I have to compare three independent groups on a single parameter for 15 years, for which one-way ANOVA could be applied, but the distribution of data is not normal for a few years. Can I apply the Kruskal Wallis test on data which is not normal and one way ANOVA on data which is normal, without making the tests incomparable? Or should I just go for the non-parametric Kruskal Wallis test throughout?
Thanks!
Darwin Alexander Pucha Cofrep
Use Kruskal-Wallis Test to compare the three independet groups, and the Wilcoxon Test for a pairwise comparison. Both methods don't need a normal distribution of your data.
0 votes
0 thanks
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA
More Nausheen Sodhi's questions See All
Similar questions and discussions