AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Do Errors Always Cancel Out?
If the judgments are identical or independent of each other, then the noise goes down with the square root of the number of observations. In that classic experiment by Francis Colton, there were more than 1,000 people and they got the weight of the ox on average was two pounds off. But even if there had been a bias, when you average a thousand judgments, noise is gone. So errors cancel out in judgments of the same object. That's where noise disappears. In judgments of different objects, every variability always causes there and they don't cancel out at all. It's a very common misunderstanding, by the way.