Thursday, July 29, 2004

previous entry | main | next entry | TrackBack (0)


Tyler Cowen gives me an assignment

Over at Marginal Revolution, Tyler Cowen makes a request:

Daniel Drezner remains on the fence, concerning the next Presidential election.

He writes about supporting Bush, Kerry, or perhaps a third party candidate (unlikely). But why should he restrict himself to "pure strategies"? Why can't he support some candidate with some positive probability? How about, for instance, "I support Bush with p = 0.63." Or "I support Kerry with p = 0.57", and so on. That way we would know how strong (or weak) his current view is.

Chris Lawrence's doubts aside, this seems fair to both me and my readers. I'll be posting my first p-value after Kerry's speech tonight. Obviously, this value will likely fluctuate over the next few months.

One thing the probability that I will vote for someone either than Kerry or Bush is zero.

posted by Dan on 07.29.04 at 09:50 PM




Comments:

Dan,

Not sure why you agonize so much about the intricate details of the two candidates' policy proposals, but then say you'll make up your mind based on the acceptance speech. Acceptance speeches are publicity events, designed to appeal to undecided voters as well as rally the faithful, but are long on symbolism and short on specifics.

by the way, don't you think that Oregon Vet is on the fast track to a cabinet post if Kerry is elected? what a great speech.

posted by: another poli scientist on 07.29.04 at 09:50 PM [permalink]



"the probability that I will vote for someone either than Kerry or Bush is zero."

I really doubt that. If one of them were to fall victim to an injury or something more sinister, and be replaced, would you in no circumstances vote for his replacement?

posted by: Bob Dobalina on 07.29.04 at 09:50 PM [permalink]



Dan, not to be pedantic, but I don't think you mean "p-value" as you wrote in your post.

(I think you know this, but), Tyler's asking for something different: a binomial distribution over the two candidates.

But a 'p-value' is something different, and oft-misused...

posted by: pedant on 07.29.04 at 09:50 PM [permalink]



Reminds of the various Slate meters that used to pop up from time to time: prob. of war in Iraq, prob. of Martha being convicted...

maybe you need a chart :)

posted by: Suresh on 07.29.04 at 09:50 PM [permalink]



Dan, you are to outsourcing as Sullivan is to gay marriage: obsessed. You're both right, of course, but fixated. The difference is that Sullivan has some justification; Bush is atrocious on that issue. If only Bush's sops to his hellfire right were as light-weight as Kerry's are to his protectionist left. Could Kerry be any more restrained on protection and still lead the party? I was delighted that Rubin sat next to Mrs. Kerry. Not much chance we'll see Laura Bush hanging with the Fab 5.

posted by: Jon on 07.29.04 at 09:50 PM [permalink]



As pedant points out, you risk some reader confusion if you call your probability of voting for Kerry a p-value.

You've also found a number of issues where you have some reasonable hope that Kerry will be better than Bush. What are the issues for which you're certain that Bush will be better than Kerry?

AB

posted by: Angry Bear on 07.29.04 at 09:50 PM [permalink]






Post a Comment:

Name:


Email Address:


URL:




Comments:


Remember your info?