Selected letters for October 28

Date: Thu, 28 Oct 2004 10:18:45 -0500

From: Bob Hall

Subject: Deciding early was easy

Dear Professor Wang,

Thank you for your website. As a recovering politician, and PolySci major (U-Mass, 72), I find it immensely interesting, and appreciate your attempt to both declare your bias, and keep it out of the site. I also find the letters amusing. Anyone who pretends to be fatidic about victory in war, sport or politics, before the event is over, has small appreciation for historical example or the vagaries of fate.

I do not have Jim G's problem, as my mind was fully made up the moment it was clear that John Kerry would be the Democratic nominee. Since then I have spent a lot of time campaigning against him, knocking on doors, and will work hard through Election Day here in Wisconsin.

I am a Republican, but not an obdurate one. I was elected to five terms in the Massachusetts state senate, from a 4-1 Democratic district, starting in 1972, from the district next to the one where Kerry was defeated for Congress. But I did not vote for George Bush in 2000, casting a write in for John McCain. It was a jerseyvote, so being a mugwump meant little to the race.

I've been on Kerry's TV show, debating the death penalty. (He gave me four days notice to debate a guy from the ACLU who did it for a living, but never mind). I do not agree with Bush on many issues. I suspect, for example, that when I spoke for Barney Frank's gay rights bills on the floor of the Massachusetts senate in 1973, I may have been the first state legislator in the country to do so. And I was a bachelor, holding a seat I'd won by nine votes. But at least Bush has things he believes in, while Kerry suffers from terminal abulia.

Had the Democrats nominated Bob Graham, I'd be voting Democrat. Had they nominated anyone but John Kerry, I'd be a spectator, rather than an active campaigner. So why am I splenetic about Kerry?

I am a Marine Vietnam veteran. Following a call from John Kennedy, I and tens of thousands of decent Americans put on the Republic's uniform, and at the behest of a Democratic president and Congress, went to Vietnam, a war then supported by the leaders of both parties and a majority of Americans.

We returned to find that, due to our service, we were permanently marginalized in American society. I did not torture, rape, murder or wantonly destroy villages, as Kerry testified (based on second-hand perjured reports) that we all did, nor did I see any of that. We who served have lived with that stigma for over 30 years. Now Medal of Honor recipients can't even get their story, Stolen Honor, on TV. Yet John Kerry refuses to sign a 180 form, releasing all his military records, as Bush, Gore and McCain did, and gets a pass from the media. Had Bush refused to sign the release, it would be the lead story on CBS every night. We who served think Kerry is hiding something, but we'll never know.

War is not ping-pong. Franklin Roosevelt coolly ordered the firebombing deaths of perhaps a million Japanese children, to terrorize Japan into surrender. Atrocities happened in Vietnam, but we believe they were the established policy of the enemy, while they were aberrations by Americans. Certainly they were not supported by officers "at all levels of command," as Kerry claimed.

Kerry now says we "must always separate the warrior from the war," a howler larger than his claim that he and John Edwards support tort reform. He was quite willing to traduce his comrades to launch his political career, so now calling us his "band of brothers" rings very hollow. In my view, there is no position John Kerry will not take, nothing John Kerry will not say, if polls say it will increase his chances of victory. If he is elected president of the republic, I will know I did all in my power to prevent it.

Robert A. Hall

Cpl, USMC, 1964-68

SSgt, USMCR, 1977-83

Massachusetts Senate, 1973-82.

Madison, WI

Date: Thu, 28 Oct 2004 09:12:09 -0700


Subject: A non-partisan activist?


I wanted you to know that your website is making a difference across the nation. Because of your site I saw that I really could make a difference in this election. I've been volunteering with my local ACT here in Las Vegas, NV for about a week now. Funny thing is most of the other volunteers are driving in from CA. I've never heard of this happening before... is this normal?

I thought this message would be interesting to your other readers because I am a registered non-partisan. I don't really consider myself a Democrat, but I will never vote Republican. I've voted Libertarian or Green Party before, but this year I'm voting for Democrats. I urge any undecided voters reading this to vote Democrat in this election too. I didn't like Kerry at first either, but the more I learn about him, the more I think he will make an excellent president, even if he isn't very photogenic. I think most of you would agree that our current president is not up to par, but also think about who he is really representing. Although the GOP may sound good on paper to some, the Republican party gets their money from, and therefore represents, large corporations and fundamentalist extremists who call themselves Christians. Unless you own a large corporation (not a small business), or are a fundamentalist yourself, the GOP will be unlikely to be in your own interest. Think about it...


working to get out the vote for anyone who's not a Republican

Date: Thu, 28 Oct 2004 09:57:11 -0700

From: Paul Bovarnick

Subject: Your terrific site

Thanks for your terrific site. I am a mathematically challenged liberal arts graduate who is hopelessly addicted to news of this election. I have found your site enormously helpful in thinking about the horse race part of the election, and surprisingly insightful in explaining the behavior of my fellow citizens. Thanks. Paul S. Bovanrick, Portland, Oregon

Date: Thu, 28 Oct 2004 13:20:13 -0400

From: "Ackman, Dan"

Subject: The late break

Professor Wang--

It's "conventional wisdom" that late deciders "break for the challenger. But does data back up the wisdom?

From the limited numbers I've seen, the late break was for Gore (an incumbent by party) and for Dole (challenger). But in 1992, the break was fairy even with Clinton gaining 1.76%, GHW Bush gaining 1.45% and Perot gaining 1.16%.

Dan Ackman

Senior Columnist

New York, NY

Dear Mr. Ackman: That's interesting, though contradictory to what I have seen. I think your figures have a bit too much claimed accuracy - I would feel uncomfortable reporting anything after the decimal at all.

This link indicates that GHW Bush polled at 37% and received the same amount on Election Day. Your figures indicate that he did 1% better. Very broadly, these are consistent - though more numbers would be helpful in resolving this.

All the best,

Sam Wang

Date: Thu, 28 Oct 2004 14:08:42 -0400

From: Richard Palmer

Subject: Handling Undecided Voters, Kerry Getting 2/3 is pretty reasonable from the current data.

I took the pooling history from another website (attached), dropped USA figures, and estimated a various models of how changes in undecided voting affect changes in Kerry's percent. To keep things on the "up and up", I dropped observations where there was no sample size estimate or no MOE estimate or no Nader Share or data information was not available. Changes in undecided voting has a strong negative correlation with changes in Kerry's share: e.g. As the # of undecided voters drops, at least 60% of the voters seem to move to the Kerry Camp. I have estimated models with and without an intercept and a model with separate state intercepts. The only other variable of significance is a trend variable (measured as the # of days from the midpoint of the earliest poll to the midpoint of the next poll. This requires I drop the earliest poll in each state). The coefficient is small and positive -- trends among the undecided voters is toward Kerry. Given the existing public poll information it is very reasonable to expect Mr. Kerry to pick up 2/3 of the undecided votes.

Richard Palmer

Date: Thu, 28 Oct 2004 02:11:55 -0400

From: John Huber

To: sswang@Princeton.EDU

Thanks for creating the site and running it in such an open fashion.

On the undecideds, two thoughts. First, on how to predict their behavior, with all your historic data from so many different polling firms, could you exploit any systematic differences that might exist in the number of undecideds each poll reports to extrapolate what the undecideds are up to? If the undecideds are 1 -(Bush+Nader+Kerry), then eyeballing it, it seems like some firms have consistently fewer undecideds than others. If this is true, perhaps some polls have polling strategies that make it more or less difficult to express "undecided." If a polling firm with consistently more undecideds produces results that are consistenly biased toward a particular candidate (vis-a-vis a poll with fewer undecideds), you might be able to make inferences about which way the undecideds are going. For example, if Zogby makes it tough to say "undecided", and Zogby results favor Bush (or Kerry) compared to other polls, then we have some information about which way voters go when "pushed" to name a candidate. I understand that other differences in polling strategies will underlie these differences across polling firms, but there still might be information there.

Regarding how to present the results, perhaps instead of making an arbitrary assumption about which way they go, you could present information about what percent of undecided voters a particular candidate needs to receive to pass a particular threshold in a state. For example, if a 85 percent probability of winning is a number people care about, you could report what percent of undecided voters Bush or Kerry must receive to pass that threshold. It's basically the information you present in reverse (i.e., you give the probability of winning given an assumption about undecideds, and you could give the percentage of undecideds necessary to surpass a particular probablility of winning). Doing this would underline that the undecideds don't matter much, except in a small handful of very important states.

Thanks again for all the time you've devoted to the project.

John Huber

Date: Thu, 28 Oct 2004 08:48:08 -0400

From: Edward Witten

Dear Sam,

On, I read yesterday a rumor that a NYT poll of Florida showing Kerry ahead by +9 percent was buried as being implausible. I don't know if the rumor is true, and if it is I am sure the poll was flawed, as Kerry is surely not leading Florida by that amount. But to me it illustrates the fragility of trying to predict the election from the available state polls. Including or excluding a single, undoubtedly flawed, poll showing a +9 percent lead in Florida for Kerry (or Bush) would probably have a significant impact on your overall assessment of the outcome of the election.

- Edward

Date: Thu, 28 Oct 2004 15:07:17 -0400

From: Matthew Bell

Subject: exit polls and absentee and early voting

I'm looking at today's NYT article about Florida and the emphasis on early voting. If it is true, as some are claiming, that there has been very heavy early voting and that it is biased towards Kerry, what will that do to the reliability of exit polling conducted for television on election day? Do you know if polling organizations are making plans to correct for that bias (if indeed there is such a bias)?

Excellent site. Thank you for the thought and hard work you've evidently put into this.

Matthew Bell

Date: Thu, 28 Oct 2004 16:49:01 -0400

From: Thomas R. Jackson

Subject: More information on undecideds

This study still doesn't answer the great undecided question, but I though it might be of interest. This prospective study shows more movement toward Kerry than the Rasmussen retrospective poll:

Pew study

Of course neither can answer the question about last minute movement.

The goat entrails are still probably better predictors. :-)

Thomas R. Jackson.