A control experiment is an experiment where the variable that is being investigated is kept constant. It acts as a comparison for the actual experiment . For example, suppose a scientist wishes to investigate the saying "Boys don't make passes at girls who wear glasses" He might take a girl who wears glasses, sit her at a bar on her own and see if any boys do in fact make a pass at her. Now he might well find that no boy did make a pass at her, but he can't be sure that the reason was that she was wearing glasses. She could just be incredibly ugly.

So the scientist sets up a control experiment. He puts the same girl, in the same bar, on the same day of the week, but this time she is not wearing glasses. Now he can compare the results of this experiment with that of the previous one and come to a meaningful conclusion.