It may be evidence, but it still leaves as an open question whether political affilitations are slanted as a result of greater rationality or a political bias. Without some sort of controlled experiment this would be hard to tell.

If they aren't slanted, it either means that what we discuss is not related to politics (implausible) or that Lesswrong doesn't have an impact on such matters.