Do we formulate a priori codes before the interview or just during the analysis? I have heard of some researches using a priori codes to come up with items for their interview guide. But most sources consider a priori codes for the coding process rather than in constructing questionnaires. Could the a priori-emergent dichotomy be applied both in questionnaire construction and in coding/analysis process?
Either approach is possible in coding. Using a priori codes is frequently referred to as a "deductive" form of analysis, while building the codes during the analysis would be "inductive." A third option would be a "hybrid" approach, where you begin with a set of a priori codes and then add to them as you examine the data inductively.
Whether you apply a similar approach in constructing the interview guide depends on your research goals. If you have a set of a priori topics that you want to examine in depth and detail, then you certainly would want to ask about them directly in your interview guide.
One possible equivalent to a "hybrid" approach is to use a so-called funnel structure to your interview. that would begin with one or two, broad open questions, followed by a set of narrower, a priori questions. Beginning this way has two advantages. First, it can start the interview by asking the participants questions that relate to their own thoughts and experiences, which can help establish rapport. Second, it gives you a chance to hear the extent to which your a priori topics are already part of the participants's thinking before you explicitly cue them to think in those terms.
Either approach is possible in coding. Using a priori codes is frequently referred to as a "deductive" form of analysis, while building the codes during the analysis would be "inductive." A third option would be a "hybrid" approach, where you begin with a set of a priori codes and then add to them as you examine the data inductively.
Whether you apply a similar approach in constructing the interview guide depends on your research goals. If you have a set of a priori topics that you want to examine in depth and detail, then you certainly would want to ask about them directly in your interview guide.
One possible equivalent to a "hybrid" approach is to use a so-called funnel structure to your interview. that would begin with one or two, broad open questions, followed by a set of narrower, a priori questions. Beginning this way has two advantages. First, it can start the interview by asking the participants questions that relate to their own thoughts and experiences, which can help establish rapport. Second, it gives you a chance to hear the extent to which your a priori topics are already part of the participants's thinking before you explicitly cue them to think in those terms.
David has given you some really good advice above. The only thing I'd add is that, strictly speaking, the words a priori are Latin for something like "from beforehand". Therefore your a priori "codes" are the ones that you'd have created (quite reasonably in some instances) in establishing your interview guide.
To remain with the Latin, any codes that you generated in light of the data that emerge from your subsequent analyses might be considered post hoc or a posteriori. Those codes would certainly not be a priori.
Robt.
The codes should be obtained from the data during the analysis, a posteriori. Qualitative research is based in an inductive approach, therefore you should not use codes previously designed.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire