ABSTRACT
Everything in the last chapter applied to properties of a population. But we almost always have to deal with samples with limited and varying information. With a random sample of births, how do we examine the relationship between birthweight and infant mortality? What kind of accuracy can we expect and how does it depend on the size of the sample? Does it matter how we select the births to be sampled? We must confront the following issues:
• How do we estimate a measure of association from a random sample from the population of interest?
