To a layman, all experiments might seem the same. But to the scientist, each type of experiment has important differences that make certain experiment designs more suitable for different kinds of research. Two such experiments, the controlled experiment and the comparative experiment, seem similar on the surface but have one important dissimilarity.

Scientists Are Control Freaks

In scientific research, a controlled experiment keeps all the factors, or variables, in an experiment consistent, except for one variable known as the “treatment.” In such an experiment, one group of subjects gets the treatment while the other doesn’t. In medical studies, the non-treatment group often gets a placebo. The word “controlled” comes from the experiment’s ability to control all of the possible effects that might influence the dependent variable-- the value in question, such as the effect of a new medication in treating a disease condition-- of the study. In this way, scientists can be almost certain that the different results between the treatment groups are due only to the difference in treatment type. For example, if scientists were testing a new drug’s ability to improve short-term memory, they would give it to one group of mice but not another. The mice would be the same in every other respect, from the size of their changes to their diets. This way, when scientists put the mice through mazes and observe the treatment group finishing the maze more quickly, they can almost certainly attribute the speed increase to the new drug.

Comparing and Contrasting

Comparative experiments are designed to determine the differences between different forms of treatments. A comparative experiment includes two or more different treatment types. For example, scientists might already know that drug A works at improving short-term memory in mice. They might want to test whether drug B also works at improving short-term memory but are especially interested in how it compares to drug A. Drug B might work better or not as well. To test this, the scientists would run an experiment similar to the controlled experiment, comparing mice injected with drug A to mice injected with drug B. They would still control all other factors in the experiment, such as the environment and diet of the mice, but would be entirely focused on comparing the effects of the two drugs on the mice rather than looking at the effect of an individual drug.

It’s All about Control

The comparative experiment is nearly identical to the controlled experiment on the surface. But an important difference exists between it and the controlled experiment. The controlled experiment always has a control group. A control group is a group of subjects that receives no treatment at all. This allows scientists to know whether a treatment has any effect. While it might be tempting to leap into a comparative study that compares the effects of two different treatments, scientists often will perform a controlled study first to ensure that the treatments make any difference when compared to no treatment at all. Sometimes neither treatment has any effect, but a scientist performing a comparative experiment might conclude that “both treatments are equally effective” when in fact the better choice of words would be “neither treatment is effective.” But a scientist won’t know this unless she or a fellow researcher has already performed controlled experiments on the treatments.

The Best of Both Worlds

No one says that you have to choose one of these two experiment types. Though it might be more costly, time-consuming and difficult, a scientist can perform a controlled comparative experiment. In this type of experiment, the scientist includes both multiple treatment types and a control group. In this way, she can compare the treatments to the control group to see if either treatment has an effect. If she finds that both treatments influence the results, she can then compare the two treatments. The advancement in statistics has allowed scientists to easily perform data analyses for this type of experiment through computer software that runs all the comparisons at once. Today’s computer programs can take in a scientist’s data and spit out results answering the two questions: “Do the treatments have an effect?” and “Which treatment has a stronger effect?”