Thursday, November 07, 2013

REALLY? WE'RE "UNSKEWING" POLLS AGAIN?
(updated)


A Cuccinelli operative thinks polls are a liberal plot -- all of them, apparently:
Multiple public polls put Terry McAuliffe ahead by double digits in the closing weeks of the Virginia governor’s race. But when the results were tallied, he won by a mere 2.5 points.

... Already, many operatives are asking why the numbers were so skewed -- and whether they helped shape the outcome.

Republican Ken Cuccinelli's chief strategist said Wednesday that the polls were worse than wrong: He believes they badly hobbled his candidate's viability and made a late comeback more difficult.

"Public polling is essentially a suppression tool used to demoralize our voters and dry up money, especially the latter," Chris LaCivita told POLITICO.
Yeah, right -- public polling is a Democratic suppression tool. Those polls in New Jersey showing Chris Christie leading by as much as 43 points over his Democratic opponent? Big liberal plot.

And in your race, Chris, do you recall which poll gave McAuliffe the biggest lead, 17 points? It was a Rasmussen poll. Rasmussen is one of yours, not one of ours.

And you know who else was in on the liberal plot, Chris? Your own campaign:
Republican internals also had Cuccinelli down much further than Democratic internals.

At the end of September, even before the shutdown started, Cuccinelli trailed in the high single digits in his own survey. A poll for the Republican Governors Association conducted in the middle of October, just before the shutdown ended, had Cuccinelli down 10.

The Cuccinelli campaign didn't like the numbers, so they ran their own poll around Oct. 18 -- after the shutdown ended -- and discovered that Cuccinelli was down 8 points among those most likely to vote. Worse, he was only winning whites by 3 points and had locked down just 57 percent of soft Republicans.

These numbers persuaded the RGA to not pour much beyond the $8-plus million investment already made....
Well, this talk of a voter-suppression plot is what you'd expect from a guy like Chris LaCivita:
LaCivita at one time worked for the Republican political consulting firm DCI Group, which had close ties ... to ... Swift Boat Veterans for Truth. During the 2004 US Presidential campaign, LaCivita served as consultant and principal media advisor to the Swift Boat Veterans, writing and producing the group's memorable (and controversial) commercials in association with Rick Reed....

Interviewed in March, 2007, LaCivita told reporters that the GOP could silence the press uproar over the sudden dismissal of eight US Attorneys by promoting stories about the most extreme among anti-Bush activists. "When are we going to make it about Code Pink and the rest of the liberal weirdos controlling the Democrat agenda?" he asked.

During the 2008 presidential campaign, LaCivita and Tony Feather launched a new 501(c)4 issues advocacy group, the American Issues Project, after a series of meetings with other Swift Boat donors. In August 2008, AIP began airing ads in battleground states seeking to raise questions about Democratic Party presidential nominee Senator Barack Obama's ties to former student radical William Ayers....
If you've spent a significant part of your career treating voters the way Barnum treated the rubes, you might assume everyone else is trying the rig the game, too.

*****

I don't know why the polls in this race were inaccurate. It looks as every pollster was polling the wrong sorts of white voters:
"Our sample of white voters was too liberal," [Tom Jensen of Public Policy Polling] said. "We had Cuccinelli winning whites by about 10. He won white voters by closer to 20 points. That kind of thing pretty much happened in all of the polls."
I have a theory. Maybe if you run a teabagger like Cuccinelli in a swing state, and his base is made up of people who think all polls are LIES from the LIE-beral media, those base voters will disproportionately refuse to do the surveys, and pollsters will turn to other demographically similar but ideologically less conservative voters -- which would mean, Chris, that the results in this race were being skewed by your own voters. If I'm right, your base was demoralizing itself by not participating.

I don't know if that's true, but it wouldn't surprise me.

****

UPDATE: Read this thread's first comment, which is from Victor, who has some experience in these matters. He did some polling work a couple of years ago and says that right-wing respondents really did express reluctance to cooperate if the pollster wasn't right-wing.

11 comments:

Victor said...

More people than I expected were poll-savvy when I did some part-time work calling for a Republican polling outfit, for almost a year, before the '10 election.

Before answering my questions, they'd ask who had commissioned the polls - and, if it was someone like John McLaughlin, or Rasmussen, they'd answer.

If not, then I often got turned down, even though the poll was commissioned by some Conservative group or politician or other.

Certainly, not everyone - but enough so that I was taken by surprise by the question, at the beginning.

When I answered polls over the years, it had never occurred to me to ask who had commissioned the polls.

Maybe that was something circulated in Conservative circles, because I'd never heard of, before answering, asking who was commissioning the polling from any Liberal groups.

So, from that, I suspect that Republicans do self-skew polls - their own, and others.

And I won't even mention the number of people who hate being called, and then openly lie, to skew the poll on purpose.

Never Ben Better said...

Hmmmm..... interesting point, Victor. I've never asked who's commissioned any polls I've answered, but I can at times tell whose agenda it serves by how the questions are phrased.

Of course, I don't answer polls now anyway, since all my phones have caller ID and I refuse to pick up anything with an 800 prefix or "unknown caller".

Steve M. said...

Thanks, Victor -- I added an update directing readers to your comment. Very interesting to learn (from someone who knows) that some right-wingers actually do this.

Anonymous said...

I assumed the lopsided polling ended up hurting McAuliffe by dampening the sense of urgency that off-year Democrats tend to need to turn out.

Victor said...

You're welcome.

The first time someone asked me who commissioned the poll, you could have blown me off the plastic stool with a feather.

I always checked who did, for myself, to see just how stupid and dishonest the poll I was calling about that day was.

And to give credit where credit is due, many of the polls we did were very straight-forward, and unbiased (those weren't the McLoughlin or Rasmussen ones, though - those were biased trash, for the most part).

But the ones that weren't unbiased - hoo-boy!!!

I've written about them before, so I won't bore people with the details.
Especially, the 'Push-polls!"
Vile.
Really, really, vile.

aimai said...

I've definitely asked who was sponsoring a poll, or privately speculated to myself and answered accordingly and combatively. I don't feel any obligation to help right wing politicians figure out what to do to win an election, and I prefer to do my best to spread disinformation and even attack the premises of the poll. And that goes double for polls that purport to be about one thing but arre really about some other corporate or political goal.

I'm rather sympathetic to the possibility that as the public has become more savvy about how they are manipulated by polls and by politicians that they push back and try to take some control over the process. Its very annoying and disheartening to be bombarded with polls that are used to tell you how " normal" or "average" people think about a topic when you feel so strongly that you have right or principle on your side. And I don't think that feeling is at all limited to right wingers.

Victor said...

Of course, aimai, YOU would! ;-)

Many of the rest of us aren't as smart as you are.

aimai said...

I also have very funny conversations with fund raisers. But I think I'm just a difficult person.

The New York Crank said...

As the ex-husband (for 25 years) of a notable and highly respected market and opinion research maven, I can tell you with some assurance that the answer-takers for most legitimate polls have no idea who the sponsor is. They are sub-subcontractors, hired by a subcontractor, who only knows what research firm hired them, and not who the ultimate client is. One reason for doing this is to keep the answers (and the questioners) bias-free. whether you're asking about political candidates or pitted prunes.

Sorry, Steve and Victor. You have a neat theory, but it may not fit the facts.

Crankily yours,
The New York Crank

Steve M. said...

I only did survey work very briefly one summer, a long time ago (1980), but I don't recall that being privileged information.

Examinator said...

As an ex professional marketeer my experience is similar to NY Crank.

Often the questioners simply don't know or care.
That doesn't mean that some or a lot of party polling isn't conducted by people who know who the client is
As for answering the questions I am sort of like Aimai as in I'm deliberately unhelpful. I play games/ waste their time and ask for lots of clarifications, constantly putting overly lengthy qualifications to answers. In short use anti sales techniques ensuring that my answers will skew their error factor.Bounce answers all over the spectrum.
In fact when surveys ring home my wife hand them over to me with the comment "for you... Play Time !".