I am considering undertaking a quant + qual research by providing teachers with a questionnaire to measure their feelings and practice with regards to a type of music.
I was thinking of gathering the values from the questionnaire, and, based on the results, I will draw a sample from respondents and interview them. That is pretty straightforward mixed methods, right?
I was thinking, conversely, of meeting a small sample of participants (20-25) one on one, and handing them out the questionnaire, and based on their responses, conduct a semi-structured interview. My question is then; is this a functioning design? Is this second scenario multi-method instead of mixed methods?
Most definitions of mixed methods simply emphasize the collecting and analyzing of both qualitative and quantitative data, but I would add that there needs to be some integration of the sets of results. The process of integration is obvious is the first study you describe, where you collect the quantitative results and then pursue them further in an explanatory sequential design (QUAN --> qual). How the two sets of results would related to each other is less obvious in your second design.
David L Morgan Thank you for your response David. Just to clarify, in the second scenario, I would provide the questionnaire one on one, and based on the results of the questionnaire, I would ask questions to get larger context of the answers.
Dear, either way, it is a mixed-methods approach. The method is not only concerned with the ways of data collection, but you will also describe it in your analysis section. You will describe the results in your quantitative and qualitative sections. Then you will integrate your both results, i.e., those obtained from the quantitative and qualitative sections.
One issue you will run into is trying to produce meaningful quantitative results from a sample of only 20-25. The problem is that you will have such large standard errors that it will very difficult to produce statistically significant results.
Yes. That is what I thought, the quantitative results might be useful if the target population is relatively small. Thank you David L Morgan
There's two cases that I know of where a small sample, say 20-25, is totally justified in a mixed methods study like you stated: A sample difficult to access or a highly differentiated group. One can be upfront about the power analysis, transferability, and generalizability. There's also the question of how the small n will be used in analysis.
The statistical significance of results depends largely on the sample size, and is only weakly affected by size of the larger population. So, if you draw a sample of only 20-25 you will not have enough "power" in your analyses to produce significant results, regardless of the population size.
Great points! First, there are many small n samples in mixed methods, such as cohort studies in medicine and healthcare plus social sciences intervention studies. This points to moving beyond overly simplified rules of thumb. Easily verified in an academic search. Secondly, don’t be seduced by only statistical significance— practical significance is more important.
Here is a good reference for a mixed-methods explanatory sequential design:
Creswell, J. W. (2009). Research design: Qualitative, quantitative, and mixed methods approaches (3rd ed). Sage Publications.