Analysing a nominal and ordinal variable
When you have one nominal and one ordinal variable you might be curious if any of the categories in the nominal variable, score different on the ordinal variable. To analyse this we go over the following steps.
1) Get an impression from the sample data by creating a cross table.
2) Visualise the sample data using a stacked bar-chart
3a) Determine if the nominal variable has an effect on the ordinal by performing a Kruskal-Wallis H test
3b) If it does, check which categories score different by using a post-hoc pairwise Mann-Whitney U test
3c) Also determine how strong the effect is using an effect size (Epsilon square)
4) Combine all the result into one report.
For each of the steps above a separate page was created. Let's begin with getting an impression from the sample data on the next page.
Nominal vs. Ordinal