We all know that delegates, not cumulative popular votes, elect the nominee.  And while it's also true that electors, not popular votes, elect the president in November, the primaries are different from state to state and held over almost half a year, so the many reasons for caring about the popular vote in the general election are lessened in the primaries.  Still, the idea that Clinton received more popular votes than Obama is a very compelling one to those of us who believe strongly in democracy, even in a primary contest.  I'm inclined to agree that it's absurd for the person who received less votes to win any election, and it certainly diminishes that person's legitimacy as winner.  So I don't think that "delegates elect" is a proper response to the claim by Clinton and her supporters that she won the popular vote.  And while Eric Kleefeld posted about the final math, including a paragraph saying that Obama won the popular vote unless you count Michigan and don't give Obama the uncommitted votes, I think this point needs to be emphasized more clearly to end any question about Obama's legitimacy as the Democratic nominee going into the general election.  The mainstream media seems to have settled on the idea that Obama won because he played by the rules and that Clinton's claim that she won the popular vote is at least plausible (on CNN last night Wolf Blitzer showed 3 different ways of counting the popular vote and concluded that in 2 of the 3 Clinton did indeed win, without comment as to which of those 3 was most accurate).  But it is not plausible.  Any reasonable tally of the popular vote, even in the light most favorable to Clinton, shows that Obama won.
    According to RealClearPolitics, Obama was the clear winner of the popular vote without counting Michigan.  Setting aside the fact that the Michigan primary was not a legitimate reflection of voter preferences, let's assume that the 328,309 people who voted for Hillary did intend to vote for her over all the other candidates.  Then she leads the cumulative popular vote by 176,465 (including the caucus estimates).  Let's round that up to 180,000 to allow for possible favoritism towards Obama in the caucus estimates.  There were 238,168 votes for uncommitted in Michigan.  On top of that, we learned at the Rules and Bylaws Committee hearing that there were approximately 30,000 write-in votes that could not be counted.  Let's round that down a little since it was only an estimate and say that there were about 265,000 votes for someone other than Clinton, Kucinich, Gravel or Dodd.  Is it reasonable to assume that over 180,000 (68%) of those votes were intended to go to Obama?
    According to the exit polls, had all the candidates been on the ballot, Obama would have received 35% of the total vote.  594,398 votes were cast altogether in the Democratic Primary (not including write-ins).  So according to the exit poll, he should have received 208,039 votes.  But only 79% of those votes would have come from the uncommitted total.  That's 164,351.  Again, let's round down a little to favor Clinton, especially since exit polls are not always accurate.  Let's say 160,000.  So he still needs 20,000 votes.  In the exit poll data, the columns showing where the intended votes came from all add up to 100% (for example, Obama would have received 18% of his votes from Clinton's total, 1% from Gravel's, 2% from Kucinich's, and 79% from uncommitted's).  So it obviously doesn't include the 30,000 write-in votes.  Were 2/3 of those intended to go to Obama?  Since the exit polls show that Obama would have received almost 3 times as many votes as Edwards (35-12) if they'd all been on the ballot, it's more than reasonable to assume that 2/3 of the write-ins were for Obama, even if we assume some write-in votes were for non-candidates, some for Bill Richardson (who apparently would have only received just 1% if his name were on the ballot), and some for Joe Biden, who had already dropped out.
    In sum, any reasonable tally of popular votes, even in the light most favorable to Clinton (counting all her Michigan votes even though the exit polls show that she wouldn't have received them had the other candidates all been listed, rounding all numbers in her favor), shows that more people intended to vote for Obama than Clinton.  Some Clinton supporters may still argue that Obama himself needlessly removed his name from the Michigan ballot and that only actual votes for him should count.  But when talking about the popular vote, the value we are considering is voter preference, irrespective of the actions, wise or unwise, of any candidate.  It would be difficult to argue that on the one hand people are being disenfranchised by not having their votes counted in an unofficial primary, but that on the other hand people in that primary could not possibly vote for a given candidate.  By any reasonable estimate, more of the people who voted in all the Democratic Primaries intended to vote for Obama than for Clinton, and his legitimacy as the Democratic nominee is therefore not tainted by the popular vote.  This is, I think, very important going forward.