Remember Pennsylvania? It was to be the last mega-state, the primary of primaries that would cap off almost four months of contests. A victory there by Obama would certainly end the race. A victory by Clinton would allow her to soldier on.

Polls and many pundits pointed out that Pennsylvania was packed with people who didn’t fit Obama’s voter profile—the elderly, Catholics, etc.—and that the state, no matter where it fell on the primary calendar, would be rough sledding for the Illinois senator.

Then some funny things happened on the way to the voting booth. We had video of Jeremiah Wright; Mayhill Fowler; some whiskey shots and some bowling. Suddenly—and with ample prodding from the Clinton camp—Pennsylvania became a test state, where the superdelegates and the media sifted the entrails of exit polls and wondered why Obama couldn’t win white working-class voters.

But what if Pennsylvania had never happened? Would the same working-class test have come in late April? If it did, how would it have played across the differing demographics of Indiana and North Carolina, two weeks later?

It’s not an irrelevant question. On March 13, 2007—yes, that’s two-oh-oh-seven—Pennsylvania Governor Ed Rendell asked the legislature to move the state’s primary to February 5, which was already shaping up to be the 2008 Super Tuesday. For whatever reason—it was a busy legislative season, and there was some debate about the best earlier date—Pennsylvania stuck with late April.

Had it moved to February 5, the contest would have blended in with the twenty-three other Democratic contests held that day.

The point is that the narrative of the primary season is based on the pegs provided by the voting. And the Democrats’ nomination calendar is almost the definition of arbitrary, especially once it lurched past Super Tuesday.

Clinton’s closing case to the superdelegates rests on some shaky pillars. Among them is that she’d be the stronger general-election candidate (polls are at best inconclusive), that she’s likely to win more of the popular vote (depends a lot on how you count).

So Clinton will urge the superdelegates and the press handicappers to concentrate on Pennsylvania and, coming up, West Virginia, Kentucky, and Puerto Rico. But again, what makes those states likely Clinton states? It’s not that they came in the wake of the Reverend Wright tempest, or the economic skid, or BitterGate. It’s that the demographics of those states, which didn’t change a bit through the campaign’s ebb and flow, favor Clinton. Period. And it’s mere happenstance that they are among the few states left to vote.

Look back to early February, when Obama, as expected, strung together a ten-contest sweep, and Clinton was on the ropes. The press gushed at Obama’s “momentum.” Clinton’s campaign hung on, even though the delegate math was fixed. She knew there was favorable turf yet to come, and hoped to deploy another argument: that the coming late victories might be used to convince the superdelegates that she was better situated to beat John McCain. She’d like to do the same now.

Demography is not always destiny. (That’s why we still have elections.) But as the press and the commentariat inch forward from Indiana and North Carolina, we hope that they remember—and remind their readers and viewers—that any “momentum” harnessed from these late contests deserves a heavy footnote.

If you'd like to help CJR and win a chance at one of 10 free print subscriptions, take a brief survey for us here.

Clint Hendler is the managing editor of Mother Jones, and a former deputy editor of CJR.