ABSTRACT
Ordinal data are common in practice. The values of ordinal or ordered categorical variables can be arranged according to size. However, calculations such as addition and subtraction are not possible (see appendix). An ordinal random variable usually has only a few mass points. The three examples with ordinal data presented by Brunner and Munzel (2002) have four, five, and ten possible values, respectively. The sample sizes are often much larger; hence ties are not only possible, but inevitable.
