I doubt it will be a tossup. Obama has gained support faster than Clinton in every state he has campaigned in, without exception. Given how well he's doing in IN before he has even campaigned there, I expect that he will surpass Clinton easily.

As a social scientist, I am disappointed in TPM's breathless and over-interpreted poll reporting (not to mention the silly concentration on he said/she said nonsense such as the Richardson flap). When polls are within the margin of error, the results should be described -- especially in headlines -- as "ties." And when dealing with horce-race polls months in advance before one of two parties has settled the nomination, why should a lot of interpretive fire-power be applied at all, especially to margins around 3-5%?

On the "race gap" question, not just TPM but also others need to take into account that how voters sort out in a primary has little predictive power for how they will sort out in a general election months later. Alternatives and context matter. The only compelling evidence I see is that white downscale voters in rust belt states are understandably very worried about economic issues. This may affect their current stance toward Obama, but only because he is now running against a "Clinton." Nostalgic for the 1990s, many of them see "Clinton" as a package deal: vote for her and you get him back, too, and maybe the better economic times will roll again.

In the general election, if he is the nominee, Obama will need to address those concerns consistently AND pick a Vice President and line up other potential officials who are reassuring to downscale Democrats (and to national-security voters). Just as Clinton, should lightening strike and she becomes the default nominee, will need to mobilize and re-inspire blacks and the young. All these steps are in the future for either nominee, just one more reason why horce-race polls now are virtually meaningless. TPM should spend more of its space and investigatory abilities on the medium-term meaningful stuff, not just the within-margin-of-error or entirely speculative short-term poll blips.

You may feel you are justified because of what "the Clinton camp will argue to superdelegates." But, after all, they are not winning, and probably won't, and all they can do is try to play games with media outlets and blogs. This is the politics of trying to get free media headlines by saying colorful or threatening things, not predictive of much of anything beyond one or two days.

TPM owes its intelligent readers better -- or else, we will check in much less often!

Theda Skocpol, Cambridge, MA

With a sample size of 400 and an moe of 5% this is pretty much worthless...

I'm not one to be hyper-sensitive about polling headlines, but it is interesting to compare Eric's three headlines this morning.

"Clinton Up By Three Points In Indiana Primary"

"CBS/NYT Poll: Obama Narrowly Leads Hillary"

Both of the above polls register 3-point leads. Eric characterizes Clinton's lead in Indiana as "up", which connotes such things as "upbeat" or "exhilarating" -- even though an 8 point lead, if it were real, would be bad news for Clinton at this stage of the game. Eric's second headline, by contrast, characterizes Obama's national lead as "narrow", which suggests such things as "insignificant" or "pinched," even though it, too, is a 3-point lead. Consistent diction would avoid any sense of bias here. Both leads should be characterized as either "up" or "narrow." In general, for all polls, and regardless of the spread, "leads" and "trails" would be appropriately neutral terms.

Especially noteworthy is Eric's third headline:

"Hillary Ahead Of Obama By 8 In Pennsylvania, Obama Trails McCain For The General"

Here Hillary is characterized as "ahead" by 8 points - which is fine, nothing wrong with that. But then note the peculiar switch that happens when it comes to the GE match-ups. Even though the poll shows that Obama runs 1 point stronger than Hillary against McCain, the emphasis shifts to Obama's trailing McCain, with no mention at all of Hillary's also trailing. Thus the overall effect of the headline is "Hillary is ahead, Obama trails" -- a double hit against Obama, so to speak. It's also worth noting that Eric keeps us on a first-name basis with Hillary while he keeps Obama on a last-name basis and thus at a formal distance. (This is Eric's habit generally, I believe.)

Again, consistency would make all the difference. The headline might better read:

"Hillary (or Clinton) leads Barack (or Obama) by 8 points in PA. Both trail McCain in GE match-ups."

How petty we are...