If you want the scholarly version, skip the blog post and click through to the original paper (which was discussed on HN previously but I can't find the thread).

Page 26 of the paper does cite a study showing that libertarians have higher SAT scores, but that's probably not the only reason Caplan blogged it.