New statistical analysis?

[Angry Girl]

You may already have this data. This is from a fellow by the name of Robert Millman. I don't know if this is new stuff or not but it looks rather interesting, leading him to pose the following questions:

"Why did Democratic votes do better than registration only in
counties where Democrats lost vote share?"

"Is it logical that Democratic vote share under-performed in
every county where they increased registration and only
surpassed registration shifts where Republicans increased in
registration?"

Data and methodology given:

http://liberty.hypermart.net/voices/2004/text/Comparison_of_Democratic_and_Republican_Populations_and_Vote_Counts_in_Florida_from_2000_to_2004.htm

Cheers!

[berniew1]

1. I've looked at it but can't understand it; more understandable version at:

www.flcv.com/fla04EAS.html

concludes very strange pattern of unexplained Bush increases in big touchscreen counties in Florida. Based on 2000 and 2004 data by county and new registrations by party, etc.

[Hobbes199]

3. It's all those Dixiecrat wannabees.

They register as Democrats, then vote Republican. It's all the rage in the Ukraine.

[EMunster]

2. huh?

[Eloriel]

4. Kick -- certainly deserves more attention

:kick:

[Melissa G]

6. kick!

[ClintCooper2003]

5. This is hard to follow.

[MarkusQ]

7. As simple as I can make it

<summary>
The percentage of people who voted for Kerry over Bush in 2004 was greater than the percentage of people who voted for Gore over Bush in 2000 in some counties, but only in those where the majority of new voters registered Republican.

You would have expected this in the counties that saw an increase in Democratic registration, but there Bush did better.

It seems odd.
</summary>

--MarkusQ