To sort out the data now coming in from the various media recounts of the presidential vote in Florida, you have to take care in framing your inquiry. The question "Who really won Florida" is much too vague. By manipulating the data and indulging various contrary-to-fact scenarios, it's possible to get to any answer you want. But there are some more specific questions that help to hone in on the big issue, namely George W. Bush's legitimacy as president. In ascending order of importance, my questions are:
1) Who did more people in Florida attempt to vote for?
2) Who would have won if the U.S. Supreme Court had not stopped the recount?
3) Who would have won if every legally valid vote had been counted?
We've known the answer to question No. 1 for some time. More people went to the polls intending to vote for Al Gore. Stephen Doig, a specialist in computer-assisted research, demonstrated this in a Miami Herald story published in early December (and unfortunately no longer archived on the paper's Web site). Based on a precinct-by-precinct analysis of 185,000 uncounted Florida votes, Al Gore would have won the state by 23,468 votes if every voter had succeeded in voting the way he or she intended to vote. Put another way, there is solid statistical evidence that more people in Florida left the voting booth thinking they had voted for Gore than left thinking they voted for Bush. But to my mind, that interesting conclusion has only a minor impact on Bush's legitimacy--no more than does the fact that Gore won the popular vote nationally. If a larger number of Floridians didn't cast legally valid votes for Gore, Bush still won the election under the rules of the game.
The Miami Herald's big recount story published yesterday answers question No. 2. If the Supreme Court hadn't ended the recount, George W. Bush probably would have won. To understand why the Herald story successfully answers this question, we must venture back briefly into voting-counting arcana. My colleague Mickey Kaus has faulted the Herald for examining only "undervotes"--those ballots registering no apparent choice for president--while ignoring the larger category of "overvotes"--those registering two or more apparent choices. As we'll see, the overvotes are highly relevant in answering question No. 3. But the undervotes were the only ballots being recounted when the U.S. Supreme Court intervened. At that point, the whole issue was clouded by a widespread, erroneous assumption that the overvotes didn't matter. They weren't going to be included in the final tally in any case. So the Herald tells us everything we need to know to answer question No. 2.
What it tells us is that under every plausible alternative recount scenario, Bush would have won the election anyhow. Say the U.S. Supreme Court held, on equal protection grounds, that there had to be a common statewide standard for the recount. Under almost any standard chosen, the result would have been the same. Under a "liberal" standard (dimples OK), Bush would have won by 1,665 votes. Under a "conservative" standard (penetration necessary), he would have won by 363 votes. Had the U.S. Supreme Court not insisted on a universal standard, Bush would probably have won by some figure between those two. Only under a super strict "clear punch" standard would Gore have won by a negligible factor of three votes. But the clear punch standard doesn't accord with the "voter intent" requirement of Florida law. In short, by attempting to stop the counting of valid votes, Bush was trying to steal what it turns out he would have won anyway. So shame, but no diminished legitimacy on him.
Forgive me if I digress for a short victory lap. I arrived at a similar conclusion--that any real-world recount was likely to make Bush the winner--four months sooner and $500,000 cheaper, using only highly amateurish statistical extrapolations. I pointed out at the time that the liberal, dimpled chad standard that Gore's lawyers were demanding would actually help Bush. That turns out to have been correct.
Back to business. It's question No. 3--who would have won if every valid vote had been counted--that I rate as the most important. It's a less realistic hypothetical than No. 2 since given the role of Katherine Harris, the composition of the U.S. Supreme Court, and the fact that no one except Mickey Kaus understood the importance of the overvotes, there was never much chance of a fair and complete recount in Florida. Neither Gore nor Bush wanted one. Furthermore, according to the Herald, election judges in Palm Beach and Broward counties miscounted according to their own standards--mistakes unlikely to be corrected in any real-world recount.
Even so, question No. 3 seems to me to have the greatest impact on Bush's legitimacy as president. If we find out that Al Gore got more legally valid votes, it's going to be hard to look at the Republican in the White House the same way, even if we know that some of those votes stood no chance of getting counted.
So did Gore get more? The Herald's conclusion is that Gore would have won by either 393 or 299 votes if every county in Florida had recounted accurately using a common standard. But that conclusion is based only on an examination undervotes. It's going to be a few more weeks until the newspaper consortium that includes the New York Times and Washington Post finishes its work of examining the 110,000 overvotes as well. Any posthumous moral victories are going to have to wait until then.