Understanding the Types of Variables in Research

Every research study, whether it explores reading habits in a library or income patterns across a city, begins with measuring something. That “something” is a variable. A variable is any characteristic, attribute, or property that can take different values across the people, objects, or events you study. Choosing how to classify your variables is not just an academic exercise. It decides which statistical tests you can run and how confidently you can interpret your findings. This post breaks down the main types of variables you will meet in research, explains how they differ, and clarifies the often-confused line between a variable and the data it produces.

Qualitative vs. quantitative variables 🔗

The first and broadest way to classify variables is by the kind of value they hold. This single distinction shapes almost everything that follows, from the scale you use to measure to the analysis you can perform.

Qualitative variables 🔗

Qualitative variables, also called categorical variables, represent non-numeric data. They describe a quality or attribute rather than a quantity. Marital status, religion, gender, and eye colour are classic examples. When you record a qualitative variable, you are placing each case into a category, not measuring an amount. According to the SAGE Encyclopedia of Social Science Research Methods, qualitative variables differ from quantitative ones because their categories matter as distinct attributes, not as greater or smaller magnitudes. Labels like “Hindu,” “Muslim,” or “Christian” simply identify a person’s religious affiliation; none represents more or less of anything.

Quantitative variables 🔗

Quantitative variables represent numeric data. They express differences in amount or magnitude, and you can perform arithmetic on them. Age, income, height, and the number of books borrowed in a month are all quantitative. The key feature is that the values reflect a genuine quantity, so it makes sense to add, subtract, or average them. Knowing whether a variable is qualitative or quantitative is the first decision point because, as Scribbr notes, this directly determines which statistical tests are appropriate for your study.

One subtle point trips up many students. Numbers are sometimes used to code qualitative variables. A researcher might assign “1” to women and “2” to men in a dataset. Despite the numerical labels, gender remains a qualitative variable, as Stats and R explains, because the numbers are just convenient codes, not real quantities.

Nominal and ordinal variables 🔗

Qualitative variables can be divided further based on whether their categories can be ranked. This gives us two important sub-types: nominal and ordinal. The distinction rests on a single question: is there a natural order among the categories?

Nominal variables 🔗

Nominal variables are the simplest type of categorical variable. Their categories have no inherent order or ranking. The values are simply different from one another, and no value is “higher” or “lower” than another. Gender, religion, blood group, and state of residence are nominal. You cannot meaningfully arrange eye colours from worst to best, and the categories are mutually exclusive, since a person cannot belong to two at once. The word nominal comes from naming, and that is precisely what this scale does: it labels without ordering. As Statology describes, a nominal scale labels variables that carry no quantitative value and no natural order.

Ordinal variables 🔗

Ordinal variables also place cases into categories, but here the categories can be arranged in a meaningful order or rank. Educational qualification is a good example: matriculation, graduation, and post-graduation follow a clear sequence of increasing attainment. Customer satisfaction ratings such as “poor, average, good, excellent” are also ordinal. The catch is that while you know the order, you do not know the exact distance between ranks. The gap between “poor” and “average” may not equal the gap between “good” and “excellent.” GraphPad highlights this limitation as the defining feature that separates ordinal data from higher scales.

A note on the four scales of measurement 🔗

Nominal and ordinal scales are the first two of four levels of measurement that statisticians use. The other two, interval and ratio, apply to quantitative variables. An interval scale orders values and keeps equal distances between them but has an arbitrary zero, such as temperature in Celsius. A ratio scale has all those properties plus a true zero, such as weight or income, where zero means a genuine absence of the quantity. These four scales were proposed by the psychologist Stanley Smith Stevens in a 1946 paper, and as Statistics By Jim explains, the level of measurement you use affects exactly what you can learn from your data and which analyses you can run. Where you have a choice, it is generally wise to record data using the scale closest to the ratio end, because it preserves the most information.

Discrete and continuous variables 🔗

Just as qualitative variables split into nominal and ordinal, quantitative variables split into discrete and continuous. The dividing line here is whether the variable is counted or measured.

Discrete variables 🔗

Discrete variables are countable. They take on distinct, separate values with clear gaps between them, and typically these are whole numbers. The number of books on a shelf, the number of students in a class, and the number of children in a family are all discrete. You can have 20 or 21 students, but never 20.5. As Statistics How To puts it, discrete variables are countable and distinct, like the number of letters in a word or traffic accidents in a day. They might take a long time to count, but they remain countable, which is the defining test.

Continuous variables 🔗

Continuous variables can take any value within a range. Rather than being counted, they are measured, usually with an instrument such as a ruler, weighing scale, or stopwatch. Weight, height, temperature, and the time taken to finish a task are continuous. Between any two values there are infinitely many other possible values. Between 60 kg and 61 kg lie 60.5 kg, 60.55 kg, 60.555 kg, and so on without end. As Outlier notes, because the possible values are infinite, we measure continuous variables instead of counting them. In practice, the precision of your measuring instrument sets the limit on how finely you can record a continuous value.

Why this distinction matters 🔗

The discrete-versus-continuous split is not just terminology; it changes how you analyse and present data. Continuous variables lend themselves to descriptive measures such as the mean, median, and standard deviation. Discrete variables, with their fixed values, are often better summarised through frequency counts and proportions. Coursera points out that this difference also affects visualisation, since the type of variable guides whether a bar chart or a histogram is the more honest representation of your data.

Variable vs. data: what’s the difference? 🔗

Students frequently use the words “variable” and “data” as if they mean the same thing, but they refer to two distinct stages of research. Getting this right sharpens how you talk about your own studies.

A variable is the characteristic or attribute you intend to study. It is the concept, such as age, income, or marital status, that can vary from one case to another. Data, on the other hand, are the actual measurements or values you record for that variable. As Scribbr describes it, data is the specific value you write down in your data sheet for a given variable.

Consider a simple example. “Years of schooling” is a variable. The specific numbers you collect from respondents, say 8, 12, 15, and 16 years, are the data. The variable is the container; the data are what fills it. Outlier frames the relationship cleanly: a variable is any characteristic that can be measured, while data refers to the observations collected for that variable. A variable also carries a type, such as discrete or continuous, and the data inherit that type. So a discrete variable produces discrete data, and a continuous variable produces continuous data.

Bringing the classifications together 🔗

It helps to see how these layers connect rather than treating them as separate lists. At the top, every variable is either qualitative or quantitative. Qualitative variables then branch into nominal (no order) and ordinal (ordered). Quantitative variables branch into discrete (counted) and continuous (measured). Underlying all of this are the four scales of measurement, which give a more precise vocabulary for the same ideas. A single concept like income can even shift between types depending on how you record it: ask for an exact rupee figure and it is a continuous, ratio-scale variable; offer fixed bands such as “below 20,000,” “20,000 to 50,000,” and “above 50,000,” and you have converted it into an ordinal variable. The choice is yours as a researcher, and it should be driven by what your study genuinely needs to know.

This is why classification is the quiet engine of good research design. Identify your variable types early, and the right summary statistics, charts, and tests follow naturally. Misclassify them, and you risk calculating an average where a frequency count was needed, or treating ranked categories as if the gaps between them were equal.

What do you think? Think about a study you might want to conduct in your own field. Which of your variables would be qualitative and which quantitative, and how might recording one of them on a different scale change the conclusions you could draw?

References

1. https://methods.sagepub.com/ency/edvol/the-sage-encyclopedia-of-social-science-research-methods/chpt/qualitative-variable
2. https://www.scribbr.com/methodology/types-of-variables/
3. https://statsandr.com/blog/variable-types-and-examples/
4. https://www.statology.org/levels-of-measurement-nominal-ordinal-interval-and-ratio/
5. https://www.graphpad.com/support/faq/what-is-the-difference-between-ordinal-interval-and-ratio-variables-why-should-i-care/
6. https://statisticsbyjim.com/basics/nominal-ordinal-interval-ratio-scales/
7. https://www.statisticshowto.com/probability-and-statistics/statistics-definitions/discrete-vs-continuous-variables/
8. https://articles.outlier.org/discrete-vs-continuous-variables
9. https://www.coursera.org/articles/discrete-vs-continuous-data