A crucial question that a researcher addresses is: how many subjects are needed to assure a given probability of detecting a true relationship between exposure and health-related event when it really exists?