Through this series, I hope to keep the Dkos community informed in a single location on what the numbers and experts say about the state of the midterms. At the same time, it’s important to keep perspective. Polls are going to be off; pundits are going to make mistakes in viewing the political environment.
There are new tea leaves to read: now we have initial early voting data — what people are actually doing, rather than what they say they will do and what political observers think they will do. A good review of early numbers, particularly in Georgia, can be found here. The short version for Georgia is that total early voting (in-person and mail-in) is well ahead of where it was in 2018, the last midterm, but well behind 2020. Of course 2020 was a presidential election year, which always brings a higher turnout, but in addition 2020 had a huge mail-in vote and this election does not (the mail-in vote has dropped an eye-popping 90%+). So bad news for Democrats? Well, no. The 2018 midterms were quite successful for Democrats, and the total early vote in GA is running nearly 200,000 ahead of that election. The % of vote attributable to Black voters is running ahead of 2018 (good for Democrats) and the female vote is outnumbering the male vote (also good for Democrats). Meanwhile, in Pennsylvania (where there still is a lot of voting by mail) ballots from registered Democrats are outpacing those from Republicans 73.1% to 19.4%. This is exactly what Fetterman and PA Democrats need — a very large advantage in early votes banked. In another summary across PA and 4 additional states, the Democratic share of the early vote is outpacing 2020, by margins ranging from 1 pt (in Ohio) to 11 pts (in Michigan). From last week, also check out this summary of why 2022 could turn out to be another record-setting year for midterm turnout. There is no certainty that early voting/mail-in advantages will remain as strong or that they will be enough to overcome the expected GOP advantage on election day, but there is nothing in this early data that should alarm Democrats, and some room for optimism.
Just for reference, below are brief descriptions and the links that can take you right to the sources for the polls and forecasting I summarize.
FiveThirtyEight: Senate ratings and House forecast. Provides estimates of election-day vote based on both a polling average and a “deluxe” model that includes historical trends and expert ratings. (538)
Electoral-vote: Provides a simple last-week polling average for Senate races. (E-V)
RealClearPolitics: Another polling average, but broader than Electoral-Vote’s. (RCP)
270towin: Yet another polling average and algorithm. (270)
Cook Political Report: Expert ratings of each race (solid, likely, lean, or tossup). (Cook)
Sabato's Crystal Ball: Also expert ratings (safe, likely, lean, or tossup). (Sabato)
Inside Elections: Yet more expert ratings (solid, likely, lean, tilt, or tossup). (Inside)
This week’s numbers and ratings were current as of Saturday 10/22.
Senate Races
|
E-V |
RCP |
270 |
538 (polls) |
538 (deluxe) |
Sabato |
Cook |
Inside |
Az |
D +3 |
D +2.5 |
D +4.4 |
D 51.2-46.7
|
D 50.9-47.0 |
lean D |
lean D |
tilt D |
FL |
R +5 |
R +5.7 |
R +6.0 |
R 51.6-46.2 |
R 53.0-44.8 |
likely R |
likely R |
likely R |
GA |
D +3 |
D +2.4 |
D +3.2 |
D 50.9-47.5 |
D 49.8-48.6 |
tossup |
tossup |
tossup |
IA |
R +3 |
R +7 |
R +9.4 |
R 54.2-45.8 |
R 56.8-43.2 |
likely R |
likely R |
likely R |
NV |
R +1 |
R +1.2 |
R +0.7 |
R 48.9-48.3 |
R 48.9-48.3 |
tossup |
tossup |
tossup |
NH |
D +7 |
D +5.4 |
D +6.4 |
D 52.3-45.7 |
D 52.0-46.2 |
lean D |
lean D |
tilt D |
NC |
R +3 |
R +2.8 |
R +2.2 |
R 49.9-48.0 |
R 51.1-46.9 |
lean R |
lean R |
tilt R |
OH |
tie |
R +2.3 |
R +1.4 |
R 50.6-49.4 |
R 52.2-47.8 |
lean R |
lean R |
lean R |
PA |
tie |
D +2.4 |
D +2.6 |
D 50.5-47.3 |
D 49.7-48.0 |
lean D |
tossup |
tossup |
UT |
R +5 |
R +4 |
R +10 |
R 49.2-41.1 |
R 51.8-38.5 |
likely R |
likely R |
likely R |
WI |
R +3 |
R +2.8 |
R +3.6 |
R 51.5-48.5 |
R 52.1-47.9 |
lean R |
tossup |
tilt R |
Senate overview: It’s not possible to tell the direction a race is trending just by looking at weekly numbers in isolation (and the numbers may of course be incorrect — more on that later). So let’s consider that longer-term context. I’ve been tracking the numbers on most of these sites since the first week in September. Compared to 2 or 3 weeks ago, the races definitely appear to have tightened for some candidates, such as Kelly in AZ and Fetterman in PA. However, over a longer time frame, this simply leaves some Democratic candidates back where they were in early September, while others have experienced a definite decline over that time. For Mark Kelly in AZ, 538’s deluxe model forecast a win by 4.8 pts on 9/2 and now forecasts a win by... 4.5 pts. In between Kelly’s advantage increased to as much as 6 pts, so this doesn’t suggest his race has fundamentally changed much. Numbers on the purely poll-based sites mostly show the same pattern (electoral-vote, for instance, showed Kelly with a 2-pt lead on September 10; now it shows a 3-pt lead).
Unfortunately, John Fetterman’s numbers show a definite recent decline. Fetterman enjoyed a substantial lead in polling averages throughout September, typically 6-8 pts. His current margin is around 2 pts, It’s noteworthy that the GOP and it’s propaganda mouthpieces, such as Fox, have focused on this Senate race more than any other, with more attacks on Fox than for any other candidate and more money from the GOP and conservative PACs. These haven’t necessarily made Oz any more popular, but they have definitely pumped up negative perceptions of Fetterman. Similarly the race in Nevada has shown Cortez Masto consistently losing ground over time to Laxalt, moving from a lead of several points to a small deficit. In Wisconsin, Barnes held a small advantage over Johnson in the early September polling averages, but has steadily lost ground and is now running behind.
However, there are also races where the Democrat is improving their numbers or at least remaining consistent. In Georgia, Warnock has consistently led Walker over this time frame. In Ohio — which Trump carried by 8 pts in 2020 — Ryan has shown an improvement in his numbers, moving from a 3-4 pt deficit against Vance to a 1-2 pt deficit. Overall, in terms of forecast Senate control, we are still where we have been all summer: slightly favoring the Democrats, with 50 or 51 seats the most likely outcome across the various polling and pundit sites. NV and PA remain the seats most likely to change hands, with a Democratic pickup in PA offset by a loss in NV.
House overview: Little movement. Forecasts across sites continue to show the GOP on the brink of outright control or favored to gain it, with the Democrats having to win almost all tossup seats to keep control.
RCP: R 225, D 175, T 35
538: R 219, D 205, T 11
270: R 217, D 206, T 12
Sabato: R 214, D 195, T 26
Inside Elections: R 212, D 205, T 18
Cook: R 211, D 193, T 31
Gubernatorial races
|
RCP |
270 |
538 (polls) |
538 (deluxe) |
sabato |
cook |
inside |
AZ |
R +1.6 |
R +2.4 |
R 50.5-49.5 |
R 50.6-49.4 |
tossup |
tossup |
tossup |
FL |
R +9.2 |
R +6.0 |
R 53.2-44.6 |
R 54.3-43.5 |
likely R |
likely R |
likely R |
GA |
R +5.6 |
R +5.8 |
R 52.3-46.3 |
R 52.7-45.9 |
lean R |
lean R |
tilt R |
KS |
D +2.0 |
NA |
D 50.5-46.2 |
D 49.2-47.5 |
tossup |
tossup |
tossup |
NV |
R +1.8 |
R +1.5 |
R 49.0-48.2 |
D 48.8-48.4 |
tossup |
tossup |
tossup |
OR |
R +2.8 |
R +0.6 |
R 41.1-41.1 |
R 41.1-41.1 |
tossup |
tossup |
tossup |
PA |
D +9.6 |
D +10.0 |
D 53.6-44.4 |
D 53.5-44.5 |
likely D |
likely D |
lean D |
TX |
R +8.7 |
R +6.8 |
R 53.0-44.9 |
R 54.8-43.0 |
likely R |
likely R |
solid R |
WI |
tie |
R +0.6 |
D 49.3-49.1 |
D 49.5-48.9 |
tossup |
tossup |
tossup |
Gubernatorial Overview: Very little change, again. Most races show a very slight shift in favor of the GOP candidate. A notable exception is Oregon, where all sites show a tightening in favor of Democratic candidate Kotek, possibly due to supporters of independent (and former Democratic) candidate Johnson coming back tot he actual Democrat in the race. Abrams and O’Rourke remain well behind in Georgia and Texas, and if they lose these races we’ll need to be talking about which way these generally acclaimed rising stars in the Democratic party go next, as they will have both lost two statewide races at that point. Arizona looks like the best chance for a Democratic pickup, despite Lake’s continuing lead in the polls, but even if Hobbs wins it may be offset by a loss of a Democratic governorship in Nevada, Oregon, or Wisconsin.
With just over 2 weeks to go, and early voting increasingly under way, I’ll take a closer look at that in next week’s edition. More broadly, I’ll consider why the likely voter models used to weight polling results may be wrong.