Crossing the presidential Rubicon: Now Trump is taking pollsters to court
President-elect Donald Trump’s lawsuit against veteran pollster J. Ann Selzer for an off-target poll released just before last month’s election may be the first of its kind in a modern U.S. presidential election.
In researching my book, “Lost in a Gallup,” a narrative history of prominent polling failures in presidential elections since 1936, I uncovered nothing akin to Trump’s litigation, which accuses Selzer of “brazen election interference.”
The lawsuit may not survive what are likely to be sharp legal challenges. But its apparent unprecedented nature could have an uneasy effect on pollsters, given their checkered record of accuracy. It is not difficult to imagine their wariness and caution should pollsters face the risk of legal action when pre-election surveys go sideways.
Trump’s lawsuit centers on Selzer’s poll, released three days before the Nov. 5 election, indicating that Vice President Kamala Harris had opened up a 3-point lead in Iowa, which Trump had easily carried in the 2016 and 2020 elections. Harris “has clearly leaped into a leading position,” Selzer said of the poll results.
The implications were clear: If Harris had actually forged ahead in a state with the pro-Republican makeup of Iowa, then her chances of winning the vigorously contested swing states — and thus the presidency — would be highly favorable. After all, the poll bore the imprimatur of Selzer and her stellar reputation for accuracy in Iowa. In fact, a fellow pollster this year had referred to her as the “oracle of Iowa.”
Then Trump won Iowa by 13 percentage points, meaning Selzer’s poll erred by 16 points. This was a humiliating misfire for a veteran and well-regarded pollster.
Trump’s lawsuit, however, claims the poll “was no ‘miss’ but rather an attempt to influence the outcome of the 2024 presidential election” by projecting “a false narrative of inevitability” about Harris’s prospects.
Trump further claims that Selzer’s “huge platform and following” allowed her “a significant and impactful opportunity to deceive voters.”
While not directly challenging the merits or logic of Trump’s lawsuit, the country’s largest polling organization, the American Association for Public Opinion Research, said in a statement: “Differences between polling results and election outcomes can and often do occur for reasons unrelated to misconduct or fraud. Such variations highlight the complexity of capturing public opinion and the importance of interpreting polls within their limitations.”
Selzer, in a recent interview with a PBS station in Iowa, denied her poll was deliberately off-target. That, she said, would be inconsistent with her ethic. Selzer also said she has not determined why her poll erred so badly. Thinking about it “does sort of awaken me in the middle of night,” she said, adding, “We don’t know. Do I wish I knew? Yes, I wish I knew.”
Shortly after the election, Selzer said she had previously decided to retire from poll-taking and had so informed her principal sponsor, the Des Moines Register.
One line of defense for Selzer and her co-defendants — the Register and its owner, Gannett — may be found in a longstanding, if lazy, cliché of survey research: that pre-election polls are but “snapshots in time.” A “snapshot” defense could argue the poll was accurate at the time it was conducted in the final days of October, and then Trump gained 16 points in the days that followed.
Although it may seem far-fetched, it is unclear how Trump would disprove such a claim.
Perhaps a stronger line of defense rests in First Amendment protections, that pollsters, like journalists, should be excused for good-faith errors. Without such flexibility, or what the U.S. Supreme Court in 1964 termed “breathing space,” robust public debate and discussion could be stifled.
Criticism of election pollsters has been frequent, even harsh at times, since George Gallup, Elmo Roper and others began pioneering quasi-scientific survey techniques in the mid-1930s. Polling methodologies have become more sophisticated in the decades since then, even if polling accuracy has been uneven.
Polls, for example, collectively understated the extent of Trump’s support in each of the three most recent presidential elections — despite modifications intended to reach and interview previously elusive Trump-backers.
The final pre-election polls this year underestimated Trump’s support by an average of 2.4 percentage points, according to an analysis by NBC News. In 2020, polls overall understated Trump’s backing by 3.3 points, their worst collective performance since 1980.
This year’s understatement of Trump’s support “ran through the polls in states across the political spectrum,” NBC said. Trump’s support, moreover, was collectively underestimated by polls conducted in each of the seven battleground states where the election turned.
Although polls have generated popular skepticism over the years, taking pollsters to court over their survey results in presidential elections was unheard of before Trump’s legal action against Selzer.
Eighty years ago, Gallup did go before a committee of the House of Representatives to testify about polling techniques and the discrepancy between his polling and the outcome of that year’s presidential election. Gallup had estimated the race between President Franklin D. Roosevelt and Republican Thomas E. Dewey would be closer than it turned out.
As he was inclined to do, Gallup engaged in what these days would be recognized as “spin,” telling the House committee during what was a cordial hearing that the polls in 1944 “came through the … election with flying colors. While their record falls short of absolute accuracy, it does represent a degree of accuracy found in few fields outside the exact or physical sciences.”
Four years later, however, Gallup and other pollsters erred dramatically in predicting Dewey’s certain victory over President Harry S. Truman, Roosevelt’s successor. In perhaps the greatest shock election in presidential history, Truman won reelection by 4.5 points. Gallup figured Truman would lose by 5 points, meaning his polling error was 9.5 points.
That error, no doubt, was on Gallup’s mind during a speech he delivered in Cleveland in late December 1948. “A doctor,” he said, “can bury his mistakes. A lawyer can rationalize his. ... But a public opinion researcher must stand naked before the world, his shame recorded for posterity.”
Exaggerated or not, it’s an observation with resonance today.
W. Joseph Campbell is a professor emeritus at American University and the author of seven non-fiction books on a variety of public affairs and media history topics.