- This topic has 2 replies, 3 voices, and was last updated 7 years, 1 month ago by
.
-
Posts
-
Hello,
l have a question about qualitative data analysis (constructing codes), which l hope would encourage discussion among members of this forum.
Does the process of constructing categories or codes have to be based on: 1) how many people talked about the issue (i.e., repettion or frequency) or 2) on the researcher’s judgement?
I have encountered some issues that have been mentioned by only a small number of participants yet they seemed important. On the other hand, the qualitative data analysis textbooks often suggest that the researcher should have enough data to support her or his interpretation. I have read a number of qualitative and research method textbooks but unfortunately I did not find an answer to the above question.Ideas, thoughts, and previous experiences on both argurments would be much appreciated.
Hello Fatmah,
Indeed, a good question! But with no simple answer. There is no single right or wrong way to code, and each researcher could code the same data differently. The important thing is that you reflect on why you’ve made the choices you have and you incorporate this into your analysis.
Have you tried Saldana’s book on coding? http://www.amazon.co.uk/The-Coding-Manual-Qualitative-Researchers/dp/1446247376/ref=dp_ob_image_bk I would really recommend reading Saldana’s book – it’s the best guide there is to coding.
Best,
Katie
Fatmah,
As Katie mentions, there is no one best answer to this question and Saldana’s book is one of the best I’ve seen too.
The approach you take to coding starts with the research question you are trying to answer, your methodology and the type of data. However, whether your data results from semi-structured interviews or an open ended ‘lived experience’ question, doing a key word count in NVivo (or whatever program you’re using) can help show which issues were discussed most frequently and that has the potential to act as a guide to the most important issues from participants’ perspectives. It is also something that takes about 2 seconds to run so a good place to start. It might tell you that one participant used a term 50 times, or that 50 participants used the term once. It’s up to you to work out if that’s important. Assuming that the terms/issues that were mentioned most frequently are terms/issues that you are interested in, the next step is to read, read, read and look for patterns in the data around those terms. This is the hardest and most time consuming part of the process.
I am not sure what field you’re in, but from a social sciences perspective, one of the best articles I have seen that describes/demonstrates the iterative coding process is Maitlis, S. & Lawrence, T. (2007) Triggers and enablers of sensegiving in organizations, Academy of Management Journal 50 (1): 57–84.
Good luck!
- You must be logged in to reply to this topic.
