State Navigate, a polling outfit working Iowa State Senate District 35 in eastern Iowa, ran a survey the first week of October and found Democrat Mike Zimmer leading Republican Joe Stutting by 4.6 points. Standard weighting shaved that to a 3-point Zimmer lead. Then they did something almost nobody in the polling industry does: they asked whether their own numbers were wrong.
They were.
When State Navigate reweighted answers to account for how respondents actually voted in 2024 — a year Donald Trump carried that district by 21 points — Zimmer's lead collapsed to 2 points. In the governor's race, where Democrat Rob Sand had been showing a comfortable 9-point cushion over Republican Zach Lahn, a similar recalibration produced a dead heat: 45% to 45%. A 5-point swing, buried inside a spreadsheet formula.
Chaz Nuttycombe, executive director of State Navigate, laid out the problem. His team identified two structural flaws in Iowa polling that nobody in the mainstream press seems interested in discussing. First, more Trump voters in that district are misreporting their 2024 presidential vote than in any other area State Navigate has polled. Second — and this is the one that should keep every Democratic strategist up at night — rural Republicans in Iowa are, in Nuttycombe's words, "very unresponsive to surveys."
That's not a minor methodological footnote. That's the foundation the entire "blue wave" narrative is built on. CNN ran polls this week showing Iowa suddenly in play for Democrats in both the governor's race and the U.S. Senate contest, where Democrat Josh Turek was clinging to a 1-point lead over Republican Ashley Hinson. The coverage treated these numbers like hard news. State Navigate just explained why they might be fantasy.
The district in question — covering Clinton, Camanche, DeWitt, and Maquoketa, plus rural stretches of Clinton, Scott, and Jackson counties — flipped to Democrats in a special election on January 28, 2025. That flip became exhibit A in the case that Iowa was trending blue. What the coverage left out was that special elections have notoriously low turnout, favor the party with the energized base at that moment, and tell you almost nothing about a general election with a fully mobilized electorate.
Nuttycombe's findings echo what Echelon Insights' Patrick Ruffini has been flagging about Republican survey responsiveness nationally. The voters most likely to show up for the GOP on Election Day are the same voters most likely to hang up on a pollster. That creates a structural tilt in every survey that relies on who picks up the phone — which is most of them.
Skepticism is mounting about the narrative that a blue wave is set to crash in the midterms. The polling industry that gave Hillary Clinton a 98% chance in 2016, missed the red wave in Florida in 2022, and had Ann Selzer's final Iowa poll showing Kamala Harris winning the state by 3 points before Trump carried it by 13 — that industry is now asking America to trust its Iowa numbers again.
State Navigate didn't have to publish its reweighting analysis. Most pollsters would have released the topline showing Zimmer ahead and moved on. Instead, they showed their work — and the work showed that the Republican was closer than anyone in the press wanted to admit.
The methodology that produced a 5-point Democratic lead and the methodology that produced a tie used the same raw data. Same voters. Same answers. The only difference was whether the pollster accounted for the fact that Trump voters don't like talking to pollsters.
Four weeks before the midterms, in a Trump +21 district, the best-case scenario for Democrats requires assuming the polls got the electorate right. State Navigate just told you they didn't.