Since Donald Trump emerged as a political force, U.S. pollsters have struggled to explain the three-time Republican nominee’s enduring popularity. They underestimated him in 2016. Despite his eventual loss, they also underestimated the extent of his support in 2020.
With that history, polling organizations are taking extra care this year to make sure that they are taking into account Trump’s base—a group that may be less likely to answer their phone calls and more disillusioned by the system that polling organizations represent.
Along comes the oracle of Iowa. Over the weekend, J. Ann Selzer dropped her closely watched Des Moines Register poll showing Democratic nominee Kamala Harris leading Trump by 3 percentage points in a state that was seen as a shoo-in for the former president. Selzer’s firm has correctly predicted results in Iowa in four of the last five presidential elections. If Selzer is right, her poll could also imply a similar swing toward Harris in Wisconsin or Michigan.
Speaking on MSNBC’s The Weekend, Selzer said her philosophy was “to allow our data to reveal to us what’s happening with the future electorate.” In the latest poll, her team wasn’t looking to the past, and it wasn’t putting its thumb on the scale to correct for past errors. They were just presenting the data as it was—and the data showed that women voters were likely reacting to the abortion ban that went into effect in Iowa in July.
We’ll find out if Selzer is right or wrong in the coming days. But the results will also go some way toward answering whether pollsters have gotten better at predicting how Americans think about Trump—or if this time they have overcorrected in predicting the extent of his support.
This post is part of FP’s live coverage with global updates and analysis throughout the U.S. election. Follow along here.