Thursday, October 23, 2008

Presidential Election Results/Margin Prediction

This is the first in a series of three posts. I'll update my predictions one week from today and the day before the election. I'm making these educated guessed based on what I've been seeing at several different poll tracking sites as well the IEM.

I predict Obama will win the popular vote by 8 percent. As to the electoral college, I foresee a massive blowout. Obama will secure 381 electoral votes to McCain's 157.

3 comments:

taco said...

As an espoused libertarian, I understand why you point to the IEM to illustrate people's political allegiances.

However, it's not a scientific poll and and the winner take all market you always select to post on this blog has consistently had the widest margin between candidates I've ever seen. (Except for the asinine unsourced electoral college maps reposted from Kos that show virtually every state across the country's breadth from Navada to Pennsylvia voting for McCain and all the costal states going to Obama.)

I advise you to give more weight to other polls as you post your result predictions. 8% is pretty high and hints at approximately 90% of undecideds going to Obama. If you really believe that, you don't have a sufficient appreciation of how many religious right voters there truly are in this country. Gallup and Rasmussen's likely voter polls (with a greater margin of error obviously) are looking at a margin of more in the 5%-7% range.

Christopher Patton said...

Taco, maybe I'm not understanding you correctly, but your comment seems to imply that the IEM market graph I link to represents expected vote margins. It doesn't and I've never said it did. (Read the market's prospectus here: http://www.biz.uiowa.edu/iem/markets/pr_Pres08_WTA.html)

The numbers represent the probability of a candidate winning. Thus, the market puts the odds of an Obama victory at around 85%. The margin of victory is not a factor in the market.

taco said...

You're right, I've been interpreting the IEM WTA graph haphazardly to date. I suppose the closer we get to election day even if public opinion polls hold perfectly steady, IEM's margin should continue to widen.

How exactly are you going to incorporate IEM data into a margin prediction though, reflect different types of confidence and are tending so much differently? What "different poll tracking sites" are you using that makes you estimate a popular vote margin nearly double what Gallup polls find?