Obama's Post-Convention Bounce Is Gone

Earlier today, I had an experience I honestly never thought I would have. It involves a former member of Congress and violating the law, but that's not what I wanted to blog about.

If you watched any news show with a panel or round-table discussion in the past week and a half, then you've probably heard something about the Obama-Biden ticket getting a "bounce" in the polls following a hugely successful Democratic National Convention. Well, as of today, we can definitively state that that post-convention bounce is g-o-n-e.

As of today, Obama is back to his pre-convention poll positions in both the Gallup and Rasmussen tracking polls. (As of right now, Gallup and Rasmussen are the only two pollsters conducting tracking surveys of the pesidential race and publicizing the results.) The president leads Mitt Romney by one point, 47%-46%, in the Gallup poll and trails him by two, 45%-47%, in the Rasmussen poll. That's exactly what these two polls showed two weeks ago, right as the Dems were starting their convention in Charlotte.

Beyond these national poll numbers, the news today was mixed: Rasmussen also published the results of a survey showing Romney up two points on Obama in Colorado, 47%-45%, which is not good news for the president, who won the Centennial State by nine points in 2008. Gravis Marketing, which has been regularly polling voters in select swing states more frequently than any other independent pollster since the conventions, has Romney leading Obama by a statistically insignificant point in Florida, 48% to 47%. More significant than that, perhaps, is that their poll shows the U.S. Senate race there to be a dead heat, with Rep. Connie Mack (R., Fla.) leading Sen. Bill Nelson (D) by one point, 43%-42%. (This is one race that may hinge on which party's presidential candidate carries the state.)

The good news for Obama came from a Washington Post poll that gave him an eight-point lead in Virginia and an NBC News/Wall Street Journal poll that shows him leading Mitt Romney by five nationally (50%-45%). The Virginia results are especially stunning--and suspicious--as they are out of line with what every other poll of the race in that state has shown. Rare is the case when such a salient outlier is correct, but the same could be said about the Gravis Marketing survey that gave Romney a five-point lead over Obama, 49%-44%, in Virginia last week.

In my view, two states have been severely underpolled in the last week: Iowa and Wisconsin. Only once in the last six presidential elections has a Republican ticket carried either of these states, but the most recent polling data we have shows that they are both very much in play this election cycle. Now that the effects of the Democratic National Convention have worn off--at least as far as polls reflect--survey results from these and other swing states should be considered more accurate. Ordinarily I advise people with strong preferences in an election to avoid fixating on poll numbers, but from this point forward, I think it's worth it.