Better "analysis," please

Washington Post reporter Dan Balz on last night's elections:

The most significant change came among independent voters, who solidly backed Democrats in 2006 and 2008 but moved decisively to the Republicans on Tuesday, according to exit polls. In Virginia, independents strongly supported Republican Robert F. McDonnell in his victory over Democrat R. Creigh Deeds, while in New Jersey, they supported Republican Chris Christie in his win over Democratic Gov. Jon S. Corzine.

For months, polls have shown that independents were increasingly disaffected with some of Obama's domestic policies. They have expressed reservations about the president's health-care efforts and have shown concerns about the growth in government spending and the federal deficit under his leadership.

Tuesday's elections provided the first tangible evidence that Republicans can win their support with the right kind of candidates and the right messages. That is an ominous development for Democrats if it continues unabated into next year.

Ten paragraphs later:

[David] Axelrod warned against extrapolating into the future the shift among independents. He said he believed that many people who called themselves Republicans in the past now call themselves independents but are still voting for Republican candidates. "I don't think they portend long-term trends," he said.

Gee, wouldn't it have been nice if Balz gave readers some indication of whether Axelrod is right that "many people who called themselves Republicans in the past now call themselves independents"?  

I mean, that would certainly have some impact on the significance of Balz assertions about independents becoming "increasingly disaffected with some of Obama's domestic policies," wouldn't it?  It could even mean that "independents" aren't "increasingly disaffected," but rather that people who are disaffected are increasingly calling themselves independents rather than Republicans.  Those two things are very, very different. 

And, indeed, various polls this year have shown the percentage of the public that self-identifies as Republican is at or near an all-time low, which lends some support to Axelrod's claim.  

This is exactly the kind of question Dan Balz is supposed to resolve, isn't it?  His article is billed as "analysis," after all.  Wouldn't it be great if he provided some?