Respondents said they preferred Republicans to control the Senate, but individual matchups showed a different story.
Eight days before the election, we have our final* midterm surveys: polls of the four states likeliest to determine control of the Senate.
New York Times/Siena College polls
Pennsylvania: John Fetterman (D) 49, Mehmet Oz (R) 44.
Arizona: Mark Kelly (D) 51, Blake Masters (R) 45.
Nevada: Adam Laxalt (R) 47, Catherine Cortez Masto (D) 47.
Georgia: Raphael Warnock (D) 49, Herschel Walker (R) 46.
All considered, this is a pretty good set of numbers for Democrats. If they win three of the four Senate seats, they hold the Senate if everywhere else goes as expected. Here, they lead in the magic three of four while remaining highly competitive in the fourth — though it’s very important to caution that most of this poll was taken before the Pennsylvania Senate debate.
There’s also a bit of good news for Republicans: Respondents said they preferred Republicans to control the Senate. Democrats led anyway, presumably because of their distaste for some of the Republican nominees or their affection for Democratic incumbents. As we head down the stretch, Republicans can hope to lure some of these voters to their side.
How the polls compare
If you compare our polls with the polling averages, they look even better for Democrats. Consider the current FiveThirtyEight averages:
In Pennsylvania, Mr. Fetterman leads in the FiveThirtyEight average by one point, compared with our six-point lead (after rounding).
In Arizona, Mr. Kelly leads in the average by 3.6 points, compared with our six-point lead.
In Georgia, Mr. Warnock leads by 1.2 points, compared with our three-point edge.
In Nevada, Ms. Cortez Masto leads by 0.4 points, compared with our tied race.
There is a twist: Although our polls may look better for Democrats than the average, they look about the same as the other traditional polls that used to be considered the gold standard in survey research, like ABC/Washington Post, CNN/SSRS, Fox News, NBC, the university survey houses, and so on.
In many of these states, our surveys are the first such poll in weeks. Consider the last such polls in the four states, and how much better they look for Democrats than the average — and how similar they look to our survey:
Arizona: CNN/SSRS showed Mr. Kelly +6 about a month ago, in a poll conducted from Sept. 26 to Oct. 2.
Nevada: USA Today/Suffolk showed Ms. Cortez Masto +2 in a poll taken from Oct. 4 to Oct. 7.
Georgia: Quinnipiac showed Mr. Warnock up seven, in a poll taken from Oct. 7 to Oct. 10.
Pennsylvania: Franklin and Marshall showed Mr. Fetterman +4 in a poll conducted from Oct. 14 to Oct. 23. In a survey fielded over a partly overlapping period, CNN/SSRS showed Mr. Fetterman up six from Oct. 13 to 17.
The absence of surveys from reputable pollsters is remarkable. The drought is partly because of rising costs — our October national survey was eight times as expensive as our final polls in 2016, on a per-interview basis. But it’s also because of a crisis of confidence among the traditional pollsters — Times/Siena included — who don’t have a great explanation for the poor results in 2020 and are understandably treading a little lightly.
The State of the 2022 Midterm Elections
Election Day is Tuesday, Nov. 8.
- A Pivotal Test in Pennsylvania: A battle for blue-collar white voters is raging in President Biden’s birthplace, where Democrats have the furthest to fall and the most to gain.
- Governor’s Races: Democrats and Republicans are heading into the final stretch of more than a dozen competitive contests for governor. Some battleground races could also determine who controls the Senate.
- Biden’s Agenda at Risk: If Republicans capture one or both chambers of Congress, the president’s opportunities on several issues will shrink. Here are some major areas where the two sides would clash.
- Ohio Senate Race: Polls show Representative Tim Ryan competing within the margin of error against his G.O.P. opponent, J.D. Vance. Mr. Ryan said the race would be “the upset of the night,” but there is still a cold reality tilting against Democrats.
The flip side: Most of the polling over the last few weeks is coming from partisan outfits — usually Republican — or auto-dial firms. These polls are cheap enough to flood the zone, and many of them were emboldened by the 2020 election, when their final results came close to the election results even as other pollsters struggled.
A couple of the nontraditional firms are worth taking seriously — CBS/YouGov in particular — but a lot of the polls that are filling up the averages just aren’t underpinned by credible survey methods.
They may have come close to the results in 2020 — and could easily come close yet again 2022 — but it’s not because they have a representative sample of the population. Imagine, for instance, a poll that is subject to the same pro-Democratic bias as the higher-quality surveys, but simply doesn’t call cellphones and misses people under 35.
In that case, the headline results could be “right,” but it’s not because the pollster has some special sauce or has uncovered the secret to reaching Trump voters.
To be clear: The point isn’t that our polls are right and the others are wrong. There’s plenty of reason to think Trump-era challenges still plague the survey industry. If so, the pollsters once considered gold standard may struggle yet again. And if so, the dearth of gold standard polls and the surge of partisan polling might just leave the polling averages closer to the results than the last election, even if it’s just as tough for pollsters as it was two years.
Right or wrong, there’s not much question that Democrats would hold a more comfortable lead in the Senate if the pollsters who dominated the averages in the past were a bigger part of the averages this year.
Note: *We will have one last set of findings for you: the results of a multi-survey study of Wisconsin. This study began in September, so it won’t exactly count as a final poll. But hopefully it will shed some light on the challenges facing pollsters and maybe even offer a path forward. I started to dig into this data only yesterday — and progress has been slow — but my goal is to report a few preliminary findings before the election.
Source: Elections - nytimes.com
