CHARLOTTE, NC - SEPTEMBER 4: Women Democrat members of the U.S. House of Representatives gather onstage after speaking at the 2012 Democratic National Convention on Sept. 4, 2012 in Charlotte. (Photo by Bonnie Jo Mount/The Washington Post)

One of the most interesting numbers in the new Washington Post-ABC News poll is buried deep within the cross tabs.

The poll shows that, on the generic ballot, the GOP leads 50-44 among all likely voters. But among unmarried women, Democrats have a 48-43 edge.

That might seem to be a good thing for Democrats, except that they count on single women in a much bigger way -- in fact, so much so that this demographic is routinely highlighted as the key to Democrats keeping the Senate.

"To Hold Senate, Democrats Rely on Single Women," wrote the New York Times in July. The group Women's Voices Women Vote Action Fund commissioned a poll from a Democratic pollster around the same time and headlined it, "New Poll: Unmarried Women Voters Could Make the Difference in Senate Battleground States. Then, when the support for Democrats among single women in that same poll doubled, Real Clear Politics reported, "Unmarried Women Fuel Pollster's Optimism for Dems."

It makes sense. After all, this is a big portion of the so-called "Obama coalition" -- voters who are less apt to vote in midterms but turned out big for the president in 2008 and 2012 and voted for him by huge margins. Also in that coalition are African American, Hispanics and young people. Unmarried women, though, comprise a significantly bigger share of eligible voters than the others: about 25 percent.

Over the last four elections, here's how the unmarried-woman vote has looked:

As you can see, all of these margin are much larger than five points. Democrats have won unmarried women by between 25 and 41 points each of the past four elections. Obama won 70 percent of them in 2008, to just 29 percent for John McCain.

Of course, we shouldn't read too much into one poll, and breaking out small sub-samples (this poll interviews 228 registered unmarried woman voters and 151 who are considered "likely" to vote on Nov. 4) results in high margins of error. The idea that unmarried women will only favor Democrats by five points next week is probably far-fetched, and the margin of error means that five-point margin could be significantly higher.

Indeed, recent polls from AP-GfK and the Pew Research Center show unmarried women favoring Democrats by 14 and 20 points, respectively. And the Democratic pollster referenced above, Democracy Corps, showed the gap at 22 points. (Correction: The 22-point gap was actually in battleground Senate states only, and it was 32 points nationally.)

But these polls still show smaller margins than have exited over the last four elections. In other words, they show this group is a precarious one for Democrats, though none of them indicate single women will definitely favor Democrats by less than in recent races.

In addition, the WaPo-ABC poll suggests this group is really not that interested in the election, with 58 percent of registered single women voters following it very closely. That's lower than every demographic tested except one -- unmarried men (57 percent).

If these numbers are close to accurate, they're a troubling development for Democrats, who have spent much of the past two years emphasizing the GOP's so-called "war on women" for the purpose of turning out this particular demographic.

If unmarried women (who, again, are about one-quarter of the electorate) somehow don't deliver Democrats a substantial margin come next Tuesday, it would be very difficult for Democrats to hold the Senate, much less avert further losses.