Why Kruskal-Wallis? Why?

Michal_Kolasa · February 6, 2019, 1:42pm

Hi everyone!
I have a quation about Kruskal-Wallis test.
Maybe it is stupid question, but how should I check if assumptions for this test are good in case of my data?
Cheers,
Michał

jwdebelius · February 6, 2019, 1:44pm

Hi @Michal_Kolasa,

This is probably a better question in general for a forum like Cross Validated. Do you have a question about using the KW test for some aspect of microbiome data in particular?

Best,
Justine

Michal_Kolasa · February 6, 2019, 2:04pm

Hi Justine!
Thanks for quick answer. Differences in alpha diversity were statistically assessed with use of this test. My colleague asked me a question if I checked normality of the distribution before using it. And I couldn’t find way to do that in QIIME, but maybe there is some obvoius exlanation that I am not aware of.
Cheers,
Michał

Nicholas_Bokulich · February 6, 2019, 2:21pm

@Michal_Kolasa,
Kruskal-Wallis does not assume normality, so there is no reason to test for normality, but you can always look at the barplots in that visualization to assess whether the distributions look normal.

Michal_Kolasa · February 6, 2019, 2:25pm

OK, thanks. Could you tell me if there are any limitations that should prevent for using this test?

Nicholas_Bokulich · February 6, 2019, 2:25pm

system · March 9, 2019, 8:25pm

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.