Hello Patrina,

Is 11 complete cases (or 28 complete/partial cases) sufficient for a survey? That depends on your specific research aims and how confident you might be that the responses you do have are representative of what you might have obtained had the complete group which you solicited all responded.

Some general observations:

1. If complete cases are the only ones you can use, then the sample size is obviously very small and you have two issues: (a) lack of precision in any estimates you might derive (and/or the associated low power for any hypothesis testing you might undertake); and (b) concern about generalizability beyond your set of respondents.

2. You could compare responses of "complete responders" (the 11 cases) to the "partial responders" (17 cases) to determine if there were any obvious differences on any of the variables with respect to the two batches. If not, that might offer some solace that self-selection out of the complete responder category didn't materially affect results. If you have any information about the 72 non-responding cases, and can compare that to the full/partial responding cases, that might offer some hope that (at least as far as you can ascertain) the non-responders might not be qualitatively different from responders.

3. You can always characterize your results as "exploratory," given the amount of self-selection that occurred among those you asked to participate.

Good luck with your work.

Read papers Babi(1990), and Klin(2010) for N: q ratio. for sample size read kerji.Morgan(1970).

Small sample size could render a quantitative study useless because of low statistical power and increased error margin. However, one way to go around the issue of a small sample is to combine qualitative and quantitative methods (i.e., mixed methods) via cluster analysis (CA). You might go through the review article by Crowther et al. (2021) and the chapter by Péladeau (2021) for illuminating inputs on CA.

Crowther, D., Kim, S., Lee, J., Lim, J., & Loewen, S. (2021). Methodological synthesis of cluster analysis in second language research. Language Learning, 71(1), 99–130. https://doi.org/10.1111/lang.12428

Péladeau, N. (2021). Cluster Analysis for Mixed Methods Research. The Routledge Reviewer’s Guide to Mixed Methods Analysis, 57. https://www.google.com/books/edition/_/SMY1EAAAQBAJ?hl=en&gbpv=1&pg=PA57&dq=cluster+analysis+mixed+methods

Good luck,

Of course it's a sample. It's a small one and it may be biased in some ways. Your job is to lay it all out, all the weaknesses, in the report of your work. Neither you nor your work can be criticized for what you've already criticized. You may also frame your work as a basis for further study, especially if your results are interesting, novel or important.

Hello Patrina Bevan!

The sample size is always critical factor in research. Considering the situation you are in now, I am of the opinion that not being able to collect the response from the 100 respondents that you intended could alter the results. If you have alternate samples that can be sourced, you may collect responses from the same, assuming that you will have positive response from the new set of samples.

If doing so is not possible or feasible because of the time frame or lack of resources, then the same can be justified as a finding that needs to be explored as scope for further research.

say universal sampling

Thank you so much, everyone, I really appreciate your responses.

I am unable to gather more responses now, so that isn't an option.

I will check out the recommended articles.

I think this may be regarded as a voluntary response sample...

Ravishankar Suryanarayana for letting me know about universal sampling, I'd never heard of it!

Mohialdeen Alotumi thank you for suggesting CA, I will look into this

David Morse thank you for suggesting I compare completed responses to partial responses, I think that would be interesting

James Riggle Thank you for your kind, reassuring words