Today’s new IBD/TIPP national survey is an outlier (a big one) but warrants a post if only because of the trendline. A week ago they had Biden’s lead growing from three points to 8.5, putting it in line with a raft of other polls showing a big Biden lead. This morning they published their new survey letting all of the air out of that balloon, with the Democrat’s lead now down to just 2.3 points. To put that in context, Biden’s *average* lead in national polling at the moment is 8.6 points. And IBD’s number is part of that average, otherwise his lead would be even higher.

Every outlier poll is potentially the start of a trend, though. And given how stable and boring the national numbers have been lately, any outlier showing a tightening race is newsworthy.

Besides, 2.3 points almost exactly matches the margin of Clinton’s popular-vote win four years ago. If Trump can hold Biden to that spread this time, he certainly can win the electoral college again. If he’s starting at an eight-point national deficit, though, then no way.

Since the IBD/TIPP 2020 Presidential Election Tracking Poll launched on Monday, Oct. 12, Biden’s support has slipped 3.8 points, from 51.9% to a new low-water mark of 48.1%. Trump poll numbers have gone the other way, rising 2.4 points to his new high-water mark of 45.8%. Biden’s lead peaked at 8.6 points in the Oct. 13 presidential poll…

Joe Biden’s lead among likely independent voters narrowed to a 45%-41% margin in IBD/TIPP’s latest 2020 presidential poll update. Biden’s lead had ranged from 6 points into the double digits.

Meanwhile, Donald Trump is now having a much easier time corralling Republicans than Biden is having in nailing down Democratic votes. Biden leads 90%-7% among Democrats. Trump leads 94%-4% among Republicans…

Senior citizens are another key group Trump carried in 2016 that have swung toward Biden, but the gap has narrowed. Seniors back the former vice president, 48%-46%, in IBD/TIPP’s latest Trump vs. Biden poll, but his lead had been in double digits. Trump won among senior citizens in 2016, 53%-44%, Pew says.

Gotta say, I’m suspicious of that boldfaced part. One constant in polling this year is that Democrats tend to be almost unanimously behind Biden due to their animosity towards Trump, or at least more solidly behind him than Republicans are behind Trump. (Never Trump righties could be the deciding factor if you believe Scott Rasmussen.) For instance, the New York Times national poll that was published this morning and conducted over nearly the same period as the IBD poll found Sleepy Joe winning Dems 93/4 while Trump wins Republicans 91/7. YouGov’s last national poll, conducted a week ago, had Biden up 85/1 among Democrats and Trump up 83/3 among GOPers.

Other numbers from IBD jibe with the polling consensus, though. They have Biden improving considerably in the suburbs compared to Hillary in 2016 and see him ahead by two points among senior citizens, a group Trump won comfortably four years ago. Although, even here, other pollsters are seeing much bigger leads for the Democrat among older voters than IBD is:

Here's my latest running average of demographic splits in national live-interview polls w/ new NYT/Siena added in. Biden's surged w/ 65+ between Sept. and Oct., but just a big a deal is the steady decline in undecided/third party since June/July (and vs. '16, of course). pic.twitter.com/TZDBta582B

— Dave Wasserman (@Redistrict) October 20, 2020

But forget the national polling. What we care about is battlegrounds, right? And there’s one battleground lately — a crucially important one — where the trend has been towards Trump. At least if you believe RCP:

Pennsylvania has been seen as Trump’s best shot at a victory in the Rust Belt, and lo and behold, here he is making his move. Winning PA wouldn’t outright guarantee him a second term; Biden could still get to 270 via Wisconsin, Michigan, Arizona, and North Carolina. But if Trump wins Pennsylvania, odds are he’s going to win at least one of those other four as well. Eight days ago he trailed Biden by 7.3 points. Today the lead is nearly half that, just 3.8. PA is in play.

Isn’t it?

One thing I noticed when looking at the numbers RCP used for its average is that two of the most recent polls come from Republican-aligned firms. One is Trafalgar, which you know all about, and they have Biden up just two points. The other is Insider Advantage, which conducted its poll for the pro-Trump Center for American Greatness. They have Biden up three. The fact that those two firms have a partisan rooting interest doesn’t mean their numbers are wrong, of course, it just means that we bear in mind the possibility of a “house effect” the same way we do when a Democratic-run poll shows a gaudy number for a lefty. (Case in point: Doug Jones released an internal poll a few days ago showing him up a point on Republican Tommy Tuberville in their Alabama Senate race, which, uh, no.) That is to say, the “tightening” that’s going on in PA at the moment may be a mirage driven by the fact that there are a higher-than-usual number of Republican-leaning polls affecting the average at the moment.

That’s Nate Silver’s theory too, in fact:

Folks, Biden's lead didn't shrink from 7.3 points to 3.6 points in PA in a week (as per RCP) at the same time it was steady or slightly growing nationally. This is why you need poll averages that take a longer time horizon and/or adjust for house effects. https://t.co/wKbW02AGO5 pic.twitter.com/daWEKcLlAm

— Nate Silver (@NateSilver538) October 20, 2020

I love many things about RCP, but if you have an average and 1/3 of it consists of Trafalgar and InsiderAdvantage and 0% of it consists of live-caller polls, it's not going to be a very reliable average. pic.twitter.com/3LTQ7gEwAd

— Nate Silver (@NateSilver538) October 20, 2020

Silver’s polling average of Pennsylvania gives Biden a lead of 6.4 points today and his model of the state’s election gives the Democrats an 87 percent chance of winning, which is several percentage points higher than on October 1. It *would* be odd for Biden to be nosediving in a key swing state at a moment when his national lead is holding steady in the eight- or nine-point range.

But then, the candidates are paying lots more attention to Pennsylvania than they are to most other states. And in fairness to Trafalgar and Insider Advantage, their recent surveys aren’t out of whack with what other pollsters are showing. Earlier this month the Democratic firm Change Research had Trump within four points in PA. The most recent poll of the state, from Reuters, had Biden’s lead … at four points. Two other polls published within the last two weeks have seen Biden up by five. By no means is Pennsylvania off the board for Trump, especially if he has a good debate on Thursday.