If the population of interest are few (eg 50 subjects), will it be appropriate to use cross sectional study design? What kinds of statistics are more strong in such sample size to make group comparisons? Can the report be strong enough to be considered as original research or will it be better as commentary?
If the population is so small, then why not just use the whole population and avoid the need for any inferential statistics at all?
- Thank you Stephen
- I just did that. All 51 subjects were studied. But can i still make group comparisons eg via Mann whitney U test
If you've measured the whole population, there is no need for inferential statistics, p-values, any of that. What do you gain by attaching a p-value to these population-level statistics?
Trying to find out significant group differences eg based on sex...not possible?
Given the various limitations of the cross-sectional study, a sample size of 50 subjects wouldn't be scientific if cross-sectional study design is used.
I would argue this sample size is appropriate to experimental study, imaging study, RCT, and longitudinal study. The use of statistical analysis will depend on your research question and study design. The result can be strong for original research if the above studies are employed. If cross-sectional study design is used it would be nice to make it commentary.
Statistical significance is a concept that only has meaning with respect to making indirect inferences about a population based on a sample. (Technically, inferences about the probability of a sample given a population: if a difference between two groups is significant, all that means is that if there were no difference in the populations then there's a
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics