Definition

A controlled variable is a commonly used term in the field of scientific research, where finding evidence to support a theory is rarely straightforward. In the case of the natural sciences, some research features are constant, but the majority of these have inconsistencies. These inconsistencies are known as variables.

Overview

For an experiment to give statistically useful results, every aspect of the study subject and the environment must be the same, or as similar as possible. Studies are made up of independent variables (the effects of a specific change the researcher wishes to observe), the dependent variable (the measurement of this change), and the controlled variable. A controlled variable creates a similar environment across the board, so that the change that is being studies is not influenced by multiple, uncontrolled factors.

If seedlings are being tested for their rates of growth at two different light levels, the results of the independent variable (light levels) and the dependent variable (millimeters of growth) will be much more accurate if the seedlings are exactly the same. This does not only refer to their genetic make-up (the size of the seed, the parent plants, the species), but also external variables such as temperature, moisture levels, soil mineral content, air quality, position, and many others.

Controlled Variables
Sprouting seeds on growth medium

By using genetically cloned seeds in a carefully prepared growth medium placed inside a closed and highly controlled environment, and by following exact schedules for sowing and measuring times – as is the case in the image above – this study may then come to the conclusion that any changes in growth are due to light levels, rather than to other changes. Controlled variables should make study subjects and their environment as similar as possible. The perfect experiment controls all variables except the dependent variable – the result.

Controlled Variable Examples

In science, and in basic and applied research, variables are innumerable. From the simplest of elements to the most complex organisms, any number of differences can change the results of a line of study. The conclusions of an experiment carried out in one facility can differ to that of another, even when the same methods are applied. Living organisms are often too complicated to be expected to react in exactly the same way, whether this refers to the research subject, or to the researcher.

Non-Living Materials

Controlled variable examples in non-living materials are easier to implement than in research on living organisms. Research that looks at the reaction of one non-living material to another has the potential to implement near-perfect experimental controls. One example of a study on non-living materials could be the testing of two different smoothing processes on four different brands of dental cement. Testing can be carried out ‘in vitro’, meaning outside of a living organism, and thereby removing countless potential variables.

Controlled variables of this experiment would include application method and materials, light-curing intensities on the cement, specimen storage (temperature and duration), the length of time of the polishing process, the settings of the electron microscope, and the rotation speed of the polishing device. The addition of a control group would be a subdivision of the controlled variable. A control group is a group that undergoes the same preparation and is kept in the same environment as the tested samples, but is not exposed to the independent variable. In the above example, the control group features cement left unpolished. This removes the possibility of natural processes such as oxidation or air humidity that might affect the smoothness of untreated cement being counted towards the effect of the mechanical smoothing processes.

What this experiment would find difficult to control would be how identical each cement sample would be, as manufactured samples can differ. The distribution of ingredients in manufactured compounds can not be considered to be identical unless stringent tests are carried out before research commences.

Living Organisms

Controlled variable examples for living organisms are much more complex than in the majority of research upon non-living materials. In more complex and naturally produced living organisms, variables are predominantly uncontrolled. This is the primary reason why very simple organisms like fruit flies, or very similar organisms such as genetically cloned mice and rats, are used in testing environments. Once statistically relevant results are available in non-human models, human testing is initiated on groups that are as non-diverse as possible. The graphic below shows the steps all FDA-approved medications must go through. From the pre-clinical stage to stage III of the clinical trial, the possibilities for implementing controlled variables drastically diminish.

Typical Drug Development and Approval Process
FDA approvals process schematic

Advertisements for research subjects often ask for people of a certain age group, gender, or body mass index. They also refer to medical, behavioral or lifestyle variables such as no cigarette or alcohol consumption, no medication use, no co-morbidities, and medium to high levels of exercise. By implementing controlled variables early on, a researcher can create a group where results are more generic.

If a potential study subject fits this initial brief they are usually invited for further analysis. In the case of a weight-loss drug, for example, this might include insulin resistance and glucose testing, endocrine function, blood count, heart and lung function tests, and medical and familial history taking. Examples of controlled variables at this phase might be candidates without any family history of diabetes, or those who pass a specific psychological test.

Initial trials are able to study the effects of treatment in a very generic and similar population. The importance of psychological variables – something less influential in animal models – is never underestimated. However, once a drug, a chemical, or a therapy must be tested on the population for which it is designed, controlled variables become more difficult to achieve. In the case of the weight-loss drug example, a morbidly obese adult with a binge-eating disorder, sedentary lifestyle, anxiety and diabetes may not respond in the same way as a slightly overweight adult who gained weight after breaking a leg and has no comorbidities. But to which variables can a difference in response be attributed?

In all types of research, limiting other factors which may change either the action of the independent variable, or the result of the dependent variable is essential to obtain the best quality data. As organisms become more complex, the ability to limit these factors progressively decreases. By implementing as many controlled variables as possible, scientific evidence becomes more accurate and is a more solid and trustworthy foundation for the next generation of researchers.

What Exactly are Variables?

In experimental processes, variables can influence the final results. Researchers must attempt to limit these variables to the specific changes they are studying. A variable represents anything that undergoes change. Variables may be temperature fluctuations, comorbidities, behaviors, environments, diet, air quality, stress levels, metabolism, or allergies. Even seasonal or global events may have an effect upon the final results.

For research purposes, variables are categorized into three groups. The first is the independent (or manipulated) variable – the change that is consciously made in order to study a particular action or reaction, or change that is independent of our control, namely time and the ageing process.

The second variable is the dependent (or responding) variable, which the researcher measures in order to come to the final result. For example, a study may look at the effect a serving of blueberries has to the results of a color-coded memory test. The independent variable is the dietary change (blueberries). The dependent variable is the memory test used to measure whether blueberries affect the memory. It is easy to envision how potential variables can limit the accuracy of the researcher’s findings. Did the subject get a good night’s sleep? Did the subject feel unwell at the time? Did the subject understand the concept of the game? Is the subject color-blind? To limit these variables, this study requires a third type of variable – the controlled variable.

Quiz

1. Which of the following is a definition of a dependent variable?

A.
B.
C.

2. Which of these is not a controlled variable?

A.
B.
C.
D.

3. The further along the research route a clinical drug trial, the less the controlled variables.

A.
B.

4. Which of the following provides the most controlled variables?

A.
B.
C.

5. A control group:

A.
B.
C.
D.