Election Day is the political Super Bowl. However, the post-game analysis is much better in politics than in sports.

I began my career as a research geek. At heart, I still am. I love poring over election data and exit polls and examining how the pollsters did. Like this election season, the aftermath is nothing short of amazing.

You may have heard how Democrats stayed home and some of the rationalizations or conspiracy theories from the left and right about what happened. 

Digging into the numbers the day after the election, there appeared to be at least eight million fewer votes than in 2020. The final number will be closer to 4.4 million. In 2020, well over 158 million Americans voted. In 2024, the total dropped to a little over 154 million.

Harris received nearly 6.8 million fewer votes than Biden in 2020, while Trump improved his total by almost 3.4 million from four years ago. 

The exit polls didn’t ask respondents who they voted for in 2020. Therefore, we can’t estimate how many 2020 Biden voters switched to Trump this year. 

Trump’s additional votes plus the number of fewer total votes comes out to about 7.8 million, 1.1 million more than the number Harris lost from Biden’s 2020 total.

Thirty-three states had lower vote totals than in 2020, while eighteen had higher totals (it adds to 51 because of Washington, D.C.).

All seven battleground states (Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania, and Wisconsin) saw vote totals increase over 2020.

Of the eleven other non-battleground states that saw increased voter totals, slightly more red states, seven (Idaho, North Dakota, Oklahoma, South Carolina, South Dakota, Tennessee, Texas) had increased totals than blue states, which are four (Delaware, Maine, New Hampshire, Virginia). The red states with increases also represented a more diverse geographical area.

The data suggests there was no conspiracy or massive fraud in either 2020 or 2024. In the battleground states where the media and talking heads told Americans the election would be decided, vote totals were up across the board.

Harris’s totals were down in 46 states and up in five, four of which were battleground states (Georgia, Nevada, North Carolina, and Wisconsin). The final state where Harris improved over Biden’s total was, interestingly enough, Utah.

Trump was up in 39 states and down in 13. He was up in all seven battleground states by margins that ranged from 81,000 in Nevada to over 200,000 in Georgia. The thirteen states where Trump lost votes were a mix of blue and red but were all states where he either won or lost decisively in 2016, 2020, and 2024.

Pollster Report Card

Most national polls showed a race so close that the pollster can legitimately claim their result was within the margin of error of Trump’s win. The same is true of most of the state polls, too. 

Nonetheless, some polls showed Harris trending up in the final days, and their last results suggested a Harris win. Other polls showed the opposite result, suggesting a Trump win. 

I looked at the results from every pollster that made the results public. I also examined the results of the three largest aggregators tracking the races: Real Clear Politics (RCP), Nate Silver, and 538. 

RCP strictly aggregates the numbers of the polls it includes, while Silver and 538 weight polls and include other factors they consider important to reach their conclusions about races.

Of the three, RCP fared best in the 2024 presidential and key senate elections. Its final average showed the presidential race a 48.5 to 48.5 tie. Silver rated it 48.5 to 47.7 for Harris. 538 predicted a 48.1 to 46.8 victory for Harris. Trump is at 50.0 and Harris at 48.4 for a 1.6 margin.

In the seven battleground states:

RCP predicted five correctly and two incorrectly.

Silver also went five and two.

538 was four and three.

All three incorrectly predicted Michigan and Wisconsin for Harris. 538 also had Pennsylvania narrowly for Harris, while RCP and Silver gave a slight edge to Trump in the Keystone State.

So much for Nate Silver and 538’s superior analysis.

Among individual pollsters, two stand out. The winner of the gold medal is Atlas Intel. Atlas called the presidential race 50-48. They predicted the margin would be 1.7 points. On election night and the first few days after that, it appeared precisely correct – dead on. Since a few states have continued counting votes nearly three weeks after Election Day, the margin has dropped by one-tenth of a point to 1.6.

Atlas accurately predicted Trump would win all seven battleground states.

Atlas missed two Senate races: Dave McCormick’s upset of Bob Casey in Pennsylvania. Atlas called the race between Ruben Gallego and Kari Lake in Arizona even — Gallego won it. 

Of the fifteen races Atlas Intel polled, it called thirteen right for an 87 percent success rate, the best of any polling firm measuring more than two races. Further, the average margin they were off across those fifteen races was 1.10, the lowest of any political polling company in the 2024 cycle.

The Silver medal goes to Emerson College Polling. Emerson’s final presidential survey was 49.3 to 48.6 for Harris, so they missed the national vote but called six of seven battleground states – missing only Michigan. Emerson also polled Iowa. The weekend before the election, respected pollster Ann Seltzer made headlines with her Des Moines Register Poll showing Harris up by three in Iowa. Emerson showed Trump up nine. He won by thirteen.

Of eight Senate races Emerson measured, they got seven right. Emerson called the Pennsylvania Senate race even.

Overall, of seventeen races Emerson College polled, they got fourteen correct and three wrong (including the national presidential result) for an 82 percent success rate, second only to Atlas Intel. The average difference between their predicted margin and the actual result was 2.72 points.

Others who were successful (60 percent correct or better with thirteen or more races) included Rasmussen, Insider Advantage, and Trafalgar.

The two biggest losers in the 2024 cycle are the New York Times/Siena College poll (often considered the gold standard in polling). Of fifteen races, they got seven right and eight wrong, a 47 percent success rate. Their average margin was off by 3.14 points.

Likewise, Marist College, which measured thirteen races, went six and seven (46 percent). Their average margin was off by 3.68 points.

Polls I haven’t mentioned here were 50/50, at best, or didn’t poll enough races to be able to judge fairly.

The data from the 2024 election is compelling. Although the nationwide vote total is down by nearly 4.4 million votes, it increased in all the battleground states. People vote when they know their vote matters.

Remember these results as new polls tell the public how voters feel about Trump and the new Republican Congress.

As the Democrat’s postmortem continues, they will blame the polls, racism, misogyny, and specific groups who voted for Trump in record numbers. Still, even the polls that were off the most told them they were wrong on the issues and that their candidate wasn’t answering voters’ concerns. Until they accept these facts, they’ll never learn.

Andy Bloom is President of Andy Bloom Communications. He specializes in media training and political communications. He has programmed legendary stations including WIP, WPHT, WYSP/Philadelphia, KLSX, Los Angeles, and WCCO Minneapolis. He was Vice President of Programming for Emmis International, Greater Media Inc., and Coleman Research. Andy also served as communications director for Rep. Michael R. Turner (R-Ohio). He can be reached by email at andy@andybloom.com or you can follow him on Twitter at @AndyBloomCom.