Monday, June 05, 2006

Bad Data 

The legacy of lousy data is embarrassment and loss of reputation. Just ask the American Medical Association. It did what unfortunately far too many companies and PR practitioners still do -- conducted a non-random online survey. It then trumpeted results only to learn its data was wrong.

Its irksome to know that an organization dedicated to the science of health could be so unscientific in its approach. On the other hand, with apologies and abundant discussion of the correction, the AMA has learned a lesson. What about others who continue to use nonrandom methods -- and worse, to defend the approach?

Polling has been suspect for quite some time because of an inability to get random results whether through phone, mail, e-mail or direct contact. People don't cooperate anymore. But randomness is an essential element to projecting a universe. Without it, all other manipulations, summations and cross-tabbing have no validity. Anyone who has taken a basis statistics course knows this -- or should know it.

So why does junk data keep reaching the public and who is being dishonest?


Post a Comment

This page is powered by Blogger. Isn't yours?