My dad says he got surveyed yesterday by the Marist poll and that after asking him a bunch of questions about the primary campaign they asked him some questions about the Super Bowl and then some demographic questions and then a question about whether or not he uses any home health remedies. What do you think that could be about?
« Mandates: Eh? | Main | Rankings »
Strange Polling
01 Feb 2008 12:15 pm
Comments (19)
Obama is undefeated like the Patriots, as he's scored the most delegates from every contest so far.
I'd like to see the correlation between team preference and candidate preference.
Cost of polling.
Firms know that polling respondents are going to answer between x and y number of questions. So, what do you do when you have clients that are only interested in paying for answers to a handful of questions? If you field a poll that only answers 5 questions when your respondents were willing to answer 30, you're wasting contact time with valuable respondents, driving up your cost of polling.
So, companies routinely bundle together polls from different clients --the soda company wants to know if 35 yo men drink their product or not, the political firm wants to know if 35 yo men are voting for their candidate or not. When their callers get a hit, they ask their "questions our various clients have of 35 yo men right now" suite.
The questions probably didn't have aything to do with each other--in fact the polling firm may have been hired by several research companies. What they have in common is that someone out there wanted each of those questions asked of your dad's demographic.
I've been asked random questions, like what is the color of my eyes, that the pollsters say are "for verification purposes," maybe to confirm that they actually did the call and didn't just make up the data. (Although why couldn't they make up the answers to the verification questions too?)
The research company, through previous research, may have found that they can segment people in some meaningful way with that question. They don't care that you use home health remedies, they care that you have unwittingly self-identified yourself as part of a cluster that has other intersting attributes (wants to raise taxes, likes activist judges, exercises a lot, believes things easily, etc.)
Re Matthew's question "a question about whether or not he uses any home health remedies. What do you think that could be about?"
------------
They're push-polling Hillary's Medicare plan.
adding to what anon said, i worked one summer at a polling company and we often had random current events questions unrelated to the main body of the poll which we would then try to sell after the fact to whomever might be interested.
Wasn't there something about the 2004 Bush campaign figuring out what kind of alcohol their voters drank and what kind of cars they drove?
Klug -- yes, but that was matching voter rolls against psychodemographic data available from commercial vendors. Much more sophisticated than tacking on a few questions to the end of a poll.
Sounds suspiciously like some Marketing surveying slipped in.
anon is correct.
Political polling isn't actually very lucrative or very sophisticated (from a statistical/market research technique point of view). Most market research outfits do the political polling for the PR and brand building. Work for actual paying companies pay the bills.
Political polling can be a loss leader to gain name recognition for the polling company in order to attract business clients who pay the actual bills.
anon is right. Go anonymity!
Around these parts, a "home health remedy" is something you smoke using your "water pipe."
to SP: I suppose that the verification question at least theoretically allows for the possibity of spot-checking - some people could be called back and the correctness of the recorded eye-color answer verified.
Were I confronted with a question about home health care remedies my question would be what they meant by that. Are they asking if I have aspirin in the house, or it I wander through the woods seeking medicinal herbs to mix into a mustard plaster? Heck if I know. I tend to annoy pollsters by asking for them to define their terms. If there is no definition provided with their script I give the most neutral "don't know" answer allowed.
I tend to annoy pollsters by asking for them to define their terms.
Or you tend to annoy the call-center employees who are hired to do CATI interviews. I think most polling/market research firms outsource their data collection. The pollsters themselves don't get on the phone.
I agree with Muttrox; this is probably a segmentation question (or some consultant somewhere thinks, a priori, that it would be a useful segmentation variable and it got thrown in to the questionnaire to see whether it matters).
The number of liberals/conservatives of various demographics who might be swayed by home health ads showed during the Super Bowl?
Pretty sure your dad got hit with what they call an "omnibus" suryey. They cobble together clients who can't afford to field a fully sampled survey on their own, each of whom contribute a handful of questions. Economies of scale in action.
Comments closed February 15, 2008.
| 
Heads in the Sand
TALON
Posted by steve duncan | February 1, 2008 12:38 PM