What DOES correlation imply?
— Karl Pfleger (@KarlPfleger) August 9, 2022
A 🧵 (readable by non-specialists):
If A & B correlate in a dataset, then either:
A->B (A causes B)
B->A (reverse causation)
C->A & C->B (a Confounding variable causes both)
or it's a statistical fluke.
Let's take each of these in turn...