Politics

GOPers may fare better with absence of Research2000/Kos polls

DailyKos founder Markos Moulitsas is asking poll-trackers to remove from their sites what he calls the “bunk” Research 2000 polls he once sponsored, after he expressed doubt on the accuracy of the company’s numbers. And if poll-trackers comply, some Republicans could see a bump in their poll averages on those sites.

For example, a Research 2000 poll commissioned by the liberal blog for the Nevada Senate race from late May to June shows Democratic Sen. Harry Reid leading Republican Sharron Angle by six points. With that poll included in the average, one poll-tracking site, Real Clear Politics, shows Angle with a +1.3 average lead over Reid.

If Real Clear Politics removes the Research 2000 poll showing Reid with a six-point advantage from it’s tracking, Angle’s average would increase from +1.3 to a +5 average, according to the calculation by The Daily Caller.

On Tuesday, Moulitsas said he has reason to believe the polls they commissioned from Research 2000 are not accurate, and that he plans to file suit against the polling company.

“We were defrauded by Research 2000,” Moulitsas wrote on his blog Tuesday, “and while we don’t know if some or all of the data was fabricated or manipulated beyond recognition, we know we can’t trust it.” Research 2000’s president, Del Ali, denies the charges.

A media representative at Real Clear Politics did not immediately return a request for comment on whether the company still plans to use the Research 2000 polls in its tabulations. Yet, another example on their poll-tracking site is the Pennsylvania Senate race between Republican Pat Toomey and Democrat Joe Sestak.

If the Daily Kos/Research 2000 poll is included, Toomey leads with a +1.4 average. If the poll is taken out of the average, his lead over Sestak increases to +2.5.

Daily Kos has published the polls for a year and half. Earlier this month, FiveThirtyEight.com rated Research 2000 as among the least accurate pollsters in predicting election results, leading to Daily Kos terminating their relationship.

If the numbers really are bogus, it’s embarrassing for Moulitsas, who used the numbers to support views on his liberal website. One example is last summer, when he commissioned a poll that said 11 percent in the country does not believe President Obama was born in the United States. Most of those 11 percent, according to the poll, are Republicans from the South.

Moulitsas used the poll to say that, “Once again, Republicans find themselves outside the American mainstream. And reality.”

Isaac Wood, the House races editor for Dr. Larry Sabato’s Crystal Ball, said he’s reserving judgment on Research2000, though said this episode underscores the problem of polling averages and the need to strengthen polling standards to ensure transparency and accuracy.

“We don’t yet know if the allegations of impropriety are true and even if they are, we do not yet know the nature of the problem,” Wood said. “Were they cooking the books to please their clients? Were they merely inventing results to save money on conducting the polls? Are they pushing a particular political agenda that they hoped their polls would boost?”

He continued saying that, “This whole situation should serve as a ‘shame on you’ moment, not only to whatever pollsters end up being discredited, but also to us, the naive public, as our love of poll numbers trumped our duty to think critically.”

E-mail Alex Pappas and follow him on Twitter