Levels of Measurement

Introduction

The levels of measurement mentioned here refers to how we categorise data in statistics. The two important terms used in this article will be variable and value. These two words are best understood when given in an example.

You receive a questionnaire after purchasing coffee from your favourite cafe. The question is "How often do you order this drink?". You are given a number of choices ranging from 1-5. The variable is be Favourite Drink and the value is the number of orders.

In other words, a variable is simply the category and whatever is inside is the value.

Not all variables are made equally. There are four level of measurements that we use in statistics:

1. Nominal

2. Ordinal

3. Interval

4. Ratio

Nominal Scale

How I prefer to understand the nominal scale is that we use numbers to name or nominate.

Example: What type of property do you own?

1. House

2. Town house

3. Apartment

4. Bungalow

There are four levels of property to this question and the values have been nominated by the numbers 1-4 respectively. It would not make sense to calculate the average type of property.

Ordinal Scale

Ordinal scale is a combination of nominal and order, where numbers are used to both label and order the value. The term order refers to the hierarchy of the value, such as in medium is valued higher than low, and so on.

Example: Perception of your body weight:

1. Underweight

2. About right

3. Overweight

The highest order is overweight and the lowest order is underweight.

Interval Scale and Ratio Scale

The interval scale and the ratio scale is the pretty much the same. They both use numbers to describe their respective values. The most important difference between the two is:

0 in the interval scale does not mean absence of the quantity being measured. 0 is just another number in the scale.

Lastly, the reason why we need to know the levels of measurement is because we need to know how we will categorise our data in statistics software applications such as SPSS. SPSS registers nominal or ordinal data as categorical and interval and ratio as metric.