The new Senate polls are grim. But can they be trusted?

(AP Photo/Susan Walsh)

Two things are true and must be borne in mind when assessing the state of the polls as we continue to endure “bedwetting season.” One: There really has been meaningful movement towards Democrats across multiple polls lately. Too many pollsters are seeing it for it to be a mirage. The midterm race has tightened. If you doubt it, sift through this data compiled by a Democratic analyst.

Two: The polls were hot garbage in 2020 and may once again be underestimating Republicans considerably.

Fox News dropped a pair of stinkers last night on the GOP. In Arizona they have Mark Kelly leading Blake Masters comfortably by eight points, 50/42. Needless to say, having a Democratic incumbent already on the cusp of a majority in August is no bueno. You can blame Trump for that, at least a little:

Up north in Wisconsin, they have progressive Mandela Barnes leading Ron Johnson, 50/46. It’s hard to believe that a two-term Republican incumbent might lose in a 50/50 state in a favorable national environment for Republicans but multiple polls lately have had Barnes ahead, including the most respected pollster in the state:

If that’s not enough, a lefty group just dropped another polling nuke on Dr. Oz in Pennsylvania:

What’s going on? Why have things suddenly gone south for Republicans in so many races?

Even Mitch McConnell is nudging GOP voters to start lowering expectations:

Yeah, “candidate quality” is a tactful way of explaining one of the problems here. None of the GOP nominees in the three races I mentioned are what we might call “generic Republicans” who are well-suited to surf the national tide to victory. Oz is a hapless squish whom the base distrusts; Masters is a nationalist weirdo who’s too close to Trump; and RonJohn has drifted towards fringy positions like vaccine skepticism over his last few years in office. It’s easy to imagine a bland garden-variety Republican being more competitive in each of these races.

Another problem is that, as I said, the national environment really has shifted. Fox’s poll found Wisconsin voters at 37/54 on whether they approve or disapprove of Roe being overturned. Meanwhile, gas prices are down, inflation has stalled (temporarily?), the Dems just passed their big progressive-pleasing climate-change bill, and the last jobs bill was a blockbuster. Everything’s coming up Brandon — for the moment:

But there are complicating factors here for Democrats. An obvious one in Wisconsin is that Johnson is already well-defined among voters whereas Barnes, despite being the lieutenant governor, is likely a blank slate to many. That makes for a likability gap, but likability gaps can be closed. And probably will in this case once the RonJohn ad team goes nuclear on the Democrat:

I’m not as certain that Oz and Masters can negatively define John Fetterman and Kelly, respectively, but the sheer volume of ad spending to come this fall will help somewhat. It’s how voters feel about the candidates in mid-October, not mid-August, that’s important.

There’s a bigger complicating factor. After the ginormous polling miss in 2020 that saw supposedly tight Senate races end in comfortable wins for several Republicans, how much can we trust this new data? Again, the fact that there’s a uniform shift towards Democrats across polls lately strongly suggests that a shift really is happening — but that’s different from saying that the polls showing candidates like Barnes ahead must be accurate. Fox News is a respected pollster but their track record in the 2020 cycle was … not the best:

Remember that the leading explanation for why the polls missed so badly in 2020 was “nonresponse bias,” the theory that a small but significant percentage of Republicans have stopped cooperating with pollsters on principle, refusing to take their calls due to perceived anti-Trump bias. If that’s true then there’s no way for pollsters to accurately gauge the true extent of support for Republicans among the electorate. And there’s also no obvious reason to think that the problem has since been solved in the two years since.

On top of that, RCP’s Sean Trende reminds us that polls of the midwest have been especially terrible over the last few cycles. Remember the last time RonJohn ran for Senate in Wisconsin? His odds looked so thin that Republican groups all but gave up on spending in his race. Oops:

That was a six-point miss. If Fox’s new poll is off by six points in the same direction, Johnson is actually leading his race against Barnes.

I encourage you to scroll through this thread by Patrick Ruffini of Echelon Insights and remind yourself how bad Senate polling generally has been over the last few cycles. And not “neutrally” bad either, in which the polls sometimes miss badly by underestimating Republicans and other times miss badly by underestimating Democrats. There really is a partisan skew, possibly due to nonresponse bias.

I wouldn’t bet against Fetterman at the moment when he’s sitting on a double-digit lead against a conspicuously weak candidate like Oz. But I sure wouldn’t bet on Barnes in Wisconsin. And my guess is that the Kelly/Masters race will be a lot closer than eight points. Stay tuned.