3 ms·

I've dug further into this and have become less impressed. I'll leave just one example (I had many but I won't hog this space). Apparently this is "high end sc

I've dug further into this and have become less impressed. I'll leave just one example (I had many but I won't hog this space). Apparently this is "high end science journalism". Kai Kupferschmidt tweets this "Some people in the US actually seem to think that rising #covid19 cases won’t lead to rising deaths eventually. It’s like they are watching lightning flash across the sky thinking the thunder won‘t come crashing." Cases don't measure infections, they measure measured infections. Ramping up testing will ramp up cases even if infections do not ramp up. There is a credible theory that more cases are due to more testing, not due to more infections, and if that is true, the death rate will not follow. I don't personally accept that theory, but it's disingenuous and not "high end science journalism" when you fail to actually address the alternate theory.

raphlinus 6y ago
This might be a good opportunity to make some distinctions. @kakape is a professional journalist, and his work product is the articles he publishes in Science. The Twitter feed is not the journalism, but is participating in a conversation. I find following Twitter feeds of journalists particularly useful, because they tend to have their ear to the ground about what is interesting, but of course it's no substitute for vetted, edited articles. That particular tweet resonates poetically for me, and speaks to major themes in my essay. The belief that you can have rising cases and mortality will stay low is, as we speak, a major narrative in what I refer to as "idiot Twitter." What I get from the Tweet is that a basic theoretical understanding of the underlying processes should be enough to dispel the notion. Just as a child might not know that thunder follows lightning, the people espousing this theory do not understand basic epidemiology. It's not a scientific proof that thunder follows lightning, or that there might not be edge cases such as the flash being so far away that the sound dissipates. Increased testing is absolutely a factor in explaining why we're seeing this puzzling pattern, but doesn't by itself explain what we're seeing. Part of the answer is lead time bias, which from what I can tell is a reasonably obscure epidemiological concept but relevant in this context. Dr. Ellie Murray explains it in this linked "tweetorial": https://twitter.com/EpiEllie/status/1280305393516904448 https://twitter.com/EpiEllie/status/1280305393516904448 Simpson's Paradox is another major factor: to really gain an understanding of what's going on, you need to look at individual states or even finer regional divisions, and aggregating them can obscure important patterns. The effect is especially important these days because different states are very different.

raphlinus 6y ago

This might be a good opportunity to make some distinctions. @kakape is a professional journalist, and his work product is the articles he publishes in Science. The Twitter feed is not the journalism, but is participating in a conversation. I find following Twitter feeds of journalists particularly useful, because they tend to have their ear to the ground about what is interesting, but of course it's no substitute for vetted, edited articles. That particular tweet resonates poetically for me, and speaks to major themes in my essay. The belief that you can have rising cases and mortality will stay low is, as we speak, a major narrative in what I refer to as "idiot Twitter." What I get from the Tweet is that a basic theoretical understanding of the underlying processes should be enough to dispel the notion. Just as a child might not know that thunder follows lightning, the people espousing this theory do not understand basic epidemiology. It's not a scientific proof that thunder follows lightning, or that there might not be edge cases such as the flash being so far away that the sound dissipates. Increased testing is absolutely a factor in explaining why we're seeing this puzzling pattern, but doesn't by itself explain what we're seeing. Part of the answer is lead time bias, which from what I can tell is a reasonably obscure epidemiological concept but relevant in this context. Dr. Ellie Murray explains it in this linked "tweetorial": https://twitter.com/EpiEllie/status/1280305393516904448 https://twitter.com/EpiEllie/status/1280305393516904448 Simpson's Paradox is another major factor: to really gain an understanding of what's going on, you need to look at individual states or even finer regional divisions, and aggregating them can obscure important patterns. The effect is especially important these days because different states are very different.