The Analysis of Proportions
This chapter is concerned with testing ideas about proportions. It utilizes the analogy between the analysis of proportions and the analysis of sample means. The simplest example of the importance of estimation from proportions is the need for a seed grower to estimate the germination percentage of his seeds. With data consisting of frequencies, we can produce a test to compare the proportions of the two possible outcomes for a number of situations. Before going on to the appropriate method of analysis, it is interesting to consider a few other situations that may give rise to the same kind of data. Although it has been convenient to talk about contingency tables in terms of situations and outcomes, the idea of contingency tables is more general than this implies. The concepts of contingency tables can be extended to more than two classifications although clear thinking is required to identify the hypothesis being tested.