The Worst Polling Failure in American History

Ryan Fan

Clifford K. Berryman cartoon — From the US National Archives Record — Public Domain

“We need a rehabilitation of the entire system,” one pollster told NBC.

I shook my head all week long. The polls were wrong in the 2020 election — again. I lamented how wrong they were as I anxiously refreshed my phone for news notifications about any states declaring and how the vote was being counted in battleground states, and then made a resolution:

I am never trusting a poll again.

Polling is now facing a reckoning this year, as it did in 2016. Most polls predicted a Biden landslide over Donald Trump, as well as a Democrat domination of the Senate and House of Representatives. Unless you’ve been living under a rock, you’ll know the election was not a landslide. While Trump is projected to be defeated in close battleground states and Joe Biden has been declared winner by the writing of this article, they’re by slim electoral margins — not a repudiation of Trumpism in America that myself and most on the left hoped for. Biden will likely win by as much as Trump won by in 2016, electorally. Republicans have also actually gained seats in the House of Representatives and will likely hold the Senate.

I will avoid buying into any large, sweeping narratives about the election, like I did in 2016, when I stared at the TV at 1 a.m. on election night, shellshocked and thinking “what the hell just happened?” But one thing for certain is that polling needs to answer fundamental questions — the landslide didn’t happen, and public polling missed the mark.

According to Lauren Feiner at CNBC, an NBC News’ poll predicted Trump would be up an average of 7.4%, and Biden would win Wisconsin by 6.7%. Biden won Wisconsin by less than 1% and around 20,000 votes, with the race not being called until the last votes were counted in Wisconsin.

“I believe in stats, but it’s also undeniable that the polls have been off for two presidential contests and they’ve been off in the same direction: undercounting Donald Trump,” UMass Boston professor Erin O’Brien said.

But this isn’t a piece about why the polls were wrong — I expect a lot of people will analyze that, and a lot of people, like myself, have lost their overall trust in opinion polls. Instead, let’s take a look at the history of polling in America — and the worst polling failure in American history that puts both 2016 and 2020 to shame.

The 1948 Presidential Election by the Associated Press in 1948 — Taken by Byron Rollins
“Thomas E. Dewey is almost as good as elected…. I can think of nothing duller or more intellectually barren than acting like a sports announcer who feels he must pretend he is witnessing a neck-and-neck race.” — Pollster Elmo Roper, September 9, 1948

No election defied the expert opinions and sampling methods of pollsters as much as the 1948 Presidential Election, where American voters chose between Harry Truman of the Democratic Party and Thomas Dewey of the Republican Party. Even before the actual election, Truman looked very weak — the support for the Democratic Party had fractured among two third parties: the Dixiecrats led by Strom Thurmond, who were displeased by Truman’s firm stance on civil rights, and the Progressive Party led by Henry Wallace, who thought Truman was too hawkish on the Soviet Union.

Truman had angered politicians in the South, including Thurmond, the famous segregationist in South Carolina. According to author Dewey W. Grantham in The South in Modern America, Truman advocated for broad civil rights legislation in February of 1948, including passing an anti-lynching law, the abolition of the poll tax, outlawing segregation in interstate transport.

In July of 1948, Truman outlawed racial segregation in all federal agencies, most notably the armed forces. Public opinion polls in 1948 showed that only 23% of white southerners through the federal government should intervene to prevent lynchings and only 22% of white respondents in the South thought Congress should enact legislation to guarantee the right of Black people to vote.

Truman angered the conservative South not only in civil rights but in labor rights. He vetoed the Taft-Hartley Law, which would have helped control union power, showing that he was pro-union and pro-civil rights. At the same time, a progressive wing of the Democratic Party thought Truman wasn’t going far enough who wanted more sweeping reform in civil rights and joined Henry Wallace’s Progressive Party instead. According to Irwin Ross in The Loneliest Campaign, Henry Wallace left the Truman cabinet due to Truman’s aggressive stance against the Soviet Union in the Cold War. Wallace famously refused to disavow communism and often openly complimented Stalin during World War II when he was Roosevelt’s Vice President.

Going into the 1948 Presidential Election, Truman was very unpopular. On one occasion, according to the New York Times, Truman had a private conversation with popular general, Dwight D. Eisenhower, to seek nomination with Truman as his running mate. The main concern of Eisenhower and Truman was that General Douglas MacArthur, then the military governor of Japan, would win the Republican nomination, and Truman wanted to counter with the Democrats also having a military general as their nominee. However, Eisenhower declined. And after Eisenhower declined, there was no significant challenge to Truman’s nomination.

Running counter to Truman for the Republicans was Thomas Dewey, who ran against Roosevelt in 1944. Dewey was then the governor of New York. Dewey was seen as a clear favorite to win the 1948 Presidential Election because of Truman’s unpopularity.

Part of what made Truman so unpopular was the popularity of his predecessor, Franklin D. Roosevelt. Truman seemed doomed, but he was determined to go down swinging. Dewey, on the other hand, had a campaign policy of avoiding saying anything remotely controversial and speaking in platitudes to maintain his lead. In the words of one editorial in the Louisville Courier-Journal:

No presidential candidate in the future will be so inept that four of his major speeches can be boiled down to these historic four sentences: Agriculture is important. Our rivers are full of fish. You cannot have freedom without liberty. Our future lies ahead.”

John Baldoni at Forbes wrote that Dewey ran a campaign then very similar to what Hillary Clinton did in 2016. Dewey avoided the press. Dewey didn’t respond to political attacks from Truman, who called him a part of the “Do Nothing” Republican Congress. He wanted to take the “high road.” He avoided talking about any controversial issues, and thought that simply not being Truman would get him elected.

Truman went in the opposite direction for his re-election. Because he was trailing by so much in his polls, Truman held nothing back and ran a relentless campaign. According to Jules Abels in Out of the Jaws of Victory, Truman openly insulted Dewey and called him out, and launched intense criticism at the Republican-controlled Congress for gridlock. He likened the GOP to “gluttons of privilege” and called them “bloodsuckers with offices on Wall Street,” lines that age well for Democratic critics of Republicans today. At one point, Truman said:

“The Communists are rooting for a GOP victory because they know it would bring on another Great Depression.”

Truman mainly campaigned to four demographics: laborers, farmers, African-Americans, and consumers. Without much of the South voting for him in 1948, Truman often courted large crowds of farmers, blaming the ills of farmers on the Republican Congress. He portrayed the Republican Congress as conservative and obstructionist, despite the fact that Dewey himself was very liberal.

His fiery rhetoric never stopped. A popular slogan for Truman’s campaign was “Give ’em Hell, Harry!” The slogan would later be the subject of a biographical play about Truman, starring James Whitmore. He campaigned relentlessly across the country, and pollsters all thought the attempts were useless — Dewey was consistently polling with a lead almost in the double digits. In the words of W. Joseph Campbell, author of Lost in a Gallup: Polling Failure in U.S. Elections:

“The pollsters were saying it was a done deal; the press was buying into that narrative and the pundits were saying to everyone it was a sure thing that Thomas Dewey would be elected president.”

On October 31, 1948, the New York Times reported that Dewey would carry 345 electoral votes and 29 states. Even people close to Truman doubted he could win, including his wife, Bess Truman. According to William Manchester, author of Glory and the Dream, Truman was the only person who was confident in his chances of victory, and he proclaimed he would win and that his chances were high to anyone within earshot.

But crowds at Truman’s events would consistently be larger than crowds at Dewey’s events. Dewey’s advisors urged him to keep taking the high road and that he would effectively waltz into victory. But according to biographer David McCullough in Truman, some polls showed Truman cutting into Dewey’s lead. From late September to the end of October, Dewey’s national lead went from 17 points to five points.

According to Abels, in the last days of the campaign, Newsweek spoke with 50 experts, who all predicted that Dewey would win. According to McCullough, professional gamblers had betting odds against Truman around 15 to 1. Some places had his chances at 30 to 1. McCullough also notes that Truman had a lot of the most influential newspapers in the country supporting Dewey through endorsements. The Detroit Free Press called Truman “intellectually unqualified,” and the Chicago Tribune called Truman “an incompetent.” The Los Angeles Times called him “the most complete fumbler and blundered this nation has seen in high office in a long time.” In the words of McCullough, all odds were stacked against Truman:

“Besides the gamblers’ odds, the opinion polls, the forecasts by columnists, political reporters, political experts, Truman by now had the majority of editorial opinion weighed heavily against him.”

But Truman remained undeterred despite the forecasts of the experts. His advisors on his campaign called the campaign utterly exhausting, but Truman, according to McCullough “seemed indefatigable” and his outlook was “entirely positive,” and would frequently take 20-minute naps between speeches. Truman’s confidence never faltered. According to one of his advisors, Jonathan Daniels:

“Strain seemed to make him calmer and more firm.”

Truman’s staff struggled to keep up, working from 7 a.m. to past midnight on most days during the campaign. Truman’s attacks against Dewey rapidly escalated, and he at one point said that voting for Dewey was akin to voting for fascism. Dewey maintained his cautious and “high road” campaign and didn’t respond to the insult, which was clearly bait from Truman. At the last minute, Eleanor Roosevelt also endorsed Truman, previously refusing to do so because he was “such a weak and vacillating person.”

On Election Day, the Manchester Guardian, a newspaper in the United Kingdom, put out an article titled: “Harry S. Truman: A Study of a Failure.”

On Election Day, however, Truman won 28 states and carried 303 electoral votes, winning the popular vote by 3,100,000 votes. Dewey won 189 electoral votes, and Strom Thurmond carried four of the southern states. Dewey conceded, and Truman won three states by a very small amount of votes: Ohio by about 7,000, Illinois by 33,000, and California by 17,000. McCullough credits Henry Wallace being on the ticket in New York as part of the reason why Truman didn’t eke out a victory there, as well. Some analysts, including H.V. Kaltenborn at NBC, anticipated that Truman would lose after late votes came in, which never happened.

Analysts, reporters, and pollsters all made a look inward to see what could have possibly gone wrong. In the most egregious case in American history, the polls were wrong. Truman overwhelmingly carried Black voters, farmers, laborers, and Catholic voters.

As for the pollsters, Gallup said he didn’t know what happened. Elmo Roper said, “I couldn’t have been more wrong. Why, I don’t know.” Gallup had stopped polling in mid-October, and a massive shift had taken place between mid-October and Election Day. Essentially, the polls had been accurate until mid-October, but Truman’s relentless campaigning won a seismic shift in people who turned out to vote for him. According to Manchester, most pollsters believed in Farley’s Law, which deemed fall campaigning effectively useless, and dictated that the election would be decided by the conventions of the summer, and Farley’s Law essentially had to be scrapped.

Four years later, pollsters predicted that the 1952 Presidential Election between Democrat Adlai Stevenson and Republican Dwight Eisenhower would be a nail-biter. Gallup and Roper wanted to be more cautious given their mistake in 1948.

It was a landslide for Eisenhower.


I think Biden ran a campaign more like Dewey than Truman, but was able to still eke out a close victory over Donald Trump in the end. The takeaways from the 1948 Presidential Election are to maintain a healthy amount of skepticism in the polls. If polling shows your candidate is ahead, don’t stay home and say “they’re probably going to win anyway” as a reason not to vote. Likewise, if polling shows your candidate is way behind, don’t say “they’re probably going to lose anyway” as a reason not to vote.

Democrats this year didn’t trust Biden’s lead in the polls and maintained a certain level of anxiety — which, at the end of the day, was good for the Biden campaign since voters voting Biden never got complacent or overconfident. I held a near certainty that Trump would defy the polls again and confidence that Trump would win, and I’m glad that I’m wrong — and I never used that as an excuse not to vote for Biden.

Regardless of how wrong polls are in any election, they are not going away. If 1948 wasn’t the death of polling, and 2016 wasn’t, 2020 will not be the death of polling either. Whether you trust them, don’t trust them, love them, or hate them, one can argue that polls being wrong leads to more attention for polls in the future. There is always a consumer market for polls, and no credible replacement or alternative for them. Polling is a multi-billion dollar industry, and according to Campbell, shoe-leather journalism is not more reliable — extrapolating one rally to a whole demographic is more fallible to error.

Our desire for polls reflects a human desire to want to foresee and predict the future. People try to predict the future in other ways, like the stock market, with varying degrees of success, but clearly, predicting human behavior can never have 100% certainty. The experts may have better predictions and models than most of us normal people. But as we can see, expert models and analyses are never perfect either.

Until then, politicians and voters would do well not to be lulled into either a sense of complacency or defeat based on the polls— just ask Thomas Dewey how that worked out.

Originally published at Frame of Reference on November 7, 2020

Comments / 3

Published by

Believer, Baltimore City special ed teacher, and 2:40 marathon runner. Diehard fan of "The Wire," God's gift to the Earth. Support me:

Baltimore, MD

More from Ryan Fan

Comments / 0