What type of data is used in the Mann Whitney U test according to its conditions of use?

The Mann-Whitney U test is a non-parametric statistical test used to evaluate whether there are differences between two independent groups. Its conditions for use specify that the data should be at least ordinal. This means that the data can be ranked, allowing for the assessment of the relative order but not the exact differences between values.

In the context of establishing rankings, ordinal data provides a basis for determining whether one group tends to score higher or lower than another without making any assumptions about the distribution of the data or the equivalence of distance between ranks. For example, a survey with rankings from "not at all satisfied" to "very satisfied" generates ordinal data, making it suitable for analyses like the Mann-Whitney U test.

While nominal data (which categorizes without implying order), interval data (which has meaningful differences between values but lacks a true zero), and ratio data (which has all the properties of interval data plus an absolute zero), are important concepts in statistics, they do not fit the conditions required for this specific test. The Mann-Whitney U test is specifically designed for situations where the data can be ordered or ranked, reinforcing that ordinal data is indeed the correct type for its application.