Glossary

Ordinal Variable

An ordinal variable is a type of categorical variable whose categories have a meaningful order but where the intervals between categories are not necessarily equal. Analysed with rank-based methods or ordered logistic regression.

Definition

Why It Matters

Ordinal variables occupy a middle ground between nominal and interval data. Their categories can be ranked (e.g., Low < Medium < High), but the distance between "Low" and "Medium" is not guaranteed to equal the distance between "Medium" and "High." This property has direct analytical consequences: treating ordinal data as interval by computing means or running t-tests can produce misleading results, especially with small samples or skewed distributions. Non-parametric methods (Mann-Whitney, Kruskal-Wallis) and ordered logistic regression are designed to respect the ordinal nature of the data, producing valid inferences without assuming equal intervals between categories.

Example

A patient satisfaction survey uses the response categories Very Dissatisfied, Dissatisfied, Neutral, Satisfied, and Very Satisfied. These categories are clearly ordered, but the difference between "Dissatisfied" and "Neutral" may not be the same as the difference between "Neutral" and "Satisfied." The analyst uses ordered logistic regression to model satisfaction as a function of wait time and treatment type, rather than assigning numeric codes and running ordinary least squares.

Related Terms

Software Notes

SPSS: Define variable type as "Ordinal" in Variable View; use Analyze > Nonparametric Tests for rank-based methods; Analyze > Regression > Ordinal for ordered logistic regression
R: Convert to ordered factor with factor(x, ordered = TRUE); use polr() from MASS or clm() from ordinal for ordered logistic regression
Stata: Encode ordinal data with encode; use ologit y x1 x2 for ordered logistic regression; kwallis y, by(group) for Kruskal-Wallis