In all scientific disciplines from physics to psychology, experimental design is a cornerstone of the scientific method. Investigators formulate a hypothesis and then test it through experiments which are carefully designed to be replicable by other researchers and to yield statistically provable results. These goals are achieved through the use of experimental and control groups.
When an experiment is carried out, the subjects or specimens which are the focus of the experiment are divided into two groups. The experimental group is subjected to an experimental procedure, but the control group is not, and the results from both groups are then compared. For example, a group of laboratory rats is divided into two. All the rats experience identical living conditions, receive identical food and all are in good health.
An experimental drug is then administered to one group of rats (the experimental group) but not to the other (the control group). Changes in the health and behavior of the rats in the experimental group are observed and compared with the health and behavior of the rats in the control group. Since the two groups are identical in everything except the administration of the drug, any change that is observed in the experimental group but not in the control group will almost certainly be caused by the drug.
Controlled experiments are standard practice in the physical sciences, where it is relatively easy to identify and control for all possible variables. However, in medicine, psychology and the social sciences, which deal with the complexities and unpredictability of human behavior, establishing a viable control group is more complex. Psychological and sociological experiments, for example, need to control for variables such as age, sex and experience, and frequently use random samples of people to ensure that subject groups are representative of the general population.
Research may also need to control for the placebo effect, for instance when testing a new drug on human subjects. To ensure that changes observed in members of the experimental group when they are given a pill are real and not psychological, members of the control group are given dummy pills to closely mimic the experience of the experimental group.
To avoid subjects’ suspecting they are receiving a placebo, perhaps based on unconscious clues provided by the experimenter, such as tone of voice or body language, double-blind procedures are frequently used to control for experimenter or subject bias. These are procedures designed to ensure that neither researchers nor subjects know who belongs to the experimental group and who belongs to the control group until after the data have been collected.
Control groups may be either negative or positive controls. Negative controls are groups where researchers expect to see no results, such as a group which is not expected to respond to medication because they receive placebos. Positive controls are groups where a result is expected, and may be used, for example, to assess the validity of a chemical test by using a neutral substance in the experimental group and a reactive substance in the control group.