There are two interesting findings in the new NBC/Wall Street Journal poll relative to voter enthusiasm, which is often exaggerated as a factor in voting outcomes.

The first is that overall voter enthusiasm seems to be declining as we approach Election Day, which is unusual given the greater media attention being paid to politics down the stretch.

But that’s on average: there’s a different picture, and one more favorable for Democrats, if you just focus on the states with competitive Senate races, which is where an awful lot of the money and effort are being expended. Here’s how WSJ’s Reid Epstein puts it:

In the 11 Senate race states, people surveyed by the Wall Street Journal/NBC News poll are more enthusiastic about voting, more likely to support Democrats and report a high interest in the midterm elections.

In the 11 most competitive states – Alaska, Arkansas, Colorado, Georgia, Iowa, Kentucky, Kansas, Louisiana, Michigan, North Carolina and New Hampshire – 54% of people said they are more enthusiastic about voting than in past years. Nationwide, that number falls to 44%. In the Senate states, 56% said they have a high interest in the election, compared to 51% overall.

And in those 11 Senate states, seven of which voted for Mitt Romney in 2012, more people say they’d prefer Democrats (50%) in control of Congress than Republicans (44%) than among the general population (46% to 42%).

That’s true even though the president’s approval level is lower in these states (which you’d expect given their Republican tilt) than it is nationally.

So no matter what happens on November 4, Democrats seem to be doing something right in Senate races. But the overall uncertainty over turnout makes predictions even more perilous than is usually the case.

Ed Kilgore

Ed Kilgore is a political columnist for New York and managing editor at the Democratic Strategist website. He was a contributing writer at the Washington Monthly from January 2012 until November 2015, and was the principal contributor to the Political Animal blog.