How to perform something similar to a two way ANOVA when one of the gruops does not fit the normality hypothesis?

Asked Jul 17 '23 at 14:10

Active Jul 18 '23 at 01:06

Viewed 32 times

I have a question about basic statistics: I have a set of data divided in four groups, defined by two variables. I ran the normality test (shapiro_test {rstatix}) and one of the groups does not fit the normality hypothesis (p=0.02).

In this case, can I perform a two way ANOVA or should I run another test? In that case, it is the Aligned Rank Transform (ART) ANOVA the best option, and what package and function do you recommend?

edited Jul 18 '23 at 01:06

kjetil b halvorsen

77,844

asked Jul 17 '23 at 14:10

DanielCG

1

Check out Mann-Whitney test, Wilcoxon test, Friedman test and Kruskal-Wallis test. – G. Grothendieck Jul 17 '23 at 15:06

How to perform something similar to a two way ANOVA when one of the gruops does not fit the normality hypothesis?

0 Answers0