SPSS Tutorials


Measurement Levels – What and Why?

The first step in data analysis is simply understanding what the data mean. This is facilitated by classifying each variable according to its measurement level. Standard textbooks distinguish four such measurement levels: nominal, ordinal, interval and ratio.

Measurement Levels - Classical Approach

Measurement Levels Quick Overview of Measurement Levels

1. Nominal Measurement Level

The nominal measurement level implies that there's no undisputable order in what's represented by the data values. An easy example is country. We may sort countries alphabetically, according to their sizes or to numbers of inhabitants. Different orders make equal sense for a list of countries.
Now don't overlook the phrase “what's being represented”. It's not the data values that matter. It's what they represent in the real world. So a tricky example are ZIP codes. These actually represent geographical areas, which don't have an undisputable order. This renders ZIP code a nominal variable.

2. Ordinal Measurement Level

The ordinal measurement level implies that there's a clear order but no fixed unit of measurement in what's represented by the data values. Typical examples are attitude scales like

  1. Bad;
  2. Neutral;
  3. Good.

known as Likert scales. There's a clear order in the sentiments represented by these values. "Neutral" is clearly between "Bad" and "Good". However, we've no clue just where between those. For instance, a demanding respondent who rates almost everything "Bad" may feel "Neutral" is actually closer to "Good" than to "Bad". This is illustrated by the figure below.

Ordinal Measurement Level Intervals between answer categories are unknown for ordinal variables.

A different way to state this is that the sentiments represented here don't have any fixed unit of measurement. This renders their data values rather arbitrary. We used "1", "2" and "3" to represent "Bad", "Neutral" and "Good" but we may just as well use "1", "1.5" and "3" instead.
Note that this arbitrary choice dictates the outcomes of calculations on this variable. This is why calculations on ordinal variables are not meaningful. Less strictly though, calculations on ordinal variables are quite common under the Assumption of Equal Intervals.

Note that monthly income measured as

  1. Less than 1000 Euros;
  2. 1000 to 2000 Euros;
  3. 2000 Euros or over.

is ordinal as well. These three values really represent income categories and these don't have a fixed unit of measurement. We'll say that “income was measured at the ordinal level”.
We could measure it at ratio level too by asking the exact income in Euros. The point here is that a single variable (income) can sometimes be measured at different measurement levels.

3. Interval Measurement Level

The interval measurement level implies a fixed unit of measurement but no natural zero point. True interval variables are rare but one of the few examples is temperature in degrees Celsius.
The measurement unit is fixed because every degree Celsius represents the same amount (“interval”) of temperature. However, zero degrees Celsius does not mean “nothing” with regard to temperature.With regard to temperature, we'd probably say that "nothing" corresponds to -273.15 degrees Celsius.
Without a natural zero point, multiplication is not meaningful. It doesn't make sense to say that 40 degrees Celsius is “twice as warm” as 20 degrees.

4. Ratio Measurement Level

The ratio measurement level implies a fixed unit of measurement and a natural zero point. An example is weight in kilos. A kilo is a fixed unit of measurement because it always represents the exact same weight.
So what is meant by "a natural zero point"? A natural zero point means that zero represents “nothing”. This holds because zero kilos corresponds to “nothing” with regard to weight.
We rather avoid “absolute zero point” because ratio variables may hold negative values; the balance of my bank account may be negative but is has a fixed unit of measurement -Euros in my case- and zero means “nothing”.
This natural zero point is what really renders multiplication - and therefore ratios - of values meaningful; it makes perfect sense to say that € 2000,- is two times € 1000,- or four times € 500,-.

Classical Measurement Levels - Shortcomings

We argued that measurement levels matter because they facilitate data analysis. However, when we look at the most common statistical techniques, we see that

In order to fix these issues, we propose a more useful classification below.

Measurement Levels - Modern Approach

Ordinal Measurement Level

Well, that's about it. Whether or not you agree with my opinion on measurement levels, I hope you found it useful anyway. Thanks for reading!

Previous tutorial: Statistical Significance – What Does It Really Mean?

Next tutorial: SPSS Variable Types and Formats Tutorial

Let me know what you think!

*Required field. Your comment will show up after approval from a moderator.

This tutorial has 21 comments