JOSH MARSHALL ON RECENT POLLS

My friend Josh Marshall (since back when TPM was a page on Blogger) takes another look at the recent polls that keep not showing Biden crushing Trump and gives a solid analysis - and a reason not to pay attention to them so much.

From today’s Editor’s Blog (subscribers only)

I routinely tell people not to look at every single poll but to focus on trends over time. That is, if you want to look at them at all. We’ll go into Election Day with the polls tight and the outcome still uncertain. I can say this because I actually watch them very, very closely … like unhealthily closely. It’s characterological. I don’t advise it for anyone else. But if you must, it’s okay, and I can relate.

This morning there’s a new batch of swing-state polls from Bloomberg/Morning Consult showing Trump ahead in all but one of those states and growing his lead versus the last of these polls a month ago. That’s not great at all. But as usual I would not invest too much weight in a single poll. These numbers are not in sync with other recent swing-state polls, though actually we have pretty few quality swing-state polls recently. But the overall trend over the last six or seven weeks still seems like what we’ve discussed in the last several posts on this topic. After several months of being behind by a small but real amount (2-4 percentage points), Biden has moved into roughly a tie.

Just to attach some particular numbers to this, the last surveys in reverse chronological order are Morning Consult, Biden +1; Marist, Biden +3; Emerson, Trump +3; Economist, Even. Average them out and you couldn’t be much more tied.

It remains the case that a popular vote tie probably won’t be enough for Biden. He’ll probably need at least a one or two percentage point lead to win the Electoral College, and maybe more.

The only additional thing I think we can glean from the last couple months is that while Biden has erased Trump’s popular vote advantage, those numbers seem to have stabilized there. He’s not moving into a popular vote lead.

There are a couple additional points from under the hood that are worth mentioning.

There have been a number of polls recently which show that interest in the election is lower than it’s been in a couple decades. I’m always a bit skeptical of these questions because just what they’re polling is difficult to know and subtle differences in the wording of the question can yield very different results. You can say you don’t like either candidate or you’re just not excited about the race. But you may end up being really motivated to make sure one guy loses. You hear all sorts of measures of dissatisfaction with the political system and yet rates of voter participation have been going up for years now.

Having said this, there are numbers which suggest this will be a low-turnout election relative to recent elections. And polls do seem to show with some consistency that if you make the voting screen really right, Biden wins. In other words, if you go from registered voters down to likely voters and keeping tightening down to people who say they are definitely going to vote, Biden wins. This computes demographically. More educated, affluent and older voters have always been the most consistent voters. And those groups have been migrating into the Democratic coalition.

Another permutation of this under-the-hood question turns on disaffection within the Democratic coalition. You’ve probably seen lots of polls saying that Black voters or younger voters are turning their backs on the Democrats. There’s a lot of reasons to believe those things are significantly overstated. But if you take them on their own terms, a lot of those findings are based on polls which suggest that many in this disaffected bucket didn’t vote in 2020 or 2022. So some people look at these numbers and say, can it really be that this is going to be (relative to recent elections) a low-turnout election and people who didn’t show up in 2020 or 2022 are going to show up now and make the difference? One of these premises probably has to be wrong. Either these people aren’t going to show up anyway or the low-turnout election prediction is wrong.

In theory you can make that work: a low-turnout election in which the never voters are actually finally going to show up. But that’s a stretch.

Final point. No one is paying attention any more since the primary is over. But last night, 17%, or 156,956, showed up to vote for Nikki Haley in the Pennsylvania GOP primary. It’s a closed primary. These are registered Republicans. Does that matter? I’m not sure. But that’s a sizable number of voters showing up well after the nomination has been decided to make a protest vote against Donald Trump. Presumably many of them are already Biden voters from 2020, from Dem-trending suburbs. They just haven’t changed their party registration yet. But how many? Interesting question.

I think the best way to look at recent polls is first to be reminded that they remain fluid. In the Great Biden Panic of January and February, one of the big reasons to dump the incumbent in favor of a hypothetical candidate to be determined later was the fact that not only was Biden running behind but that those numbers seemed locked in. They’d been that way for months and months. But then they started moving. We’ve seen clear movement in Biden’s direction. Over the next six months we need to see another two or three points of movement in his direction.

Just Keep on Keepin’ On.

All the new Paid Subscribers have given a big, positive, much needed bump to That’s Another Fine Mess. You can join them for only $7/month or $70/year.

Comments are for paid subscribers.