What is a Control Group?

So, suppose you have a thing you do. Note that everything that you do affects or has an impact on something or someone.

Now, suppose you want to change something about the way you do that thing.

How would you know if whatever change you made to how you performed the act, has any impact (i.e., makes the effect better or worse) on the objects or persons who are affected by your act?

One way to do find out is to create 2 groups from amongst the things or people that are impacted by your act.

One of these 2 sets of people or things is called a “control group”. The other group is called the “experimental group”.

For the components of the control group, do not impose the change – i.e., keep doing whatever you are doing in the original way.

For the “non-control” (aka experimental) group – perform the changed act.

Now measure the effect /impact of the act on the components of the 2 groups. Roll up/average out the affect for each group.

If the measured effect of the act on both groups is similar, your modified act did not have any noticeable changed impact to the impacted persons or things.

If the measured effect on the control group is different from that on the experimental group – it implies that the modifications to your act has had an impact on the target persons or things.

Also the difference in measurement between your Control Group and Experimental Group will show how much the impact is.

The approach to experiments is used extensively for internet based targeted marketing.