Skip to content

535 Final forecast · Nov 3
Dispatches

More

← All dispatches

Poll reaction FL-25

NRCC poll puts Singer ahead in FL-25, but our forecast barely moves

A Republican-sponsored survey shows Scott Singer up 7 points, yet our model still projects Jared Moskowitz winning by about 20 points on average.

Written by the newsroom · anthropic/claude-opus-5.5 · October 21, 2026 at 4:02 PM EDT

Correction: the House forecasts used in this article predate our redistricting correction. They used 2024 House results on earlier district lines. The current forecasts use presidential results recalculated on the 2026 lines for affected districts. Read the methodology.

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 National Research survey sponsored by the NRCC, fielded October 12-15 among 400 likely voters, shows Republican Scott Singer at 49% and Democrat Jared Moskowitz at 42% in Florida's 25th District. It is the only new poll of the race in this update.

The poll did not change our forecast in any meaningful way. Our model gives Moskowitz a 100 in 100 chance of winning the seat, the same as in our previous run an hour earlier. Our projected average margin remains D+20.2. The range of likely outcomes moved only at the edges. Our model's 5th-to-95th percentile range is now D+7.3 to D+33.2, compared with D+7.4 to D+33.0 before the poll was added.

The gap between this survey and our model is wide. The poll has Singer ahead by 7 points, while even the Republican-leaning end of our range has Moskowitz winning by about 7. A single 400-person survey from a party-sponsored pollster was not enough to pull our estimate toward it. That alone does not make the poll wrong. It does mean the rest of the evidence in our model points strongly the other way.

The national picture is also unchanged. Our generic ballot average stands at D+8.6 across 39 polls from the past 45 days. Our model gives Democrats a 99% chance of winning a House majority, with an average of 246.4 Democratic seats against the 218 needed for control. That figure is somewhat firmer than it should be. Our simulation correlates errors nationally, by census division, by state and within each district, but it does not link demographically similar districts in different regions, because we lack the data to do so.

Polls this dispatch was written from

  • National Research · Oct 15, 2026 · n=400 · FL-25

Every constant behind these numbers is on the methodology page.