Ron Johnson said people shouldn’t trust early polling. His position is more nuanced than that.

It’s not much of a surprise that President Joe Biden and former President Donald Trump are locked in a close race in battleground Wisconsin.

Trump had a 2-point lead among both registered and likely voters in Wisconsin, the latest Marquette University Law School Poll found in April. 

Other polls released in May showed Biden with a lead in Wisconsin. A New York Times/Siena College poll found Biden led in Wisconsin, but not five other battleground states. 

In that poll, Biden was up two percentage points among registered voters in Wisconsin. Trump, however, had a 1-point lead among likely voters.

A Quinnipiac poll showed Biden with a much larger, 6-point lead: 50% to 44% over Trump among registered voters in Wisconsin. 

Sen. Ron Johnson, R-Wis., was asked about those Wisconsin polls in a May 19 interview on NewsNation’s "The Hill Sunday." 

"As somebody who has run statewide three times, and seen polls wildly incorrect all three times, I just would not trust the early polls," Johnson said.

Johnson added there is no doubt that "Wisconsin is going to be a very close state."

PolitiFact Wisconsin was interested in Johnson’s claim, especially because polls of the Biden-Trump rematch in Wisconsin have shown some variation.

First, we will look back at polls during Johnson’s three runs for U.S. Senate to see where they fell. Then, we’ll add some context about what to keep in mind when looking at early polls.

When we asked for backup, Johnson spokesperson Kiersten Pels shared a few poll examples, including a poll in which Johnson’s lead was much higher than his final margin and some in which his opponents were polling quite a bit above him, even in the weeks before the election.

"Simply put, the polls were significantly wrong, especially in the last two races. He instinctively knew those races would be close, he ignored the polls, and was proven correct," she said. "The elections were close." 

Since some polls are better than others, let’s look at averages of multiple polls collected by RealClearPolling.com during Johnson’s three U.S. Senate runs – in 2010, 2016 and 2022. 

In 2010, Johnson ran against Democratic incumbent Russ Feingold. 

According to RealClearPolling’s averages, Feingold was about 2 percentage points ahead of Johnson in July and August. Then Feingold was up about 1 point through mid-September.

After the primary, Johnson pulled ahead of Feingold, polling 6 to 10 percentage points above Feingold for the rest of the race. Johnson won that election with 52%, compared with Feingold’s 47%.

This election was a rematch of Johnson and Feingold. By that point, Johnson would have had more recognition as an incumbent. S,o if that was a factor in early polls in 2010, it was less so in 2016.

Feingold had a lead over Johnson the entire race, according to RealClearPolling’s averages. Johnson trailed by more than 12 points between April and August of 2015 — which were very early polls.

Feingold was still leading by upwards of 10 points for most of the campaign, until it narrowed in the last month or so. Johnson pulled off an upset that year, winning with about 50%, compared to Feingold’s 47%.

Those averages are a little easier to parse. Democratic challenger Mandela Barnes had a 4-to-5 point lead until mid-September. 

Around the time of the primary, their leads switched. Johnson stayed about three points ahead of Barnes for the rest of the campaign. Johnson won by just a percentage point: 50.5% to Barnes’ 49.5%.

To summarize: Polling conducted in the early months of the election cycles did show Johnson trailing, though he won all three elections. That’s most obvious in 2016, when Feingold led essentially the entire time. 

Featured Fact-check

Robin Vos

stated on November 6, 2024 in News conference

“Every single county in America, every single county, Kamala Harris did worse than Joe Biden did.”

But does that mean the early polls are "wildly incorrect," as Johnson said? It’s a little more complicated than that. 

Polls are a snapshot in time rather than a predictor, Johnson’s 2016 race is a good example

PolitiFact Wisconsin reached out to Charles Franklin, director of the Marquette University Law School Poll, for context. 

"It’s an old line, but polls are a snapshot in time," Franklin said. "A poll really is a measure of where opinion stands at the time that the poll was taken."

Although it’s inevitable that people will look at polls today as a prediction of where the race will end up months from now, that’s not the right way to look at it, he said. 

Instead, polls "capture the dynamics of a campaign," he said. Johnson’s run in 2016 is one of Franklin’s favorite examples of that. 

"I think the polling captured the dynamics of that race extremely well, showing Johnson starting out trailing, but showing him running an effective campaign that closed the margin consistently, and ultimately put him in a position to win," he said.

Polls do get "better as predictors as you get later in the campaign," Franklin said, and especially in the last two or three weeks. But even then, they’re not infallible. 

Johnson said he saw early polls that were "wildly incorrect" during each of his three statewide runs. 

The heart of Johnson's claim is accurate: Polls showed him trailing in the early months, despite pulling off a victory each time. That was most clear in 2016.

When Johnson said not to trust the early polls, he wasn’t suggesting there was something wrong with the polling methodology itself. Instead, he’s getting at the idea that his polls didn’t match the final outcome.

Still, that doesn’t necessarily mean that the polls were "wildly incorrect." Polls aren’t supposed to predict the final election outcome, but show a snapshot of the campaign in time. 

Our definition of Mostly True is "the statement is accurate but needs clarification or additional information." That fits here.