# The Levels of Measurement Explained (With Examples)

By Indeed Editorial Team

Published June 18, 2022

The Indeed Editorial Team comprises a diverse and talented team of writers, researchers and subject matter experts equipped with Indeed's data and insights to deliver useful tips to help guide your career journey.

Measurement in research involves assigning numerals or other symbols to objects, events, properties, or other data points according to specific rules. This occurs on four levels and is used to identify or categorize values within a data set. A thorough understanding of the four measurement levels can help you assess datasets and be more effective in your research. In this article, we discuss the four levels of measurement, review why these levels of measurement are important, and share some examples to help guide you when conducting your own research.

## What are the 4 levels of measurement?

Levels of measurement, also known as scales of measure, indicate the precision of variables measured and organized in a dataset. Here's a list of the four measurement levels researchers typically use to identify and organize research results:

### 1. Nominal

At this level of measurement, researchers typically use arbitrary numerals or other symbols, including words, letters, and alphanumeric characters, to classify people, objects, characteristics, or values in a data set. This is the lowest of the four levels, as it classifies variables into named categories that differentiate them yet don't imply superiority of one over the other. This means the numbers or tags you assign to the categories don't hold value, and one category of variable isn't necessarily higher or lower than another.

The nominal level is often relevant for research situations where only labels and tags are significant. This level can also be appropriate when it's necessary for researchers to conduct limited statistical manipulations on research data. You can calculate only the mode and percentage distribution at this level. For example, if a company is trying to determine the geographic location of their customers, they can use nominal measurement.

Related: How to Do Market Research With 6 Guided Steps (With Types)

### 2. Ordinal

At this level, researchers rank items along a continuum. The ordinal level typically categorizes values in a data set according to an ordered relationship, and this process indicates rank. This classification means that the numbers or tags you assign to the categories may hold some value, and each category of variable either ranks higher or lower than another. For example, a survey might ask you to rate your recent shopping experience from one to 10, where lower numbers mean you are less satisfied, and higher numbers mean you are more satisfied.

As ordinal measurement defines items based on rank, ordinal scales can evaluate rank-based or value-based questions. With ordinal variables, you can use all the statistical manipulations appropriate to nominal variables such as mode and frequency distribution, and others such as the median, percentile, and other non-parametric statistical tests.

Read more: How to Find Critical Value in Two Steps (With Examples)

### 3. Interval

The interval level is one where the difference between two categories is equal. It's the first genuinely quantitative level of measurement, and it possesses all the characteristics of nominal and ordinal variables. An interval measure can form a scale along which variables occur at equal frequencies, and researchers can ascribe meaning to the difference or distance between each variable on the scale.

The interval measurement level is appropriate in situations where it's necessary for researchers to perform the broadest range of statistical manipulations on the data they collect. It can allow researchers to apply all the statistical analyzes possible with the previous two levels and additional measures, such as mean, standard deviation, and many other parametric statistics. Researchers can use the interval scale to evaluate variable number differences. Interval scales can process averaging data, unlike ordinal and nominal scales. Interval data can also be used to determine numerical values from statistics results.

Related: Examples of Negative Correlation and How to Calculate It

### 4. Ratio

This measurement level has all the properties of an interval scale in addition to a real zero point, indicating the absence of the variable of interest. Because the interval and ratio levels are similar, the rules for assigning numbers are the same. The only difference between both levels is the zero point. For example, zero isn't the lowest possible temperature on either the Celsius or Fahrenheit scales, so they're both interval scales. In contrast, a speedometer is a ratio scale because the zero point implies the lowest point on the scale, as it indicates no movement. Because ratio scales indicate an absolute zero, negative numbers aren't applicable at this level of measurement.

Read more: Research Skills: Definition and Examples

## Why are measurement levels important?

The different measurement levels are essential in research because they indicate which methods a researcher uses to analyze and interpret data from a variable. The various levels illustrate which descriptive metrics you can use to gain an overview of your data and which inferential analysis approaches you can use to verify or disprove your research hypotheses. They can also determine how you present your data. In many situations, you can measure your variables at multiple levels. It's often necessary to select the appropriate measurement level for your research prior to beginning data collection.

## Examples of variables at each measurement level

### Nominal scale

When collecting data about people belonging to different gender categories in a population, a researcher can assign tags or placeholders for each category using the nominal scale. For instance, the researcher can classify every person who identifies as female as F, individuals who identify as male as M, and every non-binary person as N. The researcher may use any other figure or symbol of their preference to represent these categories. For instance, M can be 1, F can be 2, and N can be 3. These tags only identify the different categories and have no other implication.

Here is a list of other examples of nominal data:

• Personality type: Such as introvert, extrovert, ambivert

• Nationality: Such as French, Indian, Canadian

• Employment status: Such as unemployed, part-time, full-time

• Eye colour: Such as blue, brown, green

• Place of residence: Such as Toronto, Quebec, Canada

### Ordinal scale

Suppose a researcher asks respondents about their attitude towards a recent change in the design of a product. The researcher can assign numbers to the responses as follows:

• 1: Strongly approve

• 2: Somewhat approve

• 3: Somewhat disapprove

• 4: Strong disapprove

The numbers in this instance identify the various categories and rank them according to the level of approval, with 1 representing total approval and 4 representing disapproval . Here is a list of other instances of ordinal data:

• School grade systems: Such as A, B, C

• Agreement: Such as strongly agree, agree, disagree, strongly disagree

• Level of education: Such as high school, bachelor's degree, master's degree

• Satisfaction: Such as a Likert scale on a grade of 1–5

• Position at work: Such as entry, mid, senior-level

### Interval scale

A thermometer displays an interval level of temperature measurement. For instance, the difference between 10 degrees and 11 degrees is the same as the difference between 20 degrees and 21 degrees. Both instances differ by one degree. Here is a list of other examples of interval data:

• IQ test scores

• Time on a clock

• Voltage

• Dates

Related: Data Maintenance vs. Data Cleansing: What's the Difference?

### Ratio scale

When a researcher asks for your age and gives you four different age ranges to choose from, they're measuring at the ratio level. In addition to the interval between the ranges being constant, this measure also has a true zero value, as it's possible to be zero years of age, but not possible to be -2 years of age. Here is a list of other examples of ratio data:

• Weight in grams

• Number of employees in a company

• Speed in kilometres per hour

• Income in dollars

• Sales made in one month