NV-2 forecast unchanged after the only poll in the race

The only poll in our NV-2 file remains a GBAO survey of 500 likely voters, sponsored by the Nevada Alliance and fielded August 31 to September 3. It has Democrat Teresa Benitez-Thompson at 38 percent, Republican Dave Flippo at 37 percent, and 14 percent for other options. That is a close result, and with a sample that size the one-point gap is not a lead.

Our forecast has not moved. The model puts the mean margin at R+25.3 and gives Benitez-Thompson a chance that rounds to 0 in 100. The 5th-to-95th percentile range runs from R+40.7 to R+9.9, so even the Democratic end of our simulated outcomes is a double-digit Republican win. Those figures are identical to the ones from our previous run at 2:32 PM ET on September 13: same margin, same range, same probabilities.

The gap between the single poll and the model is the story, and it is the same gap we described on September 8. Nothing in this run narrowed it. Our estimate for this district leans on the fundamentals and the national environment rather than on one 500-person survey with 14 percent unallocated, and a second look at the same poll produced no change.

Nationally, our average of 22 generic-ballot polls over the past 45 days stands at D+6.8. In the House, our model gives Democrats a 95 in 100 chance of winning control, with a mean of 241 seats against the 218 needed. One caveat is ours to own: the simulation draws correlated error at the national, census-division, state and district levels, but we have no data to add a term linking demographically similar districts across state lines. Without it, our House control number is marginally firmer than a fuller accounting would leave it. How much firmer is not measured in this run.