Gallup also examined presidential "horse-race" numbers in the swing states, and it is these results that suggest possible oversampling of Republicans or other house effects in a GOP direction (as has happened previously with Gallup, here and here). Here are Gallup's recent swing-state horse-race numbers, compared with those of two other swing-state polls taken in the past few weeks (below the bar graphs are the states included in each respective poll, with grey indicating omitted states).
As can be seen, USA Today/Gallup is the only mid-February swing-state poll (at least that I'm aware of) showing President Obama trailing his possible GOP opponents. Rick Santorum leading Obama by 5 percentage points in the swing states (and by 3 nationally) is inconsistent with virtually any other poll.
The other two swing-state polls, by Fox News and Purple Strategies, show Obama leading the GOP contenders by varying margins.
Assuming that preference for Obama in the presidential election would be positively correlated with favorable attitudes toward health care reform, one can raise the following question: If Gallup's methodology has systematically depressed Obama's numbers in the swing states, has it also depressed support for the new health law?