It's locked up behind a paywall, but people confuse "correlation" with "equivalence." For this to be a catch-22, you'd need an amazingly strong correlation.

Also, most psychology results fail to replicate. A single paper has about a 60-70% chance of being nonsense, probably more.