Why Do Experiments Usually Require a Control?

What a Control Is

• A control can be a sample of one subject or an entire group. This control is typically similar to or the same as the other subjects or groups used in an experiment. The control does not come into contract with variables that have been manipulated by the experimenters. Other samples are exposed to these variables. The experimenters can observe the change in the exposed samples, then compare the differences to the unchanged control.

What a Control Does

• A control in an experiment determines whether or not the system being tested is behaving as it should. It represents how a system or all of the samples are expected to normally behave, based on the experimenters' previous experience and observation of that system or the samples. That way, the control serves as a point of comparison between normal behavior and the behavior of other samples subjected to testing. Comparison with a control also helps point out any errors in an experiment. This helps the experimenters avoid those errors the next time tests are conducted.

Positive Scientific Control Groups

• Some experiments are designed to have a positive scientific control group, which can eliminate the chance of false negatives. An example of this type of testing can be seen in research that tests the abilities of different drugs. A researcher can assemble several samples of the same pathogen, including a control group. The researcher can use a drug that is known to overcome the pathogen on one sample, a different drug on the other samples and no drugs on the control group. If the different drug fails its test but the first drug does not, the different drug is likely ineffective. However, if the control group also somehow fails, that shows that there was an error in the researcher's design.

Negative Scientific Control Groups

• Controls also help to prevent false positive in lab experiments. If, in the example above, all of the drugs administered were successful in overcoming the pathogen, but the control group also demonstrated a decrease in the pathogen's vitality, that means that an unplanned variable effected the results. Therefore, the experiment would have to be done over because the first results were invalidated by the unplanned variable. Negative control groups help point out confounding variables or biases in experiments.