******
I have always considered that when one asks the public for an opinion, it is best to know what answer one is seeking. Or said differently, if you are trying to impress your company's management or the Government or some other body (preferably one that is paying you for services rendered) then it is well if you know what that agust group actually wants you to find for them. Suppose you go running to them with the results of your latest poll and it is diametrically opposite what they want. You probably will never be asked to do another "study" for them.
Now how do you avoid this disaster, and it is a disaster indeed if you now are a member of the unemployed? There are several rules to follow:
a) Always determine what the objective is.
b) Being armed with the objective, now design of the question becomes necessary.
c) Remember that in a long and difficult to understand question, the respondent will almost always grasp the last portion of the question as that which they understand and therefore will respond accordingly.
d) If you use the words "whatever, whenever, how-some-much, &c.) you will cause the respondent to take the opposite tact. As example: Do you agree with the following statement? Whenever septic tanks are used, the bay becomes more polluted.
The respondent, wanting a pristine bay will jump to the conclusion that the bay is being polluted by the septic tanks and therefore septic tanks should not be used.
The problem with this statement is the word "whenever", if "when" had been substituted instead, the meaning of the statement would be unambiguous and the answers would more correctly reflect the beliefs of the respondent. So if you want to condemn septic tanks, the "whenever" statement will get you the most support.
e) Now if you have formatted the questionnaire to get the response you want, you must select your audience. A random sampling is perhaps the worst you can do. You must identify a target audience then go after them with a vengeance. Remember that if you are asking about septic tanks, you sure don't want to interview someone that has one. No, it's much better to have someone connected to the sanitary sewer (whatever that is). And a person just beginning to worry about the environment is much better than someone too busy to care or an older person who has heard all this before and just may have actually thought through the problem.
Finding this target pool of respondents is easy. Take exit interviews at a PTA meeting, or in a Mall on a day when you can expect the majority of your respondents to be present.
f) Get a good response rate. A poor response rate will show that you aren't good at surveys so how do you get a response of 40 plus percent. Again identifying the target respondent so that they participate. This is as easy. People have a natural curiosity. Let's assume that you go to a Mall. Take along a number of shills who will appear to be anxious to participate in your survey. Now when just the right profile of a respondent comes along and appears eager to join the crowd and participate. Set the hook and reel 'em in.
I'm sure that the above are old hat to those "professionals" in the survey business, but this exercise is intended to assist those that want to become proficient in this matter. (Or perhaps who have a degree of skepticism in accepting the survey results so often the subject of politicians.)
Here's an example for the recent press: One Poll, Two Very Different Conclusions:
"An opinion poll shows that most of Iraq's people don't want an Islamic government and don't hate the U. S." (Karl Zinsmeister, American Enterprise Magazine* as reported in Sept. 10, 2003 Wall Street Journal)
"Poll: Iraqis Say U. S. to Do More Harm" (Lisa Anderson, Chicago Tribune, as reported in Sept. 11, 2003 Ledger.)
"Asked whether they would like to see the American and British forces leave Iraq in six months, one year or two years or more, 31.6% wanted them out in six months while 24 percent favored one year and 25 percent said it should be in two or more years." LA
"...when asked how long they would like to see American and British forces remain in their country: Six months? One year? Two years or more? Two thirds of those with an opinion urged that the coalition troops should stick around for at least another year."
* Sponsored poll.
****
It is well to remember that:
A pollster can and does phrase questions to get desired the answers.
In reading the questions and resulting answers one can and does interpret the conclusion(s).
The press can and does select what to "report."
February 8, 1999/September 11, 2003
****
Joe Wortham's Home Page, About Joe Wortham, Directory of Web Pages
Questions? Comments? [email protected]