According to my sample size estimation, the sample size comes to be 14. Would results from a trial with such a little number of subjects be valid? On the other hand I cannot increase the sample size beyond my calculated sample size except for adjustment for loss to follow up. Please suggest me.
Under usual situations 14 appears very small. I would refer you to a biostatiscian that has clinical trials experence and suggest that you follow her/his advice. Best wishes, D. Booth PS Your university probably has such a person.
I assume you are expecting very large effect sizes, so that the sample size calculation is very small. The problem with this is that the measured parameters, e.g. mean values, are not very stable in small samples. Single unusual participants have huge influence on these values and may cover actual effects or may produce unsual large effects (which tends to be more often according to simulation studies). So, despite of your power analysis/samle size estimation, I would not trust results from such small samples and would try to get large samples with a good sampling procedure. Do not only look at the p-values, since null hypothesis testing is flawed in the first place. Calculate and interpret effect sizes.
small sample size like this mean you introduce or expecting an abnormally large effect size . However, results of small sample like this are not dependable. i suggest you to revise the effect size you entered
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size