kruskal wallis aflatoxins


New Member
I have been analyzing aflatoxins in various nuts. The whole population contained about 130 nuts. In only 9 (5 almonds, 3 hazelnuts and 1 pistachio) of those 130, there was a high level of aflatoxins.

I would like to test if there is statistically significant difference between those 3 groups. I know the sample size is a bit small, but i am not sure if this is the right way to do, because of the 1 pistachio? My approach would be to try a kruskal wallis nonparametric anova. Can anyone correct me? :)

almond 6
almond 6.2
almond 5.1
almond 4.4
almond 4.2
hazelnut 4.6
hazelnut 4.3
hazelnut 4.2
pistachio 4.1


TS Contributor
does the remaining sample of 121 nuts also contain almonds, hazelnuts etc.? If you test these 9 nuts, you might get the answer to the research question, whether of all the nuts that contained aflatoxins almonds conained more than hazelnuts? That might not be the question you are interested in.

if yes, then I would leave the pistachio out of the sample, or use almonds vs. other to test.



New Member
Thank you for the answer. Yes, they remaining sample contains the other types of nuts. everything above 4ug/kg contains too much aflatoxins, therefor i am only interested in the difference between those 9 nuts. (3 categories)