When dealing with any real life data set, the data values so obtained are divided into four different kinds of variables. These are also known as the four scales of measurement. The variables are:
- Nominal – Used for categorical/qualitative data.
- Ordinal – Used for categorical data that can be put in ascending or descending order.
- Interval – Used for quantitative/numerical variables without a “true zero”.
- Ratio – Used for numerical variables with a “true zero”.
Is Eye Color a Nominal Variable?
Eye Color is indeed a nominal variable. This is because eye color gives us information about the qualitative characteristics of a person. Since eye color cannot be represented by numerical values, eye color certainly cannot be said to be an interval or ratio variable.
Of course we can arbitrarily assign numerical values to eye color but that assignment will not have any inherent significance and it will merely serve as a placeholder. For example, suppose that we are given data about the eye color of a group of 100 people. Suppose 60 of then have black eyes and 40 of them have blue eyes. We can simply assign the number 0 to those who have black eyes and the number 1 to those who have blue eyes. The our data set will consist of a sequence of 0’s and 1’s. But notice that the numbers merely serve as placeholders. We could easily have used the number 0 to represent blue eyes and the number 1 to represent black eyes. This means that the numerical assignment has no inherent significance. Therefore, eye color is not an interval or ratio variable.
Why is Eye Color not an Ordinal Variable?
Eye Color is not an ordinal variable. Eye Color is a qualitative variable and we know that nominal and ordinal scales are used for qualitative variables. But since there is no inherent notion of “rank” as far as eye color is concerned we conclude that eye color is not an ordinal variable. Eye color is a mere categorical classification and hence it is considered to be a nominal variable. On the other hand, qualitative characteristics such as singing ability can be “ranked” and hence these kinds of variables are ordinal variables.
Is Political Affiliation a Nominal Variable? Is Marital Status a Nominal Variable?
Qualitative characteristics such as political affiliation, marital status, etc. are all nominal variables. This is because they are categorical variables which are not given by numerical values. They are not ordinal variables since there is no objective notion of “rank” as far as political affiliation or marital status are concerned. Since we are dealing with categorical data that cannot be arranged in an ascending or descending order we conclude that they are both nominal variables.