The New York Times’ chief political analyst, Nate Cohn, expressed concern that pollsters may not have addressed the issues they faced in the 2016 and 2020 elections and could be underestimating Donald Trump once again. Cohn highlighted the potential for survey bias due to the fact that white Republicans, a key demographic for Trump, are less likely to respond to polls compared to Democrats. The phenomenon of nonresponse bias was a factor identified in the polls underestimating Trump in the past.
Cohn, an expert in election data, emphasized the challenge of tracking nonresponse bias and the need to analyze how Democrats and Republicans respond to pollsters in order to gain insights. He pointed out that white Democrats were more likely to respond to polls than white Republicans, raising the possibility of Trump being underestimated once more. Cohn also noted that Democrats seem to be losing their early-voting advantage, potentially benefiting Republicans on Election Day as GOP voters are expected to turn out in higher numbers.
Trump’s campaign has been encouraging Republicans to vote early in an effort to maximize turnout among the base. A recent poll conducted by The New York Times and Siena College in battleground states showed that six out of the seven key close contests were within the margin of error or too close to call, with only Arizona leaning towards Trump. Late-deciders were found to be favoring Vice President Kamala Harris, which could have a significant impact on the election outcome.
There is uncertainty surrounding the outcome of the election, as highlighted by Cohn. Concerns have been raised about pollsters potentially herding survey results to avoid being outliers, leading to less variety in polling of the 2024 presidential contest than expected. Trump campaign pollster Tony Fabrizio raised objections to the methodology of the latest survey, arguing that it skewed toward the left and had difficulty reaching Republican voters, similar to past inaccuracies in polling.
In the most recent RealClearPolitics aggregate of national polls with multiple candidates, Trump and Harris are tied. Trump is leading in most of the RCP’s battleground state polling averages, indicating a close and uncertain race. The concerns raised by Cohn and Fabrizio suggest that the accuracy of polling in the 2024 election remains uncertain, with potential biases that could impact the outcome. These factors add to the unpredictability of the election and raise questions about the reliability of polling data.
