Rantworld Offers

Blog powered by Typepad

« Demand a Recount in Ohio! | Main | Rove Feared Dean More than Kerry »




I replied briefly at Washington Monthly about your concerns. I stated that looking at party affiliation really isn't a good indicator of how someone will vote--especially in Florida. Many of these voters probably associate themselves as Democrats in the same way Zell Miller does. In other words, their not really Democrats. Also, I checked 2000 election results for a few of these counties in question and the numbers are pretty similar, with an increase for Bush in 2004. What really surprises me, however--and it may be worth looking into--is the numerical increases for Bush votes in 2000 to 2004 vs. Gore votes and Kerry votes. What I'm trying to say is that in 2000, there were 5610 votes for Bush in Baker Co. This year there were 7738--a net gain of 2128. In 2000 there were 2392 votes cast for Gore; in 2004 there were 2180 votes for Kerry--a net loss of 212. That's a 2340 vote swing to Bush in a county with only 9955 votes. Calhoun showed a gain to GOP of 907 and a DEM loss of 39. Jackson Co. shows a GOP gain of 2977 to a DEM gain of only 684. Liberty Co. GOP gain 610, DEM gain 57.
Did the GOP improve their turnout from 2000? Yes. Did rural areas vote for GOP over DEM? Yes. These are true, but the gains seem significant--suspiciously significant. But then again, I'm a conspiracy theorist at heart.
The numbers for Alachua county favor Dems slightly, so maybe the GOP simply shook the trees harder in the rural counties and more Repuglicans came out to vote. I'm still working on the numbers (from the Florida election board), and it is Friday night, so I do have better things to do.
I don't know if there's anything to it or not. Keep looking into it!



Sorry to keep pushing the subject--- but I'm really surprised by the numbers comparing 2000 voters vs. 2004.
So far I have calculated the following:
County/GOP vote diff. '04-'00/DEM vote diff. '04-'00/overall GOP vote gain(loss)
Citrus/13973/-493/14466!!! (NOTE: only 69,462 votes cast in Citrus Co.)

I will say that I developed an extensive spreadsheet for my home state of Wisconsin to help me make a strong prediction for this election (which, BTW, I predicted correctly w/in 1%). I easily calcuated the same GOP gain/loss in WI and the largest GOP gain was 12,500 votes in a heavy republican county that cast 230,000 votes. The counties with the biggest votes gains favored DEMS.
Again, I don't know if there's anything to this, but I'm increasingly concerned. Tell me I'm calculating something wrong or don't know what I'm talking about.

Rob Dixon


Thanks for the additional analysis. I agree that many of the Democrats counties must vote Republican at election time. Someone else on the Wash. Monthly blog suggested that these "old school" Southern Democrats might never have changed their official party affiliation, which sounds plausible. Even so, it seems highly unlikely that so many Democrats would vote against their party affiliation, especially in counties like Jackson that have a high percentage of African-Americans.

You're onto something with the large increase in votes from 2000 to now. I've been looking along the same lines. In fact there was an even larger percentage increase from 1996 to 2000. Of course 2000 was the first election that Jeb Bush and Katharine Harris were in charge of the counting.

For example, Baker county went from 3,658 votes for Dole in 1996 to 5,610 votes for Bush in 2000. Then, as you pointed out, to 7,738 for Bush in 2004. The vote for the Republicans more than doubled in that time.

The stats in Calhoun country are just as interesting. It went from 1,717 for Dole in 1996 to 3,780 for Bush in 2004. In that time they only registered about 500 new Republicans in that county, so new registrations don't make up for it.

All these numbers need to be adjusted for things like the Perot factor in 1996 and new registrations in each county since then. But like you I think the magnitude of the gains for Republicans in this time frame is just not plausible.


I finished my quick analysis of the Florida returns. As much as it pains me to say, I have determined that there isn't anything grossly suspicious about the data. The reason Bush won Florida is because he got more of the vote out. The evangelicals answered the call from their savior, GWB.
My analysis of my Wisconsin data shows that the Dems got 16.5% more votes than 2000, and the GOP got 16.2% more. Gore barely won WI in 2000, and Kerry barely won it in '04.
In Florida, the GOP increased their votes by a whopping 26.4%, where the Dems increased only 19.5%
Yes there are counties where the GOP had a significantly larger numerical increase in votes between 2000 and 2004. 59 counties had more GOP votes than Dem (no surprise), and 17 had a 20% or more change.
There were 7 counties that actually had FEWER votes for Kerry than Gore received.
Now whether there is some reason Kerry didn't see a reasonable proportional increase in votes, I can't say from the data. Why, for example did he get 493 fewer votes in Citris Co. than Gore while Bush got 13,973 more than in 2000, the data can't reveal. Statistically, I think the Bush win looks legitimate.
It's certainly something worth looking into, however.
I think it's just a trending to the right politically that is guiding Florida's (and the U.S.) election outcome.
My Wisconsin data surprised me by showing huge gains in counties that were strong Republican, but never overwhelmingly so. Those counties turned out voters that may not have voted in years. If it weren't for the strong Kerry/anti-Bush feelings in Madison, Milwaukee, Stevens Point and Superior, Bush would have mopped up in Wisconsin.
I'm a political junkie and a geographer as well, so I tend to get wrapped up in this kind of stuff. Sorry to run so long with this comment.
Thanks for reading.

Rob Dixon


I get kind of wrapped up in this stuff as well. :)

I agree that the Florida stats don't necessarily indicate widespread tampering with the voting results. However the trending to the right has been so pervasive and extreme, even in Democratic-leaning counties, that something still seems fishy. I guess my goal in reviewing the Florida data is to find indicators of where to look first to find evidence of fraud.

Here's another stat I've been checking out: counties where the Democratic turnout was far less than expected in light of registrations trends.

Take Broward county, which added 77,187 new Democrats since 2000. Yet the Democrats lost 13,369 votes this year compared to 2000. There were only 16,907 new Republicans added to the rolls since 2000, but the Republicans gained 48,309 votes!

That would require a massive defection of Democrats and independents to Bush this year, which doesn't match the exit polls and doesn't seem likely. There's no way you can say these new Democrats are the old school Southern ones who prefer to vote Republican. Broward, of course, is a no-paper-trail e-voting county.

Anyway I'll keep combing through the numbers. Let me know if you find anything new too.

Thanks, Rob


I didn't find the same numbers you did for Broward Co.
I see:
2000 Bush=177,902 votes Gore=387,703
2004 Bush=243,699 votes Kerry=452,360
A net gain of 1,140 votes in favor of Bush. Maybe my numbers are wrong.
Part of why I was intrigued by the thought that there was foul play in the Florida election is because it's so believable. Bush Co. is so capable of it and they are masters of the craft. Also, part is wishful thinking that half the country didn't just re-elect the worst president in the history of our country.
But alas, they did.
If you look at the Florida counties that showed the greatest swing in votes from 2000 to 2004 in favor of Bush, they are all rural areas. They are the gulf coast counties and panhandle areas (see the counties around Calhoun and the counties around Levy). Also, any Census map will show that these are predominately white counties--so that tells me there wasn't any major black vote suppression going on. And even if these counties were somehow fraudulent, they wouldn't deliver Florida to Kerry, and we all know that nothing ever sticks to Bush anyway.
While the U.S. map seems to indicate a "north vs. south" election, it really was a "urban vs. rural" election. Look at the county-by-county map of the election and you will see that.
I think we need to divorce ourselves from the the party you're registered as and the party you vote for. People switch teams all the time. Many people don't know the difference between the parties anyway. Also, many may have registered Democrat so they could participate in the spring presidential primary or other primaries. I think Florida is a partisan primary state. I know in Wisconsin you don't have to register for any party and you can crossover to vote in primaries.
I enjoy the blog and will check in to see what you have to rant about from time to time!

Rob Dixon


I described the Broward county findings poorly I'm afraid.

Here's what I have:

2000 Bush=177,902 votes Gore=387,703
2004 Bush=243,118 votes Kerry=451,521

Very similar to yours (though they must still be counting...)

But look at the differences in new voter registration:

2000 REP=266,829 new reg DEM=456,789
2004 REP=283,736 new reg DEM=533,976

Increase REP=16,907 DEM=77,187

So Republicans added 16,907 new voters, and got 65,216 additional votes for Bush in 2004.

Democrats added 77,187 new voters and got only 63,818 new votes for Kerry.

Democrats lost ground even though their GOTV efforts for new registrations were nearly 4 times as successful! That seems somewhat implausible to me.

Anyway, thanks for your comments.


Mike Hicks

I've also focused on Liberty County for an entry on my site. The thing that strikes me but that a lot of others seem to have missed is that common wisdom usually tells us that rural counties tend to be more Republican than Democrat. This leads me to ponder whether the registration info is correct or not.

I thought that it might be a simple swap of the two party identities, but I tried looking deeper at the numbers and noticed that there are over 300 blacks registered as Democrats in Liberty County, so if registrations were somehow swapped, they would most likely be among the whites in the county (again, going by the common wisdom that blacks tend to vote 90% Democratic). This seems like a weird error to me. However, all of the documents for races in different years from 1994 onward seem to show the same pattern. I guess the tables have just been generated over the last two years or so, though.

Anyway, I had swapped the numbers of white voters in the county just to check on it, and things still didn't seem to add up all that well. You end up with the county then being about 75% Republican and 15% Democrat, with independents and other parties in between. For a county with that many Republicans, the final tally of 64% Bush to 35% Kerry would be a remarkably good showing for Democrats. According to a statewide exit poll, 7% of Republicans voted for Kerry, and 14% of Democrats voted for Bush (though I don't know if exit polls can be trusted at all these days). Maybe party IDs in the registrations got switched in a few subcategories of white voters, like an age group or something.

At any rate, there are serious errors in the numbers, but I'm not sure if they're in the vote tallies or in the registrations.

The comments to this entry are closed.