One of the most important concepts in all data analysis—whether we’re working with quantitative or qualitative data—is the data matrix. There are two basic kinds of data matrices: profile matrices and proximity matrices. Figure 15.2 shows what these two kinds of matrices look like.

FIGURE 15.2.

Two kinds of matrices: Profiles and Proximities.

Profile Matrices

Across the social sciences, most data analysis is about how properties of things are related to one another. We ask, for example, ‘‘Is the ability to hunt related to the number of wives a man has?’’ ‘‘Is having an independent source of income related to women’s total fertility?” ‘‘Are remittances from labor migrants related to the achievement in school of children left behind?’’ ‘‘Is the per capita gross national product of a nation related to the probability that it will go to war with its neighbors?’’

This is called profile analysis. You start with series of things—units of analysis—and you measure a series of variables for each of those things. This produces a profile matrix, or, simply, a data matrix. A data matrix is a table of cases and their associated variables. Figure 15.2(a) shows the shape of a profile matrix. Each unit of analysis—each row—is profiled by a particular set of measurements on some variables—the columns.

The units of analysis in a profile matrix are often respondents to a questionnaire, but they can be dowry records, folk tales, interview texts, churches—even time periods (1980, 1981, 1982, . . . 1991).