Here's something you don't see every day: The liberal blog site Daily Kos is accusing its own polling firm, Research 2000, of "fabricating or manipulating" poll results in such a way to make news and shape races. Kos founder Markos Moulitsas wrote on Tuesday:
“While the investigation didn't look at all of Research 2000 polling conducted for us, fact is I no longer have any confidence in any of it, and neither should anyone else. I ask that all poll tracking sites remove any Research 2000 polls commissioned by us from their databases. I hereby renounce any post we've written based exclusively on Research 2000 polling.”
Research2000 president Del Ali is denying any wrongdoing or number-fixing.
Politico's report this morning on this matter focuses on the firm's polls showing Lt. Gov. Bill Halter up over Ark. Sen. Blanche Lincoln in the June run-off in which she won and it seemed like a surprise. And in two pieces on the matter, Politico also pointed out curiosities in two other races, an Alabama Democratic primary for governor that showed one candidate up by 7 points a week before he lost by 24 points and another showing Carly Fiorina losing by 15 points a couple weeks before she won by 34 points to capture the Republican nomination for Senate in California.
But none of these reports noted the Nevada twist, the fact that Research2000 was also responsible for the widely disseminated and reported poll that appeared the week before the GOP primary that showed Sen. Harry Reid leading Sue Lowden, Danny Tarkanian and Sharron Angle. Here's Kos' raw data, and here's some of the headlines from (in order) the Huffington Post, Talking Points Memo and the Examiner:
The only news outlet to examine this that I could find was the Daily Caller, which noted that it's that Daily Kos poll that is keeping Reid appearing a lot closer than he may actually be to Sharron Angle in popular Real Clear Politics's tracking poll average. She's presently shown as ahead in by about 1.3 points, but remove the Kos/Research2000 poll and she's suddenly got an average 5 percentage point lead. That shows that just one poll with out-of-the-mainstream results can have a pretty dramatic impact and can make a race seem a lot closer than it is. That influences donors and press coverage, of course.
Interestingly, the Vegas media appears not to have bit on this one even though all have reported on Research2000 results extensively in the past. The Las Vegas Sun didn't note the Reid-leading element of the poll in its Angle-surging piece and the Review-Journal blog ignored the Reid-leading bit in referring to the poll's show of Brian Sandoval's growing strength over Gov. Jim Gibbons and Democrat Rory Reid. I can't be sure because his archives aren't online or searchable, but I have it in the back of my mind that Jon Ralston was dismissive in his daily email blasts. Nevada News Bureau noted the poll but did not analyze it.
To be fair, on some of the other races the Research2000 poll turned out pretty close to accurate. Sandoval crushed Gibbons 55-27, and the poll had predicted 48-27. Angle crushed Lowden and Tarkanian 40-26-23, and their poll had predicted it would be 34-25-24. Those variations can easily be explained by surging momentum in the closing days of the campaign that lifted Angle and Sandoval a little higher yet.
Meanwhile, the Huffington Post has now attached this startling notice to the top of their coverage of the poll:
EDITOR'S NOTE: The story below includes references to polling conducted by the firm Research 2000. The reliability and accuracy of Research 2000's polling has since been called into serious question by a report published in June 2010 by a group of statistical analysts.
Ouch.
