« The coming mandate debate | Main | Jim Stimson on the stability of Obama's lead »

October 24, 2008

Comments

Yes, a campaign manager must try to keep hope alive, but there is no hope. One reason Bush didn't match his polls was that his old drunk driving conviction dominated the news the last 2 days before the election. In retrospect, the DUI told us nothing important. We already knew that Bush had once been a heavy drinker and we knew he had quite drinking many years ago.

Could Obama be defeated by a last minute surprise? I'd say no. Not only is his lead considerably larger, but IMHO the media wouldn't publicize a negative story the same way they did Bush's DUI.

One reason Bush didn't match his polls was that his old drunk driving conviction dominated the news the last 2 days before the election.

I think it's a stretch to blame any news story on a difference between the final polling results and the actual outcome. My guess is the drunk driving story didn't change a single vote.

Polls are estimates of public support, not actual measurements of it and have fundamental uncertainty outside of their margins of error (like what the composition of the actual voting population will be.) That's where much of the scatter in the polls resides and explains why different polls taken on the same day can fall outside each other's MOE. You can see that on the charts Brendan posted above.

In retrospect, the DUI told us nothing important. We already knew that Bush had once been a heavy drinker and we knew he had quite drinking many years ago.

Drinking isn't illegal. Drunken driving is. Lying about it on the campain trail said something about his (lack of) character (though I agree that was something we already knew).

What is with the first chart above? The label on top says it's about 2000. The legend on the left axis says it's measuring Kerry minus Bush (which, some of us remember was in 2004).

And in no way did that story 'dominate the news'. It got some coverage, but a friend who was a Gore staffer vented to me only days later about how little coverage it got, and I distinctly remember commiserating but thinking 'with as little detail as was available, it was surprising it got any at all.'

(Oh, and go Devils. That story took place somewhere on Ninth Street.)

The problem with the chart at top is that it is mislabeled -- Not the Kerry thing, that's correct -- it's the "2000" label. That looks like the chart for the 2004 campaign.

Here's the link to a chart that plots 2000, 2004 and 2008 on the same graph, courtesy of pollster.com:
http://www.pollster.com/blogs/Pres08vs04aand000verlay.png

It take it all back. The top chart on this post is the 2000 confidence interval. My apologies. But it was Gore.

The comments to this entry are closed.