Controlling variables in research: What rationale behind it?

If you’re my frequent reader, you’ve read a post on the role of theory in data analysis. If you’re new, it’s fine; welcome. Well, in this post, I am going to walk you through the rationale behind controlling some variables in one’s research. Just stay on this post and scroll down gradually. It’s going to be helpful, I guarantee. Here we go!

Experimental Research

Variable control is frequent in experimental research. To refresh your mind, you should remember that a(n):

===> experimental research refers to study of the effect of the systematic manipulation of one variable(s) on another variable.

===> variable is a characteristic that takes on different values across people or things.

How to begin? Well, for your information, experimental studies originally were used in laboratories (for hard-science) of sciences such as biology and physics; it was subsequently adopted to apply to areas of social and education sciences.

Experimental research is research that is done deliberately by researchers with a means of providing treatment/treatment against certain of the subjects to see/know its impact on the subjects examined.

The notion of control group

As opposite to an experimental group (which given the treatment), a control group is not given treatment; it is a group serving as a witness one to which an experimenter refers to explain the effect of his/her treatment on the subjects observed.

Experimental research is the most reliable research in its scientific aspect (most valid) as it is done through controlling strictly variables outside the experiment, (Borg & Gall, 1983).

Experimental research is the method used to allocate specific treatment to a group of subjects just to see the effect of such a treatment.

But how come in education, any real experimental research there? Well, there is what is termed quasi-experimental research since the so-called experiment is carried out not in lab but on human beings.

Experiments in the field of education are done in order to innovate, improve the quality of learning, determine the effectiveness content, media, method, or practice in terms of learning outcomes.

How does it work?

In most experimental research design, there is a use of a single independent variable:

====>is manipulated (made of different condition), then the effect of having this “condition-made-different” is observed against one or more dependent variables.

====>the manipulated variable is called: treatment, experimental or independent variable.

====>a variable that represents a result/impact is called: the dependent or impact variable.

====>free/independent variables (treatment/experimental variable) is the variable that will be seen to affect the dependent or impact variable.

Let me make it easily graspable: the manipulated variable is known as the experimental treatment or the independent variable while the observed/monitored and measured is often called the dependent variable. Hope the difference is clear now.

Controlling variables in an experimental research

What is a control of variables? Uh…it is simply eliminating the effect of variables (often bound to the independent one) other than the variables of the researcher’s interest. The influence of a variable not interesting a researcher at the moment of his/her research is removed/controlled,(Jaedun, 2014).

By controlling variables, a researcher eliminates all other possible explanations as the influence of irrelevant variables is removed. On the contrary, without such a control it is impossible for him/her to evaluate unambiguously the effects of an independent variable or to make inferences about causality, (Ary, Jacobs, Sorensen, & Razavieh, 2010).

But as I stipulated in another post entitled “What is the role of theory in research and its Implications in Data analysis?“, theory plays a paramount of role in avoiding design or data analysis related fallacies. In behavioral sciences where the subjects of experiment are human beings, it’s almost impossible to have a 100% experiment that meets all the requirements to be called so. However, we social or behavioral scientists minimize the influence of variables other than those interesting us.

It is theory that is likely coming to our mercy when comes to experimental research conclusions. Otherwise, social or behavioral sciences experimenters have a lot of other variables bound to the variables they seek to observe.


Suppose that an educational experimenter titles his article “The Effect of the Discussion teaching technique on Teacher Students’ Practicum Performance.” If he wanted to compare the discussion teaching technique with the discovery one in his EFL Teaching Simulations Class and that the students’ performance in practicum is indicated by their scores in teaching practice (Y1) and teaching plan writing (Y2), there is a number of related variables other than the teaching techniques s/he used during his simulations classes.

Let’s suppose s/he has a record of students’ Emotional Quotient/Intelligence (EQ) but never considers it in his/her experiment, other researchers who would read his/her publication would be stunned to see that he/she did not control students’ EQ. Even before the publication any of his/her article, the reviewers might question the epistemological contribution of the work if it’s stained by such a flaw!

When he/she’s denied publishing the experiment-related article, after a thorough such, theory might reveal how EQ has swayed his/her experiment findings, displaying then how inconclusive are the conclusions of his/hers. But it is astonishing to see how EQ can influence the teacher students’ teaching practice and plan writing scores. For your information, look at how Thaib (2013:384) defines EQ. It refers to people’s “ability to recognize and manage emotions, motivate themselves, recognize people’s emotions and the ability to build relationships with others

What weight does it (EQ) have in explaining the teacher students’ teaching practice and plan writing scores? Very likely to influence the research results. So, it’s worth controlling this variable as much as possible to minimize errors.

As you have seen, controlling variable in an experimental research is inevitable if one needs to measure the real effect of an independent/manipulated variable on the dependent/observed variable. If not, other variables can influence what is being observed and the experimenter may fall into a fallacy of jumping into conclusions that are scientifically NOT sound.

There is then a good rationale behind making control of variables in an experimental research: it is meant to minimize the influence of other variables, not interesting to the researcher at the time of his/her experimental research.


Hello dear visitor/user,….I would be more than grateful to read from YOU. What may you say about this post? Did it meet your expectations? Anything I would change to serve others better? Leaving a comment would be a form of encouraging me or helping other visitors to be served well here.

Please follow and like us:

Leave a Comment

Your email address will not be published. Required fields are marked *


Enjoy this blog? Please spread the word :)

Scroll to Top