Mark Blumenthal examines the ongoing argument as to which of the Democrats have the most popular votes. In addition to the traditional counting issues, he notes that a bigger problem is that the very notion fails to meet the test of “concept validity.”

That’s a wonky term that simply means that before measuring something we ought to define clearly what it is we are trying to measure.

[…]

Clinton supporter Gov. Ed Rendell, D-Pa., for example, has condemned caucuses as “undemocratic.” Appearing on “Meet the Press,” he argued that in caucuses, “older people can’t vote, older people who vote by absentee ballot… if you’re a shift worker and a lot of our workers, because they’re low-income workers, are shift workers, you can’t vote in a caucus.”

Obama supporters, on the other hand, point to a paper by author Glenn Hurowitz and business school professor Gregory Nini asserting that a popular vote tally “dramatically devalues the popular will” of voters living in caucus states precisely because of the much lower turnout in caucuses. The 13 caucus states (which Obama won by an average margin of 35 percentage points) represent 15 percent of eligible voters and 14 percent of Democratic delegates, but only 2 percent of the popular vote cast. Had the caucus states held primaries, Hurowitz and Nini argue, Obama’s victory margins would have been diminished, but his raw popular vote tally would be significantly increased because of much higher turnouts.

It goes beyond that.

• In the early states, where other candidates drew significant numbers of votes, merely counting the number who voted for the two who remain standing is invalid unless we presume that those votes would have been equally divided between Clinton and Obama. Those who would have preferred, say, John Edwards but would rather have Obama than Clinton are retroactively disenfranchised in this process. (In the actual delegate contest, though, they’re actually more influential than those in later states, since they help shape the race to come.)

• Additionally, it’s quite likely that many of those who voted for Clinton or Obama in January or February, when much less information was known, would prefer the other candidate if allowed to vote this past Tuesday when the residents of Pennsylvania did.

• Some states had elections on Saturdays; others, during the week. This impacts different people in different ways but certainly influenced turnout and vote margins.

I’m sure readers can come up with other points along these lines. The bottom line is that, unless there is a single election, held at a single point of time, pitting the exact slate of candidates against one another and employing the same counting rules, any reference to a “popular vote” is meaningless.

This doesn’t mean, of course, that the candidate trailing in the delegate count can’t trot it out to see if it sways the superdelegates.