Speaking as someone who did upvote the comment because the data was interesting- there are serious issues here. I've seen stuff on VDARE which involved very poor presentation or evaluation of data, and the motivation for it was pretty apparent. The signal to noise ratio there is extremely bad, and there's a lot that's just out and out racism in the strong sense of thinking that members of minority groups are of less moral value.. The post you set was quite interesting. But frankly, whenever I see something from there I have to doublecheck the data to make sure they aren't screwing up in some way. For simple matters of signaling, if I were to link there I'd include a note of the form "yes, this website is full of racists but in this case their summary data is pretty accurate." Steve himself seems to be pretty tonedeaf- I suspect a lot more people would listen to him if he didn't post stuff on such overtly racist locations.
The fake data usually is the most interesting, in so much as 'interesting' is a measure of surprise or confirmation. Also before saying 'it is pretty accuarete', the actual accuracy (vs the sources etc, or vs whenever they did the study at all) got to be evaluated. Usually people say 'pretty accurate' about stuff like this whenever it simply looks plausible to them.
Post by fellow LW reader Razib Khan, who many here probably know from the gnxp site or perhaps from his debate with Eliezer.