Those of you who follow politics closely may have noticed a curious feature about the current state of discussion about the presidential election.
One is that the national polls and the so-called objective fundamentals models favored by political scientists show a close race between Mitt Romney and Barack Obama as to who will get the majority of the national vote.
On the other hand, the poll aggregators look at the polling in individual states and use those to predict the outcome in the electoral college, and those show Obama getting an average of around 300 electoral college votes, well above the 270 he needs to win.
One position for which there is no evidence at all is that Romney is cruising to an easy victory, even a blow-out. And yet, many Republican politicians and party hacks are being increasingly vocal in pushing this point of view. The question is why.
There is the suggestion that people like to vote for the winner and thus talking up the chances of your candidate fires up the enthusiasm of your supporters and increases the chances that they will actually go to the polls. Conversely, you seek to depress the morale of your opponent’s supporters and thus increase the chances that they will mope around and say there is no point voting. This is what drives all the ridiculous talk about ‘momentum’.
On the other hand, you can make the opposite case and say that people get complacent if they think their candidate is going to win easily and will go shopping instead of voting, so putting the fear of losing into their minds will increase their chances of getting to the polls.
What surprises me is that all this seems to be based purely on guesswork. You would think that there would be some empirical studies that would shed some light on this but I am not aware of any.
I suspect that it is a wash, with some voters behaving according to the first model and others according to the second, so there is no significant net effect. In the absence of empirical evidence either way, that seems like a reasonable assumption. It may also be that researchers have looked into this and there was no net effect. Such null results tend to be not attention-grabbing.
My guess is that people are far less influenced by such ephemeral things than the pundits think they are. Whatever the beliefs, a 55%-42% majority of the RAND panel thinks that Obama will win while Gallup has the margin at 54%-34%.