Hendrick Hertzberg has an interesting breakdown of the various ways of calculating the popular vote numbers of Clinton and Obama in the Democratic primary. According to one method, Obama's ahead by about 500,000 votes; according to another, Clinton leads by 121,000. The numbers vary due whether or not Florida and Michigan are counted, whether some caucuses are counted, and how the "uncommitted" votes in Michigan are counted (Obama was not on the ballot in Michigan--only Clinton was--so, even though Clinton won a majority there, there are hundreds of thousands of "uncommitted" voters who may have been favoring Obama). Hertzberg lays out one scenario in which claims about the popular vote could be muddled enough to give Clinton an opening to argue for her popular legitimacy:

I calculate that roughly another four and a half million people will vote in the remaining Democratic primaries. If Clinton wins these by the overall margin she piled up in Pennsylvania—i.e., around ten points, 55-45—she will net another 450,000 votes.

Here’s what that would do to the margins in our eight categories, in the order I’ve discussed them, starting with just the D.N.C.-approved primaries down to the two “corrected” (by me) Michigan-included ones:

Regular primaries:

Obama’s margin: 51,298

Regular + four caucus states:

Obama’s margin: 161,520

Regular + Florida:

Clinton’s margin: 243,474

Regular + Florida + caucuses:

Clinton’s margin: 133,252

Regular + Florida + Michigan:

Clinton’s margin: 571,783

Regular + Florida + Michigan + caucuses:

Clinton’s margin: 461,561

Regular + Florida + Michigan + uncommitted:

Clinton’s margin: 371,783

Regular + Florida + Michigan + caucuses + uncommitted:

Clinton’s margin: 261,561

In other words, Clinton would have a case. Obama would have one, too—he’d still be a little bit ahead in the popular vote according to to the rules everybody agreed upon in advance, and he would definitely be ahead in elected delegates. She would have a popular vote lead in all the count-Florida-and-Michigan categories. But neither candidate could any longer plausibly claim that he or she was unambiguously the people’s choice.

Of course, that presumes she can win by ten points in every state, and a ten-point victory doesn't seem too likely in North Carolina for starters....Based on his speculations, Hertzberg says that, for every point less than ten in Clinton's overall post-PA lead, Obama nets 45,000 votes. Timothy Noah's claiming that any hope of a Clinton win is a "fairy tale," but Hertzberg thinks that there's a possibility of a "rough" popular vote tie (and thinks that the popular vote will matter a lot) and seems to suggest that the race may not be over yet.