Analysing a binary vs. ordinal variable
You might be interested if one group scored different on an ordinal variable, than another group. The variable that defines the groups is then a binary variable, while the variable with the scores could either be ordinal or scale. In this chapter we’ll look at the situation where the scores are ordinal. If you have more than two groups you want to compare, then see the nominal vs. ordinal section.
The analysis can then be done using the following steps:
Part 1: Descriptive statistics.
Use descriptive statistics to get an impression of the data, using:
- A cross table to show the sample results.
- Visualise the data with a multiple compound bar charts.
Part 2: Inferential statistics
After the first impression determine what can be said about the population based on your sample data by:
- Performing a test to see if the medians in the population (Mann-Whitney U test)
- Determine how big the difference will be (effect size with Rosenthal coefficient)
The example used in this chapter will be using as a binary variable gender (which had the options male or female), and the ordinal variable ‘5.3 The amount of online activities (Blackboard: video’s, online resources, exercise generator, forums) was …’, which had the options 1 = far too little to 5 = far too many.
Let's get started with getting an impression of the sample data.