Friday, September 01, 2006
About Time To Say It Again
This article touches on a topic I have written about many times here -- junk surveys. Howard Kurtz, Washington Post media critic, takes after journalists who unthinkingly print surveys that seem to contain surprising or alarming data. Kurtz feels journalists ought to know better by now. What he doesn't say is that PR practitioners ought to know better by now. No one wins when one issues inaccurate surveys. There might be short-term advantage, but in the long term, one destroys credibility.
It's time to go back to statistics 101 and to start over with the concepts of universal populations, random samples, careful question construction and accurate correlations. And once again, internet surveys ARE NOT random unless they draw from a carefully constructed audience. There are only a few polling companies that do this, as far as I know.
It does the PR business no good to engage in fraudulent surveying. We are correctly accused of "spinning" when we do. Resist the notion of doing a survey unless you are willing to spend the time and money to do one correctly. If you don't know how, then engage the services of a professional polling company to do it for you. Be aware. It is difficult to do a statistically accurate survey -- and often expensive.
I'm sure I'm going to write about this issue many more times before this blog is retired -- and so am I. It would be nice to see even a small bit of progress toward honesty in polling, but I have no expectations it will occur. It is part of human nature to take the easy way out.
It's time to go back to statistics 101 and to start over with the concepts of universal populations, random samples, careful question construction and accurate correlations. And once again, internet surveys ARE NOT random unless they draw from a carefully constructed audience. There are only a few polling companies that do this, as far as I know.
It does the PR business no good to engage in fraudulent surveying. We are correctly accused of "spinning" when we do. Resist the notion of doing a survey unless you are willing to spend the time and money to do one correctly. If you don't know how, then engage the services of a professional polling company to do it for you. Be aware. It is difficult to do a statistically accurate survey -- and often expensive.
I'm sure I'm going to write about this issue many more times before this blog is retired -- and so am I. It would be nice to see even a small bit of progress toward honesty in polling, but I have no expectations it will occur. It is part of human nature to take the easy way out.
Comments:
Post a Comment