Skip to content

535 Final forecast · Nov 3
Dispatches

More

← All dispatches

Poll reaction NH-1

St. Anselm poll trims Shaheen's NH-1 edge and widens the range of outcomes

A 4-point margin in the new survey cuts our projected margin by more than 5 points, and our range now reaches a narrow Republican win at its low end.

Written by the newsroom · anthropic/claude-opus-5.5 · October 22, 2026 at 5:03 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.

A St. Anselm survey of 861 likely voters in New Hampshire's 1st District, fielded October 19-20, puts Democrat Stefany Shaheen at 46 percent and Republican Anthony DiLorenzo at 42 percent. That 4-point gap is narrow enough that the poll shows a close race. It is also far tighter than our model had projected for the district.

Our forecast moved after the poll was added. Our model now gives Shaheen a 95 in 100 chance of winning the seat, down from 99 percent in the run earlier today. Our projected mean margin fell from D+18.1 to D+12.5.

The bigger change is at the edges. Before this poll, our model's 5th-to-95th percentile range ran from D+5.3 to D+30.9, entirely on the Democratic side. It now runs from R+0.2 to D+25.3. A narrow DiLorenzo win now sits inside that range, though only at its far end. DiLorenzo's chance in our model rose from 1 percent to 5 percent.

Shaheen remains the clear favorite in our forecast. The St. Anselm result is a single poll, and our model weighs it against everything else it uses for the district. But it has moved the race from one our model treated as nearly settled to one with a measurable Republican path, with 12 days until the November 3 election.

Nationally, our model gives Democrats a 99 percent chance of winning a House majority, with a mean of 246.6 seats against the 218 needed. That figure is somewhat firmer than it should be: our simulation correlates errors nationally, by census division, by state and by district, but it does not link demographically similar districts in different places, because we lack the data to do so.

Polls this dispatch was written from

  • St. Anselm · Oct 20, 2026 · n=861 · NH-1

Every constant behind these numbers is on the methodology page.