Thursday, January 24, 2008

Polling Trends

I am following Pollster's South Carolina poll reporting fairly closely, not because I think HRC will win in SC (though it is now a faint possibility) but more to watch Golden Boy Barry's long deflation after Iowa. It's kind of like watching the housing bubble flatten out - lots of irrational exuberance, people taking out lines of credit that aren't really supported by the fundamentals, and a lot of people left holding a property that is no longer worth what it was when they first bought it.

First, one perpetually silly thing that keeps happening is people are trying to compare apples to rutabagas in the polls. A poll is really only directly comparable to a poll previously done by the same operation and then only if they use the same polling methodology. Even then it's not so good. You can't compare something done by Zogby (notoriously unreliable, as I've pointed out before) and a poll conducted by Clemson. A collection of polls can help to establish trends, though no one data point is completely reliable. On Pollster's South Carolina page, what you can see are polls over time conducted by various organizations.

One thing to look at first is how candidates do over time. While Clinton had some forays into the low 40s, she has consistently been polling in the 30s in SC. Her worst recent numbers are right after Iowa and she shows a small but steady upward trend. The Clemson poll is not really an outlier. Compare it to the previous Clemson polls and you will see that all the candidates have low numbers from that poll. When I see her numbers in SC, I think we're seeing the support she has always had there (a solid 30% ) without much of anything else. She's not increasing her margins significantly, but neither is she losing them. Think about NH. I see her drop as wavering supporters kinda thinking maybe someone else, then seeing her win three in a row and fighting fiercely in the debate and this brings them back. I don't think she'll break 35%.

Obama started a real climb in early December, when the anti-HRC/pro-Obama press orgy was winding up into a fever pitch to make Iowa more important than it actually was. Before then, his numbers veered erratically between high-teens and mid-thirties. Why? As I've said in other posts, Barry's support is soft. Many people think about voting for him, but few commit and many uncommit just as quickly. I do not think this is primarily because of race because his race demographics among his supporters are diverse. He does pull a bigger percentage of the Black vote, but his key demographics are youth, income and Independent status, not race. AA voters are actually the most sceptical about The Golden One if interviews and past polls are to be believed. It is the solidification of that demographic support (previously in the "Undecided" column) that has given him his abrupt rise in SC.

Edwards has made some clear gains in the last week, due in great part to the debate performance. He didn't say anything different, but he came across as "not extreme", and so has shifted a good portion of Undecided to his camp. I think he could, with a truckload of luck, squeak past HRC for 2nd.

So what will make the difference on Saturday? Turn-out. If the youth vote decline continues, plus if a significant portion of the soft Obama vote goes all wobbly because of the Crown Prince's recent hissy fits, he's in trouble, even with a majority of AA votes. I doubt the attrition will be enough to cause a loss, but this will be a closer contest than many have presumed. However, HRC may see significant defections to (or, rather, lack of pick-ups from) the Edwards camp. This split is what will prevent HRC from winning. They must think it is a possibility or they wouldn't waste the Big Dog there when he could be wowing the crowds in Arizona and Colorado.

Take away message: February 5th is the next real primary, and it will put HRC permanently ahead.


No comments: