Brief methodological reminder: I'm combining these polls with a weighted average so that polls with higher sample size are more important than those with small sample size. Also, my method allows me to calculate a new standard error so that we can have a margin of error.
There have been no fewer than 17 new public polls (see here) in New Hampshire that were conducted entirely after the Iowa caucuses on January third. I've used only these post-Iowa poll. Here are the weighted polling averages for the Dems:
Obama - 37.1%
Clinton - 29%
Edwards - 18.8%
Richardson - 6%
Other/DK - 9%
There are two questions that we need to answer to get a better sense of what's going to happen. First, how will those undecideds break? If they break mostly for Clinton, she could win. If they break mostly for Edwards, he won't win but he will beat expectations and claim the mantle as the number two. If they break for Obama, then it'll be a landslide, a 10 point or more victory. I think it is really unlikely that they'll break for Clinton. She's been the "front-runner" for a year now (she was polling in the 40's as recently as October), and if these folks haven't planted themselves in Clinton's camp by now, they probably never will. My guess is that they'll break mostly for Obama but some for Edwards.
Second, we need to ask if there is some micro-trend going on even within the very limited time frame of these polls (as happened with the Iowa mega-poll...see earlier post). The answer to this is easy: no. There is no statistical difference between the earliest five polls and the most recent five polls. Even beyond statistical difference, there is almost no nominal difference. Here's the data:
First Five Polls:
Obama - 37.1%
Clinton - 28.3%
Edwards - 19.4%
Last Five Polls:
Obama - 37.6%
Clinton - 28.1%
Edwards - 19.5%
So, what's the prediction. I'd say that the eventual vote totals will look something close to this:
Obama - 40%
Clinton - 30%
Edwards - 22%
Richardson - 5%
No comments:
Post a Comment