We like to emphasize that scientific knowledge is often superior to common sense merely because scientific knowledge is more precise (thus allowing us to say precisely how much is "too little," how much is "too much," and how much is "just right"). Multiple independent group experiments allow us to find these specific relationships, whether it be between amount of sleep and productivity, amount of alcohol and violence, amount of exercise and creativity, number of others and conformity pressure, amount of repetition and recall, or use examples of the Yerkes-Dodson Law (e.g., relationship between motivation and performance).
We also like to point to the value of multi-level experiments when, as often happens in real life, you aren't simply arguing whether a treatment is better than no treatment, but instead are arguing about which treatment is best. In addition, we like to point out that since no one control condition is perfect, it is often a good idea to use more than one control condition to improve construct validity. To help students realize the value of appropriate control groups, you may want to have students practice improving studies by adding control groups.
To more actively involve students, you can have students design a simple experiment and then ask them to show how it would be improved by adding a third group--or, if time is short, you can give students simple experiments to modify. Since it is difficult to find simple experiments in the literature, we often take an experiment that uses three levels of an IV, but describe it as though it used two levels.