Sample Size and Power
In this chapter, in addition to the formal method of computing the sample size and power, we will discuss ways to compute sample size and power based on normal approximation through a simplified and conservative approach. As pointed out previously in assessing agreement, the traditional null and alternative hypotheses should be reversed. The conventional rejection region is now the region of declaring agreement, and is usually one-sided. Asymptotic power and sample size calculations will proceed by this principle.