Wyatt Earp would have holstered his Nielsen
Erwin Ephron rounds up some of the everyday fallacies caused by sampling error
'It wasn't his lightning-fast draw or his uncanny accuracy that impressed the town. It was his cool judgement about when to holster his Colt and let common sense handle the problem.'
About Wyatt Earp
Recently, I was at the TV Advertising Bureau talking shop. I suggested buyers could use household ratings to post-evaluate, because there is good evidence that variations in household ratings explain most of the variation in viewer ratings. One researcher fiercely disagreed. Local was different, he said. He had examined trends in lots of weekly data, and the viewer and household ratings often went in different directions. I suggested this was sampling error. 'Yes', he replied, 'But it's still Nielsen - the real world.'