Measurement scales

Variables may be measured on different types of scale:

1. Nominal - where classification is based on a descriptive characteristic (e.g. colour). This is the only scale for qualitative data.
2. Ordinal - this classifies by numerical rank, from smallest to greatest, but with no assumption of equal spacing between ranks.
3. Interval - for certain quantitative variables, where numbers on an equal unit scale are related to an arbitrary zero, e.g. temperature in °C.
4. Ratio - similar to the interval scale, except that the zero point represents an absence of that character, i.e. it is an absolute zero.

You should aim to do quantitative measurements using a ratio scale whenever possible, to allow you to use the broadest range of mathematical operations and statistical procedures. For example, if you are measuring the surface area of a sample of graphite you could give the area as 200 m² but, if you know the mass of the sample (10 g), you should quote the surface area as 20m² g⁻¹.