Let's say you're testing a new drug to see if it helps with depression. We'll have 200 people split into two groups of 100 each. The first 100 people get the new pill, while the remaining 100 people get a placebo (sugar pill). Each participant does not know whether they are taking the actual drug or the sugar pill. This is important.
The control group is the group of people taking the sugar pill. This is to counterbalance the experimental group that takes the actual drug. If we just had one group, say the experimental group, then we wouldn't be able to compare those results against anything else.
Now let's say that everyone, in both groups, reported better moods and more positive attitude. This would strongly imply that the new drug is not effective because people who took the sugar pill got better. In other words, the sugar pill is as effective if not more so compared to the new drug. It may be the sugar pill, the psychology of taking any pill, or some other factors that contributed to those people getting better.
If only the people in the experimental group got better, then we could conclude that the drug is effective in some way. Again, we're making a comparison between the two groups, and the control group is used to reduce the placebo effect.
