A Proud Member of the Reality-Based Community

Acknowledgements

SEARCH eRiposte!

For the eRiposte Election 2004 home page, click here. 
For a database on vote fraud/suppression in Election 2004, click here.

It's Not the Machines:  
Where the Voters Live in Florida

by Susan Goya

I found a map of Florida divided into counties on www.floridacountiesmap.com.  I used the data from ustogether.org/Florida_Election.htm.  I outlined the county in its appropriate color if the presidential vote tally followed the voter registration.   This means if a county is outlined in red, a plurality of voters registered as Republicans and the vote tally showed a Republican majority.  Likewise  a county outlined in blue means that a plurality of voters registered as Democrats and the vote tally showed a Democratic majority.   

Circles indicate counties with results contrary to the expectation that the vote follows the registration.  A blue circle means that more voters, according to registration figures, would have been expected to vote for the Democratic candidate than the Republican candidate, but did not.  A red circle (only one, Monroe County) means that more voters would have been expected to vote for the Republican candidate but did not.  

When I had finished coloring the map, I found three well-defined regions of contiguous problem counties.  The cohesiveness of the two anomalous regions in the panhandle and the one region in the peninsula suggests a central tabulator problem or other localized human interference.  Indeed, Bev Harris (blackboxvoting.org) has interviewed experts who say it is much easier to compromise the machines in the smallest counties.  Op-scan counties where voting followed registration (mostly Republican) did not have the astounding statistical anomalies, lending credibility to the idea that counties that did not need to be “fixed” were left alone.  The discrepant Democratic region in the middle of the panhandle suggests that the ability to make such adjustments depended on access to a central tabulator.

In the panhandle, the problem area is  two regions of small contiguous counties separated by a three-county region that registered and voted Democrat. The problem area is bounded on the west by a cohesive region of five counties that registered and voted Republican, and bordered on the east by an odd “land-locked” Democratic op-scan county (Alachua, #28) which did not cross-over and seems to mark a boundary between “fair” Republican op-scan counties and the so-called “Dixiecrats”.  Notice too that “fair” Republican regions are similarly cohesive. It is possible that each region of cohesion reflects election administration areas, each with its own central tabulator.  All of this undermines the Dixiecrat hypothesis, which, of course, remains to be tested.  In the absence of other data, there seems to be no reason to conclude that panhandle Democrats are really secret Republicans.

Nassau County (#30), a Republican e-touch (e-touch counties are shaded gray) county, is interesting. Its percent change figures are similar to “fair” op-scan Republican counties just south of it, suggesting that un-tampered-with machines of either type recorded similar “fair” results.   Duval County (#31), just below Nassau County could possibly be a genuine cross-over county with a Democratic registration of  46.2%, far less than many of the cross-over counties.  Percent change figures for Duval are also reasonable.

And what do we make of Gadsden County(#11) with a Democratic registration of 82.9% and a significant percent change toward Republicans of  165.7%?  Several other counties crossed-over with a smaller percent change, but Democrats held the lead in Gadsden anyway. Similar story in Jefferson County (#16) but with less magnitude.  Are the voters overly influenced by their proximity to that Democratic bastion, Tallahassee? Or more likely,  are Florida voters who are registered as Democrats really Democrats, not Dixiecrats?  After all, in Leon County (#13), containing Tallahassee, bastion or not,  Republican registration is 26.6%, and Democratic registration is 57.1%, significantly fewer Democrats than in many of the surrounding counties.  With such comparably low Democratic registration, effecting a cross-over vote in Leon County should have been a piece of cake.  But maybe we are talking about a different, and honest, central tabulator here.

As for the peninsula, we see another region of contiguous problematic Democratic counties with a “fair” Republican county right smack in the middle.  On the western border is e-touch Democratic Hillsborough County that may have genuinely crossed-over to the Republicans as did the two op-scan counties (Polk and Osceola) next to it.  All three have Democratic registrations in the low 40%'s.  The percent change figures for these three counties, whether e-touch or op-scan, are comparable to other e-touch counties.  What we see is that counties that may have genuinely crossed-over had a relatively low Democratic plurality in the first place, casting more doubt on the Dixiecrat idea that says overwhelmingly Democratic counties voted overwhelmingly Republican.   We need historical data and the other varied measures.

Simply by eyeballing, it was obvious that many of the problem counties were small and in the panhandle.   If we remove the three Democratic counties in the peninsula which have the likely potential to cross over, we are left with five problem counties in the peninsula and twenty-one problem counties in or  next to the panhandle.  Because the problems are so pervasive in the panhandle, perhaps location is a more salient variable than size, warranting separate analysis of the panhandle and the peninsula.  As Kathy Dopp has noted in her compendium of data and reports of  the Florida anomalies, (www.ustogether.org), mathematical analysis (and other analytical perspectives) often reveal what eyeballing cannot. 

One positive outcome of this mess is the wide acknowledgment that American democracy would have been better served by open source software in machines managed by other than private corporations.  Yes, there were machine problems, and plenty of them.  However, the wild numbers revealed by the Florida data imply direct human manipulation, like hacking.  Just last week the FBI arrested a teenager for hacking a corporation computer.  It seems to me the FBI should be aggressively investigating the possibility here.

Susan Goya is an education writer who became alarmed at the shallow “he said, she said” media campaign reporting and the scarcity of nonpartisan analysis.