Why did polls blow the election? They failed to account for independents

White, suburban, college-educated mothers. Second-generation Hispanic males in the Rust Belt. There are many demographics that Donald Trump overperformed this election on his way to victory. The polls, by and large, did a good job of picking up this sentiment.

But pollsters — myself included — modeled likely voter turnout wrong.

We did not model an electorate with both major parties in decline and independents surging to overtake the Democrats. The result was that Trump won decisively while most polls suggested a dead heat.

Don’t be fooled by pollsters who flipped a coin and guessed Trump — they were wrong, too.

I have yet to find a pollster with an Election Day model that showed these improbable movements: Democrats turning out 6 points less than in 2020, independents turning out more than Democrats (up 8 points from 2020), and Republicans making up a smaller percentage of the electorate than in 2020.

Consider Pennsylvania. From 2016 to 2020, the percentages of independents and Democrats decreased by one and two points, respectively, while Republicans increased by two points.

Then, the political winds shifted wildly. From 2020 to 2024, Democrats decreased by four points and Republicans by one point, while independents increased by about the difference, 4 points.

The Republican decline there was similar to that in other swing states and nationwide.

The notion that Trump won because Republicans turned out en masse for him is incorrect. Additionally, Republicans didn’t vote for Trump with nearly the same intensity as Democrats voted for Kamala Harris.

Here again, Pennsylvania is indicative of the toss-up states and the nation: 91% of Republicans voted for Trump, while 95% of Democrats voted for Harris. In other words, Republicans saw less turnout and less support for Trump among those who voted.

What put him over the top? Independents.

Arizona illustrates this well: From 2020 to 2024, Republicans decreased as a percentage of the state’s electorate, and Trump only received 93% of the Republican vote.

In comparison, Harris won 99% percent of the Democratic vote. Once again, Republican turnout and support for Trump were slipping.

Yet Trump won the support of Arizona’s largest political group, independents, by 9 points, and beat Harris 52%-47%.

This data is essential for us pollsters to understand what happened and improve in the future.

The surge in independents was unprecedented, but it should have been considered in hindsight.

This highlights the problem with polling: We have no idea who’ll actually show up to vote. We can ask and gauge sentiment, but we never truly know.

It becomes tricky when, for example, Democrats support Harris with such intensity, but their share of the electorate plummets. High intensity among likely voters typically translates to a large voter turnout on Election Day. However, that wasn’t the case this time.

Remember, the Republican share dropped, and Trump’s support was lukewarm.

This was not a “Republican tsunami.” The toss-up states Trump won are almost identical to Biden’s path to victory in the last election. Was 2020 a “Blue Wave?” No, and no one believes that despite Biden winning the popular vote with the largest vote total in American history.

Notably, Democrats won Senate races in Arizona, Michigan, Nevada and Wisconsin — all states that went for Trump. They also won the governor’s mansion in another Trump state, North Carolina. This suggests more split-ticket voting, which may be due to independent voters.

Bottom line: Any pollster who didn’t account for the surge of independents and the drop in traditionally Republican and Democratic voters got it wrong.

Independents now represent the second-largest voting political affiliation, just one point behind Republicans and three points ahead of Democrats. No one accurately modeled that.

Whenever I’ve been asked to make predictions during election season, I always gave the same answer: Whoever wins the independents.

Yet this election reshaped the landscape so much that I can no longer rely on my standard four-word answer. It’s more complicated than that.

Will the two major parties continue to lose ground to independents? Will split-ticket voting become the norm? Are independents threatening traditionally red or blue congressional seats? Is there a better measure of voter turnout?

These questions should guide our future models.

Pollsters understand the value of asking good questions better than anyone else. The answers to those questions will ultimately determine who learned from 2024 and who did not.

Brett Loyd, the Republican National Committee’s polling director from 2008 to 2014 and a pollster for Trump’s 2016 and 2020 campaigns, is president and CEO of the nonpartisan Bullfinch Group and a contributor for the Independent Center.