But In Other Polls....
There's another major presidential poll that was released today: the highly regarded "Battleground" poll jointly conducted by the Democratic firm of Lake Snell Perry, and the Republican firm The Tarrance Group, sponsored by George Washington University. The data's a little over a week old.
Among LVs, the poll shows Kerry up 44-43, with Nader at 1 and 12 percent undecided. But after quizzing respondents about the candidates and then including leaners, their "aided ballot" shows Kerry up 49-47 without Nader in the mix, and 48-47-3 with Nader in.
Interestingly enough, the poll has Bush winning a big 5 percent of Democrats, as opposed to the 15 percent reported by the LA Times poll. It also shows Kerry up by 5 among households with veterans, and only down 3 among vets themselves.
One of the fun things about the Battleground polls is you get to read "strategic analysis" memos from both Celinda Lake and Ed Goas: in other words, Democratic and Republican spin on the same data. And it's about what you'd expect, with Celinda emphasizing Bush's failure to get majority support after four years and Ed suggesting that Kerry's missed an opportunity to seal the deal. With the particular relish of so many partisans, both predict a result determined mainly by turnout patterns. It's so much neater than the messy process of actually persuading voters rationally, doncha know.
At any rate, this poll may offset some of the insider buzz about the LA Times survey. In the past, Battleground polls have been known for an unusually tight "screen" for likelihood to vote, and have often shown results more favorable to the GOP than other major surveys.
And finally, it should be noted that Zogby Interactive released a big batch of polls of 16 battleground states the other day, all conducted from August 16-21. They show Kerry leading in 14 of these states. But Zogby's erratic rep in state polling makes that scorecard questionable, and the specific results don't exactly inspire confidence. He's got Kerry up by 2 in Tennessee, and Bush up by more than 5 in Ohio, and by nearly 8 in West Virginia.
In general, state polling in a presidential race is a very tricky thing, with small samples and widely varying methodologies complicating the picture. Much as I appreciate RealClearPolitics' effort to provide all the polling, their effort to summarize the race each day by averaging poll results is definitely misleading, especially when they include blatantly partisan polls by outfits like the GOP firm Strategic Visions.
Maybe we'll be better able to separate the sheep from the goats among state polls when we get a little closer to November 2.