Ordinal data: definition and survey examples
Ordinal data is data in categories that have a clear order, but where the distance between two categories cannot be measured. A survey question with the answers ‘Poor’, ‘Fair’ and ‘Good’ gives ordinal data.
What is ordinal data?
Ordinal data lets you sort answers from low to high. You know that ‘Good’ is higher than ‘Fair’, but you do not know how much higher. Wikipedia describes the ordinal type in the same way: it allows rank order, and the real differences between adjacent ranks may not be equal. Source: Wikipedia: Level of measurement.
Sullivan and Artino use the same idea for survey answers: in an ordinal scale, responses can be rated or ranked, but the distance between responses is not measurable. Source: Sullivan and Artino: Analyzing and Interpreting Data From Likert-Type Scales, Journal of Graduate Medical Education (2013).
5 examples of ordinal data in surveys
These examples are illustrations. Each one has ordered answers.
- Likert items. ‘The form was easy to complete.’ Strongly disagree / Disagree / Neither agree nor disagree / Agree / Strongly agree.
- Satisfaction ratings. ‘How satisfied are you with support?’ Very dissatisfied / Dissatisfied / Neutral / Satisfied / Very satisfied.
- Rankings. ‘Rank these 3 features from most to least useful.’ The answers are 1st, 2nd and 3rd.
- Education level. ‘What is your highest level of education?’ Secondary school / Bachelor’s degree / Master’s degree / Doctorate.
- Frequency. ‘How often do you read our newsletter?’ Never / Rarely / Sometimes / Often / Always.
Likert items are ordinal. Read our Likert scale guide for more labels and question examples.
Nominal, ordinal, interval and ratio data
Statisticians sort data into 4 levels of measurement. Each level allows different operations. Source: Wikipedia: Level of measurement.
| Data type | What it tells you | Survey example |
|---|---|---|
| Nominal | Names only, with no order | ‘Which channel did you use?’ Email, chat or phone |
| Ordinal | Order, but no measured distance | ‘How satisfied are you?’ Very dissatisfied to Very satisfied |
| Interval | Equal distances, but no true zero | A temperature in degrees Celsius, for example in a survey about room comfort |
| Ratio | Equal distances and a true zero | ‘How many minutes did sign-up take?’ |
Wikipedia gives the Celsius scale as an example of interval data, and mass, length and duration as examples of ratio data. Source: Wikipedia: Level of measurement.
How to summarize ordinal data
Choose a summary that does not need equal distances between answers.
- Median. For ordinal data, Stevens held that the median is the appropriate measure of central tendency, and the mode is also allowed. The mean is not. Source: Wikipedia: Level of measurement. Sullivan and Artino describe the argument for the median for Likert data. Source: Sullivan and Artino: Analyzing and Interpreting Data From Likert-Type Scales, Journal of Graduate Medical Education (2013).
- Mode. The mode is the answer that most people chose. It works for nominal data too. Source: Wikipedia: Level of measurement.
- Percentage of each label. Sullivan and Artino write that a frequency distribution of responses will likely be more helpful than a mean for Likert data. Source: Sullivan and Artino: Analyzing and Interpreting Data From Likert-Type Scales, Journal of Graduate Medical Education (2013).
- Chi-square for a crosstab. A chi-square test of independence checks whether two variables in a contingency table are independent of each other. Source: Wikipedia: Pearson's chi-squared test. Sullivan and Artino list chi-square tests among the methods that some experts recommend for ordinal data. Source: Sullivan and Artino: Analyzing and Interpreting Data From Likert-Type Scales, Journal of Graduate Medical Education (2013). Use our chi-square calculator to run the test on a survey crosstab.
To get the median, the mode and the percentage of each label from your answers, use our Likert scale calculator.
Collect ordinal answers with Tickbox
Tickbox is a form builder with voice and video. You can add rating and choice questions to a form, and AI analyzes the responses so they are easy to read.
Sources
We checked these sources on 18 November 2026. Each one supports the statements that link to it.
- Wikipedia: Level of measurement: the ordinal type, rank order, the median and the mode, and the examples for each level.
- Sullivan and Artino: Analyzing and Interpreting Data From Likert-Type Scales, Journal of Graduate Medical Education (2013): ordinal responses, frequency distributions, the median and chi-square tests.
- Wikipedia: Pearson's chi-squared test: the chi-square test of independence for a contingency table.