Unpopular Science 231001
A less popular way to interpret COVID deaths
We read a ton, especially about issues and ideas relevant to personal science. Our paid subscribers will get this curated list of ideas that you might have missed.
Unpopular Science
When too many scientists become aligned with a specific political viewpoint, an all-too-human result is that some types of data and evidence will be easier to publish than others. One way to tell if you’re heading into unpopular science territory is to ask whether something would have been published if it had shown the opposite result.
One recent example is the idea that COVID kills more Republicans than Democrats. This fact could very well be true, but imagine that it’s not. Would the opposite conclusion have been scrutinized as much, or more?
This week statistician Nate Silver posted the 2 key facts about US Covid policy that everyone should know, concluding:
Until vaccines became available, there was little difference in COVID death rates between blue states and red states.
After vaccines became available, there were clear differences, with red states having higher death rates, almost certainly as a result of lower vaccine uptake among Republicans.
Nate Silver’s case is pretty straightforward and seems to simply confirm the results of JAMA-published researchers.
So how could you have an unpopular take? You may not like the data, but science doesn’t care. And it’s not like this conclusion is even necessarily saying something “bad” about Republicans, who might respond that the effect anyway is small and that there are always tradeoffs.
There are good reasons why the JAMA article is bad science, but who’s going to publish that headline? It would be much less popular.
Nate Silver’s piece is popular science. After all, who doesn’t want to see “proof” that their political compatriots are “better”. And data is data.
But even if you accept (as I do) that the data is correct, there’s another reading of the result that you likely won’t hear about, because it’s not as popular.