[Torie]

Interesting analysis, but older elections are not relevant in the media age (what poll showed FDR ahead in 1936?), and Veeps are not incumbents, and the dynamic is totally different, and Dales is relying on just one poll apparently, and the last poll, and right now, what we are seeing is not the last polls, but polls two weeks out. Dole moved up smartly from what the polls showed in 1996, and so did Reagan in 1984, from where we are now in any event. I stand by my post. I think Bush should sweat because of the probably larger turnout. He should sweat a lot, with his current numbers.

[HawaiianGecko]

	Personally, I don't think there "really" are undecided voters at this point in time this year. People have had four years to make up their mind about President Bush and a year and one half to do the same with Genghis Khan.  It is my true belief that anyone still undecided is entirely too nuanced, picky or stupid to find a polling precinct on November 2nd, therefore, changing the heading of that proverbial third column heading from undecided to MENTAL-VOTER-ONLY!

[Dales]

", and Dales is relying on just one poll apparently, and the last poll, and right now, what we are seeing is not the last polls, but polls two weeks out"

Nope. As a matter of fact, I ignored the last poll (due to the fact that Gallup tries to allocate undecideds in their last one). I looked at it from both a month out and the second to last poll for that writeup, and again later (not written) for other timeframes such as two weeks.

What I found is that the undecideds at any of those periods either 1) break for the incumbent in a substantial way, 2) break evenly, or 3) break ever so slightly towards the challenger-- and the times that #3 was the case happened to be in races that were already blowouts.

[AHerald]

Interesting analysis, but older elections are not relevant in the media age

A seemingly compelling argument at first blush. But I would argue that these "older elections" actually point to a remarkable relevancy when one considers that they show a consistent trend in spite of the fact that they spanned an era of history that featured other very dramatic developments related to mass media (from radio to tv, for example) and the public's access to news. Not to mention wars and any number of other incredible demographic, cultural, and techologica changes that took place during this same time.

In spite of all the incredible changes in America between 1936 and the year 2000, these polls show a consistent trend regarding the undecideds breaking toward the incumbent. As different as the world today--the media age--seems to be, I doubt it's so different as to be removed from the historic trends that span over 6 decades.

We'll find out come Nov. 3. But my money is on the historical trend.