Analysis of Categorical Data I: Contingency Tables and the Chi-Square Test
In many cases, the researcher collecting categorical data is most interested in assessing how pairs of categorical variables are related, in particular whether they are independent of one another. Cross-classifications of pairs of categorical variables, that is, two-dimensional contingency tables, are commonly the starting point for such an investigation. Such tables and the associated chi-square test should be familiar to most readers from their introductory statistics course, but for those people whose memories of these topics have become a little faded, the next section will (it is hoped) act as a refresher. Subsequent sections will consider a number of topics dealing with two-dimensional contingency tables not usually encountered in an introductory statistics course.