I am conducting an exploratory study of a teaching innovation. I see that the sample is dividing into two groups. One group expresses it is benefiting from the innovation while the other group is not. In reporting my results, how many case study students from each of the two groups would be sufficient to write up? Minimum would be one per group, but since there is diversity of reasons for membership in each of two groups which I know I should somehow present, how many case studies should I write up in each group to produce a compelling, informative but manageable narrative?
This is an important methodological question in research. It does not seem to me that because the study is exploratory it can be restricted to one case. One of the pillars of research is generalizability; one cannot generalize one's findings from one case even if it is an in-depth investigation. And even if one is not after generalizing, it is safe to deal with a representative sample from each group to be able to find a guiding pattern justifying the thesis behind the raison d'etre of the study.
I appreciate my fellow RGrs up-to-date on this methodological question. And while I am hoping in my study to produce some insights that will contribute to my, and others understanding of the issue, I am not striving for generalizable results from this study. Also, Ahmed, while I appreciate the link, the guidance in it seems to be less evidence-based and more authority driven. As those of us who regularly use qualitative methodology come to know, even doing one case study can be intense, so limiting the number to a manageable number is necessary. Key question remains: how few is enough? Thanks for your thoughts!
Thanks, Toh, for your contribution. How did you decide on the selection of the 4 from the original 11, and was that sufficient for your purposes or did you find it perhaps too many?
Perhaps you might elaborate a bit, Toh, in this RG discussion on how you thought the 4 participants you selected "would be sufficient." That type of reasoning would contribute to our conversation regarding the methodological question I have raised. Could it be that the 4 in a cumulative manner, for example, represented most if not all of the diversity of your total sample?
Thank you, Toh, for your elaboration. Agree with you that a prerequisite for case study selection would be completion of the entire study. In your situation, sounds like that was 4 out of your total sample of 11. In my situation, I have a total sample of approximately 40, and 36 completed the entire study. So, for sheer feasibility issues, I cannot write up 36 indepth case studies. My problem is that I, and other qualitative researchers who face similar circumstances, need some additional rationale for how many to select to report as case studies.
Other RGers' thoughts welcome to join this conversation!
In terms of case study analysis, 40 participants is too many. Case study is done to highlight the anomaly in a given area of research or highlight a significant aspect of a case that can be impacted to an area. I think would assess your 36 cases and look for those that present anomaly, uniqueness, and significance -- perhaps three or four and this becomes a strong case study sample size.
Darrin, sorry if what I shared earlier in a response confused the issue. I was NOT saying I have nearly 40 case study participants in my study.(I have N=36 that completed it). From that total sample of approximately 40, what is being discussed is how many case study participants would be advisable to select from each of two groups that make up that sample--one group perceives the innovation is beneficial and a second group does not. I know at least one case study selected from each group is necessary, but are more? And if so, for what reason(s)? Your suggestion of selecting those that may "present anomaly, uniqueness, and significance" is promising, but how does one know if any would be "significant" or adequately "unique?" "Anomalous" ones would be the easiest to identify, but I believe in this instance, other researchers would also be interested in cases of the most common individuals in each group, as well. Thanks to all who are engaged in this discussion!
I would suggest the question of why? Why do individuals believe a particular way? Then consider answers for their level of deviation from common answers among your participants. For the many "typical" or similar answers create an aggregate case example. This takes into account both issues. You could also use demographic information collected from your participants, for example age. Choose any outlier ages to do case examples with. Perhaps most of the participants are married then aggregate the married into a single case example and do another related to the few people that are not married? These of course are just examples of ways I might consider the data to discover uniqueness.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing