A positive control is used in experiments to demonstrate that the scientist and her/his laboratory are capable of carrying out the experimental design.

A classic toxicological experiment to determine the effect of compound X will include three groups: a standard control group, a positive control group, and the experimental group.

The standard control group is the set of individuals (mice, cells, people, etc.) that were exposed to a substance known to cause no effect.

The positive control group is the set of individuals exposed to a substance known to cause an effect, based on previous, established and accepted research.

The experimental group is exposed to compound X.

For the experimenter to conclude that compound X has a real effect, the impact of compound X should be statistically distinguishable from that of the standard control.

If the impact of X is not different, then the crucial question becomes whether the impact of the positive control is different from the control. If it is not, then the experiment has failed. No conclusion can be reached about the impact of X because the positive control should have been different from the control. The implication is that there has been a procedural error or design flaw in the experiment.

There are many possible sources of such mistakes. For example:

  • An unknown contaminant may be present, either preventing the response altogether or provoking it in all groups simultaneously. For example, if the control group is affected by the unknown contaminant in a way that mimics the expected result of the positive control, then the effect in the positive control and the experimental group may be indistiguishable from the control.
  • The laboratory procedures may be too sophisticated for the lab conducting the studies because of a lack of training, inadequacy of equipment, etc.
  • The sample sizes used in the experiment may be too small to yield statistically significant results.

Whatever the cause, experiments in which the positive control fails cannot be used to reach conclusions about the impact of the experimental treatment.