I have been clear all along on whom I think will win the U.S. Senate race between Chris Murphy and Linda McMahon. I’m not going to go there again.
I do want to discuss the McMahon’s campaign’s reaction to the recent slew of negative polls for her. It’s indicative of a terrible habit among campaign flaks to attack polls. If you are going to adopt that strategy, you had better not miss. The McMahon campaign did.
Last week, four polls, conducted by Siena University, Rasmussen Reports, Public Policy Polling, and UConn respectively, all showed leads for Chris Murphy ranging from 2 to 6 points. In response, the McMahon campaign issued a statement that indicated a fundamental misunderstanding — or deliberate obfuscation — of the composition of the Connecticut electorate.
The campaign’s primary objection is that the polls’ samples did not reflect the fact that “independents represent the largest voting bloc in the state.”
At first glance, this seems an unobjectionable claim. Connecticut’s most recent registration is as follows: Democrats are 37 percent of registered voters, Republicans are 21 percent of registered voters, and unaffiliated are 42 percent of voters.
The McMahon campaign took particular issue with the composition of the UConn poll for oversampling Democrats. That’s true. It did. The UConn poll’s sample was 50 percent Democrat. But it also oversamples Republicans at 34 percent. Each party gained 13 percent over its registration numbers and thus each party was “overrepresented” equally.
That doesn’t mean, however, that the poll undervalues independents. The truth is that what people describe to pollsters as their party affiliation is not necessarily what their actual registration reflects. Just because the largest bloc of registered voters in the state is unaffiliated does not mean the largest bloc of voters think of themselves as independents. Exit polls of voters in even years going back to 2004 show that in every year Democrats have been the plurality party in terms of self-identification.
In addition, the Democratic nominee for president won “independents” in 2004 by 17 points and in 2008 by 18 points. Simply put, true partisans are clearly hiding out somewhat in the “independent label.”
It is also true that partisans participate in elections at higher rates. Much of the 5-point advantage for unaffiliated voters over Democratic voters in registration is likely to evaporate at the polls. Since both parties equally take advantage of this movement toward partisanship in actual voting having polls include more party-affiliated respondents doesn’t skew the results at all. It just reflects who’s more likely to be voting.
The McMahon campaign tried another tack when the Quinnipiac poll was released earlier today. After the UConn Poll, she complained bitterly about using self-identification not registration to sort the parties. The campaign insisted on registration. Now for this round of spinning registration is terrible. The self identified exit is what matters, and Republicans are undersampled when of course in this poll Dems are too. I suggest if you are going to have a complaint don’t make conflicting ones each time.
The campaign is also missing on the big things.
First, saying only the Q-poll and the Democratic Senatorial Campaign Committee poll had her trailing by a few points. Both margins are six, which is not actually a few, and four earlier polls also had her trailing. She then cited two polls that showed her doing better. The key thing about her “best” polls is that they both have Obama leading by fewer than 10 points in Connecticut.
This could be right but it doesn’t square with what others are finding, nor does it square with the race being tied nationally (when George W. Bush won by 2.5 percent nationally, John Kerry won Connecticut by 10.4 percent), nor, quite frankly with McMahon running an ad basically hugging Barack Obama.
There’s another problematic number for McMahon in these polls — the percentage of support she’s getting. In all the polls, McMahon is averaging about 43 percent. Going into the election in 2010, she had an average of 44 percent in the polls and got 43.3 percent of the vote. She trailed by 8.7 in the polling average; she ended up losing by 11.8. Murphy doesn’t yet have U.S. Sen. Richard Blumenthal’s numbers, but he is holding McMahon to her old number. Most undecided voters now were almost certainly Blumenthal voters in 2010. She hasn’t closed the deal and in Blue Connecticut, if you haven’t closed the deal yet, when will you?
Here’s the key point. If the McMahon campaign wants to defend its claims that the polls are inaccurate, it has a simple way to prove it. With its war chest of $77 million, the campaign has clearly conducted hundreds of polls over the last three years. The campaign probably polls every single day. If her campaign was ahead or closer, McMahon could release the results of her polls. If she doesn’t, it is clear that Murphy has the lead reported in public polls. McMahon can either release a poll to dispute it or she can stop spinning.
Jason Paul is a Connecticut Democratic political operative from West Hartford and a University of Connecticut Law School student.
