Stevens' four scales — nominal, ordinal, interval and ratio — describe how much information a measurement carries.
Why it matters
The scale limits which calculations are meaningful and which tests are valid.
Key ideas
- Nominal: names only (blood group)
- Ordinal: order but unequal gaps (pain: mild, moderate, severe)
- Interval: equal gaps, no true zero (temperature in °C)
- Ratio: equal gaps and true zero (weight, income)
- Higher scales allow all operations of lower ones
Real-life research example
Satisfaction on a 5-point scale is ordinal: the gap between 1 and 2 need not equal the gap between 4 and 5. Weight in kg is ratio: 80 kg is twice 40 kg.
When to use
- Use the scale to choose summaries and tests (median for ordinal, mean for interval/ratio)
When NOT to use
- Don't call "twice as hot" for 20 °C vs 10 °C
- Don't treat Likert items as exact numbers without thinking
Quick takeaway
Nominal < ordinal < interval < ratio. The higher the scale, the more you can calculate.
Try it yourself
Classify each: blood group, pain level (mild/moderate/severe), temperature in °C, weight in kg. Then paste the pain levels below and see how they can be summarised.
Model answer
Nominal, ordinal, interval, ratio. Ordinal values should be summarised with counts and percentages (or the median), not a precise mean.