For example, at least one statistic is based on earnings 5 years out of school.
5 years out of school I was making less than 30k as a research assistant. A few months later I was making easily 100k+ more than most senior SDEs, and doing exactly what I would be doing if I had a trust fund to pay the bills.
Many of my classmates who became doctors were in similar situations -- working through residencies at T+5 years but setting themselves up for mid 6-low 7 salaries in the near future.
So earnings over even a medium-term time horizon can be extremely misleading. You really need that 35-45 year career to get the full picture. Any college grad salary number with less than a 6 or evne 10 year time horizon is going to be a poor representation of what's really going on.