So far we have focused on a fairly simple scenario—the number of errors at a hospital. It is not something that requires a graph or a table to understand. Let us now switch to a different scenario—the admission rates of men and women to different departments at a university. We now have three variables—admit/reject, man/woman, and department name. We will look at several simple graphical displays that illustrate what is going on with the data. Then, we will look at time series data and introduce the concept of indexing.
After completing this chapter, you will be able to:
Let us start with a fragment from a university database; we will also use it later in additional examples (these data are from six UC Berkeley graduate departments for the fall of 1973; they are discussed in Freedman, Pisani, Purves, and Adhikari, Statistics, W. W. Norton).
The subjects in this example were applicants to graduate school. The variables are the gender of the applicant, the department to which they applied—which is alphabetically coded—and whether the applicant was admitted. For this observational study, the data were gathered by simply reading the existing applications.
The most common and ...