New OK-5 poll shows Bice up five, nudging our forecast slightly

A GBAO survey of 500 likely voters in Oklahoma's 5th District, fielded September 17-20, puts Republican incumbent Stephanie Bice at 48% and Democrat Jena Nelson at 43%.

That is the only new number in the race, and it moved our forecast a little. Our model now gives Nelson a 14 in 100 chance of winning the seat, against 11% before the poll was added, with Bice at 86%. Our mean margin narrowed from R+10.7 to R+9.5. The 5th-to-95th percentile range now runs from R+24.0 to D+5.2, slightly tighter at both ends than the R+25.2 to D+3.9 we had earlier.

The direction of that shift is worth being precise about. A single 500-voter poll showing a five-point gap is closer than our model's central estimate, so it pulled the average toward Nelson. It did not pull it far. Our model still treats a Bice win as the more likely outcome by a wide margin, and the GBAO result on its own is consistent with a race in the high single digits once other inputs are weighed.

Nationally, our average on the generic ballot stands at D+8.1 across 30 polls over the past 45 days. In the House, where 218 seats are needed for control, our model gives Democrats a 98% chance of winning a majority and a mean of 245.3 seats. One caveat is ours to own: our simulation draws correlated error at the national, census-division, state and district levels, but not across demographically similar districts in different places, because we have no cluster data for it. That omission makes the House control number somewhat firmer than a fuller model would render it. The size of the difference is not measured in this run.

In the Senate, our model gives Democrats a 50% chance of winning control and Republicans 49%, with a mean of 50.4 Democratic seats against the 51 they would need; a 50-50 tie goes to the Republicans because the vice president is a Republican.