I don't think that would work. People here tend to be more consequentialist than I've seen from people not from here, so we'd probably not be able to see as much of a difference. Plus, the people here are hardly what I'd call normal and are more homogeneous than a more standard set of people. To effectively test that, we'd have to conduct that survey with a more random group of people. I mean, that survey would work, but the sample should be different than the contributors of LessWrong.

If the number of deontologists isn't big enough to power our inference, the stats should tell us this. There are some though.

And I think going outside LW is unnecessary. This essay is hardly aimed at people-in-general.