What Iowa's surprising polls mean

The U.S. election had a small “November surprise” over the weekend: a survey suddenly found Kamala Harris leading in Iowa. Iowa is not actually the Democratic choice in 2024. That may say something about the election results.

follow me

J. Ann Selzer's surveys are considered particularly good; they are among the best in the country. They are known for being very accurate, especially because of their performance from 2008 to 2020. That's for sure – despite Donald Trump's recent claims otherwise and a head-on attack on the 68-year-old.

Of course, he's doing this because he simply doesn't like Selzer's latest poll. Their research institute, Selzer & Company, released the latest data on voting intentions in Iowa over the weekend. Their performance was surprising: Kamala Harris suddenly led by three percentage points.

Iowa, a 'staunchly Republican' state in the Midwest

This is surprising because Iowa is not actually a swing state. Trump won by 8 points in 2020, up from 9 points four years ago. The independent Cook Political Report rates Iowa as “solidly Republican.”

Of course, individual surveys now have to be compared to their peers. Any analysis that looks at a single survey in isolation is incomplete and potentially flawed. Because in fact, Selzer's numbers are the only ones so far that show Harris leading in Iowa. In every other poll this year, Trump leads by at least four points. All polling averages show Republicans leading by a few percentage points — even when taking into account the latest data. Furthermore, Selzer's three-point lead for Harris is still just within the statistical margin of error of 3.4 percentage points.

So why should you care about Selzer’s investigation?

J. Ann Selzer, Fugitive Queen

For one thing, the data from the institute, founded by J. Ann Selzer in 1996, are the most accurate in the country because of the facts mentioned above. On the other hand, as The New York Times reports today, this isn't the first time Selzer & Company has been in this situation.

In the final week before the 2020 election, the institute released a poll in Iowa that showed Trump leading the state by 7 points. It's an outlier: Other polls show the race is much closer, with Trump ahead by an average of just over a percentage point. Selzer's polling two months ago showed Trump and Joe Biden in a close race.

In the end, Trump won the state with six electoral votes, 53% to 45%—8 percentage points.

J. Ann Selzer shares her surprising poll with CNN's Anderson Cooper.Bild: CNN screenshot

J. Ann Selzer told The New York Times that she has been the queen of anomalies many times. “I'm not neurotic.” Surveys from 2016 also underlined this point: At the time, Selzer had Trump leading by 7 points, even though overall polls showed his lead was much smaller, at 3 points. He won the state championship that year by nine points.

The story goes back even further: In 2008, Selzer released a poll showing a relatively unknown black Illinois senator in the lead in the 2008 Iowa Democratic primary. , which even surprised experts and analysts. That senator was Barack Obama, who famously won the Iowa caucuses and overall election, clearing his way to the presidential nomination and ultimately the White House.

When even your own results are questioned

Even though J. Ann Selzer is used to generating outliers in polls, she said she, too, had to go short at first.

According to the New York Times, public opinion researchers in the United States sometimes succumb to a phenomenon known as the “herd instinct.” This occurs when a research institution releases its results only if they agree with those already established in other or previous surveys. But J. Ann Selzer's willingness to publish her results, even if they are outliers, is one of the reasons she is widely considered a trustworthy pollster.

She never believed her institute would be led by Harris. But after careful checking, analysis and discussion of the data, it was concluded that there were no obvious errors.

Instead, she even found some potential outcomes that could explain what might have happened.

More women want to vote

A possible trend has emerged in a previous survey Harris conducted in Iowa. Trump still led by four points in September. However, this is already slightly less than in June. But most importantly another trend has emerged.

The share of respondents who said they were very likely to vote (or had already voted in the previous survey) was significantly higher in this and the previous Selzer survey than in the June survey.

In particular, more women, younger and educated voters say they want to vote now.

Women may have a possible victory thanks to Kamala Harris, who campaigned in Michigan. Image: trapezoid

In recent polls, a majority of voters age 65 and older also support Harris. That's largely thanks to women: 63% of women 65 and older say they want to vote for Harris (28%) — a huge lead. Independent voters, especially independent women, also lean toward Ms. Harris in new polls.

“It’s like people stood up from the bench and said, ‘OK, I’m going to vote,’ and they were Harris supporters,” Selzer said.

Selzer's survey is likely an outlier because their survey could also be wrong. There's also a good chance Iowa will actually vote Democratic but will be an outlier nationally. Still, the New York Times concluded: “Even if Harris doesn't win Iowa, the idea that she's making headway in the Midwestern state could give some hope to Democrats' chances in battleground states.”

Selzer said she was prepared for Trump to eventually win in Iowa and wasn't concerned that her polling might be wrong. Eventually, she becomes either “gold” or a “skunk” among public opinion researchers. How does she react to this? «I'll take it like a big girl. I won't shed a tear. “