Selected letters for October 27

Date: Wed, 27 Oct 2004 23:54:39 -0700 (PDT)

From: Jim G

Subject: Undecideds

Prof. Wang,

I'm probably what most would call an undecided voter, and I've been reading your site regularly this month.

I've always been fascinated with numbers, and while I make my living in the world of professional sports (not as an athlete), election time has always been exciting for me. At least until 2000, the election represented the one time when the real world was like the sports world - the game gets played on the field, everyone starts with a score of zero.

Not that I'm saying Bush cheated to win. I'm confident it was a fair election. But millions aren't, and that has haunted us for the last four years. More than anything, I hope that the end result is decisive. We all need, at least, to feel comfortable that the people decided and not the courts or the union guys in the back room or the police intimidating minority voters, etc.

So, here we are, five days away, and I don't know what box I'll fill in. I've benefitted from Bush's tax cuts, but not so much that I wouldn't have benefitted more from a strong economy.

But I don't blame the economy's troubles on Bush. In fact, I believe the tax cuts, as they affected the very rich and corporations, may well have spurred the economy. So that issue makes me want to lean toward Bush a little.

Health care is very important. I run a small business, and all I can afford is catastrophic insurance for my family. Tort reform is a big issue for me, and Kerry's choice of Edwards for vice president was a huge disappointment. My wife gave birth to our son last year, and I almost fought with the doctors to avoid her having to have a c-section just because our son was taking his sweet time coming out.

We have a happy, healthy kid, and now I read that the rush to c-section is solely CYA medicine on the part of doctors worried about cerebral palsy lawsuits. A practice Edwards himself helped pioneer. Complete junk science, as studies have shown rather convincingly.

I was active in the Democratic primaries, voted for Lieberman (in New Hampshire, you do not have to register for a party to vote in a primary). I was very disappointed that his candidacy was all but over before New Hampshire. He seemed so sensible and competent. I would have had no trouble supporting him for president.

Anyway, on domestic issues, I'd be a reluctant Bush supporter.

But, there's the war. Literally as my wife went into labor last year, the first bombs fell on Baghdad. Kerry's right when he says we had not reached a point of last resort when the invasion began. There was no imminent danger. There was no reason to believe there was. While I think Kerry's dreaming when he says he could get France and Germany to even entertain sending troops into Iraq, Bush quite simply screwed up - and screwed up badly.

On foreign issues, I'm a Kerry supporter. Quite the opposite of these silly pundits who think Bush wins points on foreign affairs and Kerry on domestic affairs. No wonder Bush was a babbling idiot during the first debate, and sounded almost sentient in the second and third. Pundits spend far too much time listening to partisans and not enough time listening to middle-roaders.

And then there's religion. Until Kerry found Jesus during the last debate, he had a huge edge there. You see, I believe in the separation of church and state. Neither of these guys does. You'd think they were both running for the position of vicar in the local parish, not for president, the way they keep spouting off about things that should remain private.

Which brings me to why I don't like Kerry. I don't really believe he's a religious nut. But I do believe that someone on his staff told him he'd win a few votes in Iowa and Ohio if he made a few statements about religion. I have no idea who he really is or what he would do as president. He's very good at criticizing what Bush did wrong, but I have no confidence that he really has a plan for anything. He just wants people to vote for him.

So, I'm pretty torn right now. My gut tells me we need to fire the incumbent. My brain tells me that the alternative may be worse, may be better.

I see why you're giving 2% to Kerry. Ordinarily, I'd agree with you. But I think this case is different in that we have an almost uniquely unpleasant alternative. I would give about .5%-1.0% to Kerry, and that's because there are so many of us now (mostly younger and more likely to support the Democrats) who just can't be reached by phone these days (I won't pick up the phone unless caller ID indicates it's someone I want to hear from). But in all other cases, the challenger tried to make a case as to why he should be president. Kerry hasn't.

My prediction is that Bush will win the popular vote by about 1%. And that because of the structure of the electoral college, Kerry will become president due to incredibly thin margins in Florida, Ohio, Minnesota, Iowa and New Mexico. Maybe not as thin as Florida 2000, but on the order of half a percent at least in Florida and Iowa.

Sorry to ramble on so long. I get so tired of hearing commentary from the partisans and I was hoping you did, too (even if you are a Democrat). Thanks for listening.

Jim G., New Hampshire

Date: Wed, 27 Oct 2004 08:49:19 -0400

From: John Walsh

Subject: Hitting the streets

Hello Sam,

Nice job on the web site. I have been trying to figure out where to go to help Kerry in the four days before November 3.

I called DNC national HQ and they said that they needed people most in MN and Iowa. What do you suggest? (Of course the DNC knows what resources are already in place and you do not. So in that respect your suggestion is not as good as theirs. On the other hand they may not properly understand the probability of winning in a given state.)


John V. Walsh

Sam Wang says: My calculations indicate that those are good suggestions - see the October 28 entry on the main page. I infer that the DNC must think they have enough people in place already in OH and FL.

Date: Wed, 27 Oct 2004 08:49:02 -0400

From: Steve Rosen

You re-inspired some interest in statistics. My first hypothesis tested was (obvious to me) that the dumber a state was (21 factor Smart index about State's education systems and successes from the web), the redder it was (based on your projected probability array). Correlation of .38 was significant at 93% confidence.

Please turn WI blue again ASAP.... :-)

Steve Rosen

Roswell, GA

Date: Wed, 27 Oct 2004 09:48:33 -0700

From: Gail Morthole

Subject: thanks

Professor Wang,

I have been enjoying your site very much. Not having a wide math background, (only college statistics), I am wondering if your analysis is ultimately limited by the "real world" problems of effective voting methods which surely played a part in Florida last time around. The October Scientific American has a good explanation of the multiple reasons there was a huge discrepancy between pre-election polls, exit polls and actual results. If the same thing happens in Fl. again (and why couldn't it?) then in a race this close, your results can only go so far. I agree that turnout is just about everything at this point.


Gail Morthole

San Francisco

Date: Wed, 27 Oct 2004 11:46:05 -0400

From: Dana C. Hutchins

To: sswang@Princeton.EDU

Professor Wang,

I want to thank you for your efforts with your state poll analysis site. I've found it the most interesting and educational in terms of trying to predict the election.

I'm a "registered Independent" in Bucks County, PA and have had no less than 5 calls/visits by Kerry supporters, either from the Democratic Party or ActUp [she corrects this to MoveOn - SW]. No contact whatsoever from the Republicans other than direct mail. This mobilization effort probably isn't reflected in poll data, and I expect that Pennsylvania will go for Kerry by a wider margin then currently indicated. One would have to suspect that similar "get out the vote" efforts by the Democrats in other states are underestimated as well.


Dana C. Hutchins

Date: Wed, 27 Oct 2004 18:56:38 +0000 (GMT)

From: Randall A. Blake

Subject: Election Website & question

Prof., the election website is a great website. Election polling interest will die after the election, but this site should not. Please develop a subject for post-election discussion so that you will keep it up. Though past my school years, I am very interested in (and teaching myself) statistics and I find this site to be a great resource (even if a bit hard for a new comer to statistics like me to understand).

On the "undecideds": I believe that many people who claim to be undecided in presidential polls are not, in fact, undecided. You hint at this sort of analysis in your discussion of party identification. My question: Is there a way to accurately quantify the proportion of persons who at any given time (in any given poll) claim to be undecided but are not, in fact, undecided?

Keep up the great work. Your site is a wonderful learning tool.

Thanks, RAB

Date: Wed, 27 Oct 2004 14:12:19 -0700

From: Marc Benjamin

Subject: Great site


One quick note about the Hawaii polls.

I was looking at newspaper articles about them on Sunday. The Advertiser revealed the people it called were 31% Democrats, 42% Republicans and 24% (I believe) independents. I don't know if that mirrors the state's political backdrop, but I doubt it.

The other Honolulu paper, The Bulletin, did not reveal the party affiliations of those called for its survey. Both newspapers are operated under a newspaper joint operating agreement and function out of offices, I believe, within the same building with separate editorial employees only.

The competitive pressures of one newspaper having a poll and the other one not,no matter how inaccurate the polling was, likely was too great to pass up. My guess is both surveys are inaccurate, but I have only seen two other polls, none recently, that showed Kerry up 7 and 10 points.

For what it's worth.

Your site is outstanding and thank you for your diligent efforts.


Marc Benjamin

Date: Wed, 27 Oct 2004 16:22:25 -0700 (PDT)

From: Keith Pickering

Subject: Electoral tie vote

Hi Prof. Wang,

Love your site, I'm watching it daily. You are incorrect, however, in speculating that in the event of an electoral tie that the VP would be Edwards if the Senate switches to Democratic. Undoubtedly the Dems would vote in Kerry as VP. (Just a heartbeat away ...)

Keith Pickering

Date: Wed, 27 Oct 2004 13:49:35 -1000

From: Jeff McNeill

Subject: site comments

Aloha Professor,

Thank you for your kind efforts in creating and maintaining your election poll meta-analysis site. I wanted to offer a pithy quote to some of your other letter writers:

Belief is no substitute for arithmetic. -Henry Spencer


Jeff McNeill

Date: Thu, 28 Oct 2004 09:03:45 -0500

From: Marta Caminero-Santangelo

Prof. Wang,

I am an English prof at the University of Kansas, and I^“ve been avidly following your site.† I^“ve read the entire site, including your explanation of why the probability map and the ^”median outcome^‘ sometimes differ, but I have to confess I still don^“t understand this point.† (I will admit I^“m mathematically challenged.)† Could you please explain it again, in very simple layman^“s terms?† What is ^”median^‘ measuring, and what is ^”probability^‘ measuring, and why are they different?

Thanks for all the good work.

Marta Caminero-Santangelo,

SW: This is quite a good question, one that I need to answer carefully. I will do this soon.