Interesting, but I'm having trouble thinking of any possible set of figures that would indicate a Dunning-Kruger effect. What was it you were looking for?
Maybe we have a difference in terminology. To me, the Dunning-Kruger effect is where less competent people have higher opinions of their ability than more competent people. The LW survey could have shown that if, e.g.,people with lower IQ as measured by iqtest.dk had higher self-reported IQs. But I don't see how any statistical relationship between intelligence and LW karma could be an instance of D-K. What am I missing?
I was equating "inclination to contribute to LW" with "opinion of one's ability" (in this case to come up with useful and accurate insights). In other words, if D-K is correct, maybe there are a bunch of high-IQ LW readers who never contribute (because they're underestimating their ability, and they don't think they have anything useful to say) and lots of low-IQ LW readers who contribute lots (because they incorrectly see themselves as brilliant and full of insights).
Of course, voting does give pretty good feedback. But still, interesting that there's no apparent trend for folks with higher IQs to contribute more or be more willing to post in Main.
Precommitted to publishing this in Discussion to fight publication bias. It looks like intelligence (as measured by IQ, SAT scores, etc.) isn't meaningfully related to how much one posts to LW. Probably in the ideal case, they would be related and higher-IQ people would post more, but that doesn't appear to be going on either.
How well-educated you are doesn't seem to be much related to participation either. I'm not controlling for hours spent on LW for any of this, though.
Script output:
Script source here.