The Good Side of the Study by Amodio et al.

At least in this study (earlier discussed here) the researchers identified who was liberal or conservative by asking them instead of applying a psychological test that they imagine to measure political opinions (the problems with that were earlier discussed here ). According to John J. Ray, the psychological-test method is common and I think Amodio et al. should be commended for avoiding it.

That still doesn't excuse a sample size that would be considered laughable in public-opinion polling or medical research.

Addendum: There are criticisms from people who actually read the paper here and there.