Does it predict people that allegedly finished university not knowing the difference between correlation and causality?
This reminds me of a fraud risk classification model I once heard about, which ended up being an excellent income-by-postal-code classifier.

