One of the things I've noticed is that HN voters typically vote up screeds rather than evidence. For instance, there have actually been several articles about the results of real life bailouts, both successful and otherwise; these almost never seem to get as many upvotes as the "Francisco's money speech" type submissions.

But now that you mention this, maybe the problem is that it's easier to recognize articles with identical data than articles with identical ideas. So people say "oh, I've already seen an article about bailout X", but not "oh, this is just another flavor of libertarian blathering."