers over the 2022 election and influences the work, and reputations, of the partisans and pollsters (and partisan pollsters) who are trying to gauge the trajectory of a large, diverse and unruly electorate.
One partisan who has staked out a clear patch of terrain in this debate is Simon Rosenberg, a longtime Democratic consultant. Rosenberg is a particular kind of political player, both partisan and credible. While he wants Democrats to win, he also values his reputation as an astute analyst of American politics; he wants to be right.
At a time when there is growing discussion of a red wave — “Democrats Prepare for Loss of Congress as Voters Break Late to GOP” reads a recent headline — Rosenberg has been making the case that Democrats might still buck history and retain control of Congress. “I think the election now is very close and it is unknowable what’s going to happen,” Rosenberg told me. “The polling can’t get down to a level of granular detail when elections are this close.”
His bet is that much of the anti-Trump coalition of 2018 and 2020 will reconstitute itself as an anti-MAGA force, having correctly gauged that the stakes of this election are immense.
History suggests that an in-party led by a relatively unpopular incumbent president is going to have a tough time. It’s hard to gauge how the MAGA threat to democracy might alter that equation. Quality polling in swing states mostly indicates tight races for U.S. Senate and governor in the final campaign week, with few exceptions. Polling on the generic ballot for Congress has generally been tight. But most news media coverage, and many analysts, presume Republicans will win a majority in the House at least. If the polling errors of 2020 are replicated in 2022, overestimating the Democratic share of the vote, a larger red wave would result.
The challenges to accurate polling include a volatile electorate and recurring nonresponse bias, the tendency of Republican voters to shun polls at a higher rate than Democrats. But Rosenberg believes the vast changes in the electorate since 2016 make it just as possible for pollsters to undercount Democratic votes. “We have a far greater number of irregular voters on both sides. Trump brought in irregular Republicans. So trying to understand in a midterm which of those groups is going to vote, and then trying to figure out who in each congressional district, in each state, is going to vote, is incredibly hard,” he said.
The presidential electorate grew from fewer than 129 million in 2016 to more than 155 million in 2020. Meanwhile, the midterm electorate exploded from 2014 to 2018. Turnout rose from 36.4% in 2014, the lowest since 1942, to 49.4% in 2018, when a record 122 million Americans voted.
That roller coaster ride has made modeling the electorate, and parsing its affinities, more difficult. It has also led to more attacks on the integrity of legitimate pollsters. As pollster Natalie Jackson noted in a tweet, if polling and news organizations make predictions that incorrectly favor Democrats, they will not be accused merely of error: They will be attacked as partisans trying to depress Republican turnout. There is no similar countervailing pressure from the Democratic side.
Contrary to myth, national polls were pretty accurate in 2016. Donald Trump lost the popular vote by 2.87 million, roughly as expected. But Trump relentlessly attacked the vote tally and the polls that predicted it, and the MAGA assault on them has never abated.
Rosenberg is trying to counter that attack. In effect, he wants to take the built-in uncertainty that MAGA employs to cast doubt on polling (and elections) and reclaim it as a talking point for Democrats. “No one’s ever polled an electorate like this before,” he said. “We’ve never had this enormous infusion of new voters into the electorate.”
In the 20th century, elections provided an answer to which candidate received the most votes — and whose election analysis was correct. On Nov. 8 or soon thereafter, the results of the 2022 midterms will be known. But the legitimacy of those votes, and the meaning of them, will not be resolved. MAGA candidates in Arizona, Michigan and elsewhere have refused to say that they will honor election results. Attacks on the election outcome were being readied before votes were even cast.
That precarious state may have more to do with the volatile political environment than any technical issues about polling methods and data. American democracy is in deep turmoil. In that regard, shaky polls are a highly accurate snapshot.
____
ABOUT THE WRITER
Francis Wilkinson is a Bloomberg Opinion columnist covering U.S. politics and policy. Previously, he was an editor for the Week, a writer for Rolling Stone, a communications consultant and a political media strategist.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
