Ok, so I read the article (years ago, while in the midst of a social sciences PhD, so imagine how I felt :)), and I read it again.

Then, I vainly read the comments hoping that someone had made my point, but alas.

So, a couple of propositions:

1) pay in academia is less than that in industry (especially in engineering and computing) 2) more PhD's stay in academia. 3) Averages are affected by extreme values. Note I'm using the mean here, as that's what all the studies do (annoyingly enough).

Therefore, it would make sense that on an aggregate level, the salaries of people with PhD's would not show an increase relative to those with masters.

Note that clearly this does not imply that doing a PhD is an (economic) waste of time, as it may be that PhD's, conditional on being employed in industry actually earn more than those with master's degrees.

Man, I hate the mean sometimes.