The measurement scales differ in three important ways:

• Hierarchy
• Distance
• Zero point

Checklist

To find out what kind of variable you have, we suggest that you ask yourself the following questions:

Is it possible to arrange/order the values hierarchically?	Yes/No
Is it the same distance between the values?	Yes/No
Does the scale have an absolute zero point?	Yes/No

Hierarchy

What does “arrange/order the values hierarchically” mean?

If we take gender as an example, it is not reasonable to say that “Man” is less or more than “Woman”. As another example, we can take nationality: it is not reasonable to see “Danish” as less or more than “Finnish”.

For variables such as self-rated health, on the other hand, it is possible to say that “Excellent health” is better than “Good health”. Moreover, it is possible to say that the grade “A” is better than the grade “B”. 

Distance

What does “distance” mean?

If we take income as an example, we know that 1000 dollars are twice as much as 500 dollars, and 2000 dollars are twice as much as 1000 dollars. The same logic applies to variables such as age: it is the same distance between 2 years and 4 years as between 6 years and 8 years.

Thus, having the same distance between the values means that the differences between two values are the same regardless of which part of the scale you are looking at.

Zero point

What does “absolute zero point” mean?

Basically, it means that the scale cannot have negative values. It is possible for the temperature to be minus 10 degrees Celsius, but is not possible to have less than zero years of schooling or having less than zero days of unemployment.

Relating it to measurement scales

Nominal variables

Characteristics
Order values: No
Same distance: No
Absolute zero point: N/A

Examples
Yes/no questions
Gender
Nationality

Nominal variables are variables for which the values cannot be ranked, and we do not have the same distance between the values, e.g. gender or questions that can be answered with yes or no.

Ordinal variables

Characteristics
Order values: Yes
Same distance: No
Absolute zero point: N/A

Examples
Attitude questions
Self-rated health
Educational level

Ordinal variables are similar to nominal ones, but here the values can be ranked, such as for self-rated health: “Excellent is better than “Good”; “Good” is better than “Fair”; and “Fair” is better than “Poor”. Similarly, however, for ordinal scales we do not have the same distance between the values: the “amount” of better health is not necessarily the same between “Poor” and “Fair” as between “Good” and “Excellent”.

Interval variables

Characteristics
Order values: Yes
Same distance: Yes
Absolute zero point: No

Examples
Temperature (Celsius)

Interval variables are similar to ratio variables, in the sense that values can be ranked and we do have the same distance between the values, but here we do not have an absolute zero point (values can be negative).

Ratio variables

Characteristics
Order values: Yes
Same distance: Yes
Absolute zero point: Yes

Examples
Age
Income
School marks

Ratio variables are similar to the ordinal variables in the sense that values can be ranked but, additionally, we do have the same distance between the values: for example, we know that 10 years of schooling is twice as much as 5 years of schooling. We also do have an absolute zero point since no value can be lower than 0 (you cannot have a negative number of years of schooling).