All experiments must have a treatment group and a control group for comparison. No treatment is applied to the control group. Controlling all other conditions carefully is essential. If both groups are not treated exactly the same way in all other respects, you cannot be sure that the treatment variable was the reason for any different responses.