Gregory Cochran points out that early voting undermines election-day exit polling, especially if Obama voters were more likely to vote early. The result could be well-conducted, high-quality exit polls showing a McCain win in a state where ends up losing because of votes already cast. As Cochran puts it, “It seems to me that if any single network releases those exit polls without simultaneously explaining the impact of early voting – and maybe even if they _do_ explain it – a lot of rubes are going to come away with the inexpungible impression that they wuz robbed.” Since I don’t see how exit polling can be banned, this is one more reason to limit early voting only to those who can’t vote on election day because of absence or infirmity.