The hyperintelligent Jay Cost at RearClearPolitics has produced for us a most helpful spreadsheet computing the various popular vote scenarios.
But this paragraph, is, to me, a very crucial point that both Clinton and Obama campaigns would rather ignore:
We have a large number of unknown factors. For many of them, we have very little idea what values they will ultimately take. What we do know is that small changes in several of them could induce large changes in the vote count. This makes it extremely difficult to be as precise as many commentators have been. We need to be wary of all the uncertainty we face here.
So -- my fairly conservative calculation has Clinton netting about 446,000 votes between now and June 3. Under all scenarios that exclude Florida and Michigan votes -- and count the votes of Washington's primary -- Obama still retains a popular vote lead of not more than 330,000 -- or an advantage of less than one and a half percent.
Under a scenario that includes the Florida and Michigan votes for Clinton, gives Obama all of the uncommitted Michigan votes, estimates the votes for all the caucus states and includes the Washington primary, Clinton wins by about 16,000 votes -- or about a tenth of one percent.
Which scenario is "right?" Under DNC rules, until the credentials committee figures out which delegations to seat, Florida and Michigan do not exist. But the voters in those states certainly do in the existential sense -- and if we're answering the question by figuring out how many Democrats voted for Obama versus how many Democrats voted for Clinton.
Obama supporters will anchor their estimates in the worldview most hospitable to Obama's nomination, and Clinton's supporters will similarly find ways to justify including Florida and Michigan before it is DNC-legal to do so.
The media may be called upon to take a stand -- especially since the superdelegates tend to listen to the media more than other entities -- and the most reasonable answer may well be -- well, it depends on who you talk to.
Are there historical precedents? Well, Democrats like to count every vote. So -- advantage Hillary? But there has to be some tempering factor to account for Obama's name not being on the Michigan ballot. Ok, but then there has to be some tempering factor to account for the fact that Obama's campaign made the decision to stay off the Michigan ballot as least as much because they feared losing the state to Clinton as they wanted to make a statement to Iowans about the integrity of the calendar process. Obama's campaign also made the strategic decision to contest caucuses; the Clinton campaign dumbly decided to avoid them. If they had spent a comparable amount of money and resources in the caucus states, Obama's margin of victory would have been lower and he certainly would have less of a delegate lead.
These are all arguments... all persuasive in their own way... and they don't get me any clearer towards the answering the question about which votes to count and which votes to ignore.