4 ms·
But if we compare California to Florida/Texas, for example, we can at least say pretty conclusively though that the effect wasn't strong enough to overcome clim
by danenania 3y ago
But if we compare California to Florida/Texas, for example, we can at least say pretty conclusively though that the effect wasn't strong enough to overcome climate differences or whatever other confounding variables might be involved. So while we can't say for sure they aren't effective at all, we can place a limit on how effective they are based on the data, and as far as I'm aware, that points to them being marginal at best. They certainly aren't, in themselves, any kind of panacea for stopping a pandemic--that much has been proven.
- ghaff 3y agoExactly. I've not done clinical research but I have done market research. And, while you can cherry-pick apparent differences (or lack of differences), construct stories around those, and just conclude some results are weird anomalies, at some point you need to conclude there's very little evidence there's a lot of consistent signal in some particular policy/demographic/cultural difference.
- kadoban 3y agoDoes the comparison even tell us that? There's more confounding variables than there is signal IMO.
- danenania 3y agoWell that's my point. If these measures were really so effective, then there should have been a strong signal in spite of the other variables, but we didn't see it.
- kadoban 3y agoThe variables are all over the place though, why should the signal be strong?
- ghaff 3y agoAt some point then though, it's about faith. The data is noisy but we believe this hard to discern signal to be true even in the absence of statistical evidence because it's complicated.
- danenania 3y agoIt should be strong if the effect is strong. We find strong signal in noisy data in all kinds of contexts. If all the states with masks and social distancing mandates had, on average, 30% less cases per capita than states without, that would be a strong signal even if it's not conclusive given all the other variables involved. I doubt any other variable could be found that would show that much signal. But we haven't see anything like that.
- kadoban 3y ago> It should be strong if the effect is strong. We find strong signal in noisy data in all kinds of contexts And we fail to find signal, or can find whatever "signal" we want to find in all kinds of other noisy data. What says any kind of useful info is here to find?