There are many different data collection methods for both quantitative and qualitative data. You can use the myriad surveys available, personally observe, and much more. You’ll likely come across one of the four major scales used to measure data.

These scales, called nominal, ordinal, interval, and ratio scales, are the basis of all the types of scale questions you’ve come across in the past. For example, a rating scale of one to five is an ordinal scale and so is a Likert scale. A question that seeks to understand genders and provides multiple options would be a nominal scale.

Measurement scales don’t live in a bubble and each one adds to the one preceding it. All of the scaled questions that you’re familiar with are derived from these four types of scales.

In this guide, we’ll take a deep dive into the four types of measurement, provide examples, and show you when to use them in your own research.

What are the nominal, ordinal, interval, and ratio scales?

The ordinal, interval, nominal, and ratio scales are described as four scales of measurement that are utilized to get data from forms, questionnaires, surveys, etc. using multiple-choice questions. All the measurement scales have MCQ questions.

This categorization was created back in 1946 by psychologist Stanley Smith and has remained the most popular method throughout the years. The graph below highlights the type of calculation or analysis each of the scales can be used for.  

Type of calculationNominalOrdinalIntervalRatio
Frequency distributionYesYesYesYes
ModeYesYesYesYes
MedianNoYesYesYes
Addition and subtractionNoNoYesYes
Multiplication and divisionNoNoNoYes
Mean, standard deviationnoNoYesYes
Ratios, coefficient of variationNoNoNoYes
Geometric meanNoNoNoYes

Nominal scale

A nominal scale can be referred to as a categorical variable scale. This is because it deals with variables, categories, options, groupings, etc. that aren’t affected by the order and have no inherent ranking. For example, Apple Computers, HP computers, Dell computers, etc. are all types of personal computers. They also don’t have any inherent ranking. Of course, people can rank them if they’d like but the ranks wouldn’t apply universally.

When researching, you can apply a value to the variables on a nominal scale but those numbers don’t have a meaning beyond the research being performed. You wouldn’t be able to use those assigned values to calculate something like frequency distribution or the mean.

Using the above example, if you have Apple a value of 1, HP a value of 2, and Dell a value of 3, they couldn’t be used to calculate the mean or the mode. At the end of the day, nominal information can be considered labels for the information being collected.

The Apple label is used to describe a certain type of computer and the HP label is used to describe a different type of computer. None of these labels overlap and they also don’t have any numerical importance (even if you attach a numerical value to them).

Nominal scale examples

What is your favorite sneaker shoe brand?

  • Nike
  • Adidas
  • Pumas
  • Reebok
  • Converse

What is your favorite color?

  • Green
  • Red
  • Yellow
  • Cyan

What type of vacation activity do you prefer?

  • Hiking
  • Skiing
  • Ocean sports
  • Leisure

If you look at the examples above, it becomes clear that on a nominal scale only the names of answer options are important to the researcher. It doesn’t matter if hiking, skiing, or ocean sports is first on the list or last on the list. It matters what survey respondents choose.

Though nominal scales don’t care about the order, some of them may have a de facto order. For example, very small, small, medium, large, very large. Technically, you can put these options anywhere on a nominal scale but they have an inherent order. Even though they have an order, it’s irrelevant to the analysis that will be performed later.

Nominal scales are the backbone of quantitative research and are arguably the most widely used scale for measurement.

Ordinal scale

An ordinal scale builds on the nominal scale and is considered the second level of the measurement hierarchy coined by Stanley Smith. The defining feature of an ordinal scale is that the order of variables is important but the differences between those variables (also referred to as values) aren’t important.

This stems from the inability or difficulty associated with accurately expressing the degree of difference between values on an ordinal scale.

For example, how do you measure the difference between satisfied and very satisfied? You may have a vague idea and understand that being very satisfied is usually a better outcome for a brand. But, beyond that, how would you quantify the difference?

Can you objectively say that very satisfied is X units of satisfaction greater than satisfied? Unfortunately, we have no way of doing that.

Even when you can express the difference between values, it’s not as important as the way in which values are arranged on the scale.

For example, if you have an ordinal scale that has the following options:

How many times do you work out every week?

  • 1
  • 2 – 5
  • 6 – 10
  • 10+

The values are known and quantifiable but the difference between each one isn’t uniform. What’s more important is the order of the values on the scale.

An ordinal scale comes into play when you’re trying to understand and quantify information or variables that don’t have a clear mathematical aspect. This includes but isn’t limited to sadness, happiness, frequency, the intensity of a feeling, etc. An easy way to remember ordinal scales is by associating ordinal with order and ordinal scales gain their usefulness from the order of the variables on the scale.

The last thing to note is that ordinal scales don’t have a useful origin so it’s not possible to say where the scale starts or even ends.

Note: ordinal scales (and many other types of scales for that matter) can be impacted by central tendency bias so be on the lookout for it and plan accordingly.

Ordinal scale examples

How was your shopping experience?

  • Amazing
  • Good
  • Neither good nor bad
  • Bad
  • Very bad

How often do you eat red meat?

  • Very occasionally
  • Occasionally
  • Neither often nor occasionally
  • Often
  • Very often

When analyzing ordinal data, it can be presented in multiple ways. The most common way to present it is by using various graphs to show the percentage of people that selected each variable. Conversely, it can show the number of respondents as a figure.

Difference between nominal measurement scale and ordinal measurement scale

In a nutshell, the nominal scale only looks at the variable itself and ignores the order it appears on the scale. Ordinal scales consider the variable and the position it occupies on the scale. It does not look at the numerical value of the measurement scale and neither do nominal scales.

Nominal scales can have as few as two answers because there’s no need to establish a scale. Ordinal scales need at least three options so that a scale can be established.

Interval scale

An interval scale (also known as an interval variable scale) can be defined as a numerical scale where the order of the variables matters and the difference in value between them is important. It can be considered the third type or level of measurement and continues to build on the previous two. Interval scales have no true zero and can measure sub-zero values.

Even though there’s no true zero value, each point on the scale is equidistant from the previous point or the next point.

  • On the Fahrenheit measurement scale, each value has the same increment. One Fahrenheit is constant. So, if something is ten degrees and something else is twenty degrees, the magnitude of change can be measured because each individual Fahrenheit has the same value.
  • At the same time, you can have negative Fahrenheit temperatures so the zero on the scale is just another value. It doesn’t represent the absence of what is being measured. 

Interval scale examples

What was your SAT score?

  • 400
  • 500
  • 600
  • 700
  • 800

Ratio scale

The fourth type of measurement takes all the other types into consideration and builds on top of them. The ratio scale derives importance from the value between units, needs a set order, and even comes with true zero. In essence, you can call it an interval scale that has a zero.

In addition to having a zero, the proportion between values is also important. 20 meters is twice as long as 10 meters and 40 meters is twice as long as 20 meters. With an interval scale, 20 degrees isn’t necessarily twice as hot as 10 degrees.

Length, duration, weight, speed, etc. are examples of ratio scales. The zero allows a wider range of statistical, descriptive, inferential, and other types of analysis. Ratio scales are also considered more accurate than something like a nominal scale.

This can be important when you’re trying to understand market forces or make policy decisions that require the allocation of resources. Even though they’re not as widely used as nominal scales the data is considered more versatile.

Ratio scale examples

What is your height in feet?

  • Less than five feet
  • Between five and five and a half feet
  • Between five and a half feet and six feet

How long did it take to finish the exam?

  • 1 hour
  • 2 hours
  • 3 hours
  • 4 hours

Interval scale vs ratio measurement scale

Even though they’re related, there are clear differences between the two measurement scales. It’s important to understand them so you don’t try to perform the wrong type of analysis on the data you gain. The most important differences include:

  • For interval data, the ratio between values is meaningless while the ratio between values in a ratio scale is meaningful.  
  • A ratio scale has a true zero while an interval scale does not. If you have a measure of zero on a ratio scale, that means you have an absence of the value being measured. For an interval scale, zero is just another measured value. Temperature can go below zero while weight cannot go below zero.

Conclusion

Insight into the type of measurement scale that can be applied in different situations can help you make more informed decisions. You can collect more useful data and apply the right statistical analysis methods.

Remember, nominal scales are the most limited when it comes to analysis and ratio scales are the most versatile. Choose the one that fits your situation and unlock the ability to perform more complex analysis. Let me know what you think in the comments and don’t forget to share.