Thursday, September 12, 2013



Ever been asked to answer a “few simple question” for a poll?

Back in the day, when Hector was a pup and I was a “print journalist,” I had a once-a-week assignment to go out onto the sidewalks of beautiful downtown Harrisburg PA to ask random people The Question of the Week; always something topical that my boss (“Slim” Milliron) or I contrived.

Never mind if the “feels like” temperature was 0F or that snow was blowing at 30 mph; if it was Thursday afternoon, I had to hit the bricks to find three people willing to (a) answer The Question and (b) allow my tag-along (and equally suffering) photographer to shoot a mug shot of those willing to answer The Question.

The thing that prompts this exercise is a snippet on the Advisen FPN email I receive 5 days-a-week than reads:

PCI Study Finds Americans Support Federal Role in Terrorism Insurance Market

A majority of likely American voters favor a federal role in protecting against losses related to a terrorist attack, representatives of the Property Casualty Insurers Association of America said during a Sept. 11 conference call. The study conducted by GS Strategy Group on behalf of PCI found 90% of respondents supported having the federal government play...

I never found the article on the WWW, but I have to wonder

  • What constitutes a “likely American voter”?
  • How many people participated in the survey?
  • Where was the survey given?
  • How were the questions asked?
  • How were the respondents selected?
  • Were the responders anonymous?

  • How were the questions phrased?

Any pollster worth his or her salt can design a survey/questionnaire to elicit the responses the pollster or the pollster’s client desires.

It’s not WHAT you say, but HOW you say it. Old truth from my junior high days. Are you still beating your spouse?

In the poll cited by Advisen, I wonder if the question was:

    Do you think the government should pay for terrorism insurance?

But consider

    Are you willing to pay extra taxes for government terrorism insurance?

Phrased the first way, the cost of the insurance is borne by “the government,” not by the individual taxpayer, never mind that taxes will fund the insurance.

Phrased the second way, Joe Taxpayer plainly sees that he will pay the cost.

Question: Should the minimum wage be raised?

If you ask someone my age, whose first post-service job paid the minimum wage of $1, the response likely will be “No.”

If you ask a young adult with dreams of family and home, the response probably will be “Yes.”

Rephrase the question to be: Are you willing to pay an extra (??) for a hamburger or for a dozen eggs or a gallon of fuel? Smokers are the only people, as a group, who will, albeit reluctantly, bite the bullet and pay whatever the price for their pleasure.

About the only surveys to which I respond are those asking how I liked the treatment at this place or that.

Part of the problem with surveys/questionnaires is that rarely does the responder have the ability to do more than select a pollster-defined answer (Yes/No, 0 through 10, This vs. That). Reality can’t be captured with pre-defined answers; on the other hand, the pollsters would have a nightmare trying to compile verbose responses.

My bottom line: If you are planning to tell me that n percentage of y group prefer z action/condition, then also give me the details on how the information was gathered and compiled.

Questionnaires/surveys, like statistics, can be manipulated to support any desired position/point of view; the best those of us lacking input into the survey/statistic development is to demand all the information about the questions and the responders.


If I wrote it, you may quote it.

No comments: