What Are Comparative Experiments?

Many students of science understand the basic idea of the comparative experiment because the name "comparative experiment" mostly explains itself. Students would be correct in defining a comparative experiment as one that compares the effects of two treatments. However, like most anything in science, the comparative experiment has advantages and disadvantages. Students must understand these aspects at a deep level before fully understanding the comparative experiment itself.

According to Penn State, a comparative experiment starts with a question or hypothesis that asks how two or more treatments affect some response. When a scientist wants to know the difference between the effects of treatment A and treatment B on dependent variable C, he will run an experiment in which all of the conditions are the same except for one: the treatment — A or B — given to the subject. After receiving the results of the experiment, the scientist can then compare the difference in the dependent variable C for each treatment, concluding either that one treatment is more effective than the other or that both treatments have about the same effectiveness.

The keys to a comparative treatment are control and randomization. Control refers to holding constant all of the other variables that could affect the outcome. For example, a comparative experiment comparing the effects of two diets of different nutritional value on the growth of mice should ensure that the mice eat at the same time, regardless of which diet they are assigned to eat. Randomization refers to randomly assigning the experiment's subjects, such as mice, to the two or more treatment groups. This randomization allows for valid conclusions and statistical analysis across treatments.

To many students of science, the comparative experiment is a time-saver. Standard, non-comparative experiments use a "control," which refers to a group of subjects that receive no treatment or a placebo. Scientists engaging in non-comparative experiments in their research would need to run the experiment twice, once with each treatment. For many experiments, however, running just one experiment can be a remarkable expense in both time and money. Thus, a comparative experiment can save a scientist the trouble of having to allocate resources to a second run with a different treatment.

Comparative treatments do not need to include a control, which can be a problem if both treatments yield similar results. For example, if two different injections lead to a similar amount of increased activity in mice, a scientist might be tempted to conclude that both of the injected drugs are effective at inciting activity. The truth is that without a control, the scientist cannot make such a conclusion, as other factors might be influencing the enhanced activity of the mice, such as anxiety from the injection or being handled by the scientists. A comparative experiment is generally limited to conclude the relative effectiveness of one treatment compared to the other.