Well, this is a median, not a mean, so the quants can't raise it that much. A bifurcated distribution with a couple peaks could though, which is why I was thinking tenured professors + industrial researchers + Googlers on one end, with post-docs on the other.