Specious Survey Watch: Having devoted several pages of the newsletter recently to the subject of specious surveys, I wouldn’t let last week’s excellent column by Wall Street Journal “Numbers Guy” Carl Bialik pass without comment, since it includes what appears to be a prime example of the phenomenon.
Simply put, the Wine & Spirits Wholesales of America took the results of a deeply-flawed survey into online alcohol purchasing among teens (the sample being deliberately biased to over-estimate the problem) and exaggerated them (miraculously turning the fact that 2.1 said they had bought alcohol online into the conclusion that half a million kids bought alcohol online).
Once again, the survey was swallowed hook, line and sinker by gullible reporters, despite the fact that the sponsoring organization clearly had a stake in the scary outcome. Some will argue that if the media are dumb enough to buy this kind of nonsense, PR people are going to continue churning it out. I say it just makes everyone more cynical about the manipulation of data and the credibility of corporate America.
Simply put, the Wine & Spirits Wholesales of America took the results of a deeply-flawed survey into online alcohol purchasing among teens (the sample being deliberately biased to over-estimate the problem) and exaggerated them (miraculously turning the fact that 2.1 said they had bought alcohol online into the conclusion that half a million kids bought alcohol online).
Once again, the survey was swallowed hook, line and sinker by gullible reporters, despite the fact that the sponsoring organization clearly had a stake in the scary outcome. Some will argue that if the media are dumb enough to buy this kind of nonsense, PR people are going to continue churning it out. I say it just makes everyone more cynical about the manipulation of data and the credibility of corporate America.
posted by Paul A. Holmes @ 1:30 PM
2 Comments:
At 12:52 PM,
Paul A. Holmes said…
I suspect the attitude is that fact-checking might undermine a good story. If they buy the survey findings, it makes a great headline. The less frightening reality is also less newsworthy.
At 8:32 PM,
Anonymous said…
If a laywer tells an executive that a certain action will get them in trouble, they'll often stop. If a PR person does this, the executive will...actually, few PR people ever do this.
This post actually highlights two problems. The first is that few PR people have even a basic knowledge of statistics, so they are not able to counsel on the validity of a methodology or statistical significance. If statistics are such powerful tools of persuasion, shouldn't we require a basic knowledge of this tool.
The second is the result of the first. Executives are aware that PR executives often lack the above knowledge. So why would they take counsel on the topic (not that it's often given - we're usually too focused on the potential article).
Post a Comment
<< Home