How to run a Kruskal-Wallis test
When you pivot by a categorical column and a numeric column, and the categorical column contains 2 or more categories with less than 4 results, AddMaple will use the Kruskal-Wallis test rather than ANOVA.
1
Select a categorical column with some categories that have a small number of results, and a numeric column. AddMaple will run the Kruskal-Wallis test and give you a summary in the top right of the legend.
2
When you click "See more", AddMaple will take you to the stats tab with an overview of the results of the test and the detailed numbers. In this example there is a strong relationship between the 2 columns.
For the Kruskal-Wallis tests these are the outputs:
- P-Value: the measure of probability as to whether the relationship between two columns is due to chance or not.
- H Statistic: the value determines if there are significant differences among the distributions of groups. It compares the ranks of data across groups rather than the data values themselves. A higher H Statistic indicates that the differences among the groups are less likely to be due to random variation, suggesting stronger evidence against the null hypothesis that all groups have the same distribution
- Eta-Squared: this indicates the strength of the relationship between categorical groups and numerical data. It shows what proportion of the total variance in the numerical data can be explained by the differences between the categorical groups.