Monday, November 01, 2004

The Blow by Blow over the Next Crucial 24 Hours - And What the Best Polls are Predicting

Tomorrow is the BIG DAY, and though we’ll be grinding away during the workaday hours, we will be monitoring a number of excellent blogs and other resources that will provide clear insight as to how the day is going for George W. Bush.

Personally, through their tactics, I see a confidence in the Bush / Cheney team: Cheney spending time in Colorado campaigning for Coors' US Senate race, flying off to Hawaii to campaign in that state's unusually close race – a state that has not seen a top of the ticket candidate since 1960. Bush is focusing on Ohio, Michigan, New Mexico, and leaving Florida out of the mix.

I think they know something that the MSM is not getting across because it behooves the MSM to make it a close race – even if it isn’t -- maybe they sell more ads?

The one fly in the ointment is the excellent analysis from The Horserace Blog (THB). Up to this point, THB has been projecting better outcomes for Bush / Cheney than what was measured this weekend. I don’t believe that the temperament has changed all that much in a single weekend against the President. I can’t see how it has – the news has all been positive for him, and negative for Kerry. Yet the polls have tightened as you can see by the excerpt from THB below.

I provide this excerpt because it is an excellent and realistic analysis and explanation of the best methodologies in polling. It is also a clarion call to all those in the Bush camp to remain calm despite the tightening; despite the MSM continuing its unrelenting assault on the Bush / Cheney team.

You can read the entire transcript from THB here. Nonetheless, below is a very good excerpt from THB. Let's see how these predictions stand up tomorrow night. I think they will be close:

Greetings, all! Here are some news and notes for you.

1. Continue not-sweating the Fox poll. They are working off a 100% weekend sample.

2. I know many of you are nervous. I exhort you to remember the Von Neumann-Morgenstern utility function. What is the Von Neumann-Morgenstern utility function? It is a formula for calculating expected utility. It strikes me as being applicable to psychology. Von Neumann and Morgenstern argued that you could calculate the expected "utility" (or, in common parlance, benefits) from multiplying the probability of an event occurring by the expected benefits from that event. Thus, for example, suppose that you have a 30% chance of winning a game and would win $50.00 if you won. Von Neumann and Morgenstern would argue that your expected "utility" is $15.

Similarly, you are undoubtedly worried about this election, as am I. I am worried not because I think Bush is likely, but the consequence of a Bush loss would be grave. You "multiply" the former by the latter, and you wind up with a high number. This is not because the probability is high, but because the consequence of its occurrence is high.

3. My election night project is coming along swimmingly. Note that I will not be able to "call races" as the networks do. I do not believe I will have access to exit polling data, and further I am skeptical that my mathematical knowledge is sufficiently high to evaluate it in relation to exit returns to make a call.

4. Do not sweat the stories about Democrats outdoing GOPers in terms of early turnout, especially in IA and FL. The number of early voters in IA in the Des Moines Register survey was something like 90. The number of early voters in FL in the Gallup survey was something like 150. While these numbers are large enough to draw inferences about the population, the margin of error is exceedingly high. The Dems know this. They are clinging to these numbers purely for spin purposes.

5. Update on polls. Gerry Daly has convinced me to excise both the St. Cloud State University poll and the Star-Tribune poll from my analysis of Minnesota. Here are the updated results. I have also excised the Columbus Dispatch poll, as it was done by mail.

Finally, I added the new Quinnipiac and University of Cincinatti Polls. Here are the results:


Bush: 47.12% Kerry: 47.79% MOE: +/- 1.2% (Respondents: ; 6,199; Polls Used: Fox News, Cleveland Plain-Dealer, LA Times, Strategic Vision, Mason-Dixon, Gallup, University of Cincinatti). Based on these results, we can be 64.43% confident that Bush currently has a lead.


Bush: 48.12% Kerry: 46.41% MOE: +/- 1.3% (Respondents: 6,074; Polls Used: Fox News 11/01, Mason-Dixon 10/30, Insider Advantage 10/29, Strategic Vision 10/28, NY Times 10/28, Quinnipiac 11/01, LA Times 10/27, Gallup 10/31). Based on these results, we can be 97.13% confident that Bush currently has a lead.


Bush: 46.52% Kerry: 48.02% MOE: +/- 1.7% (Respondents: 3,259; Polls Used: St. Paul Pioneer Press 10/31; Strategic Vision 10/27; Humphrey Institute 10/27; Mason-Dixon 10/30; Gallup 10/31). Based on these results, we can be 88.69% confident that Kerry currently has a lead.


Bush: 47.68% Kerry: 47.85% MOE: +/- 1.4% (Respondents: 5,373; Polls Used: West Chester Univ. 10/28, Gallup 10/31, Quinnipiac 10/27, LA Times 10/27, Temple University 10/27, Mason Dixon 10/30). Based on these results, we can be 57.14% confident that Kerry currently has a lead.


Bush: 49.06% Kerry: 45.41% MOE: +/- 1.4% (Respondents: 3,790; Polls Used: Fox News 11/01, Strategic Vision 10/28, Badger Poll 10/28,Mason-Dixon 10/30, Gallup 10/31). Based on these results, we can be 99.99% confident that Bush currently has a lead.


Bush: 47.92% Kerry: 46.03% MOE: +/- 1.4% (Respondents: 4,651; Polls Used: Fox News 11/01, Des Moines Register 10/31, Research 2000 10/28, Strategic Vision 10/28, Gallup 10/31, Mason-Dixon 10/30). Based on these results, we can be 96.56% confident that Bush currently has a lead.


Bush: 48.55 Kerry: 46.37 MOE: +/- 0.9% (Respondents: 13,671; Polls Used: Marist 11/1, Newsweek 10/30, ABC News 10/31, Gallup 10/31 (sans leaners), GWU/Battleground 10/28-10/31, Fox News 10/27-11/01, Pew 10/31 (sans leaners), NBC News/WSJ 10/31, CBS News/ NY Times 10/31, Battleground 10/31). Based on these results, we can be 99.99% confident that Bush currently has a lead.

Current best probability of a minimal Bush win in the Electoral College (FL, WI, NM): 94.57%. Current best probability of a minimal Kerry win in the Electoral College (OH, PA, FL): 0.69%. Bottom Line: Kerry's best shot a winning is to sweep the "Big 3:" PA, FL, OH. Currently, the chances of that happening -- assuming that the poll averages are accurate -- is less than 1%.7.

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]