Tuesday, March 11, 2014
(Update: Check out Mike Tipping's post for a local take on this PPP mess)
The North Carolina-based polling firm Public Policy Polling has become the target of some scathing criticism from polling analysts following its decision to withhold the results of a survey that showed Colorado lawmaker Angela Giron would likely be recalled by voters on Tuesday.
Even famed statistician Nate Silver got some shots in.
It's difficult to tell who's right in this whole dust-up. After all, if PPP was being sneaky, why did they disclose that they withheld the results? Nobody would have known if they hadn't. On the other hand, their justification for suppressing the results is also puzzling, as Silver and many others have noted.
The flap is of interest to Maine politics. There's a fair amount of polling here and it's fair to say that the media -- including yours truly -- is educating itself on the fly about the impact of surveys coming from all kinds of sources with all kinds of motives.
In Maine there's been at least one example of a partisan polling outfit that withheld results that may not have been favorable to their interests. In 2010, the Maine Center for Public Opinion, a Republican-backed firm conducted a poll of the gubernatorial race and the two congressional races. It released the results of the gubernatorial race and the 1st Congressional District, which showed Republican Dean Scontras gaining on incumbent U.S. Rep. Chellie Pingree, but not the 2nd District where Republican Jason Levesque was thought to be nipping at the heels of U.S. Rep. Mike Michaud, D-Maine.
The reason for not releasing the CD 2 results? "Under-sampling," according to Mark Smith, who did the poll for Maine Center for Public Opinion.
A more cynical explanation: The results were bad news for Levesque. Levesque was eventually defeated by Michaud by nine points. But the thing with polls is that they can garner media attention and determine whether a candidate gets financial support from outside interests. In 2010, Levesque had been close enough to Michaud to receive help from the National Republican Congressional Committee; a poll a month before Election Day showing him trailing Michaud -- a five-term incumbent at the time -- by a significant margin could have killed the prospect of NRCC resources.
Did a similar scenario influence PPP's decision to withhold the Colorado results?
PPP polls here. A lot. The firm has a good track record in Maine, but its results are routinely criticized by Republicans as biased because PPP does a lot of work for Democratic clients. The firm's handling of the Colorado recall poll probably won't help.
There's also a nerdy debate about PPP's methodology and disclosure policy, which I won't go into here. I will, however, link to the email debate between Tom Jensen and Nate Cohn.
Below is a sampling of the Twitter exchange between PPP and Silver. H/T to The Fix over at WaPo for Storifying the exchange.
Steve Mistler covers politics and government for the Portland Press Herald. He spends a lot of time in the hallways of the State House.