In making the case against the Electoral College, I frequently encounter the argument that the "popular vote" not only doesn't determine the winner but is a meaningless concept in our system. The argument goes as follows. Because of the Electoral College, candidates focus their campaigns in particular states and this influences the outcome. Al Gore may have received half a million more votes than George W. Bush, but that was the end result of two campaigns that had been conducted on a state-by-state basis. We have no way of knowing what the nationwide totals would have been in a non-Electoral College system, because the campaigns would have been conducted differently, yielding different results. Therefore, Gore's apparent popular-vote lead doesn't mean anything.
I heard this argument several times from Republicans in 2000. I heard it most recently from Charles M. Kozierok, a blogger and self-described Democrat who is presumably not speaking from partisan bitterness over what happened eight years ago. Whoever makes the argument and whatever motives there may be in making it, it is rooted in flawed logic.
The essential fallacy in this argument is that it confuses the causes of public opinion with the measurement of public opinion.
The purpose of an election is to ascertain the will of the people. It cannot ever be a perfect measure, since we can't read minds. A flat tire on the way to the polling station can distort the outcome. So can a medical emergency, or bad weather, or problems in the voting machines, or any number of other factors unrelated to people's intentions. Nevertheless, an election is meant to reflect as accurately as possible what the public thinks at a particular moment in time.
Campaigns do influence the outcome, of course. But they have have little to do with the accuracy of the election in measuring public opinion. Instead, they have an effect on public opinion itself before it is measured.
Say a candidate campaigns in Missouri but not Kansas. The election may turn out differently than if the candidate were to campaign in both states. But all that means is that the candidate has influenced the voters in a particular way, before their views were measured in the voting booths. The combined vote total in both states is still an accurate and meaningful measure of the collective will of Missouri and Kansas voters. So too with the collective will of voters in all fifty states (plus DC). It may not determine the winner, but it is independently significant.
There is plenty to debate about the Electoral College. Let's retire silly arguments that distract from the obvious, which is that electoral-popular splits like what happened in 2000 do great damage to public confidence in our system. We can't conjure this problem away with sophistry.