How to Code an Interview

Making sense of interviews, focus groups and responses to questionnaires primarily involves the researcher's judgements and categorizations.

The process for discovering themes from a text is called "coding." A new researcher may feel overwhelmed at how to make sense of all these words. But the process is much like doing an inductive study of a Bible passage.

It is possible to use a variety of tools for the coding process. The tools you use are only aids. People have effectively used paper and pencil, Microsoft Word or Excel as well as specialty software like QDA Miner. Therefore the choice of tools is largely how useful you find them.

Whatever tool you use, the process is quite similar. The following are suggestion for coding.:

1. Edit the text. Correcting spelling and grammar at this point is relatively easy in Word before you start adding codes. Eventually you will want correct language usage in any quotes or appendices you have.
2. Make sure you can identify who said what. It makes a difference in your analysis if the same words, phrases or concepts are frequently mentioned by the same person or by different people.
3. You started with some basic questions in mind, if not explicitly written down and asked. This should be your first level or organization. Sometimes in an interview the respondent may have said things after the second or third question that really related most to the first question.
4. Read the responses of the first few respondents to a question multiple times to start formulating some ideas in your mind about what the important themes that seem to be emerging and what words or phrases give rise to those themes. Then you can start marking key phrases, themes and/or words by highlighting using different colors, changing text color, changing font size, italics, bolds, underlines, etc. It may be helpful to make a sheet with the codes and the associated code markings.
5. As you go through the text the first time, you will likely see themes requiring additional codes. You also may see nuances in a code that might be important to distinguish.
6. After the initial coding of a question, count the number of times each code was used. If one code was used many more times than another, consider if there is enough variation among these responses that would be preferable to break into two different codes. Likewise if a code was used only a few times, consider if it is close enough to another code to combine them.
7. There is no ideal number of eventual codes. In reporting, list the basic codes or themes, represented by a few words, and then discuss the range of phrasing included in the code with two or three verbatim examples. If it seems reasonable, put all the responses that generated each final code in an appendix in case the reader desires more detailed information.

One final suggestion, consider having one or more people code the text under study. Different people have different perspectives and catch different themes. So, using multiple coders, can result in a richer understanding. Remember, "Two are better than one, because they have a good reward for their toil." Ecclesiastes 4:9