I saw this at The Command Post and thought it read like a "Please believe me" press release. So I took some time to get the numbers from the Florida DOS website. It's nice and they have loads of info for both 2000 and 2004, but don't go to the PBC handcount, it's a huge table and will take 3 or 4 minutes to load (that an eternity, now.)

Anyway here some of what I picked up:

Broward:
2004 RVoters - 1,032,098; BCast - 709,724; Turnout - 68.77%
Vote: Bush - 244,674; 34.61%
Kerry - 453,873, 64.21%
Other - 8,200 +/-

2000 RVoters - 887,764; BCast - 587,928; Turnout - 66.23%
Vote: Bush - 177,279; 30.92%
Kerry - 386,518; 67.47%
Other - 9,510 +/-

Now here is the third bullet from the Berkeley Study summary:

"In Broward County alone, President Bush appears to have received approximately 72,000 excess votes."

So, what the Berkeley study is asserting (suggesting?) is that despite:
- A statewide distribution change from 48.8B/48.8G in 2000 to 52.1B/47.1K in 2004, and
- An increase in registered voters in Broward by 144,000+ in the intervening years
that Bush should have gotten less votes (in absolute terms) Nov 2nd than he did in 2000. The bullet notes 72,000 while the increase in the Bush tally from 2000 is 67,000. Now it is possible that that these folks are just in a hurry to publish for some reason (like, there hasn't been a vote fraud controversy yet that the Dems can sticky-post a talking head lawyer to yet) and that is why they have a summary page describing Broward as having registered a possible 72,000 excess vote for Bush while the press release uses the number 28,000.

What was it SDB said a while back ... something about going back to your desk and looking at your work again.

(Note, I did not note the increase in ballots cast from 2000 in Broward, which was 121,000 + in respect for the possibility that their study argues there should not be an increase and I have not done calcs yet because I have not found statewide RV & BC numbers. Even so, common sense would indicate that there is no possible way in hell that Bush would have gotten fewer total votes Broward County in 2004 than he did in 2000 with an increase in registration twinned with a statewide increase in % of over 3%.)

Update: Wizbang has done some analysis here.