The view that education is mostly about signaling is inconsistent with the established consensus on the returns to schooling and yet the writers at EconLog do not respond to this literature or, as far as I can tell, even acknowledge it.
I acknowledged the literature in my original post. The Card survey actually cites only a couple of natural experiments that are relevant to college. One is the paper that uses proximity to college as a natural experiment. That is, if you live in a college town, that is supposed to be uncorrelated with everything else other than your propensity to attend college. I call baloney sandwich on that one.
Another "classic" natural experiment is by Oreopolous, who uses the fact that some states have laws that require staying in school to different ages--16, or 17, or 18. It turns out that the average age at which students drop out in a state is not related to this variable. Nonetheless, he proceeds to use it as an instrument.
There is a "literature" on all sorts of bogus relationships. Consider epidemiology, for example.
I deeply resent the implication that skepticism about the returns to education is based on ignorance of the literature. Perhaps it is the believers who need to read more.
I say again: Run an actual experiment before you commit yourself to your beliefs.
[UPDATE: Arpit Gupta cites more literature. It turns out that David Card's survey is not the final word o the topic.]