Charles on… anything that comes along

I get at least one “survey” - sometimes two or three, depending on the season - from PR companies whose clients think that by asking 12 visitors to their website whether they have a pressing need for Product X, which by chance is sold on that website. Or else it’s “X per cent of people think dirty thoughts during boring meetings”. Er, are you surprised by that? News has to surprise.

Surveys, in general, have become pervasive among the stuff that gets sent out, and pervasive among the stuff that gets binned. Here’s why:
1) too rarely tries to find out stuff we don’t already know
2) doesn’t seek to highlight surprising findings; instead The Client’s desires get pushed upwards
3) not representative, because they have a sample that’s too small, or self-selecting. To be representative of the UK population a survey should have at least 1,000 randomly chosen people. That’s expensive to do. Self-selecting is asking people who come to a website “Do you like websites like this one?” Probably a lot of gambling stats are built on surveys like that.
4) not interesting. Sorry, but that’s the news test. Survey finds that dogs bite men: not surprising. Survey finds that men bite dogs: surprising.

My advice: save your money for something useful.

These posts might be related (the database thinks..):
• Oh yeah? Survey this! (27 August 2004; score: 100.23%)
• Is it because they were too busy answering surveys? (18 October 2004; score: 58.02%)
• But hang on, if they're phishing.. (9 April 2005; score: 51.47%)