Correlation between variables means nothing if you don't have some explanation for a causal mechanism from the one to the other.
That's a hypothesis test in that footnote. Don't trust those either: look at that p-value! The null hypothesis is that there's no relationship between the two variables, and that's easily rejected. Yeah, maybe there really is a relationship here, but I think it's probably a Type I error (i.e., rejecting a true null).
This is not saying hypothesis tests are bad. Instead it means they can be done prematurely and thoughtlessly. Garbage in garbage out, as they say.
In gray, it notes a website devoted to these, as well as a companion site dedicated to academic papers based on probably spurious correlations.
No comments:
Post a Comment