If there is an enthusiasm gap between the candidates—if, that is, the supporters of one candidate are much more enthusiastic about their choice than supporters of a second candidate are about theirs—it would seem to follow that polls would understate the level of support for the second candidate. Specifically, if unenthusiastic supporters are less likely to take part in polls or less likely to express their support for the candidate to whom they’re leaning, I don’t see how the polls could accurately capture their voting intentions. (I invite polling experts to correct me.)
For what it’s worth, in 1996, when there was an enthusiasm gap between Bill Clinton and Bob Dole, the CNN/USA Today/Gallup poll for Oct. 27-28 gave Clinton a 16-point margin (51%-35%), and the final poll gave him a 13-point margin. Clinton’s actual margin of victory was only 8-1/2 points. So, especially given the much larger enthusiasm gap this year, perhaps there’s reason to hope that the polls contain an even larger overstatement of Barack Obama’s support—in which case John McCain may be ahead. (Yes, yes, I admit that this is wishful thinking.)
Of course, lack of enthusiasm can also lead folks not to vote and not to persuade their fellow citizens whom to vote for. Not this election, please.