Development of coding categories Qualitative data rapidly cumulate, and even with regular processing and summarizing, it is easy to get overwhelmed. The material is unstructured and difficult to deal with. Coding provides a solution. A code is a symbol applied to a section of text to classify or categorize it. Codes are typically related to research questions, concepts and themes. They are retrieval and organizing devices that allow you to find and then collect together all instances of a particular kind.
Miles and Huberman (1994) distinguish between first- and second-level coding. First-level coding is concerned with attaching labels to groups of words. Second-level or pattern coding groups the initial codes into a smaller number of themes or pattern. The development of pattern codes is an integral part of first-level coding, where you need to be continually asking yourself, ‘what seems to go with what? And elaborating on and checking these hunches. You will probably start with a very small number of potential patterns, modify and add to them during the course of analysis, and finally be left with a small number once more as various ‘runners’ are disconfirmed by the data. The work that you do in creating these codes is central to developing an understanding of your data. It lays the foundation for your subsequent analysis. Further on coding are given below in the section on grounded theory analysis (p. 402), where a somewhat different terminology is used.
Memoing A memo can be anything that occurs to you during the project and its analysis. A particularly important type gives ideas about codes and their relationships as strike you while coding. The length is not important; they are simply attempts either to link data together, or to suggest that a particular piece of data falls within a more general category. They should be adequately labeled so that they can be sorted and retrieved. Memoing is a useful means of capturing ideas views and intuitions at all stages of the data analysis process.