I want to compare three groups of the samples (n=10). But the data of a group is not following the normal distribution.
1. Kindly suggest suitable method of comparison.
2. Can i use Friedman test for this comparison? if yes then what is the limitation of sample size?
It does not sound like this design would be appropriate for Friedman test. Maybe you want to consider Kruskal-Wallis ??
@Salvatore S. Mangiafico yes i want to apply Kruskal-Wallis but in Minitab help, it is mentioned that I should have least sample size(n=15).
I don't have any definitive advice on that, but maybe see the following site and search for "sample sizes" (no quotes): http://www.biostathandbook.com/kruskalwallis.html
10 observations per group doesn't seem too small to me.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size