I can’t recommend this enough. takes apart the idea of “absence of evidence is evidence of absence”, and explains: well, sometimes it really is!
This seems to me to be the more pragmatic question.
zeynep.substack.com/p/pandemic-the
4
36
130
Conversation
Also: when does correlation imply causation, and when does the plural of anecdote equal data?
This requires pragmatic judgment — a difficult thing to do.
1
7
Replying to
In real life business situations (as opposed to toy examples or financial markets) I find that correlation usually implies a link (if not causation) worth exploring.
Totally spurious correlation is rare.
1
1
I'd probably generalize to say, in a data rich situation, coordination doesn't imply a link. In a data poor situation, it usually does.
2
1
1
Replying to
This is actually pretty darned good, as a loose heuristic.