ABSTRACT
For the moment let's be an executive of a company that makes toothpaste. Our chemists and biologists have decided from their research that an additive called "Q" should be helpful in reducing tooth decay. We decide to test this idea on the teeth of living human beings. How should we go about it? We could try Q on our own family for six months, observing perhaps one cavity among the four of us. If in the previous six months we had two cavities, is the reduction meaningful? It is not, because there has often been more variation than this in the number of our cavities even when no change was made in our toothpaste. What we need is more people, to "average out" the chance fluctuations.
