One frequent refrain you hear in the press is that given the underlying electoral fundamentals, Barack Obama should be doing better in head-to-head polling matchups against John McCain. This line of argument was somewhat persuasive when McCain was running narrowly ahead of Obama in national polling; now that he's locked up the nomination Obama has a consistent four-to-six-point lead in national polling, confirmed again by the Washington Post poll out this morning, which has Obama up six among likely voters. November is a long way off, but a victory of this size in the general election, while modest by historical standards, would be fairly sizable by the standards of our current era in which party coalitions are more stable and ideologically cohesive. In the Democrats' "landslide" in 2006, they won the national popular House vote by seven points; Bill Clinton beat George H.W. Bush in 1992 by six.
What seems to be driving this notion that Obama should be doing better in polls is a comparison between Obama's numbers against McCain and generic Democrat-versus-Republican numbers, which have found a Democratic advantage of 10 to 15 points. There are a few reasons to be suspicious of these latter results, though. For one thing, generic ballots more or less always overstate the Democratic advantage--check out the 2006 polls, for example. There are a lot of voters who identify themselves in the abstract as Democrats for historical and cultural reasons, but almost never vote Democratic. Barack Obama won't win by 15 points, nor will Democrats will the national popular House vote by that much; even ten points, which would be a genuine blowout, is a stretch.
What's more, in the today's political climate, the Republican brand is toxic, and the current media narrative reflects that. But this has the potential to become a self-fulfilling prophecy; voters constantly hear that their fellow citizens are fed up with Republicans, and so the only people willing to admit to a generic preference for Republicans are die-hard party loyalists. What that means is that there are a lot of soft Republicans and Republican-leaning independents who voted for Bush, voted Republican in 2006, and are likely to vote for John McCain, but don't show up as Republican supporters in the generic ballot. Most people take the generic ballot as a baseline and try to explain why Obama is underperforming; I would say, on balance, it's probably more accurate at this point to take the presidential numbers as a baseline and try to explain why generic Republicans are underperforming.
Granted, Obama still has work to do to shore up support among Democrats and Democratic-leaning independents, and if the race were to drop back into a dead heat for an extended period of time, particularly if it happens after the conventions, that would be cause for concern for Obama. But if his lead stabilizes where it is now, even if the generic Democratic advantage remains in double digits, I'd say that's about par for the course in terms of what it's reasonable to expect.