The "Dixiecrat" theory does not explain the big Republican vote gains in Florida counties where Democrats had very successful voter registration drives.
In my post What's Wrong With Liberty, Florida? I showed some of the incredible variations in Bush vote counts compared to the number of registered Republicans in certain Florida counties.
For example, Liberty county with 320 Republicans gave Bush 1,927 votes. That is 6 times more votes for Bush than there are registered Republicans (602% to be exact).
One possible explanation for the variations is that these counties are flush with "Dixiecrat" Southern Democrats who haven't changed their official party affiliation in years, and who now vote solidly Republican. In fact many of these counties voted for Bush in 2000 too, though not in the same quantities. I don't think the Dixiecrat theory fully explains Bush's incredible Florida turnout, but clearly there are many registered Democrats voting Republican in rural Florida.
The Dixiecrat theory does not apply to newly-registered voters though. You would expect that someone who registered as a Democrat in the last four years has a high propensity to vote for Kerry in 2004.
When we look at the votes in counties with lots of newly-registered Democrats we see more indications of Bush vote inflation that the Dixiecrat theory can't fully explain.
Take Democratic-leaning Broward County, for example, which Gore won in 2000 by 67% to 31%. Democratic GOTV efforts were very successful in Broward. 77,187 new voters registered as Democrats since 2000, compared to only 16,907 new Republicans. Here are the details:
2000 Registered REP=266,829 DEM=456,789
2004 Registered REP=283,736 DEM=533,976
Reg Increase REP=16,907 DEM=77,187
But the vote tallies in 2004 tell a different story:
2000 Votes Bush=177,902 Gore=387,703
2004 Votes Bush=243,118 Kerry=451,521
Vote Increase REP=65,216 DEM=63,818
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 than the Republicans. This seems highly implausible. Broward of course uses new touchscreen voting machines that don't provide a paper trail for verification.
Other counties show similar anomalies. Here's an example from Pinellas county:
2000 Registered REP=231,652 DEM=223,544
2004 Registered REP=231,652 DEM=223,544
Reg Increase REP=-10,425 DEM=9,383
2000 Votes Bush=184,825 Gore=200,630
2004 Votes Bush=225,627 Kerry=225,367
Vote Increase REP=40,802 DEM=24,737
So Pinellas county lost 10,000 Republicans since 2000 and gained over 9,000 Democrats, for a net gain of 19,808 Democrats.
Yet Republicans gained over 40,000 votes in Pinellas in 2004 even though the Democrats were far more successful registering new voters there. Did I mention that Pinellas switched to no-paper-trail electronic voting machines this year too?
When you look at these numbers in light of exit polls showing a preference for Kerry, other polls showing that most citizens do not favor Bush's stances on key issues, Bush's great victory in Florida seems increasingly improbable.
Statistical anomalies like these do not prove that vote fraud was committed in Florida. But if enough anomalies and improbabilities are discovered and publicized, it should become easier to convice Democrats and election authorities that the Florida results need to be officially reviewed in detail.
More to come...