Hello everyone
How can I statically validate a translated sleep questionnaire with small sample (n=30 maybe more)?
- I have no access to previous Valid questionnaire used in the same population or to a clinical sample to verify discriminative validity.
Someone advised me to use Bayesian Factor Analysis, if it's possible to do CFA using Bayesian with this sample size can you recommend any books on the matter.
thank you.
Hello Aziz,
So, very small sample, no other measures administered so as to inform concerning convergent/discriminant validity, and you want to establish evidence for validity of a translated measure, right? Well, we all prefer a challenge!
1. There is evidence that Bayesian CFA can function well with smaller-than-usual data sets, but B-CFA is hardly a simple matter (see this link: Article A Comparison of Penalized Maximum Likelihood Estimation and ...
).2. Proponents of PLS-SEM claim that small N can still be profitably used to evaluate factor models (have a look at this link: https://www.sciencedirect.com/science/article/abs/pii/S0148296319307441)
I suspect that an N of 30 likely strains the bounds of credibility no matter what method you might apply, in the realm of CFA (especially if your measure has any appreciable number of items). However, I can appreciate that there are times when you simply have to forge ahead with whatever data you are able to collect. You might be wise to characterize the results as "tentative," or "exploratory," rather than as definitive for your target population.
Good luck with your work.
I agree with what Sir David has replied with, the sample size are too small to be conclusive.
Regards.
David Morse I'm glad you see it as a challenge rather than a problem.
I don't know the sample size yet, but I am preparing an alternative plan in case I get a small sample and it is expected in my study population.
I will look at the links.
Thank you Sir
With a small sample size, you could do confirmatory composite analysis using PLS-SEM, which can confirm your measurement model. However, the findings of a small-sample quantitative research could be challenged easily, thus rendering the investigation valueless. To enhance the validity of your small-sample study, you might adopt a rigorous design, such as mixed methods research design (MMR). In particular, you could consider explanatory sequential MMR. The following publications might be helpful.
Creswell, J. W. (2022). A concise introduction to mixed methods research (2nd ed.). SAGE Publishing. https://us.sagepub.com/hi/cab/a-concise-introduction-to-mixed-methods-research/book266037
Hair, J. F., Howard, M. C., & Nitzl, C. (2020). Assessing measurement model quality in PLS-SEM using confirmatory composite analysis. Journal of Business Research, 109, 101–110. https://doi.org/10.1016/j.jbusres.2019.11.069
Hubona, G. S., Schuberth, F., & Henseler, J. (2021). A clarification of confirmatory composite analysis (CCA). International Journal of Information Management, 61, 102399. https://doi.org/10.1016/j.ijinfomgt.2021.102399
Schuberth, F. (2021). Confirmatory composite analysis using partial least squares: Setting the record straight. Review of Managerial Science, 15(5), 1311–1345. https://doi.org/10.1007/s11846-020-00405-0
Good luck,
Thank you Wati Temjen
Mohialdeen Alotumi don't I need to establish evidence of validity for the measure before using an explanatory sequential approach?
Thank you very much
The answer is simply yes! Your tool has to be valid and reliable regardless of the study design. The point is that since you have a small sample and might opt for PLS-SEM or other procedures, which could be helpful to assess your measurement model, the ensued quantitative findings of your validated tool could be questioned based on a lack of appropriate sample size. One way to shun such a challenge is to tweak the research design by including a qualitative aspect. In other words, you could utilize an explanatory sequential MMR, which has two distinct stages. The first of which collects quantitative data, and the second stage gathers qualitative data to explain the significant or insignificant quantitative results. In doing so, your research findings can be more robust and pragmatically oriented. The clarification by Abutabenjeh and Jaradat (2018) might be of interest.
Abutabenjeh, S., & Jaradat, R. (2018). Clarification of research design, research methods, and research methodology: A guide for public administration researchers and practitioners. Teaching Public Administration, 36(3), 237–258. https://doi.org/10.1177/0144739418775787
Good luck,
Mohialdeen Alotumi I will read the article, I also ordered the book you recommended.
Thank you Sir.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size