When polls last month showed President Obama pulling ahead, Republicans set up a noisy chorus that the results were flawed. Democrats dismissed the claims.
Now the shoe is on the other foot and Democrats are finding it pinches their toes.
The Obama campaign's chief pollster called a Gallup/USA Today poll released Monday afternoon “an extreme outlier†because it found Mitt Romney ahead over all in battleground states, primarily because women are evenly divided between Mr. Romney and the president.
The pollster, Joel Benenson, cited 13 other polls of battleground states taken since Oct. 4 that show Mr. Obama leading among women.
Mr. Benenson said the poll, the first swing state survey by Gallup this cycle to measure likely voters, had “deep flaws†in how it identified likely voters. The questions the poll asked to determine likely voters are biased, Mr. Benenson wrote in a memo, because they screen out many who are young, renters, minorities a nd urban dwellers, groups that tend to vote more for Democrats. The Gallup survey showed that among registered voters, Mr. Obama leads among women by 53 to 44 percent in the 12 battleground states polled.
Over all, the poll of 1,023 registered voters and 869 likely voters, which was done for USA Today, showed Mr. Romney leading Mr. Obama by five percentage points among likely voters. The Republican nominee “has growing enthusiasm among women to thank,†according to the paper's analysis.
Mr. Benenson wrote that ahead of the 2010 midterm elections “the distortions in Gallup's likely voter screen were exposed, leaving Gallup's survey 9 points off the mark.''
Polling methodology has flared into a hotly partisan issue as the campaigns wrangle over perceptions of who has more support, with just three weeks till Election Day.
When a parade of polls last month showed Mr. Obama pulling away after the Democratic National Convention, some Republican strate gists argued that the results couldn't be accurate and were skewed. Their voices were amplified by a vociferous online conservative community. A Web site, unskewedpolls.com, “corrected†what it viewed as methodological flaws, purporting to show that Mr. Romney was in fact winning. The argument was largely that the polls with Mr. Obama in the lead did not adjust for respondents' party alignment.
Respected polling experts, including Andrew Kohut of the Pew Research Center and Frank Newport, Gallup's editor in chief, argued that party affiliation was a red herring, because it was fluid, so most pollsters did not adjust their samples to reflect it as they do fixed characteristics like ethnicity, sex and age.
Ever since polls started showing Mr. Romney closing the gap in the wake of his first debate with Mr. Obama on Oct. 3, Republicans have gotten much quieter in complaining of flawed polls.
Enter the Obama campaign. Mr. Benenson's memo was distributed by Ben LaBolt, the campaign's press secretary.
This time the disagreement over methodology is not about party affiliation but how Gallup arrived at deciding when a registered voter is likely to actually vote. The survey included Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, New Mexico, North Carolina, Ohio, Pennsylvania, Virginia and Wisconsin.
