During my two week absence from this blog, I’ve been able to do a slight bit of poking around on the internet. The sense I got from what I have been reading about the US presidential election is that there are those who feel pretty confident of an Obama win, and those who are still on the fence, believing that one can not fully trust the polling data. I find myself jumping back and forth between the two categories.
In the Washington Post, Michael Abramowitz addresses this question: are the polls really accurate? And considering that the national polls right now swing between a 2-point margin and a 15 point margin for Obama. How can that possibly be? And how do we know which polls are closest to reality?
The latest ABC-Washington Post poll might be one. This poll, which gives Obama the edge at 52-45, does not increase African American or under-30 voter turn out over the 2004 level, even though we have reason to believe that both groups are unusually motivated to get out and vote this year. We also have reason to believe that Republican leaning voters are less likely to turnout, due both to lack of interest in the ticket and also perhaps due to a sense that the election is already over.
(The sense that the race is over could also cause some Obama leaners to stay home, which might explain why Obama has been barnstorming pretty safe territory like the state of Pennsylvania and the NBC/MSNBC network, where he will sit for several interviews this week.)
The ABC-Wash Post poll does include a random sample of cell phone-only voters, which gives it a leg up on other polls that do not. And, of course, this poll falls basically in the middle of the latest poll findings. And, while the polls’ margins may be all over the map, the fact is that in more than 50 national polls taken over the last six weeks, Barack Obama has been up in all but one of them (in the first of the six weeks I examined). Shouldn’t such a long and consistent run mean this election is over?
Maybe. But, as Abramowitz notes, the New Hampshire primary really traumatized the Obama campaign and its pollsters. Obama was clearly polling ahead in that state and Hillary Clinton defeated him easily. It is hard to say what really happened. A whole new set of principles and schools of thought on polling will be minted after this year’s election is finally over. Until then, the Obama campaign is probably going to be preoccupied trying to handicap the race very conservatively.
For starters, they have to be worried about the voters who have yet to swing to Obama. If you haven’t gone over to Obama despite the Republican brand being so in the toilet, and John McCain having stumbled and bumbled for nearly eight straight weeks, then I am guessing you must really have a healthy (as in ‘a lot of’) amount of skepticism about the Democratic alternative. The campaign should brace itself for 70% or more of undecideds in these polls to either swing to McCain or stay home. Plus, there are a lot of slightly committed voters out there who are still vacillating, despite having told pollsters and friends that they think they’re going to vote for candidate X.
Of course, ask yourself how much the national polls count anyway, since the race will be won in states like Virginia, Iowa, Michigan and New Mexico.
Inform Your Consent 