Gee, why might that be? I’m not sure what point Pew Research thought they were making with this chart, but it ain’t the one that comes across in it. The headline of their analysis actually hits the nail more on the head, even if it is irrelevant: “Coronavirus death toll is heavily concentrated in Democratic congressional districts.”

This, however, is spin that might even put MSNBC to shame:

That’s how Pew titled the graphic in its report, too, even though a first-year geometry student could point out the fallacy. The obvious reason why GOP-controlled districts have a shallower decline in death rates is because they had a much shallower incline. The peak in these districts was just one-third of the peak in Democrat districts, which still have more than twice the rate of deaths at this time. This is so obvious that it speaks to a blinding bias on the part of Pew Research that they couldn’t see it for themselves before publishing it — and then promoting this hot take.

Even worse, it’s tough to understand why this data would be important at all:

The coronavirus outbreak has taken the lives of nearly 100,000 Americans. Yet since the start of the outbreak, the death toll has been concentrated in a just a few places – mostly large metropolitan areas, especially the New York City area.

The places hit hardest by the coronavirus outbreak – which have relatively large shares of ethnic and racial minorities and residents living in densely populated urban and suburban areas – are almost all represented by congressional Democrats.

No kidding? Next thing Pew will tell us is that almost all large-city mayors are Democrats, too.

This might matter in an election, but it doesn’t really have anything to do with policy — except perhaps in New York City. Pandemics spread faster in higher-density populations, which is all this actually shows. Republicans got pushed out of high-density urban districts a generation or more ago, and now represent suburban/rural voters — where pandemics of any stripe would spread more slowly.

That’s really all this actually shows, other than a non-causative correlation. Put it this way: what impact does a House member have on policy and governance specific to his/her district? None at all, which is why this analysis is entirely useless. This is correlation without causation, a very common statistical fallacy that one would ding students for creating in an undergrad paper.

To have this appear in a Pew Research publication suggests another form of causation — political bias. It undermines Pew’s credibility and calls into question all of their work in this political environment, if for no other reason the sheer idiocy involved in making that bias sooooooo obvious.