DataTerminology.comNEW
Categorical values with relative ordering but no equal intervals.
DESCRIPTION
Ordinal scales provide rank order among categories (e.g., low, medium, high) without assuming equal distances between levels. Medians and nonparametric tests are appropriate; means and differences are not strictly meaningful. Treating ordinal data as interval can bias inferences. Common examples include Likert responses and rating tiers. Model choices include ordinal regression and rank-based methods.
ALTERNATIVE DEFINITIONS
Wikipedia
The ordinal level of measurement classifies data into categories that can be ranked; however, precise differences between the ranks do not exist.
Emphasizes ordered categories without meaningful interval distances.
UCLA Statistical Consulting
Ordinal variables have categories with a natural order, but the intervals between the categories are not necessarily equal.
Clarifies order vs. unequal intervals in applied settings.
RELATED TERMS
Nominal Scale Ratio (NOIR) Likert Scale Nonparametric Statistics Ranking Median Quantile Monotonic Relationship Measurement Scale
COMPETING TERMS
KNOWLEDGE AREAS
Data Science Mathematics Data Quality
TARGET AUDIENCE
Data Scientist Data Analyst Information Analyst Data Steward
EXTERNAL SOURCES
- Wikipedia - Level of measurement
- NIST - Measurement scales definitions
- Penn State - Intro to data scales
TOP QUESTIONS
What statistics are valid for Ordinal Scale data?
Medians, quantiles, and nonparametric tests (e.g., Mann–Whitney) are appropriate; means and standard deviations are not strictly meaningful.
Category: Requirements
Can Likert scales be treated as interval data?
Often approximated as interval in practice, but strictly they are ordinal; justify the assumption and validate robustness.
Category: Opinion
What models handle Ordinal Scale outcomes?
Use ordinal logistic/probit regression, cumulative link models, or rank-based methods.
Category: Requirements
What is a common mistake with Ordinal Scale features?
Using label encoding as continuous numeric values, implying equal distances between ranks.
Category: Mistakes
How should Ordinal Scale data be encoded?
Use ordinal encoders that preserve order or one-hot encoding if order is not needed by the model.
Category: Example
How to visualize Ordinal Scale data?
Ordered bar charts, cumulative distribution plots, and stacked bars convey ordered categories well.
Category: Example
What risks arise from misclassifying Ordinal Scale data?
Assuming equal intervals can bias coefficients, p-values, and decisions.
Category: Risk
When is Spearman correlation suitable for Ordinal data?
When assessing monotonic association between two ordinal variables or ordinal with continuous data.
Category: Requirements
What is an example of Ordinal Scale data?
Customer satisfaction ratings: very dissatisfied, dissatisfied, neutral, satisfied, very satisfied.
Category: Example