Qualitative researchers often want to summarize data that involve the intersection of multiple variables, according to sociologist Earl Babbie, author of “The Practice of Social Research.” For example, political ideology represents a single dimension. However, political ideology in terms of domestic policy and foreign policy issues involves more than one dimension. In research situations involving the intersection of multiple variables, Babbie writes that analysts create typologies for categorizing the data. He cites as an example a study of the political orientations of newspaper editorials. Under this basic typology, editorial stances could be categorized as liberal or conservative on domestic and foreign policy, liberal on domestic and conservative on foreign, or vice versa.
Researchers constructing a single measure of multiple variables often end up creating a typology, according to Babbie. This occurs when researchers find that items they originally believed represented a single variable actually appear to represent two or more. Typology analysis requires careful review of the raw data before deciding in which cell of a typological matrix each data point should be classified. The categories used in a typology should be mutually exclusive to reduce ambiguity in classifying data. However, Don Ratcliff, a professor of Christian education in Illinois, points out on his Qualitative Research website that the categories often are not mutually exclusive.
Typological analysis in qualitative research serves three functions: descriptive, classificatory and explanatory, according to Colin Elman of Arizona State University. The descriptive function defines and describes the various types, distinguishing, for example, a liberal orientation from a conservative one. The classificatory function assigns cases in the raw qualitative data, classifying them under one of the categories in a given typology. The explanatory function places data in relevant categories, allowing researchers to make comparisons and assess whether the data are consistent with relevant theories.
Babbie writes that typological qualitative analysis works best when the typology is an independent variable. For example, an analyst could employ a typology to study the percentages of newspapers in each cell of the liberal-conservative, domestic-foreign policy matrix that endorse Democratic or Republican candidates. Typology analysis is more difficult when the typology serves as a dependent variable. Babbie points out that this approach would make it difficult to analyze why a given newspaper falls into a particular category in a typology.