What are the four levels of measurement?

Nominal, ordinal, interval, and ratio. Nominal data are unordered categories (such as blood type). Ordinal data are ordered categories with unequal or unknown gaps (such as pain rated mild, moderate, severe). Interval data have equal gaps but no true zero (such as temperature in Celsius). Ratio data have equal gaps and a true zero (such as weight or reaction time).

Why do levels of measurement matter?

Because the level determines which statistics are legitimate. You can count categories at any level, but you can only compute a mean at interval or ratio level, and only ratios (twice as much) at ratio level. Choosing a statistic the data cannot support produces numbers that look precise but mean nothing.

What is the difference between interval and ratio data?

Both have equal intervals between values, but ratio data have a true, meaningful zero and interval data do not. Because 0 kg means no weight, 20 kg is twice 10 kg (ratio). Because 0 degrees Celsius is not the absence of temperature, 20 C is not twice as hot as 10 C (interval). The true zero is what licenses ratio statements.

What level of measurement is a Likert scale?

A single Likert item is ordinal, because the gaps between response options are not guaranteed equal. A multi-item Likert scale (the sum or mean of several items) is commonly treated as approximately interval, which is what allows means and parametric tests. The distinction decides how you analyse the responses.

Home › Guides › Data Collection › Levels of measurement

Levels of measurement

By Dr. Rafiq Muhammad, MD, PhD · Updated June 2026

Every variable you collect sits at one of four levels of measurement — nominal, ordinal, interval, ratio — and the level isn’t a technicality: it decides which statistics you’re allowed to use. Compute a mean on the wrong level and you get a number that looks precise and means nothing. Decide the level for each variable as you design your instrument, not after the data is in.

The four levels

Level	What it adds	Example	Centre
Nominal	Unordered categories	Blood type, country	Mode
Ordinal	+ order (unequal gaps)	Pain: mild/moderate/severe	Median
Interval	+ equal gaps (no true zero)	Temperature (°C)	Mean
Ratio	+ true zero	Weight, reaction time	Mean

Each level adds a property to the one before it — order, then equal spacing, then a meaningful zero — so it’s a ladder, not four unrelated boxes.

Interval vs ratio: the true zero

The only difference is a meaningful zero. 0 kg means “no weight,” so 20 kg is genuinely twice 10 kg — that’s ratio. 0 °C does not mean “no temperature,” so 20 °C is not twice as hot as 10 °C — that’s interval. The true zero is what licenses “twice as much” statements.

Why it decides your statistics

The level caps what’s legitimate: you can count and find the mode at any level; the median needs order (ordinal+); the mean needs equal intervals (interval+); ratios (“twice as much”) need a true zero (ratio). This is exactly why a single Likert item (ordinal) should be summarised by a median, while a multi-item Likert scale (≈interval) can take a mean. It’s also the first question the statistical-test guide asks when choosing a test.

Level of measurement is step one in choosing an analysis. See how it drives the rest in which statistical test should I use? and try the Statistical Test Selector.

Get the free Data Collection toolkit

A measurement worksheet, questionnaire-design checklist, and a data-management-plan starter — from Research Design Simplified. We’ll email you the download link.

Get Research Design Simplified

Frequently asked questions

What are the four levels?

Nominal (unordered categories), ordinal (ordered, unequal gaps), interval (equal gaps, no true zero), ratio (equal gaps + true zero).

Why do they matter?

The level caps which statistics are valid — e.g. a mean needs interval or ratio data. Wrong level → meaningless numbers.

Interval vs ratio?

Ratio has a meaningful zero (20 kg is twice 10 kg); interval doesn’t (20°C isn’t twice 10°C).

What level is a Likert scale?

A single item is ordinal; a multi-item scale is treated as approximately interval.

Cronbach’s alpha & reliability → Open the Statistical Test Selector →