Qualitative Data Analysis

Qualitative data analysis (QDA) includes the strategies for summarizing and determining the meaning of qualitative research data included in this analysis. During QDA, the researcher explores the data and searches for patterns, themes, relationships, and significant statements. The researcher prepares the qualitative data, such as interviews and field notes, for analysis by transcribing verbal or written data into text data. Visual data also are important in qualitative research. Researchers are increasingly importing qualitative data into software designed specifically for QDA.

To obtain a general sense of the material, a researcher usually reads through the data several times; during this time, the researcher might jot down notes. The researcher must continually decide whether more data are needed because analysis in qualitative research often takes the form of interim analysis, which means that data collection and analysis are conducted cyclically during the research study.

Central to QDA are the activities of segmenting and coding. The researcher locates meaningful segments of data (e.g., text) and assigns a code (e.g., a symbol, category, concept, or short phrase) to label each segment. The researcher should choose codes to represent text segments that clearly describe people, places, events, activities, concepts, and processes in order to help the audience understand the study's central phenomenon. Types of codes include inductive codes (generated by the researcher during examination of the data), a priori codes (developed before examining the data), co-occurring codes (overlapping codes), and face sheet codes (codes applying to a complete document or case). Sometimes enumeration of codes is conducted, in which the frequency of occurrence of codes is counted.

During QDA, many types of “relationships” might be found in the data. One's codes/categories might naturally fall into a hierarchy of concepts, which would form a hierarchical relationship. Typologies can be cons tructed from codes. Codes/categories can be depicted in diagrams and matrices to help the analyst and reader “see” relationships in the data. Spradley (1979) suggests searching for the following types of semantic relationships: strict inclusion, spatial, cause-effect, rationale, location for action, function, means-end, sequence, and attribution.

The content of qualitative data (e.g., codes, categories, descriptive phrases) often is collapsed into a few (five or six) themes. A relatively ordinary theme would be one that the researcher might expect to find throughout the course of his or her study (e.g., “bullying in school”). An unexpected theme would be one that comes as a surprise during the study (e.g., “tolerance of bullying in school”). Social science themes are those that reflect some social construct (e.g., “cliques in school”). Hard-to-classify themes contain ideas that overlap with several themes or don't easily fit into one theme (e.g., “discipline problems with girls”). Major and minor subthemes are those representing major ideas and minor ideas, respectively, in the researcher's database. Most qualitative researchers rely on the logic of induction and therefore emphasize that the themes should emerge out of the data rather than being forced onto the data. A researcher can be assured of a good qualitative analysis when multiple perspectives are represented, good quotable material is identified, all the research questions have been covered, and responses from several people to each question can be tied together to represent the different perspectives. For more information, see Spradley (1979).

...