What to do if member checks don’t agree with my grounded theory?

Home Forums Default Forum What to do if member checks don’t agree with my grounded theory?

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
  • #4189


    This is a grounded theory question. I’d appreciate any feedback.

    I recently finished analyzing texts using grounded theory and developed a theory from the documents examined.

    I e-mail people that submitted documents and I showed them my theory.

    so far, about 25% of the people slightly agree with my emerging theory.

    They point out that I overlooked some things in texts b/c they weren’t in explicit detail.


    SO, what do I do????  Is this nomral/ok???



    My suggestion would be to enter into dialogue with those that are of opinion that you overlooked things. People are more willing to talk than to write. It is important in qualitative research to accomplish participant agreement with regard to what they believe they shared.


    in this stage you must check with other researcher not with participants.

    Hassan Rafiey

    surely it’s normal.
    It means that your theory doesn’t match the reality, i.e. it’s NOT grounded enough. so you should gather more data, analyze them, theorize again, show them again to participants, … this cycle should be iterated more and more … up to the point that your theory could be compatible to the reality.
    Sometimes the method of data-gathering could be improved. for example a face-to-face in-depth interview may be more fruitful than text analysis.
    hope my comments be useful …
    Hassan Rafiey, MD, MPH

    Roger Gomm

    I think you should seriously question whether it is reasonable to expect your subjects to agree with your analysis. When I write my diary I find that I am often not satisfied with the accounts I give of my own feelings and behaviour – and here where I am both subject and analyst/theorist I fail my own fallibility test with myself. I am very sceptical of the idea that people understand themselves and that if you want to understand them you just ask them how they understand themselves. So a first suggestion for your problem is that your analysis gets it right, and they are wrong about themselves. But I doubt whether this is an adequate response. Second then, consider that the issue is about incorrigibles or qualia about which no one can finally be right or wrong, or rather about which people can be wrong, but where there are multiple possibilities which are all equally cogent – then the resolution to your problem is that you prefer one kind of interpretation, and that a lot of your subjects prefer another. Third, consider that your subjects are less likely to be interested in what is an accurate account of them (a sensible stance if point two is correct) and more likely to be interested in how they come across as people, how your interpretation bears on their identity and reputation and such like. People are rarely grossly self-serving and in my experience are as rejecting of accounts that seem over flattering as they are of accounts which villify them, but self-presentational considerations are likely to be at play in whether they accept / articulate acceptance of your version.Of course this theme can be worked up along psychoanalytic lines referring to ‘defence mechanisms’, or along ‘critical lines’ referring to ‘false consciousness’, ‘mystification’ and so on. And note that the semantic context in which you do the memberchecking will be very different from that in which you elicited the original information, because the former contains your version which they are forced to address -among other things your version draws on your knowledge of what was said by all respondents – each respondent only knows about themselves (and of course, time has elapsed between the occasions and we know from questionnaire research that the answers people give to the same question on different occasions can vary considerably).Lastly, how do you react when some smart-arse claims to understand you better than you understand yourself ?


    The remedies suggested by folks who believe in membership checking as a sound validity test, have the danger of being self-fulfilling. They prescribe reaching ‘the truth’ through an iterative backwards and forwards exchange between subjects and researcher. Of course this can produce a consensus, but the consensus may be more a reflection of the establishment of good relationships ( cooperation bias), or of the pedagogical effectiveness of the researcher convincing subjects he is right (as in the consciousness-raising that goes on in some feminist research). These remedies often confuse the moral/political precription that people have a right to say what is true about themselves, with the actuarial assumption that people accurately know what is true about themselves (in relevant regards). The very idea of member-checking assumes that there is a correspondence between what both the analyst and the subjects find as an acceptable version, and the version which is accurate.

    bernard smith

    I understand your concern that 75 percent of the subjects are suggesting that your understanding is flawed but what do you think? Do you think that they are in fact highlighting ideas that were available in the data you have but which you overlooked or are you reading  /interpreting those texts differently from them? Are you in fact approaching the data you have (those texts) by applying grounded theory methodology? In other words, are your ideas coming from the texts or are you perhaps bringing ideas to those texts and reading the texts through those ideas?  If you are drawing your ideas from the texts you might very well be missing somethings that other people might want to talk about but nevertheless what you are saying may be very reasonable and robust. If you are bringing explanations and accounts to the texts then you may be distorting what those texts can provide…

Viewing 6 posts - 1 through 6 (of 6 total)
  • The forum ‘Default Forum’ is closed to new topics and replies.