ABSTRACT
This chapter discusses the analysis of categorical data—specifically, data that represent the counts for each category of a multinomial experiment. The statistic used for most of the inferences is one that possesses, approximately, the familiar chi-square distribution. The investigation examined data on a random sample of 437 conflict alerts. Each alert was first classified as a “true” or “false” alert. In the analysis of contingency table data, one or more of the categories may contain an insufficient number of observations. To investigate the relationship between the magnetic age of Earth’s crust on the ocean floor and the probability of finding manganese nodules in that location, crust specimens were selected from seven magnetic age locations and the percentage of specimens containing manganese nodules was recorded for each. To compare the proportions of defective impellers produced by three production lines, a quality control engineer randomly sampled 500 impellers from each line.
