As the nation approaches Election Day, a razor-thin margin separates the two main candidates in the 2024 U.S. presidential election, Vice President Kamala Harris and former President Donald Trump. With polling day just one day away, recent election polls underscore a nearly unprecedented level of uncertainty, with Trump and Harris locked in one of the closest races in recent history. Across the United States and especially in battleground states, both candidates remain neck-and-neck, sparking questions of who is winning the election and whether we could be heading toward yet another polling misfire.
Polling averages from major aggregators, including FiveThirtyEight and The New York Times, show Harris holding a minuscule lead in the national popular vote, with her edge often well within the margin of error. This slim advantage mirrors the deadlock in state-level polling, particularly in seven key swing states—Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania, and Wisconsin—where either candidate could feasibly emerge as the victor. This tight margin has led to speculation that voters may not learn who’s won the election on Election Day itself, or even in the days immediately following, especially if recounts are triggered in states like Pennsylvania, where a narrow half-percentage-point gap could legally mandate one.
The scrutiny on election polls has intensified following notable misses in the 2016 and 2020 elections, where polls underestimated Trump’s support both times. Recognizing these gaps, many polling organizations have since updated their methods, introducing changes to adjust their samples and weighting protocols. They have diversified their survey methods, including more online and phone-in responses to reach a wider demographic that reflects the voting population more accurately.
Even so, it remains challenging to gauge certain demographics accurately, especially Trump supporters who are reportedly less likely to participate in polls compared to other voters. This trend—known as “non-response bias”—remains one of the polling industry’s most significant issues, as under-sampling of this voter base can lead to inaccuracies. In Wisconsin, for example, research suggests that the state’s voter files may misclassify some voters’ political affiliations, contributing to disparities between predicted and actual outcomes.
In a close race like the one unfolding between Trump and Harris, turnout models could be the deciding factor in predicting the eventual winner. Pollsters rely heavily on assumptions about likely voter turnout, but with the candidates polling so closely, these models are likely to face intense scrutiny after Election Day. Pollsters are particularly attuned to potential shifts in the gender gap this year; they note that female voters, who constitute a majority of the electorate and typically vote at higher rates than men, could be crucial for Harris. Conversely, Trump is expected to perform strongly among male voters, potentially balancing out this demographic split in states where either candidate could win.
Historically, the results of U.S. presidential elections have sometimes been called on election night, but recent years have shown that a delayed result is entirely possible, especially in a close race like this one. The first states to close their polls will do so at 6 p.m. EST on Tuesday, with final counts and projections likely extending late into the night. In some swing states like Pennsylvania and Wisconsin, counting processes will continue well beyond Election Day due to mail-in ballots and potential recounts, meaning a winner might not be determined immediately.
In past elections, recounts, legal challenges, and even potential disruptions at polling locations have contributed to delays. For example, in 2020, Joe Biden was not declared the winner until four days after the election due to an extensive vote count in Pennsylvania. In 2000, the Florida recount delayed the final outcome by weeks, until the U.S. Supreme Court’s ruling handed George W. Bush the victory. Election experts say such scenarios are possible again this year, depending on how close the Trump vs. Harris polls remain as votes are counted.
Ultimately, while the latest election polls offer a snapshot of current sentiment, experts caution that they do not serve as definitive predictions. The models indicate that this is still anyone’s race, and that, come Election Day, both Trump and Harris have equal paths to victory. With turnout likely to play a crucial role and the potential for polling error on either side, Americans will soon see if the adjustments to polling methodologies have produced more reliable results—or if the polls will once again miscalculate in another highly competitive U.S. presidential election.