Comment by wojt_eu

This shows correlation but not causation. Hypothetically: people with certain background (social class) and traits (IQ) are more likely to go to college, and they would fare as well or better without college. I'd be very interested in stats of lifetime earnings difference that was corrected for selection bias.