I used a semi-structured interview guide to collect the qualitative data from the target sources. In addition to that, I conducted two conversational interviews ( free-flowing) with individuals from the context which is distinct from the target source. With the latter, I did not follow any interview protocols as it was not meant to be included in the data set. Obviously, no triangulation intended with these two interviews. I am thinking, it makes perfect sense to code and analyze the data of the target dataset only and use the two additional conversational interviews simply to provide the context for the original data. How feasible is that? What would be the best way to analyze and report this type of conversational data?
As long as you have ethics approval for both sets of interviews, I don't see any reason why you couldn't use your full data set. You can't really suppress your knowledge of what was said in the two conversational interviews, so if they have influenced your thinking, they should be part of your data.
if you do so, your Methods section should describe both data sets, and if you use any quotes from the conversational interviews in your Results, you should label them as such.
I agree with David, you have collected them and they exist and they will inform your thinking and your analysis but as David rightly points out they need to be also covered under your ethics approval. It is an interesting question as to why these stayed as conversations and did get brought into a more formal interview context. Something for you to discuss your Methods section.
I agree with both professor Morgan and Ms Munford about what they say about ethics approval for these interviews. But as professor Morgan point out, the interviews exist and the information from them seem important to you, otherwise you would not have posed the question.
I also wonder why you made them? What was the reason for you to perform them? In that answer I think that a part of the answer for this question depends on the answer for that. As Ms Munford states, your reasons for performing the two interviews in this way needs to be adressed both in the methods section and in your discussion.
Thank you all for your valuable suggestions. The point about the ethics is crystal clear to me. What is not clear is the methods of analysis that I should use for those two additional interviews. This ambiguity is largely due to the fact that unlike semi-structured interviews, my conversational interviews, as you can imagine, do not follow a set of structure or ask the same set of questions as semi-structured interviews. Moreover, they yield somewhat contradictory pattern. I answer to Ms. Rejno: I did those conversational interviews just to have a different view from the group I conducted semi-structured interviews with . To illustrate the point, initally I wanted to learn about the views of the teachers only, but as the data collection proceeded, the circumstances required to learn about the other side of the coin to juxtapose the perspective of the teachers and the perspective of the administrative staff . Now I am having a hard time figuring out how to analyze the two conversational interviews as they are different from my major dataset- the one which represents the perspective of teachers.
If you do not use the converstional interviews, your readers probably will have difficulties to accept your conclusions because they are actually grounded in both data sets. Personally I would find it at least irritating to be excluded from a main data set. The obvious solution seems to be to use both of your data sources. This would be a coherent approach giving your reader a simimilar access to knowledge as you have benefitted from.The problem of this will be that you need a method for analysing these heterogenous data together. Have a look into grounded theory which is based on using data which are not similar in form.
I think having interviews from a different perspective should be a strength for your data set. Most approaches to analysis relying on a coding system, so you could define a single set of codes that includes codes from both sets of data. This would allow you to keep track of things that were mentioned in only one data set or the other, of that were mentioned in both data sets, but perhaps in different ways.
If you have not already settled on a method for your analysis, take a look at Braun & Clarke's (2006) thematic analysis, which is very widely used.
I think you might need to reconsider the method you earlier adopted for your analysis. If the data require a method that will be able to accommodate more divergent views, you might consider thematic analysis, like David suggests above, and then go with a more free-flowing narrative approach that will allow you to juggle the different views raised in your data collection. That's the beauty of the qualitative approach.
- Badges
- Science topic