There are three from the West, three from the Soutwest, two from the Midwest, four from the Northeast and one from the South.

At first glance, that would appear to be apportioned poorly.

However, my apportionment is better because it is based on style of voter.

We have:

4 beat team/regional beat writers or columnists, which provides a more localized angle, obviously (Greenstein and Hamilton of the Chicago Trib, Carlson of The Oklahoman, Bianchi of the Orlando Sentinel)

5 national beat writers, which provides a bigger, though expert, picture
(Mandel, Feldman, Dienhart, Buchanon, myself)

1 national feature writer, promotes an even broader picture
(Murphy of Sports Illustrated)

2 magazine editors, which helps us represent those voters who are less involved in the day to day of the race.
(Schecter of SI, Morris of ESPN)

1 somewhat older voter who covers both basketball and football and sort of represents the old school, traditional voter
(Weiss of the NY Daily News)

So, my voters are distributed more by how in-tune they are with the race than by region. I feel it gives us a better representative sample. My own HP Heisman Watch tries to adjust the sample for the regions, but it obviously doesn’t always work out.

I do think it would be better to grab a couple more votes to round it out a bit better, but I think too big a sample just might piss off the Heisman Trust.

It seems like a weekly poll that had an even distribution, say 3 voters from each of the 6 regions, would be–relatively speaking–fairly representative of the voting population at large.