In data analysis, numbers can be used to ‘measure’ things in different ways. These measurement systems – Nominal, Ordinal, Interval, and Ratio – tell us how precisely variables are recorded, and more importantly, they determine what level of insights can be derived.

When analyzing data, it is critically important to understand how – and when – to use each type of measurement system. Choosing the wrong one could have profound consequences, especially when assessing risk.

Nominal scales are used to describe (or identify) things. They are not really numbers in the mathematical sense; they are more like labels. Objects are “measured” by determining the categories to which they belong. Eye color, blood type, and nationality are common examples.

Ordinal scales are used to rank-order objects. These systems tell us that one item may be more – or less – than another, but we cannot state by how much because there are no fundamental units we can use to compare things.  The interval between each unit may be different. A common example is the Likert Scale: the difference between “strongly agree” and “agree” is practically impossible to determine. 

Interval scales are like ordinal scales; except they have defined units. Therefore, the intervals between each value are equally split. A common example is the Celsius temperature scale. Each temperature unit – degree– is the same. 

Ratio scales are interval scales (i.e., they have fundamental units) with a true zero point. This scale allows us to do the math we learned in grade school: we can multiply, divide, add, and subtract. This is the measurement system that allows us to compare objects using ratios. Common examples are distance and money.

What Does This Mean for Hazard Scoring Systems?

Hazard scoring are ordinal measurement systems. They rank order hazards, usually on a scale from ‘1’ (the lowest category of risk) to ‘5’ as the highest category of risk. There may be many hazards in the ‘5’ category, but do they have the same level of risk? Is a flight in severe icing equal to a flight at night? Of course not! In addition, since there are no fundamental units in ordinal scales, we cannot add hazards together. A pop-up trip (level ‘3’ hazard) at twilight (level ‘2’) hazard does not equal severe icing (level ‘5’ hazard).

Treating ordinal data like ratio data can have severe consequences. By understanding the different types of measurement systems – and their inherent limitations – we can make more effective decisions. 

Question: Does 2 + 2 = 4?

Answer: Yes, IF you are using Interval or Ratio measurement systems.