Florida 2004 General Election Results By County

The table below was created using data provided at Florida's Department of State Division of Elections. The supporting data  is publicly available at their official web site. 

An indicator has been added to the right. A ratio has been taken of the registered Democrats against the actual votes for the democratic presidential ticket. A similar ratio has been created for Republicans. The latter ratio is then divided by the first to create the indicator.

If one examines the numbers with a high ratio, such as Lafayette county, the distribution of the vote compared to the distribution of the registration look highly suspect.  Another stunning example is Holmes county. You can just go on down the list and make your own observations. Also, as the counties increase in size, the swing-over changes become less pronounced.

A next data table may include early exit poll data, and comparisons with other states.

Revised note: The evidence now shows that these are simply the strange voting patterns of Florida and have happened in the past. For example, most of the counties on the top of the list have LOST Democratic registrants and GAINED Republican ones in the last four years, meaning that the counties have grown more conservative in their party registrations. This adds value to the suggestion that the registrations simply lagged behind the sentiment of the vote - though these voters would vote Republican, they never bothered to change their official party affiliation. This observation is backed by the simple fact that all but two these counties already backed Bush in the 2000 election.

The conclusion is that there was probably NO SYSTEMIC voter fraud, though there is anecdotal evidence of isolated incidents of fraud. It also means that Democrats can not be accused of massive fraudulent registration efforts in those counties, since most of them evidenced a net loss of registered Democrats. 

Please click here to see a comparison of registration patterns and the general election from 2000 and 2004. The sheet is in the same order as the one below, with the same indicator. 

Another note: A group of doctoral students together with Michael Hout at UC Berkeley have analyzed the Florida numbers more rigorously, and have concluded that electronic voting machines may have given Bush around 130,000 additional votes. Their analysis can be found at UC Berkeley here at this link. Their conclusion is disturbing, however their numbers do not yet produce enough of a vote change to alter the outcome of the election.