Why can't you compare apples to oranges? Why would it be important to match your intervention and control groups on certain characteristics in a research study to help with interpretation of the results?