Forecast in OH-1 barely moves after two Quantus polls

Two Quantus Insights surveys of 509 likely voters in Ohio's 1st District, both fielded August 21-24, show a close race. One has Democratic incumbent Greg Landsman at 47 percent and Republican Eric Conroy at 43. The other has Landsman at 48 and Conroy at 45.

Our model has now absorbed them, and the result is close to a null. It gives Landsman a 93 in 100 chance, unchanged from the 4:30 p.m. ET run on August 26. The mean margin moved from D+14.8 to D+14.4. The 5th-to-95th percentile range shifted from R+1.5 to D+31.4 into R+1.9 to D+30.8. That is a wide range, and it includes outcomes where Conroy wins narrowly.

The gap between the polls and our margin is worth stating plainly rather than resolving. Our estimate rests on more than these two surveys, and single-district polls of about 500 respondents carry enough sampling error on their own that a three- or four-point result and a double-digit expectation are not incompatible. What the new numbers did not do is change the probability we attach to the seat.

Nationally, our average on the generic ballot is D+6.5 across 27 polls in a 45-day window. In the House, where 218 seats are needed for control, our model gives Democrats a 93 percent chance and a mean of 240.9 seats. One caveat on that figure is ours to own: our simulation draws correlated error at the national, census-division, state and district levels, but we have no cluster data for a term linking demographically similar districts wherever they sit, so the House control number is marginally firmer than a fuller treatment would leave it. In the Senate, our model gives Democrats a 45 percent chance of reaching the 51 seats they need, with a mean of 50.0 seats.