Levels of Measurement in Statistics

When it comes to statistical analysis, there are four levels of measurement or types of data that you need to be aware. The levels of measurement you are dealing with will in part determine what sort of statistical test for significance you will perform on the data. Basically, the levels of measurement used in statistics and mathematics for a particular variable tell you the nature of the information or data contained within the numbers of the variable. The four levels of measurement are: nominal, ordinal, interval, and ratio. This article will help you differentiate each.

NOMINAL

A nominal variable contains categorical data for mutually exclusive, but not-ordered, categories. The categories may be coded by a number for calculation purposes, but the order of the categories does not matter. Examples of nominal (or categorical) data include marital status, ethnicity, religious affiliation, and gender.

ORDINAL

For ordinal data, the order does matter but the difference between values does not. An example of ordinal data would be on a survey where the participant is asked to rate something on a scale of 1 to 10 (i.e. pain ratings) or a Likert Scale (such as movie ratings, education level, or other rank orders).

INTERVAL

With interval data (sometimes called scale variables), the difference between values does matter, but the zero point is arbitrary. An example of interval data is a temperature scale (Fahrenheit or Celsius).

RATIO

Ratio data has similar properties as an interval variable, but there is a clear definition of 0 (zero). Examples of ratio data include height, weight, Kelvin temperature (includes absolute zero), population, and annual income. Numbers can be compared as multiples of one another.