- characteristics
- Variables according to the relationship with other variables
- -Independent variables
- Example
- -Dependent variables
- -Moderating variables
- Example
- -Weird variables
- Example
- -Variable control
- -Situational variables
- -Variable participants
- -Confusion variable
- Types of variables according to operability
- -Qualitative variables
- Dichotomous qualitative variables
- Example
- Qualitative polytomous variables
- Example
- -Quasi-quantitative variables
- Example
- -Quantitative variables
- Discrete quantitative variables
- Example
- Continuous quantitative variables
- Example
- Variables according to their scale
- -Nominal variable
- Example
- -Orderinal variable
- Example
- -Interval variable
- Example
- -Ration variable
- Examples
- -Continuous variable
- Other lesser known
- -Categorical variables
- Example
- -Active variable
- -Binary variable
- -Variable covariate
- -Criteria variable
- -Endogenous variable
- -Exogenous variable
- -Identifying variables
- -Intervention variable
- -Latent variable
- -Variable manifest
- -Mediating variable or intermediate variable
- -Moderating variable
- -Polycotomic variables
- -Predictive variable
- Statistical variables as a method to analyze empirical reality
- Operational criteria for selecting variables
- Definition of the terms of the variables
- Structures of the variables
- Parameters to consider regarding the operational use of the variables
- Denomination
- Type of variable
- Nature
- Measurement
- Indicator
- Unit of measurement
- Instrument
- Dimension
- Operational definition
- Conceptual definition
- Random variable
- References
The types of variables in research and statistics consist of a series or a set of abstract entities that can acquire different values depending on the categories and characteristics of the object of study.
In other words, statistical variables are typologies that can fluctuate or vary; this variation can be measured and observed. Likewise, a variable can be understood as an abstract construction that refers to a property or an element, which can play a specific role in relation to the object that is being analyzed.
The variables in research and statistics can be measured and analyzed. Source: pixabay.com
This means that said property or element directly influences the subject or object to be studied. The concept of variable seeks to bring together different modalities or options that must be taken into account to understand the object of study.
Consequently, the values of the variables will be inconsistent or different in the subjects and / or moments to be analyzed. Understanding this concept in the theoretical field can be complex.
However, through concrete examples the approach can be better understood: a variable can be the sex or age of a person, since these characteristics can affect the object of study if an analysis is to be carried out in patients who suffer from heart disease or other illnesses.
characteristics
The variables are characterized by two fundamental elements. In the first place, they possess features that can be observed and registered directly or indirectly, which allows a confrontation with practical reality.
Second, they have the property of varying and being measurable, since in some cases they can be classified or measured (for example: age and sex).
Statistical variables cannot be manifested in individual or isolated cases, since the existence of a group is necessary so that those characteristics or elements that are going to vary can be expressed.
If statistics is the science that collects and interprets data, it is understood that the variables of this discipline are in charge of analyzing a plurality of information and are not dedicated to analyzing an isolated or singular data.
There are many types of variables, so these can be classified according to different aspects. For example, statistical variables can be qualitative and quantitative; in turn, these can be divided into other categories, depending on their specifications.
Variables according to the relationship with other variables
In addition to the operational variables, there is also a classification according to the relationship that exists between the values of these variables. It is necessary to bear in mind that the role played by each type of variable depends on the function that is being analyzed. In other words, the classification of these variations is influenced by the object of study.
Within this classification there are independent, dependent, moderating, strange, control, situational, participant and confounding variables.
-Independent variables
These refer to the variables that are taken into account during the research process and that may be subject to modification by the researcher. In other words, these are those variables from which the analyst starts to contemplate and record the effects that their characteristics produce on the object of study.
Example
An example of an independent variable can be sex and also age if you want to make a registry of people with Alzheimer's.
It can be established that the independent variable conditions the dependent one. In addition, the independent can be called experimental or causal, since it is manipulated directly by the researcher. Independent variables are used primarily to describe the factors that are causing the particular problem.
-Dependent variables
They are those that make direct reference to the element that is modified by the variation produced by the independent variable. This means that the dependent variable is generated from the independent variable.
Examples
For example, if we want to determine depression according to sex, the latter will be the independent variable; modifying this will generate fluctuations in the dependent variable, which in this case is depression.
Another example could be found in the relationship between smoking and lung cancer, since "having lung cancer" in this case would be the dependent variable, while "smoking" is an independent variable, since it can vary depending on the number of packs consumed per day.
-Moderating variables
These variables alter or modify the relationship that exists between a dependent and an independent variable; hence their name, since they moderate the link between the two above.
Example
For example, study hours are related to academic sequelae; therefore, a moderating variable could be the student's state of mind or the development of his motor skills.
-Weird variables
The strange variables receive their name because they were not taken into account for the development of the research but they had a noticeable influence on the final results. They are also known as the intervening or puzzling variables, since they can weaken the relationship between the problem and the possible cause.
Consequently, it is a group of variables that were not controlled during the analysis of the object of study, but can be identified once the investigation is completed, and in some cases they are even identified during the course of the study.
They are similar to the moderators, with the difference that these are taken into account at the time of the investigation. Strange variables can also lead the researcher on the wrong path, so the importance of their presence will depend on the quality of the studies undertaken.
Example
For example, a variable of this type may be the fact that nervous people smoke more and have a greater tendency to suffer cancer than those who do not suffer from nervousness; the strange or puzzling variable in this case is nerves.
-Variable control
Control variables are those that a scientist wants to remain constant, and he must observe them as carefully as the dependent variables.
For example, if a scientist wants to investigate the influence of diet (VI) on health (DV), a control variable could be that the people in the study are non-smokers.
This would be the control variable; it is necessary to control it because the observed differences in health could be due to whether people smoke or not. In any case, in an experiment like this there could be other control variables; being an athlete, having other habits…
-Situational variables
A situational variable is an aspect of the environment that can influence the experiment. For example, air quality in a health-related experiment.
-Variable participants
A participant or subject variable is a characteristic of the subjects that are studied in an experiment. For example, the gender of individuals in a health study. Also known as participating variables.
-Confusion variable
A confounding variable is a variable that influences both the independent variable and the dependent variable. For example, stress can make people smoke more and also directly affect their health.
Types of variables according to operability
Statistical and research variables can be classified according to their operability, this category being the best known and most useful. When speaking of operability, allusion is being made to the ability to "number" the values of these variables. Consequently, we can subdivide them into three main types:
-Qualitative variables
Qualitative variables are those variations that allow establishing the identification of a specific element, but that cannot be quantified. This means that these variables can inform about the existence of a characteristic but it cannot be valued numerically.
Consequently, these are variations that establish whether there is equality or inequality, as occurs with sex or nationality. Although they cannot be quantified, these variables can contribute forcefulness to the investigation.
An example of a qualitative variable would be the motivation that students have during the learning process; this variable can be identified but cannot be numbered.
In addition, these can be subdivided into other categories, such as dichotomous qualitative variables and polytomous qualitative variables.
Dichotomous qualitative variables
These variables can only be considered or analyzed from only two options; hence the word "dichotomy" is present in its name, since it indicates a division present in two aspects that are usually contrary to each other.
Example
A precise example would be the variable of being alive or dead, since it only allows two possible options and the presence of one of these immediately negates the other.
Qualitative polytomous variables
These statistical variables are the opposite of dichotomous variables, since they allow the existence of three or more values. However, in many cases this prevents them from being ordered, since they only establish the identification of a value.
Example
A precise example is the color variable since, although it allows identification, it declares that there is only one possible characteristic or element assignable to this variable.
-Quasi-quantitative variables
These variables are characterized by making it impossible to carry out any mathematical operation; however, they are more advanced than those that are solely qualitative.
This is because quasi-quantitative ones allow establishing a hierarchy or a kind of order, although they cannot be quantified.
Example
For example, the level of studies of a group of people can be a variable of this type, since the completion of a postgraduate degree is located in a higher hierarchy than the completion of an undergraduate degree.
-Quantitative variables
These variables, as their name indicates, allow the performance of mathematical operations within their values; therefore, the different elements of these variables can be assigned numbers (that is, they can be quantified).
Some examples of this type of variable include the following:
-The age, since this can be expressed in years.
-The weight, which can be defined in pounds or kilograms.
-The distance between a given place and the place of origin, which can manifest itself in kilometers or minutes.
-Monthly income, which can be expressed in dollars, euros, pesos, soles, among other types of currencies.
In turn, these types of variables can be subdivided into two groups: discrete quantitative variables and continuous quantitative variables.
Discrete quantitative variables
These refer to quantitative variables that cannot have intermediate values - they do not admit decimals within their number. In other words, they must be numbered through a complete number.
Example
A precise example consists of the impossibility of having 1.5 children; it is only possible to have one or two children. This means that the unit of measurement cannot be fractioned.
Continuous quantitative variables
On the contrary to the discrete ones, the continuous variables can have decimals, so their values can be intermediate.
These variables are measured by the interval scales. In other words, continuous quantitative variables can be fractionated.
Example
For example, measuring the weight or height of a group of people.
Variables according to their scale
In addition to the previous classifications, statistical variables can be cataloged taking into account the function of their scales and the measures that are used to calculate them; However, when talking about these variables, greater emphasis is being placed on the scale than on the variable itself.
In turn, the scales used for the variables may undergo modifications depending on the level of operation, since the latter allows the incorporation of other possibilities within the range of scales.
Despite this, four main types of variables can be established according to scale; These are the following: the nominal variable, the ordinal variable, the interval variable, the ratio variable and the continuous variable.
-Nominal variable
This type of variables refers to those whose values only allow distinguishing a single specific quality without introducing the performance of mathematical operations on them. In this sense, nominal variables are equivalent to qualitative variables.
Example
As an example of the nominal variable, gender can be found, since it is divided into masculine or feminine; as well as the marital status, which can be single, married, widowed or divorced.
-Orderinal variable
These variables are essentially qualitative since they do not allow the performance of mathematical operations; however, ordinal variables do allow establishing certain hierarchical relationships in their values.
Example
An example of a nominal variable can be a person's educational level or economic status. Another example can be the ranking of academic performance by the following adjectives: excellent, good or bad.
Variables of this type are used to classify subjects, events or phenomena in a hierarchical way, considering specific characteristics.
-Interval variable
The variables that have scale in interval allow the realization of numerical relations between them, although they can be limited by the proportionality relations. This is because within this range there are no "zero points" or "absolute zeros" that can be fully identified.
This results in the impossibility of carrying out transformations directly in the other values. Therefore, the interval variables, rather than measuring specific values, measure the ranges; This somewhat complicates operations but encourages coverage of a large number of securities.
Interval variables can be presented in degrees, magnitudes, or any other expression that symbolizes quantities. Likewise, they allow to classify and order categories, as well as they can indicate the degrees of distance that exist between them.
Example
Within this classification can be found the temperature or the IQ.
-Ration variable
This type of variable is measured by a scale that operates in a total way, which does allow the direct transformation of the results that were obtained.
In addition, it also encourages the performance of complex number operations. In these variables there is an initiation point that implies the complete absence of what was measured.
Consequently, the ratio variables do have an absolute zero and the distance between two points is always the same, although they also have the characteristics of the previous variables.
Examples
For example, age, weight, and height are ratio variables.
-Continuous variable
A variable with an infinite number of values, such as "time" or "weight."
Other lesser known
-Categorical variables
Categorical variables are those whose values can be expressed through a series of categories that define them.
Example
A good example of a categorical variable corresponds to the consequences of a given disease, which can be broken down into recovery, chronic illness, or death.
-Active variable
A variable that is manipulated by the researcher.
-Binary variable
A variable that can only take two values, usually 0/1. It could also be yes / no, high / short, or some other combination of two variables.
-Variable covariate
Similar to an independent variable, it has an effect on the dependent variable, but it is generally not the variable of interest.
-Criteria variable
Another name for a dependent variable, when the variable is used in non-experimental situations.
-Endogenous variable
Similar to dependent variables, they are affected by other variables within a system. Used almost exclusively in econometrics.
-Exogenous variable
Variables that affect others, and that come from outside a system.
-Identifying variables
Variables used to uniquely identify situations.
-Intervention variable
A variable that is used to explain the relationship between variables.
-Latent variable
A hidden variable that cannot be directly measured or observed.
-Variable manifest
A variable that can be directly observed or measured.
-Mediating variable or intermediate variable
Variables that explain how the relationship between variables happens.
-Moderating variable
Changes the intensity of an effect between independent and dependent variables. For example, psychotherapy can lower women's stress levels more than men, so sex moderates the effect between psychotherapy and stress levels.
-Polycotomic variables
Variables that can have more than two values.
-Predictive variable
Similar in meaning to the independent variable, but used in regression and in non-experimental studies.
Statistical variables as a method to analyze empirical reality
The different types of statistical variables allow the human being to simplify and classify reality, since it divides it into simple parameters that are easy to measure and calculate. In this way, it is possible to isolate a group of elements that are part of a society or nature.
Consequently, the human being cannot consider that he understands the totality of the world that surrounds him through the variables, since these remain a limited knowledge compared to the totality of the universe.
This means that the researcher must choose to apply a critical look at the results obtained through the variables, in order to avoid as much as possible the approach to wrong conclusions.
Operational criteria for selecting variables
Definition of the terms of the variables
First, the variables need to be operable; In order to achieve this, they must be made measurable or comprehensible.
Then, it is necessary to assign a meaning and a definition to each term that is a fundamental part of the context of the research to be carried out. This definition must be based on the reference of the features found in empirical reality.
In addition, these definitions must be concrete and operational, based on scientific observation and using measures that refer to the indicators of reality that are directly observed.
Later it will be necessary to examine all the definitions of the term, past and present, as many as possible. Next, it is necessary to proceed to identify the variables or the group of variables that may help to explain the problem posed during the establishment of the investigation.
Structures of the variables
The structure of statistical variables can be divided into four main elements, these being the following:
-Name.
-The set of categories.
-The verbal definition.
-The procedure to group them taking into account the observation units of categories.
Parameters to consider regarding the operational use of the variables
Denomination
It refers to the name that is given to the variable during the development of the investigation.
Type of variable
It refers to the category that a variable has at the time of introducing it into the object of study to be investigated. This is established according to the location of the variable within the hypothesis of the work.
Nature
It must be established whether the variable will be quantitative or qualitative, since this classification allows to solidify the theoretical bases of the investigative process. Once the nature of the variable has been identified, it will be easier to carry out the rest of the comparisons and descriptions.
Measurement
This refers to the measurement scale that the variable will use when establishing relationships with the object of study or with the other variables.
Indicator
This parameter is the base that starts the measurement. In other words, it is the instrument that makes the measurement of variables possible.
Unit of measurement
This will depend on what the variable indicator establishes. The unit of measurement works most of all in those variables that can be quantified.
Instrument
This parameter refers to the tool that the researcher will use to collect the information and data concerning the statistical variables.
Dimension
It refers to the extension that the variable occupies within the empirical reality. For example, a variable can have a clinical dimension, a geographic dimension, a social, biological, diagnostic or demographic dimension, among others.
Operational definition
This parameter seeks to define the work carried out by the variable within the object of study.
Conceptual definition
It refers to the definition with which the variable is known or treated, taking into account the medical dictionary or another specialized in the area that the variable occupies.
Random variable
In the field of statistics and in the mathematical discipline, a random variable is called a function whose purpose is to assign a value - generally of a numerical nature - to a result that has emerged from a random experiment.
The most concrete example can be found in the game of dice, since rolling a dice twice raises two possible random outcomes: (1,1) and (1,2).
A random variable raises possible values that represent the results of an experiment that has not yet been performed. It can also represent the possible values of a quantity whose value at that moment is uncertain; in this case, it is an inaccurate or incomplete measurement.
In conclusion, the random variables can be taken as a quantity that has a non-fixed value that, in turn, can take different values. To calculate these variables it is necessary to use the probability distribution, which is used to describe what probabilities exist for the different values to occur.
References
- (SA) (sf) Types of Variables in Statistics and Research. Retrieved on April 8, 2019 from Statistics How to: statisticshowto.datasciencecentral.com
- Benitez, E. (2013) Variables in statistics. Retrieved on April 8, 2019 from WordPress: wordpress.com
- Del Carpio, A. (sf) Variables in Research. Retrieved on April 7, 2019 from URP: urp.edu.pe
- Mimenza, O. (sf) The 11 types of variables used in research. A review of the main classes of variables used in science to investigate. Retrieved on April 7, 2019 from Psychology and Mind: psicologiaymente.com
- Mota, A. (2018) Statistical variables. Retrieved on April 7, 2019 from Universo Formulas: universoformulas.com
- Carballo, M., Guelmes, C. Some considerations about the research variables that are developed in education in Scielo. Retrieved on April 7, 2019 from Scielo: scielo.sld.cu