USA TODAY/CNN/Gallup Poll shows a remarkable turnaround in 17 battleground states where polls and historic trends indicate the race will be close, and where the Bush campaign has aired TV ads. Those ads say Bush has provided "steady leadership in times of change" while portraying Kerry as a tax-hiking, flip-flopping liberal.

The ads have been one factor in wiping away an inflated lead Kerry held in those states. Most of them have had primaries or caucuses that allowed Democrats to dominate the news and Kerry to emerge as a victor. In a survey taken in mid-February, Kerry led Bush by 28 percentage points in those states, 63% to 35%. Now Bush leads Kerry in them by six points, 51% to 45%.

So this would seem to be evidence against my contention that Kerry is dong pretty well, but I have to make the point that aggregating poll results from 17 battleground states into one lump figure isn't really any more valuable, in terms of predicting the outcome of the elction, than looking at national figures, because the race will be determined by the results of 51 different elections, each weighted differently (by the electoral votes the winner will receive).

The national poll numbers are somewhat useful in that they provide fodder for campaigns to claim advantages, and thus to influence the voters in swing states which will ultimately determine the outcome of the election, but they are otherwise all but meaningless and shouldn't really be dwelt on, nor should they be touted by the media. Even less relevant are other sub-national groupings such as are being reported here. Until Gallup releases their data for each of the 17 states they're referring to, I can't really assess what this supposed change in direction means for my current counting.

I'm going to keep working on putting together my larger post, but if the Gallup results become available state by state, I'll take them into account and refigure things.

Update (3/31): Both Ruy Teixeira and Billmon counsel chilling out about the Gallup results.