A concomitant variable, or covariate, is a variable which we observe during the course of our research or statistical analysis, but we cannot control it and it is not the focus of our analysis.
Although concomitant variables are not given any central recognition, they may be confounding or interacting with the variables being studied. Ignoring them can lead to skewed or biased data, and so they must often be corrected for in a final analysis.
Examples of Concomitant Variables
Let’s say you had a study which compares the salaries of male vs. female college graduates. The variables being studied are gender and salary, and the primary survey questions are related to these two main topics. But, since salaries increase the longer someone has been in the workplace, the concomitant variable ‘time out of college’ has the potential to skew our data if it is not accounted for.
If this variable is observed, recorded for and accounted for in the final results, your conclusions will be more valid. Typically this is done by noting the concomitant variable (here, age) in the initial data gathering, and then running a regression to ‘equalize’ all of the data points to the same number of years out of college.
Similarly, in a study comparing the effects of soil composition on the growth of tomatoes over 20 different locations country-wide, average temperatures and hours of sunlight available to each tomato patch would both be concomitant variables that would need to be included in a final analysis in order to get valid results.
An Extraneous Variable is something that the experimenter cannot control, which can have an effect on the overall outcome of the experiment. The main four extraneous variables are demand characteristics, experimenter effects, participant variables and situational variables.
(i) Demand Characteristics: Environmental clues that may tell the participant what is expected of them, such as the environmental setting or the researches body language. This in turn can affect their behaviour.
(ii) Experimenter Effects: When the researcher themselves affect the outcome by giving subconscious clues about how to behave. This may involve unintentionally asking leading questions that inform the participant of the desired result.
(iii) Participant variables: Something about the participant that is out of the researcher’s control. For example, whilst researches may try and target individuals with a certain background for an experiment, existing variables such as their health, or prior knowledge, could affect the outcome. For example, a participant with prior knowledge of Milgram’s experiment would be an extraneous variable in a reimagining of the experiment.
(iv) Situational Variables: Whilst the researcher may do their best to control an experiment (for example, controlling the time of day), situational variables can still affect the results. For example, a field experiment conducted at the same time of day across a week may experience sporadic weather or unexpected noise pollution, changing the mood/actions of the participants.
Treatment group is a group that receives a treatment in an experiment. The “group” is made up of test subjects (people, animals, plants, cells etc.) and the “treatment” is the variable you are studying. For example, a human experimental group could receive a new medication, a different form of counseling, or some vitamin supplements. A plant treatment group could receive a new plant fertilizer, more sunlight, or distilled water. The group that does not receive the treatment is called the control group.
In an experiment, the factor (also called an independent variable) is an explanatory variable manipulated by the experimenter. Each factor has two or more levels, i.e., different values of the factor. Combinations of factor levels are called treatments.
Treatment Group Examples
Example no. 1: – You are testing to see if a new plant fertilizer increases sunflower size. You put 20 plants of the same height and strain into a location where all the plants get the same amount of water and sunlight. One half of the plants–the control group–get the regular fertilizer. The other half of the plants–the experimental group–get the fertilizer you are testing.
Example no. 2: – You are testing to see if a new drug works for asthma. You divide 100 volunteers into two groups of 50. One group of 50 gets the drug; they are the experimental group. The other 50 people get a sugar pill (a placebo); they are the
Control group, the standard to which comparisons are made in an experiment. Many experiments are designed to include a control group and one or more experimental groups; in fact, some scholars reserve the term experiment for study designs that include a control group. Ideally, the control group and the experimental groups are identical in every way except that the experimental groups are subjected to treatments or interventions believed to have an effect on the outcome of interest while the control group is not. Inclusion of a control group greatly strengthens researchers’ ability to draw conclusions from a study. Indeed, only in the presence of a control group can a researcher determine whether a treatment under investigation truly has a significant effect on an experimental group, and the possibility of making an erroneous conclusion is reduced.
A typical use of a control group is in an experiment in which the effect of a treatment is unknown and comparisons between the control group and the experimental group are used to measure the effect of the treatment. For instance, in a pharmaceutical study to determine the effectiveness of a new drug on the treatment of migraines, the experimental group will be administered the new drug and the control group will be administered a placebo (a drug that is inert, or assumed to have no effect). Each group is then given the same questionnaire and asked to rate the effectiveness of the drug in relieving symptoms. If the new drug is effective, the experimental group is expected to have a significantly better response to it than the control group. Another possible design is to include several experimental groups, each of which is given a different dosage of the new drug, plus one control group. In this design, the analyst will compare results from each of the experimental groups to the control group. This type of experiment allows the researcher to determine not only if the drug is effective but also the effectiveness of different dosages. In the absence of a control group, the researcher’s ability to draw conclusions about the new drug is greatly weakened, due to the placebo effect and other threats to validity. Comparisons between the experimental groups with different dosages can be made without including a control group, but there is no way to know if any of the dosages of the new drug are more or less effective than the placebo.
It is important that every aspect of the experimental environment be as alike as possible for all subjects in the experiment. If conditions are different for the experimental and control groups, it is impossible to know whether differences between groups are actually due to the difference in treatments or to the difference in environment. For example, in the new migraine drug study, it would be a poor study design to administer the questionnaire to the experimental group in a hospital setting while asking the control group to complete it at home. Such a study could lead to a misleading conclusion, because differences in responses between the experimental and control groups could have been due to the effect of the drug or could have been due to the conditions under which the data were collected. For instance, perhaps the experimental group received better instructions or was more motivated by being in the hospital setting to give accurate responses than the control group.
A control group study can be managed in two different ways. In a single-blind study, the researcher will know whether a particular subject is in the control group, but the subject will not know. In a double-blind study, neither the subject nor the researcher will know which treatment the subject is receiving. In many cases, a double-blind study is preferable to a single-blind study, since the researcher cannot inadvertently affect the results or their interpretation by treating a control subject differently from an experimental subject.