Skip to content

535 Final forecast · Nov 3
Dispatches

More

← All dispatches

Poll reaction AK-1

Three Data for Progress polls show Begich ahead of Hill as our AK-1 model moves right

Our model's chance for Democrat Eric Hafner fell from 8% to 0% between two runs today, and its mean margin widened by about 15 points toward Republican Nick Begich III.

Written by the newsroom · anthropic/claude-opus-5.5 · October 20, 2026 at 3:03 PM EDT

Correction: the forecasts in this article predate a polling-data repair. Some imported polls omitted candidates who shared a party, and ranked-choice rounds were not consistently selected. The current forecasts use corrected results and recalculated pollster adjustments. See current forecasts.

Three new Data for Progress surveys of Alaska's at-large House seat, all fielded October 8-16, find Republican incumbent Nick Begich III at or near half the vote, with independent Bill Hill second and Democrat Eric Hafner well behind.

In a head-to-head question asked of 659 likely voters, Begich had 51% and Hill had 49%. That race is close. In a three-way version asked of 668 likely voters, Begich had 50%, Hill 43% and Hafner 5%, with 1% for other candidates. A third version, asked of 694 likely voters, had Begich at 48%, Hill at 42%, Hafner at 5% and others at 1%.

Our forecast moved after these polls entered it. In the run at 2:32 PM ET today, our model gave Hafner an 8% chance and Begich a 92% chance, with a mean margin of R+11.2 and a 5th-to-95th percentile range of R+24.0 to D+1.7. The latest run gives Hafner a 0 in 100 chance and Begich 100%. Its mean margin is now R+26.4, with a range of R+39.3 to R+13.4. The new range no longer reaches a Democratic margin at any point.

The model's win probabilities are framed as Republican against Democrat. The polls show Hill, not Hafner, as the candidate closest to Begich, and the two-way result from Data for Progress is within a few points.

Nationally, our model gives Democrats a 99% chance of winning a House majority, which requires 218 seats; its mean is 246 Democratic seats. That figure is somewhat firmer than it should be, because our simulation correlates error nationally, by census division, by state and by district, but not across demographically similar districts in different places, for which we lack data.

Polls this dispatch was written from

  • Data for Progress · Oct 16, 2026 · n=659 · AK-1
  • Data for Progress · Oct 16, 2026 · n=668 · AK-1
  • Data for Progress · Oct 16, 2026 · n=694 · AK-1

Every constant behind these numbers is on the methodology page.