Every number in a research report sits on one of four measurement scales: nominal, ordinal, interval or ratio. The scale decides which statistics you are allowed to run. Averaging the wrong kind of number is one of the most common errors in survey analysis.
The classification comes from psychologist S. S. Stevens, who set it out in his 1946 paper “On the Theory of Scales of Measurement” in the journal Science.
Key takeaways
- Nominal scales name categories. You can count them but not rank them.
- Ordinal scales rank things, but the gaps between ranks are unknown.
- Interval scales have equal gaps but no true zero, so ratios are meaningless.
- Ratio scales have a true zero, so every calculation is allowed.
The four scales at a glance
| Scale | What it does | Marketing example | Valid summary |
|---|---|---|---|
| Nominal | Names categories | Brand used most often, gender, city | Counts, percentages, mode |
| Ordinal | Ranks in order | Brand preference ranking, satisfaction rating | Median, percentiles |
| Interval | Equal gaps, no true zero | Temperature in Celsius, calendar year | Mean, standard deviation |
| Ratio | Equal gaps and a true zero | Spend, units bought, age, income | All of the above, plus ratios |
1. Nominal scale
A nominal scale labels things. The numbers, if any, are codes with no mathematical meaning. If 1 = Delhi and 2 = Mumbai, Mumbai is not twice Delhi. You can report how many respondents fall in each category and which category is the most common. You cannot calculate an average.
2. Ordinal scale
An ordinal scale puts things in order. If a respondent ranks three brands first, second and third, you know the order of preference but not the distance between them. First may be far ahead of second, or barely ahead. The median is a safe summary.
Rating scales such as a five-point agreement scale are strictly ordinal. In practice many researchers treat them as interval and report means. That is a common convention, but it is worth knowing it is an assumption.
3. Interval scale
An interval scale has equal distances between points, but its zero is arbitrary. Temperature in Celsius is the classic case: the gap between 10 and 20 degrees equals the gap between 20 and 30, but 20 degrees is not twice as hot as 10. You can add, subtract and average. You cannot form ratios.
4. Ratio scale
A ratio scale has equal intervals and a true zero that means “none”. Purchase value, number of units, age and income are all ratio measures. A shopper who spends 400 does spend twice as much as one who spends 200. Every arithmetic operation is valid.
Why the scale matters
The scale tells you which analysis is defensible:
- Do not average nominal codes. Report percentages.
- Be cautious with means of ranked data. Check the distribution as well.
- Do not claim “twice as much” unless the measure has a true zero.
- Choose your statistical test to match the scale of the variable you are testing.
Checking the scale before running the numbers takes seconds and prevents conclusions that do not hold.
FAQ
What are the four measurement scales?
Nominal, ordinal, interval and ratio. Each adds a property: naming, order, equal intervals, and a true zero.
Is a Likert scale ordinal or interval?
Strictly it is ordinal, because the gaps between points are not proven equal. Many researchers treat it as interval by convention.
What is the difference between interval and ratio scales?
A ratio scale has a true zero meaning none of the quantity, so ratios are meaningful. An interval scale does not.
1 thought on “4 Measurement Scales Every Researcher and Marketer Should Know”