Welcome to our Election Update for Sunday, Oct. 21! You know things are really getting down to the wire when we start releasing Election Updates on the weekend. My colleague Geoffrey Skelley covered the House yesterday; today, let’s tackle the Senate.
As of 11 a.m. on Sunday, Republicans had a 7 in 9 chance of keeping control of the Senate. Their prospects were nearly identical in all three versions of our model — Lite (80 percent), Classic (79 percent) and Deluxe (79 percent) — showing that various indicators such as polls, fundamentals and expert predictions are all in general agreement. What’s more, all three versions of the forecast also gave Republicans similar chances last Sunday, so you might even say the Senate forecast has been a little … boring. (Boring is good for Republicans; they seem to be holding the ground they gained during the Supreme Court confirmation hearings.)
Still, several polls were released this week that caused some movement under the hood — it’s just that some races moved toward Republicans and some toward Democrats, thus having little net effect on the topline forecast.
It was a good week for Democratic Sen. Joe Donnelly in Indiana. Two polls were released showing him ahead: 44 percent to 40 percent according to Gravis Marketing and 55 percent to 45 percent according to Vox Populi. His chances of winning in our polls-only Lite model have risen from 69 percent last Sunday to 82 percent today.
You might think that the two polls released this week in Missouri — giving Republican Josh Hawley leads of 1 and 3 points — would be good news for him. But both pollsters have Republican-leaning house effects, so our model actually read each poll as giving Democratic Sen. Claire McCaskill a lead of about 2 points. Her chances of winning in the Lite model have ticked up to 5 in 9 as a result. (McCaskill fares a bit better in our Classic forecast, with a 5 in 8 chance.)
But Democratic Sen. Tina Smith’s odds in the Lite model in Minnesota dipped from 92 percent to 88 percent after a Change Research poll gave her just a 3-point lead. This race has barely been on anyone’s radar, but four successive polls taken in August and September showed it within single digits before a Marist poll (which gave Smith a 16-point lead) broke the streak earlier this month. The Change Research poll is a data point in favor of the hypothesis that Marist (which has also given Democrats really good news in other Upper Midwestern states this year) is just an outlier rather than reflecting real movement in the race. But the fundamentals remain favorable to Smith: Our Classic forecast still has her as an 11 in 12 favorite.
Four new polls released in Nevada this week helped boost Republican Sen. Dean Heller’s chances in the Lite model from 53 percent to 57 percent.1 His best result was a 7-point lead in an Emerson poll. And one of the two D+2 polls that appeared to look good for Democratic U.S. Rep. Jacky Rosen was actually a net benefit to Heller because it was a Democratic internal poll; that poll also represented a 3-point shift toward Heller from the group’s prior poll.
In the first Senate survey of Florida conducted after Hurricane Michael, St. Pete Polls found Republican Rick Scott (who, as governor, has been the face of hurricane recovery efforts) leading Democratic Sen. Bill Nelson 47 percent to 45 percent — a 2-point bump for the Republican since the pollster’s last survey of the race. A Kaiser Family Foundation poll was also released giving Nelson a 3-point lead, but buyer beware: It was more than two weeks old (and also polled registered voters instead of likely voters). Older polls like the KFF’s had been showing a slight Nelson lead, so the St. Pete poll caused only a slight uptick in Scott’s chances of winning (currently 2 in 5 in the Lite model and 3 in 8 in the Classic model). But the model doesn’t “know” about the hurricane, so this may be a case where older polls are less useful. We’re also in the midst of a polling drought in the Sunshine State, probably also because of Michael; whereas in September we were getting a poll every couple of days, St. Pete is the only poll we have so far that was conducted after Oct. 5. I’ll be looking closely to see if the same rightward shift also shows up in polls next week.