L9: Non parametric tests Part 2

Question 1

What test do we use to compare data from >2 independent groups with continuous non-normally distributed data or ordinal data (presented as median IQR)?

Answer 1

Kruskal-Wallis Test

Question 2

What is the null and alternative hypotheses for the kruskal wallis test?

Answer 2

H0: All the medians of the underlying populations are the same
H1: eg Not all the medians of the underlying populations are the same. OR the medians of at least two of the underlying populations are different.

Question 3

If we find a statistically significant result from the kruskal wallis test, what do we do?

Answer 3

Perform post hoc analysis:

• conduct pairwise comparisons using Wilcoxon rank sum test and use bonferroni adjustment
• adjusted alpha = alpha/no of pairwise comparisons

Question 4

What test do we use if we are comparing data between two paired groups with continuous non-normally distributed data or ordinal data (presented as median IQR)?

Answer 4

Wilcoxon signed-rank test

Question 5

What is the null and alternative hypothesis for the wilcoxon signed-rank test?

Answer 5

H0: The median difference between the two groups = 0
H1: The median difference between the two groups not equal to 0 (two-tailed).