Model gives Democrats 54 in 100 in the Senate, 98 percent in the House

Forty-one days from the November 3 election, our model gives Democrats a 54 in 100 chance of winning Senate control and a 98 percent chance of a House majority. The Senate mean is 50.6 Democratic seats; Democrats need 51 seats for control and Republicans need 50, because the vice president is a Republican and a 50-50 tie goes to the Republicans. In the House, our mean is 244.5 Democratic seats against the 218 needed for a majority. Our House simulation draws error at four correlated levels — national, census division, state and district — but has no term linking demographically similar districts in different places, for want of cluster data. That is a gap in our own method, and it leaves the 98 percent marginally firmer than a fuller decomposition would make it.

Our generic ballot average stands at D+7.9, drawn from 28 polls over a 45-day window. Recent surveys in that range include Morning Consult's September 14-20 poll of 18,336 likely voters, which had Democrats at 49.74 percent and Republicans at 42.62 percent, and a Quantus Insights poll of 1,389 likely voters fielded September 15-18, with Democrats at 50.3 and Republicans at 41.7. YouGov, for the Economist, found Democrats at 52 percent and Republicans at 38 among 975 likely voters fielded September 18-21, and 45 to 36 among 1,394 registered voters in the same period. ActiVote, polling 1,000 likely voters September 11-21, had the two-way split at 52.8 to 47.2. Reuters/Ipsos, fielding September 17-20, had Democrats ahead 43 to 35 among 979 registered voters and 37 to 29 among 1,277 adults.

The largest single move on the board was Iowa's Senate race, where our estimate of the Democratic chance rose 16.8 points to 60 percent, from 44 percent. Marist University polled 1,050 Iowa registered voters September 17-20 and found Democrat Josh Turek at 50 percent and Republican Ashley Hinson at 42. Our North Carolina Senate estimate rose 9.2 points to 93 percent, and Kansas rose 10.7 points to 19 percent, from 8. In the House, NY-17 moved 4.6 points, to a 67 percent Democratic chance.

Texas moved the other way. Our estimate of the Democratic chance there is now 56 percent, down 14.2 points from 70 percent. The two most recent Texas surveys disagree: Marist, polling 1,139 registered voters September 17-20, had Republican Ken Paxton at 52 percent and Democrat James Talarico at 43, while the Barbara Jordan Public Policy Research and Survey Center, polling 1,800 likely voters September 15-19, had Talarico at 47 and Paxton at 46, a margin close enough to call the race even. In Michigan, our estimate of the Democratic chance is 71 percent after a Suffolk University poll of 500 likely voters, fielded September 16-20, put Democrat Abdul El-Sayed at 46.8 percent and Republican Mike Rogers at 39.6. In Florida, InsiderAdvantage polled 600 likely voters September 20-21 and found Republican Ashley Moody at 49 percent to the Democrat's 42.