Saturday was a much quieter day for polling than we've grown accustomed to - and the FiveThirtyEight forecast was unchanged to the decimal place, with President Obama given a 76.2 percent chance of winning the Electoral College, the same as in Friday's forecast.
There was slightly more action beneath the surface. Mr. Obama's projected margin of victory in the popular vote declined slightly to 3.3 percentage points - down from 3.5 points on Friday, and from a peak of just over 4 points immediately after the Democratic convention. However, Mr. Obama's Electoral College chances held steady in the forecast because of a strong poll for him in Pennsylvania.
At the very least, Mr. Obama's lead in the nati onal polls no longer seems to be growing. If he gained additional ground following the attacks on Americans in Benghazi, Libya - or from Mitt Romney's response to it - there has been no sign of it in the most recent national tracking surveys.
Instead, the question is to what extent, if any, Mr. Obama's lead has declined. The Gallup national tracking poll now shows him ahead of Mr. Romney by four points - down from a peak of seven. And there has been a clearer reversal in the Rasmussen Reports tracking poll, which has now reverted to showing a two-point lead for Mr. Romney.
Two online tracking polls do not show any signs of decline for Mr. Obama, however. In the American Life Panel survey conducted by the RAND Corporation, Mr. Obama had a smaller bounce than in some other polls - but it has held steady over the past week, as he has continued to hold roughly a three-point lead among likely voters. The RAND poll differs from others in that it uses a panel of the sam e 3,500 respondents who are asked their opinions about the presidential race continually throughout the contest; it is therefore subject to less statistical noise than other surveys.
Mr. Obama also enjoyed his widest lead to date, seven points, in the last version of the Ipsos online tracking poll, which was published on Thursday - although the poll has not been updated over the past two days.
There's also been a large volume of state polling to sort through. Our view is that the consensus of the evidence from these surveys has shown pretty good numbers for Mr. Obama - but not great ones -Â and tends to provide the most support for the hypothesis that Mr. Obama holds a lead in the national race of just under four percentage points right now.
The lone poll from a swing state to be published on Saturday was one to show more unambiguously impressive numbers for Mr. Obama, however. That survey, from The Philadelphia Inquirer, showed Mr. Obama with an 11-point lead there, up from 9 points before the conventions.
This is not the first poll to show a lead for Mr. Obama in Pennsylvania. In fact, the polling has been quite consistent there, especially as compared with other states like Virginia and Florida, in which there has been less agreement among the pollsters. Mr. Obama has held a lead in each of the last 23 Pennsylvania surveys, dating back to early February.
Intuitively, it might seem that the polling average is subject to less error when there is more consistency among its individual components. And in fact, I have found this to be the case in my empirical study of polls; less volatility in the polls implies more uncertainty about the eventual outcome.
But to the extent there has been a trend in Pennsylvania, it has favored Mr. Obama, with his lead slowly but steadily expanding over the past few months.
This may reflect Mr. Romney's on-again, off-again interest in competing in the state. According to T he National Journal, Mr. Romney's campaign has spent nothing at all on advertisements in Pennsylvania since May 1, while outside groups backing him have spent a relative modest sum of about $10 million.
I wrote in July that I thought it was a mistake for Mr. Romney's campaign not to make a more vigorous effort to contest Pennsylvania. Moving the numbers in Pennsylvania is difficult, making it a high-risk proposition for a Republican who trails there - but it could also provide a huge reward for Mr. Romney, since losing the state would carve an enormous hole in Mr. Obama's re-election map.
But at this point, it may be too late. The forecast model now projects Mr. Obama to win Pennsylvania by about seven points. Given the consistency of his lead there, that translates into a 93 percent chance of victory on Election Day, according to the forecast.
It is in the blue-leaning states like Pennsylvania where Mr. Obama has arguably shown his best numbers since the c onvention.
Although New Jersey stretches the definition of a swing state, Mr. Obama had a safe-looking lead of 14 points in two polls published there since the conventions.
And in Michigan, two polls published last week showed Mr. Obama making gains. His lead expanded to 10 points from 3 in a poll published by the firm EPIC-MRA. In another Michigan survey, from the firm Foster McCollum White Baydoun, Mr. Obama reversed a four-point deficit to take a two-point lead. (That looks like a decent result for Mr. Romney despite Mr. Obama's gains, but it isn't really, since Foster McCollum White Baydoun has had an extremely strong Republican lean so far this cycle, showing results about 10 points more favorable to Mr. Romney than the consensus.)
All of this makes me very curious about another blue-leaning state, Wisconsin, which has not been polled at all since the conventions. Mr. Romney had appeared to draw Wisconsin to within about two points following his select ion of Representative Paul D. Ryan as his running mate, making it among the most important states in the country.
But we just don't know what has happened to Wisconsin since then - whether the improvement that Mr. Romney had in the polls following his pick of Mr. Ryan was temporary, and likewise, how Mr. Obama's convention bounce has carried into the state.
If Wisconsin behaves more as true purple states like Virginia and Colorado, where Mr. Obama's bounce has been harder to perceive in the polls, Mr. Romney will be able to advance a stronger case that the damage he took during the Democratic convention has been tolerable. If the set of Wisconsin polls shows a clearer shift toward Mr. Obama, however, as in Pennsylvania and Michigan, Mr. Romney will be working from a rather thin electoral map.
No comments:
Post a Comment