Princeton Election Consortium

Meta-analysis of Polls

The suspense in Ohio...

by Sam Wang

August 15, 2008, 9:24 am

I don't know about you, but the suspense is killing me. Where is that history trend headed? If you look at the details, the big state that hasn't reported in weeks is Ohio. Once Ohio reports, we should have a fairly clear idea of the August swing, which I ascribe to McCain's Celebrity(!)/Anti-Christ ads.

Update: ...and, the other shoe drops. PPP (D) reports a tie in Ohio, an eight-point swing from their previous poll. As you can see from the history graph, the result, Obama 291 McCain 247, with a Meta-Margin of 1.58%, is the closest the race has been since mid-June. The next few weeks (VP selections, conventions, McCain's 72nd birthday) should be interesting to watch, to say the least.

Blogospheric mentions of the Meta-Analysis

by Sam Wang

August 8, 2008, 12:14 pm

Thanks to returning and new readers Washington Monthly, The Economist, Angry Bear, Chimpanzee Tea Party, and Matt McIrvin. McIrvin's capsule history of the 2004 Meta-Analysis is particularly good.

A technical note: Non-independence among states

by Sam Wang

August 7, 2008, 7:28 am

The comments on the last thread were quite instructive, and led me to look over Silver's methods documentation in detail. Wow, that’s quite a complex procedure he has. I should probably address your questions about it before commencing with further description of the Meta-Analysis (which is not a prediction). [Read more →]