ABSTRACT
When a statistical investigation is undertaken it is often with the purpose of confirming a conjecture or a theory. This is most effectively done if one can obtain data, a model for the data, and a null hypothesis which must be rejected if the conjecture or the theory is to be confirmed.
