Definition of Measurement Level

The characteristics that a variable can take are known as measurement levels, which can be four: nominal, ordinal, interval and ratio; The first two are also known as categorical variables, while the last two are considered quantitative variables.

One of the essential characteristics of any science is the quantification of the properties possessed by the objects of study. How much mass does an object have? How much of a chemical compound is required for any medicine? How many doses of medication does a patient with a certain disease require? How much force is required for an object to acquire a certain speed? Although the measurement of the previously described elements is relatively simple, since they tend To be tangible and objective elements, in the social sciences the objects of study tend to be somewhat more subjective, so their measurement can be a bit complex; Let's think that unlike a doctor who can determine how many units of glucose per liter of blood a person has, patient with diabetes, psychologists cannot determine how many units of depression or anxiety a patient suffers from. person. However, and although the above sounds hopeless for psychological science, it is actually possible to carry out a objective and systematized measurement based on a series of rules that characterize what is known as levels of measurement.

Before delving into the levels of measurement, it is necessary to understand the concept of “measurement”. Measurement is an inherent and essential process of all research, regardless of whether it is qualitative or quantitative; It is the process in which a value or property is assigned to an object based on a series of pre-established rules. Measurement is the process of estimating the magnitude of a property possessed by some object or subject, through the support of a metric system or tool (in the social sciences these tools are usually scales, inventories or questionnaires). It is necessary to mention that when talking about properties we refer to the observable aspects or characteristics of the object of study. Finally, measurement is not limited only to the quantification of the properties, it also involves the evaluation of these, that is, the similarities and differences it presents.

Nominal measurement level

It is the most basic level that can exist, and involves the categorization of properties into one or more groups. It is necessary to have a minimum of two categories; These categories do not have an order or hierarchy, they are positioned at a similar level and are only differentiated by the criteria established to create these categories.

When a nominal variable only has two groups, it is called a dichotomous variable, and the examples The most typical of this may be the sex assigned at birth (male or female) or the vital status (alive or death). On the contrary, if the nominal variable has more than two groups it is called polytomous, typical examples are orientation sexual (heterosexual, gay, lesbian, bisexual or pansexual) or affiliation to a religious group (Catholic, Muslim, Christian or Hindu).

To create categories, two basic criteria must be taken into account:

1) Completeness. All objects can be classified into some category

2) Exclusiveness. Each object can be classified in only one of the categories.

Ordinal measurement level

In a similar way to the nominal level, at this level there are categories, however, the biggest difference is that these categories do have a hierarchy. It is from this characteristic that this variable acquires its name, as it helps to order the elements of a variable. It is necessary to mention that such a hierarchy is not characterized by being strict, that is, between one level and another there may be a great distance, while between the other two it may be minimal.

Some of the most typical examples of this level are the educational level (basic, intermediate or higher), the socioeconomic level (low, medium-low, medium-high, high).

Interval measurement level

This level is similar to the ordinal level, as it also allows the elements or properties of the evaluated variable to be ordered. The difference is that here a series of equal intervals are established between the properties being evaluated. That is, the categories have the same distance from each other that remains constant throughout the measurement. Another characteristic is that it does not have an absolute 0, so it can have negative or positive values ​​of the attribute to be measured.

A typical example of this level is thermometers, because although they have established intervals (degrees) 0 is relative depending on the scale used (Fahrenheit, Celsius or Kelvin).

Ratio measurement level

It is the most rigorous level to achieve and includes the properties of the interval measurement level, with the difference in that at this level there is an absolute 0, which indicates an absence of the attribute to be measure.

The typical example of this level is the rules, because regardless of whether it is centimeters, inches, feet or yards, 0 will always indicate a starting point.

It is important to know the level of measurement of the variables used in our studies, since depending on These will be possible to use certain analysis techniques that allow us to reach interpretations and conclusions.

References

Pardo, A., Ruiz, M.A. & San Martín, R. (2009). Data analysis in social and health sciences I. (2nd. Ed.) Synthesis.