Now, quite a bit of work is being done around these parts regarding Ohio numbers, and I'm thankful for that. But, hey, let's take a look at Florida numbers, just for fun.
The Florida Department of State has posted its first set of unofficial returns from every county. Every one has included counts of absentee ballots, and every one is marked as a final report, according to the data on the site. Now, I'm not sure about that, but let's accept it for now.
Turnout is reported for every county except Hernando (which inexplicably reports 0 turnout). Now, let's assume that every registered voter in Hernando county (all 109,656 of them) voted. That gives us total voter turnout of 7,460,556. Let's compare that with the reported total numbers of votes for Senate, President, and Amendment 5, the amendment receiving the most total votes of the 8 offered (number of votes in excess of maximum turnout in parentheses):
for Senate: 7,409,542 (-51,014)
for Amendment 5: 7,277,832 (-182,724)
for President: 7,586,995 (126,439)
Something is wrong here. A bit more below the fold...
Now, this turnout estimate is high. Let's assume that turnout in Hernando County is actually at the same rate as elsewhere in the state (excluding Hernando County), 72.1%. This gives us 79,062 votes from Hernando, and a total state turnout of
7,429,962. So, we could estimate that 20,420 people didn't vote for any of the seven Senate candidates, 0.27% of those who turned out. But we also then estimate that the number of votes for President exceeds turnout by
157,033.
Granted, the number of votes for Bush exceeds the number of votes for Kerry by 381,210, more than double this margin. But regardless of that, what is going on here? Are provisional votes to blame? If so, why aren't they reflected in votes for the equally hard-fought Senate race? The state is claiming that the turnout counts include absentee ballots--perhaps this isn't really true?
I would love to see a reasonable explanation, but I can't come up with anything.
UPDATE:
I have broken down the data by county. Several observations:
- The Florida data aren't even internally consistent--vote totals from the "First Set of Unofficial General Election Returns" (.pdf file) don't exactly match the vote totals elsewhere on the site, though they're close. Above and elsewhere, I used vote totals from the .pdf, and turnout figures from the website, since turnout isn't in the .pdf.
- There are 12 counties where currently recorded votes for Presidential candidates exceed reported turnout (not including Hernando county, which didn't provide a nonzero turnout figure), by a total of 192,397 votes. I'll provide a chart of these counties here in just a moment.
ALSO: many are confused by what I'm pointing out here. It's relatively unobjectionable for there to be significant numbers of voters who vote for some items on the ballot and not others (though some states make this as confusing as possible; e.g., my home state of NC counts votes for President separate from a straight-ticket vote, meaning voters must vote for both separately). The problem here is that
Florida is reporting more votes for president than voters. Other ballot items were provided for reference.
My analysis shows that 12 counties have reported votes for Presidential candidates in excess of reported turnout. Six have discrepancies of less than 1% of turnout (though still more than 0, obviously, and worrisome), and I'll focus for now on the other six:
- Glades: 2443 Bush/1718 Kerry/27 Other; Turnout 3446; 742 Discrepancy = 21.53% over turnout
- Highlands: 25874 Bush/15346 Kerry/271 Other; Turnout 33996; 7495 Discrepancy = 22.05% over turnout
- Miami-Dade: 358613 Bush/406099 Kerry/3841 Other; Turnout 716574; 51979 Discrepancy = 7.25% over turnout
- Osceola: 43108 Bush/38617 Kerry/453 Other; Turnout 63589; 18589 Discrepancy = 29.23% over turnout
- Palm Beach: 211894 Bush/327698 Kerry/3243 Other; 452061 Turnout; 90774 Discrepancy = 20.08% over turnout
- Volusia: 111544 Bush/115319 Kerry/1495 Other; 209052 Turnout; 19306 Discrepancy = 9.24% over turnout
FURTHER UPDATE:
Anyone interested can now take a peek at my spreadsheet. It's certainly not pretty, but it has all the numbers referenced herein.
YET ANOTHER UPDATE:
The other six counties are:
- Collier: 83485 Bush/43713 Kerry/1154 Other; 127409 Turnout; 943 Discrepancy = 0.74% over turnout
- Lake: 74382 Bush/48216 Kerry/1340 Other; 123751 Turnout; 187 Discrepancy = 0.15% over turnout
- Leon: 51594 Bush/83830 Kerry/890 Other; 136229 Turnout; 85 Discrepancy = 0.06% over turnout
- Okaloosa: 69654 Bush/19358 Kerry/695 Other; 89485 Turnout; 222 Discrepancy = 0.25% over turnout
- Orange: 192390 Bush/193217 Kerry/2145 Other; 386104 Turnout; 1648 Discrepancy = 0.43% over turnout.
Additionally, as fellow dKos user
DEA points out, Escambia county looks very strange, too. Turnout exceeded reported votes for president by 18193, 11.29% of turnout, making it a strong outlier.