> The markets? Stable? I'm assuming you mean this about Intrade (and
> Iowa) only, and not as a general proposition. But it's not even true
> about Intrade.
>
> PollingReport.com has the Gallup daily tracking polls since January
> 24. Below are some descriptive stats comparing Obama's daily support
> with the Intrade Obama price. The poll has been a lot more stable,
> both since January 24, and since March 1.
Doug,
Are you playing gotcha with me? You are making an apples-to-oranges comparison. But let's cut through the methodological differences. This is really simple. Effectively, ruling out extreme events, there are only two serious possibilities for a Dem nominee: Clinton or Obama. If we take Gallup or Intrade or the IEM as predictors of the Dem nominee, they predict either Obama or Clinton. That's it. Now, take a look at the Gallup and IEM graphs and compare them -- say, from 1/08 to 5/08. Question: Which one has Obama winning over Clinton for a week or two, and then not, and then yes again, and then not again, and then yes again -- for a total of 10 times or more over the period?
Which one has the curves crossing only once in the period?
http://www.gallup.com/poll/election2008.aspx http://iemweb.biz.uiowa.edu/graphs/graph_DConv08.cfm
It's pretty clear to me that the markets have made up their mind (Obama) and that their choice has been much more stable (since 1/08, they've changed their mind only once) than the Gallup polls suggest. You can find this in many other instances. So, conclusion: As predictors of a winner, the Gallup polls are not as stable as the markets.
Now back to putting words in Max's mouth, which is what started all this. In a post from a few days ago that I'm too lazy to find now, Max suggested (or that's what I thought he meant) that the fluctuations in Obama's poll numbers a few weeks ago was randomness playing a joke on us. It is clear to me that the polls are more news driven than the actual conditions that shape people's voting decisions. We all know that the news cycle has been shortened for the sake of profit. For obvious rea$ons$, the media has a bias towards exaggerating the importance of news events. On top of that, they have a built-in bias towards stretching the horse race and make it appear tighter than it is (which does work towards tightening it, since there are real effects of the media echo chamber), because they are the ones who ultimately end up appropriating the fantastic sums of money the candidates raise. Anyway, this is what I understood Max meant -- which I thought wasn't picked up on the thread. But maybe I misconstrued his comments.
I hope this clarifies my point better. Or maybe it won't.