We have a total sample size of N=5000. We would like to compare subsample of n=200 to the rest of the sample (n=4800) regarding some well-being variables. Distributions in both samples are asymmetrical as typical for well-being.
We are worried because of such a huge difference in sample sizes.
Hi, Tihana,
You could compare using Bayes Factors, which are less prone to be affected by such size differences (JASP is a free software that allows for both frequentist and bayesian analyses).
However, I wonder the purpose of you intended breakdown: a subsample of 200 against its sample of 4,800? Let's assume the subsample is a random selection from the larger sample. In such case, it is like having an apple (the large sample), then taking a sliver from it (the subsample), then comparing the sliver against the sample. What is there to learn, other than the sliver resembles the apple it came from?
If the subsample can be considered as a 'sample' from another population (e.g., a minority, versus the general population), then it seems appropriate to make such comparison.
Regards
Thank you very much Beatrice and Jose.
Jose, yes, subsample can be considered as a sample with specific characteristics, and that is why we need to do this comparison.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size