Types of variables in scientific research

There are many types of variables in scientific research. Before we list out all and define them, let’s define the term variable itself. Paul Spector (1981) provides a succinct definition stating, “a variable is a qualitative or quantitative entity that can vary or take on different values” (p. 11). Variables represent the concept that we are studying or measuring.

Deductive research often starts with a theory that can explain the phenomena. A theory is a formalized set of concepts (constructs) that organize observations and inferences. It also predicts and explains the phenomenon. But scientific research often relies on positivism and thus demands verification of any theoretical claims. Claims are generally verified through empiricism (systematic observations) or by logical deduction. Empirical testing demands that the theory be testable. That is, a testable theory makes specific predictions that we can evaluate using a structured, scientific approach to the research process.

Concepts or constructs that are used in the theory are generally abstract. For example, overcrowding, stress, love, intelligence, aggression, etc. are all abstract concepts We need variables to measure these constructs. Measurement is simply the process of assigning numbers or labels to variables that represent attributes or properties of subjects or treatments. We operationalize these constructs and measure them through variables. For example, overcrowding is a concept but to measure overcrowding in the neighborhoods we need to operationalize it. We can do that by creating a variable that measures overcrowding in concrete terms. For example, overcrowding can be measured by calculating the total population per city block or by the average number of people per dwelling or per city block. So variables measure the abstract concept used in a theory.

There are many different types of variables depending on the role they play in a theoretical framework. Let’s look at each one of them.

Variables in scientific research

Categorical variables measure a construct that has different categories or that varies by type or kind. Examples of categorical variables include gender, race, religious affiliation, political affiliation, etc. Categorical variables put things/places/people (whatever we are measuring) into different categories and these variables are generally qualitative.

Quantitative variables measure constructs that vary by degree of the amount and is generally quantitative. Weight, height, intelligence scores, age, anxiety, etc. are all examples of quantitative variables.

Independent variables (IV) measure construct that are considered to be the cause in a theoretical framework. It is the variable that causes a change in another variable. Researchers study independent variables to see their effect or influence on the dependent or the outcome variable. Sometimes independent variables are also referred to as factors, treatment variables, predictors, determinants, or antecedent variables.

Dependent variable (DV) measure construct that is considered to be the effect in a theoretical framework.  It is the outcome variable that changes based on other variables. It is generally affected by more than one variable. Researchers study dependent variables to see the effect or influence of other variables. Sometimes the dependent variable is also referred to as the outcome variable, effect, criterion, or consequence variable. For example, research shows that smoking causes lung cancer. Smoking is the independent variable and lung cancer is the dependent variable.

IV and DV

Intervening or MediatingAs the name suggests these types of variables measure constructs that intervene, mediate, or stand in between the cause and the effect exercising an influence on the dependent variable apart from the independent variable. That is, it operates between the two variables in a way that defines how one variable affects another. For example, if a study claims that poverty causes overcrowding, and overcrowding, in turn, causes crime, then overcrowding intervenes between poverty and crime. Additionally, we can also say that there is indirect causation between poverty and crime. Indirect causation refers to the causal path from the independent variable to the dependent variable via an intervening or mediating variable.

Mediating

Moderating variable as the name suggest moderates the relationship between the two variables. That is, it determines how the relationship between the cause and the effect changes across different levels of this variable or by its absence or presence. Sometimes moderating variables are also referred to as secondary independent variables that produce a combined effect on the dependent variable. For example, a recent study by Misra et al found that when people share a close relationship (moderating variable) the presence of smartphones (independent variable) even when you are not using the smartphone, reduces the quality of communication where people report a lower level of empathy (dependent variable). Here the combined effect of the presence of a smartphone and closeness of relationship interact to produce a negative effect on empathy. The joint effect is called moderating or interaction effect or interactive causation. It occurs when the effect of one variable (independent) on other (dependent) depends on the level of another variable (moderating).

Moderating

Confounding or Lurking variable are variables that might compete with the independent variable to explain the outcome. Confounding variables are the ‘rival explanations’ that explain the cause and effect relationship. That is, the causal path might drop to zero or disappear if the effect of the confounding variable is considered. They are called confounding as they confound the understanding of the relationship between cause and effect. A unique thing about the confounding variable is that confounding variables are linked to both the independent and the dependent variables. For example, say a study claims that an increase in shoe size leads to an increase in intelligence in children. The obvious confounding variable here is age. When you consider age as one of the variables, the spurious effect between the shoe size and intelligence disappears.

However, confounding variables are not always this simple to identify. When the Head start program was being evaluated for its effectiveness in improving the learning readiness of poor children entering school, the effect appeared to be weak when compared to the control group (those children who did not participate in the Head Start program). The result appeared to be weak because of confounding variables. The control or comparison group was composed of affluent students from high socioeconomic status (SES) families.  We know that children from high SES families receive more help from their parents, which in turn affects how well they learn. The level of intellectual stimulation in high SES homes is also very different compared to low SES homes. Thus, SES is one confounding variable. Other confounding variables include the different school environments for children coming from poorer neighborhoods compared to affluent neighborhoods and exposure to violence. All these confounding factors when taken into account nullify the low impact of the Head start program on poor children.

Confounding

Control variables are variables that are not of primary interest and thus are the extraneous variables whose influence can be controlled or eliminated. They might have an impact on the dependent variable but the researcher tries to control it statistically or through design procedures to nullify their effect. It comes from the desire of the researcher to estimate the effect of IV over DV, independent of these extraneous variables. Typically in any research study, the researchers identify potential variables that might affect the dependent variable. To isolate the effect of the independent variable on the dependent variable (to produce a stronger causal inference) researchers brainstorm possible variables that can affect the DV and try to measure them in the study so that they can later be controlled statistically. Control variables are called covariates. For example, to study the effect of room temperature on the typing speed of the typist, the researcher might want to control typing experience in years across different test conditions by selecting people with similar experience so that experience can be ruled out as an influence. In this way, the effect of room temperature on typing speed can be isolated Typically, demographic data such as gender, socioeconomic status, age, race, etc. are used as control variables. Additionally depending on the research, one can add more variables in data collection measures.

Variables Example

Variables in quantitative research – Summary

Variable Family

Bibliography

Creswell, J. W. (2012). Educational research: Planning, conducting, and evaluating quantitative and qualitative research (4th ed.). Boston, MA: Pearson Education.

Dooley, D. (2001). Social research methods (4th ed.). Upper Saddle River, New Jersey: Pretence Hall.

Salkind, N. J. (2010). Encyclopedia of research design. Thousand Oaks, Calif.: SAGE Publications

Spector, P. E. (1981). Research designs. Beverly Hills: Sage Publications.

Cite this article (APA)

Trivedi, C. (2020, November, 18). Types of variables in scientific research. ConceptsHacked. https://conceptshacked.com/variables-in-scientific-research/

Chitvan Trivedi
Chitvan Trivedi

Chitvan is an applied social scientist with a broad set of methodological and conceptual skills. He has over ten years of experience in conducting qualitative, quantitative, and mixed methods research. Before starting this blog, he taught at a liberal arts college for five years. He has a Ph.D. in Social Ecology from the University of California, Irvine. He also holds Masters degrees in Computer Networks and Business Administration.

Articles: 37