Off-Topic

Certainly two key states. The MOE is 4 points in the polls. It's good news for Obama that he's ahead but he's still well below 50 percent (45 percent) in both, which isn't so good news. Romney has more room to grow, since many people don't know him that well yet. OTOH, voters have had 3.5 years to see Obama in action and he can't get more than 45 percent of the voters. As has been discussed here before, undecideds typically break heavily for the challenger. Both of these polls have a sizable amount of undecided voters that will decide the outcome. The Romeny camp isn't happy their guy is down in two key states, although within MOE in one, but see they can make it up if the can convince the undecideds to go with Mitt. I see these two polls as essentially a wash this early in the process.

So you're either left with, "Who knows?", or Gallup is right if they're right about the black vote, as opposed to the non-white vote. This confirms my suspicion that polls are used more to move votes than to figure out where the voters actually are.

Fixed.

If you look at the fine print, the Gallup approval rating poll samples adults. The Obama vs. Romney samples registered voters. Registered voters tend to be a more conservative population than adults, and likely voters even moreso.

If the difference is statistically insignificant, why bother to filter by registered voters or create likely voter models? Of course, these differences are quite significant and all adult samples are not nearly as predictive as more restricted samples.

Below is an excerpt from FiveThirtyEight's methodology page:

The likely voter adjustment. Throughout the course of an election year, polls may be conducted among a variety of population samples. Some survey all American adults, some survey only registered voters, and others are based on responses from respondents deemed to be “likely voters,” as determined based on past voting behavior or present voting intentions. Sometimes, there are predictable differences between likely voter and registered voter polls. In 2010, for instance, polls of likely voters are about 4 points more favorable to the Republican candidate, on average, than those of registered voters, perhaps reflecting enthusiasm among Republican voters. And surveys conducted among likely voters are about 7 points more favorable to the Republican than those conducted among all adults, whether registered to vote or not.

--
and here's one from Silver's blog from Monday:

In the past eight presidential election cycles or so, the Republican candidate has done a net of about two percentage points better on average in likely voter polls than in registered voter polls. That is, if the Republican had a four percentage point lead in a poll of registered voters, it might be inferred that he had a six percentage point lead among likely voters instead. This historical trend is probably not simply a statistical fluke, and instead reflects the fact that the demographic groups that tend to vote Republican — for instance, older and wealthier voters — tend to be more likely to vote as well.

Funny, it's around October that I feel likely voter models are worthwhile. By then, it's easier to forecast who a likely voter is as opposed to months out.

Gotcha. I feel like 6 months out from an election, anyone is a potential voter. So when pollsters are using likely voter models so far out, I'm a little dubious about that being an accurate description of the current race. Once you get to October, and the campaigns have finished their voter registration drives and are working their ground game, then I feel like it's easier for them to be accurate, and get an idea about what turnout will look like.

Serious question: given that no one seems to like anyone, could we see a plurality this time?

Could be. Especially if Gary Johnson gets a little steam from the Paul-ites.

LOL. That's how we felt when we got Nadered (well, and butterfly ballotted!) in 2000.

LOL, it's like the American version of "Kim Jong Il looks at things."

Dog Lovers' Day at Fenway??