This is the fourth post in a series that began here and continued here, here, and here.
No, Elizabeth Warren isn’t declining from her recent peak, whatever RCP’s garbage methodology that bungs apples, oranges, and bananas in the same bin tells you.
According to this inaccurate RCP graph, Biden has suddenly recovered a 6-point lead after briefly dipping below Warren:
But as those of you familiar with this series know by now, that’s just a trick of the light because the Biden Bull polls have released more recent data than the Biden Bears. The overall trend-lines remain solidly up for Warren and solidly flat or down for Biden.
I want to show you a new way of cutting these polls, because there’s been a recent stark bifurcation in polling methodology that illustrates this divergence better than anything I’ve seen so far.
Eight polling outfits have released a new poll with data from September 24th or later. Exactly four of them have Biden up by double-digits over Warren. The other four have Warren ahead of Biden.
If that’s not a clear enough difference for you, here’s a graphical depiction of what a different picture of the race these eight polls present.
First, here are the charts from Politico/Morning Consult, The Hill/HarrisX, Harvard-Harris, and FOX News:
Next, here are the charts from Economist/YouGov, Quinnipiac, Monmouth, and IBD/TIPP:
To give you a sense of just how different these are (if that’s not already clear), here are all eight polls on the same graph:
As that last graph clearly shows, there are four entirely separate ranges of polling at present. The Biden Bulls’ estimation of Biden, then the Biden Bears’ estimation of Warren, then the Biden Bears’ estimation of Biden, and finally the Biden Bulls’ estimation of Warren.
What the heck is going on?
It’s one thing to say that the polls generally have it right and questioning polling methodology is problematic and derives from an over-emphasis on 2016. But even being generous to these polling operations, how do you resolve the cognitive dissonance? The Biden Bulls have him up, on average, 31-19 in their latest entries. The Biden Bears, meanwhile, have him down, on average, 29-26. That’s a 15-point marginal swing among contemporary polls. Indeed, the most extreme outliers manifest a whopping 19-point marginal swing (The Hill/HarrisX has Biden up 31-15, while Quinnipiac has Warren up 30-27), a swing larger than what The Hill/HarrisX thinks Warren is polling altogether.
In short, these polls can’t all be right. One of these groups, or perhaps both, are getting fundamental and significant things about the state of the race very very wrong.
It is worth noting, of course, as I’ve been saying all along, that Warren’s trend-lines are still all pointing up and none of Biden’s are in any sustained way. Even The Hill/HarrisX, the ultimate Biden Bull, shows a 4-point drop for Biden and a 10-point jump for Warren since late May.
It’s easy enough to split the difference and say, hey, Biden is probably up slightly. And that may triangulate to the truth. But if he is, it’s not because he was just down a fraction and then jumped back up. Looking at the RCP graph remains one of the most misleading ways to get information on this primary election.
