Skip to main content
ABC News
Sept. 10: Will Obama’s Bounce Hold?

Has President Obama’s convention bounce reached its peak? On Monday, his position declined slightly in our forecast for the first time since Aug. 27.

To be sure, Mr. Obama had a fairly strong day of polling on Monday relative to the long-term baseline. But the data was a little bit more equivocal than in polls released over the weekend — which may suggest, at least, that he will make few further gains in the polls.

Mr. Obama gained a single point in the Rasmussen Reports and Ipsos tracking polls on Monday, but held steady in two others from Gallup and the RAND Corporation. (His lead also declined in the Rasmussen Reports tracking poll on Tuesday, which was published after our forecast was updated late Monday night.)

Some of the surveys apart from the tracking polls were erratic and not necessarily so strong for Mr. Obama. He actually declined by several points in one national poll, conducted by TIPP, although a considerable amount of their data was from early last week and did not reflect the key convention speeches.

National polls out on Monday from CNN and ABC News/The Washington Post told somewhat contradictory stories.

In the CNN survey, Mr. Obama moved into a six-point lead among likely voters (and an eight-point lead when the Libertarian candidate, Gary Johnson was included in the poll) — up from a tie previously. But he made fewer gains in the survey among the broader universe of registered voters, with whom his lead expanded to eight points from seven.

In the ABC News/Washington Post poll, just the opposite happened. Mr. Obama made substantial gains in that poll among registered voters, going from a one-point deficit to a six-point lead. But the ABC News/Washington Post poll has started to report likely-voter results, and this is where Mr. Obama’s gains were not all that substantial, going from a two-point deficit in its August poll to a one-point lead now.

The state polls were also somewhat mixed for Mr. Obama. He received one of his worst polling results of the year in a North Carolina survey conducted by SurveyUSA for the Civitas Institute, which put him 10 points behind Mitt Romney. There are a few things to be mindful of in this poll: most of its data preceded Mr. Obama’s acceptance speech last Thursday, and the poll somewhat implausibly showed 30 percent of its African-American respondents voting for Mr. Romney.

Nevertheless, it’s one reminder that candidates probably do not get much, if any, benefit from holding a convention in a swing state; Mr. Romney, for his part, got some middling data in Florida in polling conducted last week.

A better state-level result for Mr. Obama came from an Ohio poll from Gravis Marketing, which put him four points ahead there, reversing a one-point deficit in a poll conducted after the Republican convention. Furthermore, Gravis has had substantial Republican lean in the other polls it has released this year, making these results look stronger for Mr. Obama by comparison.

For the most part, however, Mr. Obama’s stronger state polls on Monday were from places like Minnesota and Washington State that are not as relevant to the Electoral College. We’re lacking data so far on what, if any, effect the Democratic convention has had on Virginia, Florida, Colorado and other important swing states.

Given that the forecast model was quite aggressive about pricing in Mr. Obama’s convention bounce, we’re also going to be monitoring carefully for any signs of its decline.

The expectation of the model is that Mr. Obama’s bounce will decline some, by a percentage point or two, over the course of the next week and a half. If he holds his full bounce in the polls, it will be an encouraging sign for him.

What makes matters a bit more confusing is that we’re still seeing some polls trickle in that were conducted partly or wholly before the major speeches at the Democratic convention. In those surveys, just the opposite is true: it counts as a disappointment if Mr. Obama’s numbers fail to improve, as interviews that predate the convention are replaced with newer ones.

This is one time when it’s important to carefully check the field dates associated with a poll. It seems pretty clear that the number of polls is likely to proliferate this week, including polls from survey firms that have not been very active before. Although they are often fine polls, some surveys from newspapers and academic institutions have a longer lag time before they report their results.

But at the very least, it appears as though Mr. Obama is not on track for the large bounce in the polls that seemed possible over the weekend, when Mr. Obama’s numbers were improving at a prodigious rate in the wake of former President Bill Clinton’s speech.

Mr. Obama almost certainly had the more successful convention than Mr. Romney. But in some sense, his bounce has been fairly ordinary; conventions typically do produce bounces.

It was the very small bounce that Mr. Romney received in the polls after his convention — about two points — that is more unusual historically, and somewhat low even relative to reasonably diminished expectations.

When incumbents receive a bounce in the polls after the conventions, it can potentially be more persistent. In each year between 1988 and 2004, there was little immediate sign of decline in the incumbent party’s numbers after its convention, with its results holding in about the same place for up to five weeks.

The reason may simply be that the incumbent party holds its convention last — and so there is less to interrupt its message.

I suspect that in future years, we may see the challenging party avoid scheduling its convention so close to the incumbent’s, as it that provides it with very little time to build a sense of momentum.

Moreover, polls conducted after both conventions will reflect voters having had a chance to hear both sides of the story. It’s not good for Mr. Romney that voters just became much more informed about the election, and the numbers have shifted toward Mr. Obama.

Still, there have also been years, like 2008, when the incumbent party’s bounce faded quickly. There have also been years in which the incumbent had a long, gentle fade in the polls between the convention and the election itself. The more advantage an incumbent builds after his convention, the more he’ll be able to tolerate shedding a few points in the polls later.

In 2004, George W. Bush built about a five-point lead after his convention, and lost three points of that by Election Day — still enough to give him a two-point win over John Kerry.

Four years earlier in 2000, Al Gore had built up just a slightly smaller lead — four points over Mr. Bush — but it wasn’t quite enough for him to win the Electoral College.

The unusually low volatility in the polls this year also makes Mr. Obama’s bounce more difficult to analyze. Since the numbers have been very hard to move, does that imply that when a candidate does see a shift, his numbers are more likely to remain intact and establish a new normal?

Conversely, is the gravity in the race so strong that we’re sure to see Mr. Obama’s numbers revert to the long-term mean, with about a two-point lead over Mr. Romney, before very long?

Ultimately, the best reason for Mr. Romney’s backers to be pessimistic about his chances may be that the old normal wasn’t that great for him. With a two-point deficit in the polls, Mr. Romney could win the election in any number of ways, but he would nevertheless be an underdog, and especially so on the premise that voters were extremely resistant to changing their minds for any reason.

Still, Mr. Romney needs to take care of first things first, and seeing some steam come out of Mr. Obama’s numbers would be his first step on the road to recovery.

Nate Silver founded and was the editor in chief of FiveThirtyEight.


Filed under