|
A favorite pastime on DU is to speculate about what is going on in the minds of the public. Hey, guess what, people—if we were to put 10% as much effort into a little collective, ongoing surveying endeavor as we do into our endless and pointless speculations, we could actually find some stuff out.
It is within our collective power to do accurate and useful scientific survey work.
Here's how it might go:
First, a nuclear group comes up with a survey idea, and a small survey with a limited number of items is constructed.
Second, these items are distributed to a carefully selected fairly large group of volunteer surveyors.
Third, each volunteer makes a few phone calls, using a predetermined set of random procedures, and asks the survey questions.
Fourth, the results of the survey are returned to a designated statistical analyst (at this moment, me), tabulated, and analyzed. The results are then distributed as appropriate.
Because lots of folk find designing a questionnaire more fun than making phone calls, anyone participating in the design will also be asked to participate in the data collection.
Many questions can be answered with sufficient accuracy using relatively few respondents. For example, we might want to know how many people have heard of Scooter Libby and can tell us a little about him. With only 100 respondents, we could get results with a standard error of about 10%. If we have 20 volunteers, that would only require data collection from 5 respondents apiece.
Here is the first question for the DU Scientific Poll:
Have you heard of Michael Moore's movie SiCKO? (Y) (N) (If “NO,” terminate) Have you seen it? (Y) (N) Do you intend to see it? (Y) (N) Regardless of whether you have seen SiCKO, from what you know about the movie, Would you say that you strongly agree, somewhat agree, somewhat disagree, strongly disagree or are not sure if SiCKO presents an accurate and fair picture of health care in America?
Would you be willing to serve as a designer and volunteer data collector? Please PM me if you can help.
|