Within a day, three national polls asked Americans who they preferred in the contest between President Barack Obama and Republican Mitt Romney, and one of them produced a result that only the White House could love.
A Bloomberg National Poll showed Obama, a Democrat, with a 13-point lead in the presidential race over Romney, while an Associated Press Poll found that Obama held a 3-point lead and Pew Research Center had the president up by 4 points.
The variance can be explained by differences in methodology and by voter opinion that may be highly changeable in an election that both campaigns have forecast as close.
“Small differences in methods between reputable surveys can amount to large differences in results of the surveys when voter commitment is still uncertain,” said Andrew Kohut, president of the Washington-based Pew Research Center, a nonpartisan group that conducts polling and demographic studies. “This is a race where many people are still not committed to their choice.”
While the so-called horse-race number -- who is ahead -- attracts attention, political professionals take greater note of the answers to underlying questions, particularly five months before an election.
On questions such as whether the public approves of Obama’s handling of the economy, all three polls had ratings of below 50 percent, a troubling sign for the incumbent.
The Bloomberg Poll showed that Romney hasn’t recovered from the primaries as his favorability rating was 39 percent, unchanged from when he announced his candidacy a year ago; his unfavorable rating jumped by 17 percentage points. The AP poll found registered voters split on whether Obama or Romney would do a better job improving the economy, an issue that the Republican has made the centerpiece of his White House run.
Bloomberg’s poll gave Obama a 19-point edge in being in touch with ordinary Americans compared with a 31-point advantage in the Pew survey. When asked whether the country was on the right track or wrong track, Bloomberg’s poll found 62 percent said the wrong track, while AP’s said 60 percent.
The Bloomberg poll also asked a number of what are known as evaluative questions before seeking the respondent’s choice between Romney and Obama. They include whether they had a favorable or unfavorable opinion of Romney, Obama or former presidents Bill Clinton and George W. Bush.
Results Not Skewed
J. Ann Selzer, whose firm conducted the poll for Bloomberg News, said the questions didn’t skew the result, noting that one of them showed by 62 percent to 31 percent that the country was on the “wrong track,” not a positive finding for Obama.
Kohut said the Bloomberg questions before the horse-race query might have pushed a voter toward Obama because they asked about Clinton, who was viewed favorably, and Bush, who was viewed negatively.
“It will get people thinking in a particular way about their choice, and that may not be the case if they are unprompted,” he added.
The Bloomberg poll asked the horse race as the eighth question; a Wall Street Journal/NBC News Poll conducted May 16-20, asked it as the 12th question after other evaluative queries.
Bloomberg’s poll didn’t weight its sample for the educational level of the respondent, yielding a higher-than-average number of college-educated respondents that may have produced a higher level of support for Obama.
Selzer said that while college graduates were represented in the survey by 6 more percentage points than 2008 exit poll data, “every education subgroup votes for Obama over Romney.”
“The two groups that deviate most from the 2008 average -- some college and college degree -- are where the race is closest,” she said. “It is simply not true that the higher education of our likely voter sample, compared to the 2008 electorate, is the reason for our higher margin for Obama.”
Pew surveys weight for educational level, Kohut said, “because we know there are imbalances in surveys, that we get more educated people. You want to make it look like the public as much as you can based on census statistics.”
While Obama opened a lead over Romney in the Bloomberg poll, the survey also found the public gives the president low ratings on his handling of unemployment and jobs, which they cited as the most important issue to them.
The poll also showed that among the most enthusiastic voters, Romney led Obama 49-48.
Bloomberg’s poll surveyed 1,002 adults, with a margin of error of plus or minus 3.1 percentage points for the full sample. Questions asked of the 734 likely voters have a margin of error of plus or minus 3.6 percentage points.
The Associated Press-GfK Poll, based on interviews with 1,007 adults nationwide, including 878 registered voters, was conducted June 14-18 by GfK Roper Public Affairs & Corporate Communications. The overall survey sample had a margin of error of plus or minus 4 percentage points, and 4.2 points among registered voters.
The Pew survey of 2,013 adults, including 1,563 registered voters, was conducted June 7-17. The poll has a margin of error of plus or minus 2.5 percentage points for the total sample and 2.9 percentage points for the sample of registered voters.
When a survey says there is a 3.5-percentage-point margin of error with a 95 percent confidence level that means if the poll were conducted 100 times, the data would be within 3 points above or below the reported percentage in 95 of 100 surveys. It also means that in one survey out of 20, the results may exceed the margin of error.
“The poll is extremely different from other results we have been seeing lately,” Nate Silver wrote in his FiveThirtyEight blog for the New York Times, referring to the Bloomberg survey. Silver said he would continue to use the poll as part of his overall forecasting model for the election.
“In this case, the firm conducting the Bloomberg Poll (Selzer & Co.) actually has a good track record, and their previous polls this cycle had not shown especially favorable results to Mr. Obama. So the model uses the poll, just carefully.”
“The most likely possibility is that this poll simply represents a statistical outlier,” Mark Blumenthal, founding editor of Pollster.com, which was acquired in July 2010 by HuffingtonPost where he is senior polling editor.
“Yes, with a 3 percent margin of error, its Obama advantage of 53 to 40 percent is significantly different than the low single-digit lead suggested by the polling averages. However, that margin of error assumes a 95 percent level of confidence, which in simpler language means that one poll estimate in 20 will fall outside the margin of error by chance alone.”
Selzer said, “This is an age of instant analysis, and the reality is we won’t know whether this poll is or is not an outlier for a while. What we do know is that using the same sampling method, same weighting, procedure, same question working in roughly the same order, Obama has opened a lead over Romney when we had them tied in March.”