Time to jettison horse race polls

Time to jettison horse race polls
© Getty Images

The pollsters egregiously erred in 2016 when they all but promised that Hillary Clinton would beat Donald Trump and become the next president of the United States. The Princeton Election Consortium pegged Clinton’s chances of winning at just under 100 percent. “It is totally over. If Trump wins more than 240 electoral votes, I will eat a bug,” said group leader Sam Wang. To his credit, Wang did just that. Another compiler of polls, Nate Silver, the prominent founder of Five Thirty Eight, was less sanguine about Clinton’s prospects. But he still concluded that she had a 71 percent chance of beating Trump.

The pollsters insisted that for 2020 they had remedied their past errors. They claimed to have improved their response rate, assembled better samples, and ferreted out a previously hidden vote for Trump and Republican candidates. According to an October 2020 report published by Five Thirty Eight, only one pollster, Gravis Marketing president Doug Kaplan, said he was worried about the “hidden votes” of Trump supporters.

However, the polls were more egregiously wrong this time. The compilers correctly predicted a victory for Joe Biden, but were wrong about everything else: Senate and House results, the national popular vote, and the vote in battleground states. As in 2016, the errors were not random, but consistently underestimated support for Trump and Republicans.


The Princeton Consortium projected that Democrats would take control of the Senate with a comfortable 53 seats. Silver concluded that the Democrats would win up to 52 Senate seats. Yet, the Democrats won 48 seats and have only a longshot chance of gaining 50 seats by sweeping both Senate runoff elections in Georgia this January.

The compilers badly missed the Democratic debacle in the House elections with the party on track to lose upwards of nine seats and hold one of the thinnest majorities in modern history. The Princeton Consortium indicated that Democrats would expand their majority by five seats. Silver projected a gain of seven seats. Silver also pegged Biden’s lead in the popular vote at 8.4 percent, more than 4 points higher than Biden’s actual lead of just under 4 percent.

Silver has claimed that his presidential compilations were highly accurate because they correctly predicted the winner of 48 of 50 states. This is a badly misleading metric because the election was not seriously contested in most states. There were only 10 true battleground states, won by either Clinton or Trump by less than 5 percent in 2016.

Both Silver and the Princeton Consortium miscalled two of battleground states, Florida and North Carolina, where they incorrectly projected that Biden would win. These states accounted for 45 Electoral College votes, 37 percent of these votes in the battleground states. The pollsters also had us believe that Iowa, Ohio, and Texas, won by Trump in 2016 by 8 points to 9 points were also battleground states because their polls showed a thin, statistically insignificant lead for Trump.

Wrong again. Trump won these states by 6 to 8 points this year. Silver’s projections of the vote in the battleground states were strewn with errors, all them in the one direction of underestimating the vote for Trump. In Wisconsin, Silver was off by 7.8 points, in Florida by 5.9 points, and in Michigan by 5.2 points.


Given their false assurances for 2020, we have no reason to believe that the pollsters and their compilers will miraculously find the holy grail of accuracy for 2024. Horse race polling, moreover, misleadingly turns our attention to the daily ups and downs of the campaign, rather than the underlying dynamic of presidential elections, which is a verdict on how well the incumbent president governs the nation.

By keeping my eye on the big picture of incumbent strength and performance through my 13 Keys to the White House I was able to predict Biden’s win in early August, well before the polls were of any use. Once liberated from the tyranny of horse race polls, presidential candidates could conduct a different kind of campaign focused not on sound bites and attack strategies, but on building a mandate for governing.

In response to criticisms of his poll compilations, Silver used profanity and said Five Thirty Eight did fine. Silver needs to find another job or stick to sports betting.

Allan Lichtman is an election forecaster and a distinguished professor of history at American University. He is the author of “The Embattled Vote in America: From the Founding to the Present.” He tweets @AllanLichtman.