I have 3 different groups based on their level of resilience : low (n=60), moderate (n=35), high (n=10). After a qualitative thematic analysis of their responses, different subjects emerged, like "being super funny" for example. => I would like to know is there is a significant difference between those 3 groups about how many people said that they are funny. I initially did Chi2 tests with a contingency table 3x2 (said that they are funny/didn't say that they are funny) but I think that the Kruskal Wallis test is finally more appropriate, what do you think ? However, when I tried to do the Kruskal Wallis test on JASP, it said that the number of obs is <2 for some data (ex : 11 people in low resilience group said that they are funny, 5 in the moderate resilience group and 0 in the high resilience groupe...). Do you know how I can still conduct my test please ? Thank you !!
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
It sounds like Fisher's exact test would be suitable for your analysis. It's similar to a chi-squared test but works at lower sample sizes. You have a contingency table by the sounds of it with three levels of resilience and two levels of funny (not funny and funny).
