2 ms·
I'm surprised I need to say this on HN, but you can't make inferences from this sort of uncontrolled choose-your-own-correlation-adventure in a whole population
by hombre_fatal 2y ago
I'm surprised I need to say this on HN, but you can't make inferences from this sort of uncontrolled choose-your-own-correlation-adventure in a whole population over time because thousands of things have changed in the population over that time.
If this were your epistemic standard, you'd have to accept all sorts of absurd or conflicting conclusions.
- tzs 2y agoWhen someone is suggesting that a particular thing is causing or partly causing a particular outcome it is not a choose-your-own-correlate-adventure to look at how prevalence of that thing and that outcome have changed over time. Things tend to be affected by their causes (!) so when the outcome is changing significantly and the purported cause is not then it suggests more evidence needs to be offered before we accept that the purported cause actually has a significant affect.