Types of Variable > Qualitative variable
What is a qualitative variable?
A qualitative variable, also called a categorical variable, is a variable that isn’t numerical. It describes data that fits into categories. For example:
- Eye colors (variables include: blue, green, brown, hazel).
- States (variables include: Florida, New Jersey, Washington).
- Dog breeds (variables include: Alaskan Malamute, German Shepherd, Siberian Husky, Shih tzu).
These are all qualitative variables as they have no natural order. On the other hand, quantitative variables have a value and they can be added, subtracted, divided or multiplied.
| Quantitative Variable | Qualitative Variables |
| Fractions | Cat breeds |
| Decimals | Cities |
| Odd Numbers | Fast Food Chains |
| Whole Numbers | College Major |
| Irrational Numbers | Fraternities |
| Ordered pairs (x,y) | Hair Color |
| Negative Numbers | Computer Brands |
| Map coordinates | Beer breweries |
| Positive Numbers | Pop music genre |
| Exponents | Tribe |
As a general rule, if you can apply some kind of math (like addition), it’s not qualitative — instead, it’s a quantitative variable. For example, you can’t add blue + green (unless you’re in an art class — even then you “mix” them, you don’t add them!) — making “blue” and “green” qualitative variables.
Numbers are sometimes assigned to qualitative variables for data analysis, but they are still classified as qualitative variables despite the numerical classification. For example, a study may assign the number “1” to males and “2” to females.
Qualitative Variables and the Nominal Scale
Qualitative variables aren’t ordered on a numerical scale so they are placed on a nominal scale. The word “nominal” means “name”, which is exactly what qualitative variables are. A nominal scale is a scale where no ordering is possible or implied (except for alphabetical ordering like New York, Washington, West Virginia or Chelsea, Edinburgh, London). In other words, the nominal scale is where data is assigned to a category.
Avoiding qualitative variable misinterpretation
Sometimes, numbers are assigned to qualitative variables during data analysis. For example:
- Marital status: 1 = married, 2 = single, 3 = divorced, 4 = widowed.
- Ethnicity: 1 = white, 2 = black, 3 = Hispanic, 4 = Asian
However, the numbers assigned here (1, 2, 3, 4) don’t have mathematical meaning in the traditional sense. For example, assigning 1 to males and 2 to females doesn’t imply that males are “greater” than females. Not can they be added, divided or multiplied together. Rather, the category members are given numerical “names.”
This numerical naming convention can lead to misinterpretations such as:
-
- Confusing ordinal and interval variables: For example, assigning “1” to “poor,” “2” to “fair,” and “3” to “good” doesn’t imply equal differences between these categories.
- Assuming magnitude implications: For example, if a study assigns “1” to men and “2” to women in the catergory “education,” it doesn’t mean that men have twice as much education as women.
- Inferring causality: For example, let’s say you have a range of numbers for education, where 1 = high school and 4 = graduate school. A study might show that more educated individuals are more likely to be employed, but it doesn’t imply that education causes employment, nor does it imply that the numerical values are equated to some sort of ranking for likelihood of employment.
To avoid possible errors or misinterpretation when describing variables, be clear and explicit when describing [1]:
- The kind of variable (qualitative or quantitative).
- The reference class (the overall class of all those things for which a variable makes sense, such as calories for food or color for hair).
- The procedure used to make the measurement.