9th August 2010 at 5:37 pm #4229mandyMember
I’m using Strauss’ grounded theory for my dissertation and I’m hoping that someone could help me answer my question.
My question is about open and axial coding.
I am analyzing documents to examine how college students are prepared to teach writing. As I compare and contrast documents, I have developed axial codes from some of my open codes. However, there is one code that was developed during open coding that hasn’t been saturated.
Even after examining 23 documents, this one code has only occured 3 times throughout all the texts that I’ve analyzed.
So, what do I do with this code? Is it considered an outlier ( I know this is Quan terminology)? Should I mention this point in my theory (B/c frankly the code is important when I look at my overall discussion of results and it says something about what’s missing from these classes, but how do I make it fit and where)?
What do you think?
Thank you for your help15th August 2010 at 7:04 pm #4235Dr Thomas GroenewaldParticipant
Minority views or voices are as important as majority ones – it is one of the important aspects of qualitative research: to let the voices be heard. It is your responsibility as researcher to give voice to minorities (or ‘outliers’). Good luck!4th November 2010 at 7:09 pm #4234Muhammad ZubirMember
It’s will be the issues and gaps of the code. it might somethings to search further on that issues. That’s the beauty of qualitative research.16th November 2010 at 12:29 pm #4233Dr Enda DunicanParticipant
I’d mention it in discussion but make it clear to the reader that it hasn’t been saturated. If you report honestly then there isn’t a problem. It could also be discussed in the ‘future research’ section as an interesting and, possibly, emerging concept.6th December 2010 at 1:34 am #4232Muhammad ZubirMember
Does that code related to your reserach question?17th December 2010 at 8:38 pm #4231bernard smithParticipant
I am not clear about two things. First, how the documents you refer to were generated. Were they obtained by you or generated by you (interviews?). If you have been interviewing, would there be a methodological problem if you were to interview either the same subjects again or a set of new subjects where you then focus on the issues you see as significant but which you say have not yet been saturated.
My second confusion is about what you mean by saturation. My understanding is that this term refers not to the researcher finding dozens of examples of an idea in the data but that saturation occurs when the data yield no additional codes. So it would be perfectly reasonable for you to identify , say, only one case of X in a thousand texts and dozens or more cases in each of N different concepts from your data but as long as you can find no additional concept (N+1) then you have reached the point of saturation. The fact that you have found only 3 cases of this concept may itself be what is important, No?29th December 2010 at 2:04 pm #4230bernard smithParticipant
Guera, . I don’t pretend to be an expert in grounded theory (GT). My background is in ethnomethodology. I use GT. BUT… the codes one creates in GT are integral to the data set itself. You are saying, I think, that your codes have been established from outside the data you are using and that you are trying to code your data set with these codes. For a discussion of this kind of issue you might want to look at a brief exchange between Yvette Taylor and a group of researchers (Hall et al) who worked to code the official transcript from a government inquiry in London. (see Qualitative Research 2008 8 section 1 for Taylor article and section 5 for the Hall et al response). Hall et al were using a Delphi exercise to create the codes, not GT. Now, I am not offering myself as a “purist” and arguing that there is only ONE way to do GT (compare Charmaz and Glaser’s disdain for Charmaz’s constructivist approach) but I don’t know how externally created codes that pre-date your initial and subsequent readings of a data set sit comfortably with GT. Which is not to say that what you are doing is illegitimate. Banish that thought. But it is to suggest, I think, that the issues that GT makes available and the issues it does not may not apply to your methods. So for example, GT would suggest that if your data provides you with only a few codes in a particular theme (let’s say issues about the age of respondents and a particular set of understandings but you have only a narrow age range (because you never thought that talk of age would have any significant role) then GT would suggest that you can look for additional subjects to broaden the age range you have (and GT calls that theoretical sampling – a change to your sample in the middle of your study based upon your intermin theorizing (based upon your coding)) . That kind of active sampling cannot legitimately be done in quant studies and I don’t think it can be done in other kinds of qual studies apart from GT. Because the codes and the data are so fundamentally integrated theoretical sampling does not distort or corrupt any findings. I guess my point is are you sure that you are using GT?
- The forum ‘Default Forum’ is closed to new topics and replies.