Unstructured data is data that does not have a numerical scale or rating attached to it (or the equivalent). It is the type of data that we most often analyze qualitatively because it is difficult or impossible to analyze with common statistical methods. Unstructured data includes language, visuals and observational data. New qualitative researchers often have difficulty analyzing unstructured data because they need to be trained in the process of recognizing patterns, and they also need to develop an understanding of how to assess the validity of their analysis. Sometimes this type of analysis feels “fuzzy” or invalid because of the fact that it is not statistically supported. A good qualitative analysis, however, will have its own type of validity and rigor. In this case, validity and rigor can come from the following:
- A strong theoretical and methodological framework.
- A detailed coding guide that supports similar results from multiple coders.
- Triangulation of results across multiple methods of analysis.
- Very detailed accounting of results in the reporting of the literature (also known as thick description).
- Verification of the results by the community (if community-based or indigenous methods).
See this video, from the Yale University series on qualitative methods, to learn more about qualitative data analysis (17 mins):