ABSTRACT
This chapter describes classic and modern methods for addressing the issue stemming from the notion of a sampling distribution. The notion of a sampling distribution provides a useful perspective on the relative merits of the measures of location. It also provides a basis for understanding and appreciating a modern technique called the bootstrap. Sampling distributions help foster and understanding of the deleterious effects of nonnormality and the practical benefits of numerous methods developed during the last half century. The chapter illustrates how to compute probabilities associated with the sample mean when sampling is from a normal distribution and the standard deviation, σ, is known. But typically σ is not known, so a practical concern is finding a reasonably satisfactory method for dealing with this issue. The chapter describes a classic method for addressing this problem that was derived by William Gosset about a century ago.
