The real concern is not whether college degree holders, as a class, do better, but whether a 17-19 year old would do better by entering that class. What those

The real concern is not whether college degree holders, as a class, do better, but whether a 17-19 year old would do better by entering that class. What those statistics don't include is something for recent grads. Ideally, you'd want to look at the unemployment rates for people approximately 24-30, stratified by college degree or not. And even then, it wouldn't necessarily give you a complete picture, due to all kinds of selection bias in the people who do or do not go to college. I would predict, however, that the unemployment rate for recent grads would be closer to the non-college graduate rate for the same age than to the all-ages college graduate rate.