- Concept and characteristics
- Examples of nominal variables
- - Examples explained
- Attendees at a soccer game
- Benefit of labor policies
- Country of birth of a person
- References
A nominal variable is one that takes values that identify a class or category in which the objects of study are grouped. For example, the variable 'hair color' groups people into brown, black, blonde hair, etc.
The nominal scale identifies, groups and differentiates the study units, according to a certain quality, in clearly defined and exclusive classes, in such a way that all those that belong to a class are equal or equivalent with respect to the attribute or property under study.
Man and woman icons. Source: pixabay.com
Classes are differentiated by names or by identifying numbers, so they have no numerical value or established order. For example: the variable sex has two classes, male and female; The numbers 1 and 2 can also be used, representing the male and female categories, respectively. These numbers are just arbitrary identifiers.
In these types of measures, names or labels are assigned to objects. The name of most of the nominated specimens or definitions is the "value" assigned to the nominal measure of the object of study.
If two objects have the same name associated with them, then they belong to the same category, and that is the only meaning that nominal measures have.
Concept and characteristics
The nominal scale is the most elementary and the variables measured on this scale classify the study units (objects, people, etc.) into classes, based on one or more unique and observed characteristics, attributes or properties.
Classes or categories have a name or number, but these serve merely as labels or identifiers, they make categorical rather than quantitative distinctions, they serve a purely classification function.
They cannot be manipulated arithmetically, they do not reflect order (ascending or descending) or hierarchy (higher or lower), the observations cannot be ordered from smallest to largest or from small to large, that is, none of the categories has a higher hierarchy than the another, they are only reflecting differences in the variable.
Nominal variables with two classes are called dichotomous, such as the variable sex (male or female). Variables with three or more categories are called multichotomic or polyhotomic. For example: the variable occupation (worker, carpenter, doctor, etc).
Nominal variables only establish equivalence relationships; that is, a particular object of study either has the characteristic that defines the class or it does not.
With nominal variables, calculations of proportions, percentages and ratios can be made, and with them frequency counts or tabulations of the number of events in each class of the variable studied are carried out. The measure of central tendency that can be handled with these types of variables is the mode.
Examples of nominal variables
Examples of variables measured in nominal scale:
- Nationality (Argentine, Chilean, Colombian, Ecuadorian, Peruvian, etc.).
- Colors (white, yellow, blue, black, orange, etc.).
- Color of the eyes (black, brown, blue, green, etc.).
- Classification of students by careers (Administration - 1; Systems - 2; Electronics - 3; Law - 4; etc.). (the number is a code with no value or order)
- Marital status (single, married, widowed, divorced, common-law union).
- Profession (engineer, lawyer, doctor, teacher, etc.).
- Sex (male, female).
- Religious affiliation (Christian, Muslim, Catholic, etc.).
- Political affiliation (liberal, conservative, independent, etc.).
- Type of school (public or private).
- Race (white, black, yellow, mestizo, etc.).
- Blood groups (O, A, B, AB).
- Examples explained
Attendees at a soccer game
If a count is made of the attendees entering a soccer match, the nominal variable 'attendance by sex' can be defined. The count reports how many men and how many women attended the match, but the classification variable is gender.
Divide the public in the soccer game into two categories and neither group has a preference over the other. Finally, the categories are exclusive since there is no doubt to which group each of the attendees belongs.
Benefit of labor policies
You want to know the opinion of the people before the application of reforms in the labor policies of a country. The variable of 'interest' is the benefits of labor policies, and in the survey there are five possible positive results: More money, Better medical attention, Better retirement, Work / family balance and Others.
All responses are measured on a nominal scale with Yes or No. values. The Others result encompasses all those benefits that respondents believe they will obtain, but which are not part of the survey values.
The number of affirmative or negative answers are necessary to calculate the percentage of respondents of the total who consider that they will improve or not in any of the aspects, but these percentages do not have meaning from the point of view that one benefit is greater than another.
Finally, there is no natural order to the results, you can put Better healthcare first instead of More money, for example, and it doesn't change the result at all.
Country of birth of a person
The country of birth is a nominal variable whose values are the names of the countries. For purposes of working with this variable, it is convenient to make a numerical codification of that information, we assign code 1 to those born in Argentina, Bolivia code 2, Canada code 3, and so on.
This coding facilitates computer counting and the management of information collection instruments. However, and since we have assigned numbers to the various categories, we cannot manipulate these numbers. For example, 1 + 2 is not equal to 3; that is, Argentina + Bolivia does not result in Canada.
References
- Coronado, J. (2007). Measurement scales. Paradigmas Magazine. Recovered from unitec.edu.co.
- Freund, R.; Wilson, W.; Mohr, D. (2010). Statistical methods. Third ed. Academic Press-Elsevier Inc.
- Glass, G.; Stanley, J. (1996). Statistical methods not applied to the social sciences. Prentice Hall Hispanoamericana SA
- Cute.; Marchal, W.; Wathen, S. (2012). Statistics applied to business and economy. Fifteenth ed. McGraw-Hill / Interamericana Editores SA
- Orlandoni, G. (2010). Statistical measurement scales. Telos Magazine. Recovered from ojs.urbe.edu.
- Siegel, S.; Castellan, N. (1998). Nonparametric statistics applied to the behavioral sciences. Fourth ed. Editorial Trillas SA
- (2019). Level of measurement. Recovered from en.wikipedia.org.