Some of These Midterm Polls Really Bombed on Election Day

And the award for the least accurate polling record of the 2022 midterm cycle goes to…. keep reading below to find out.

Every year, pollsters get beat up for getting it wrong, in many cases, or for understating or overstating the support from one party or another. With most of the 2022 midterm elections now decided, it’s time to take a look over which polls got it right, and which polls were way, way off in one direction or another. We’ll examine a few races where the results are finalized and the end numbers won’t change much over the coming days.

It must be noted that most of these polls, even the ones that predicted the winner incorrectly, were still within the margin of error in most cases.

We will only be looking at races where the polling data over the final 30 days of the race sharply diverged by several points with a split decision on the possible outcome. For example, every poll agreed Florida Gov. Ron DeSantis was going to win, but the debate was by how much.

Pennsylvania Senate – Oz/Fetterman

This contest started out as a rout for Fetterman up double-digits for most of the race. In the end, the race significantly tightened but not to the extent that some pollsters were predicting.

Current Results

The Bad

The Good

New Hampshire Senate – Hassan/Bolduc

While it’s true that Don Bolduc was always a longshot challenge to Democratic incumbent Sen. Maggi Hassan, some polls near the end indicated an upset might be possible. On Election Day, no dice, as Hassan cruised to an easy victory.

Current Results

The Bad

The Good

Michigan Governor – Whitmer/Dixon

This is another race where Democratic incumbent Gov. Gretchen Whitmer wasn’t expected to have much of a challenge. The race tightened somewhat but in the end, Republican Tudor Dixon never really came as close as some polls suggested.

Current Results

The Bad

The Good

Wisconsin Governor – Evers/Michels

While Tim Michels was down most of the race, polling in the final stretch indicated that he may be able to beat the incumbent Democratic Gov. Tony Evers. It was also thought that if incumbent Republican Sen. Ron Johnson wins, which he did, it would pull Michels over the line. That scenario didn’t pan out and many polls overstated Michels’ support.

To be fair, this race was one of the closer races of the night. Evers only won by just over 3 percentage points meaning anything within the realm of a tie to +1 Michels wasn’t too far off from margin of error territory.

Current Results

The Bad

The Good

Conclusion

If there’s a trend to spot, it’s polling from Trafalgar and InsiderAdvantage overstating Republican chances in several key races. What’s clear is that the “submerged voter” effect was present on both sides, but also that some pollsters with traditional phone-heavy models were more accurate in some races.

Here’s a great exit question to consider. With the votes still being tallied, will this final Trafalgar Group poll of the Arizona Governor’s race be better or worse than some of those above:

Who should be more worried, Hobbs or Lake?

Republicans showed up to vote on Tuesday, but so did Democrats, and independents split more evenly than some polls and models predicted. The end result was a typical partisan pattern in many states that some pollsters flat-out missed or discounted in favor of the “red wave” narrative.

The races in Nevada and Arizona are yet to be decided so it’s not yet worth exploring those polls in a greater context.