Methodspace - home of the Research Methods community

Sample size and number of interviews in qualitative research

Hello:
I and a recent doc. student grad. have had difficulty with some journal review boards with the small sample size of qualitative research. I have tried to communicate the point that saturation is more important than the sample size. Does anyone have any references that may discuss sample size in qualitative research?

Thanks,

David

Views: 23837

Reply to This

Replies to This Discussion

Hi David,

Try this:
Gaskell, George (2000). Individual and Group Interviewing. In Martin W. Bauer & George Gaskell (Eds.), Qualitative Researching With Text, Image and Sound. A Practical Handbook (pp. 38-56). London: SAGE Publications.

Best,
Charlie
Hi David, if you don't mind, can I ask what number of participants you had in your sample? I am wondering because I am at that stage in my PhD at the moment of recruiting participants for qualitative interview, and it's been quite tough finding people, so I am interested in what number of participants are seen by some reviewers as too small a sample size.
Thanks and good luck arguing your case with the review boards. Karolina
Karolina:
My student had 4 participants, 2 rounds of interviews and member checks with each- for a total of 12 interviews.
Guest,Bunce, and Johnson (2006) found with their study that involved 60 interviews theme saturation was achieved after 12 interviews. I use their study as supported my student's decision. I certainly know it was one study. More studies like it actually are needed. I have found that the majority of editorial boards in the counseling profession do not understand qualitative research, have had very little training in it, and thus de-value the power of qualitative research as a means to knowledge production. Quite ironic in my mind given the idiographic nature of the counseling process and change. More with review boards, those emedded in a post-positivistic philosophy value nomothetic knowledge (sweeping statements of "norms" and laws of behavior across people and groups, generalization is king. For some if you cannot generalize the results of a study why do the study?) For me, I can allow someone to believe this if they too can allow me to exist in a post-modern/ post- structuralist/social action paradigm that views knowledge much more contextually bound and research participants as active players in the knowledge construction process.

I encourage you to stay steadfast in your interest and merely be armed with support for your methodological choices.

Good luck,

David
Hello David..
I think there's no need to search or determine the sample size in a qualitative study. The thing that must be considered is the appropriate data sources, but not in sample-population term...
Hi David,

I just had a review board meeting for my thesis, which includes a mixed methods design with focus groups (qualitative portion) and surveys (quantitative portion). The review board mentioned this same dilemma, it seems sometimes that people outside of the field equate quantitative principles with qualitative research. However, it seems like you have a great number of answers listed already here.

Best of luck,

Debbie
Debbie: The dilemma, as I see it and teach it my qualitative courses, is enhancing researchers' awareness that research inquiry is embedded in philosophies of reality. We discuss the ontological, epistemological, axiological, and methodological position of various research paradigms. So, when a committee who is embedded in a , say post-positivist philosophy toward reality and knowledge, assesses qualitative research they may "see" qualitative research through the philosophical lens that provides the conceptual infrastructure for quantitative research. Herein lies the problem. For me, I believe it essential to declare one's philosophical position toward inquiry at the outset of a study. I also then expect the following methods and procedures to be embedded in this same position. In my profession of Counseling more systematic training in not only the methods of qualitative research are needed but also the philosophical positions providing its infrastructure. Thanks for continuing the conversation.

David
Hi David,
Thanks for your reply. I think you're absolutely right, your methodological choices need to reflect your ontological position and understanding of knowledge production, and that's also where you can argue a strong case for smaller qualitative studies, as you say. This is not only a problem for certain subjects, I think it's a problem in certain departments or journals across the board of social science research, as it's a question of academic culture. I totally agree with what you said about viewing qualitative research through a 'lens', I have that problem in my department where I'm being questioned about the usefulness of a study that can't generalise its findings...

On a more general note, it might be naive of me, but I'm quite optimistic about the status of qualitative research, especially with what seems to me to be an increasing interest in complexity theories, non-representational theories (e.g. Nigel Thrift) and so on that elaborate on the postmodern/post-structuralist approach to understanding the complexities of the world we live in. The current trend in my subject (cultural geography) seems to swing that way in any case. However, I suppose the challenge at the moment is to develop methodologies that are sensitive to these theoretical developments.
Method and theory should inform each other, there has to be a lot of dialogue there, but at least for me it seems theory is running far ahead of method and it's far too easy to fall back on conventional methods like the semi-structured interview and not spend too much time questioning if this is the best way to answer your questions.
Is there something to be said for a kind of 'make-do method' that is just about doing what works, and also about letting the socio-material context and conditions of your study dictate the methods?
Accepting that you just can't impose order, sameness and rigidity upon a dynamic world, and like you say, recognise the participants as active producers of knowledge.

Karolina
Hi David,
Two references which are sympathetic to small samples are:

Small, M.(2009) 'How many cases do I need ? On science and the logic of case selection in field-based research' Ethnography 10(1) 5-38

Williams,M. (2000) 'Interpretivism and generalisation ' Sociology 34(2) 209-224

Though these are recent references they don't go a great deal further than the classic statements in this field by writers such as Stake, Donmoyer,or Clyde Mitchell in the 1970s and 80s, though the water has been muddied frequently by various philosophical currents since then.Key papers by these three (and others) are reprinted in:

Gomm,R.,Hammersley,M, and Foster P.(eds) (2000) Case Study Method: key issues, key texts, London. Sage.

In addition references to Geetz seem almost mandatory in this kind of debate; try:
Geertz,C.(1973) 'Thick description:towards an interpretive theiry of culture' in 'The interpretation of cultures:selected essays' NY Basic Books:3-30. Available online at

www.stanford.edu/~davidf/qualitative151/geetz.pdf
or
www.scribd.com/doc/267783/Clifford-Geetz-Thick=Description-and-othe...


However, rather than grabbing desperately for defences of studies with small samples, it would first be worth thinking seriously about why such studies might be regarded as unimportant. The issue is very simple really, given that you can produce a rich and authentic (whatever that means) study of a handful of people, why should any but a handful of people be interested in reading it, and or, given that there are 6000 million people in the world, and each plays a large number of roles and is involved in a large number of situations, then why should anyone be more interested in what is shown in your small study than in a study of any other small group of people within a circumscribed areas of their lives.Only two kinds of answer seem possible.One is that your subjects are important in the impact of their activities on others (they are cabinet ministers perhaps) or because a study of them will tell us something important about a much larger number of people, or about an historic moment, or about a cultural tendency. Your problem with the latter answer is that without studying (or drawing on other people's studies of) a large number of people, you can't know how your small sample illuminates the lives of the larger number.

Note also the Achilles heel you are building into your research if you combine the notion of contextually generated knowledge and theme saturation. If you lay great stress on the former, then it suggests that saturation may be the product of the context of elicitation- qualitative interviews are very 'leading', combined with the practice of analysis. Theme saturation is never convincing if it relies on the judgement of a single analyst: best practice involves some kind of inter-rater reliability test.

I think that Kalolina is mistaken to invoke complexity theory as an argument in favour of studies which cannot be generalised, or can only be generalised within a narrow temporal/geographical or cultural range. The message of complexity theory is that what appear to be (and indeed are) diverse phenomena are generated by the interaction of relatively simple mechanisms.Paradoxically perhaps, complexity theorists are not interested in complexity in the sense of richness and diversity, but in the simpler underlying mechanisms which generate this: they are interested in the climate rather than the weather. In this regard, complexity theorists proceed in a direction opposite to that of most qualitative social researchers looking for the kinds of abstraction which will accommodate the widest range of generalisations- usually abstractions in the form of equations.
Roger:
Thanks for the input and references. For myself, and the students I teach, I stress the importance of the researcher identifying the philosophy of inquiry they move forward from when conducting research, as this will impact methods, questions, interpretation, etc. For example the notion of inter-rater reliability is grounded in some very rich beliefs/values about reality and the relationship between knower and known. If a person came from a more interpretivist stance, especially sprinkled with a post-modern, post-structuralist "twist" issues of getting an "accurate" interpretation via multiple interpreters and inter-reliability ratings is incongruent with the value claims of a post-modern informed interprevist philosophy toward research and knowledge. I believe it important to develop "best practices" only if we can articulate within which philosophy of inquiry we are working. I also do not see the need to make sure my sample will "illuminate the lives of the larger number." That is not my job-from my philosophical stance. This job lay in the minds of those who choose to use the results of a study and wonder about its applicability to their setting/context.

Truly love the conversation. Thank you.

David
Hi David,

I am not all that into sample sizes in qualitative research, but I will suggest you to read:

Lieberson, Stanley 1991: "Small N's and Big Conclusions." Social Forces 70:307-20. (http://www.jstor.org/pss/2580241)

The article explores some of the basic limitations that small sample sizes (qualitative or not) imply, especially if you are concerned with relating some observed differences between your interview persons' statements (or other attributes) to other observed differences between these persons. The article is concerned with logical problems and not ontological or ethical problems.

A critique of Lieberson's position is:

Savolainen, Jukka 1994: "The Rationality of Drawing Big Conclusions Based on Small Samples." Social Forces 72:1217-24. (http://www.jstor.org/pss/2580299).

All the best,
Kristian
I have recently joined this group as I'm also a PhD student who's going to submit her proposal to a review board made up of people used to the quantitative study paradigm, and have little or no awareness about mixed methods study (which I plan to do). I'm faced with two problems: convincing the board about my methods, and finding the 'right sample size'. Could the articles referred to by Kristian be made available on this site? Where I'm located it's not possible to get access to many journals including JSTOR.

Nasreen Ahsan
thanks for the jstor links

RSS

Follow us:

© 2014   Created by SAGE Publications.

Badges  |  Report an Issue  |  Terms of Service