Cynthia is a psychologist who is interested in whether or not watching a specific television show will help preschoolers learn to count. Since the show in question has several instances of counting, she believes that it will help the preschoolers learn their numbers. But how should she test her hypothesis? How can she prove that watching the television show will lead to better counting?
In order to show whether or not the television show has an impact on preschoolers' counting, Cynthia should run an experiment. Experimental design is the process of choosing how to run an experiment to best answer the researcher's question. For example, Cynthia is interested in whether or not watching the television show will lead to better counting by preschoolers. This is her experimental question, and she has to choose how best to run her experiment to answer that question.
On a practical level, the center of experimental design is usually how to use and measure participants. Should Cynthia show the television show to all her subjects? Should she just show it to some of them? Should she have the preschoolers count before and then measure their counting after viewing the show? These are all questions that she needs to answer with her experimental design. Let's look at one important element of experimental design that is the center of many experiments: the control group.
Okay, so Cynthia wants to see if watching this television show will help preschoolers learn to count. That seems pretty straightforward - she just shows them the television show and then asks them to count for her, right?
But how will she know whether the counting was due to the television show or whether it was just normal counting for preschoolers? The truth is, if she just shows them the program and then asks them to count, she can't answer that question. If the average preschooler can count to seven after watching the show, that might be because the show taught them that, but it might be because they already knew how to count to seven.
So how does Cynthia figure out how much impact the television show has had? She'll want to form a control group, or a group of subjects that did not get the treatment. In the case of Cynthia's experiment, the treatment is watching the television show in question. But for other experiments, a treatment might be a teaching method, or a drug, or some other thing that the researcher is studying.
What Cynthia will do is divide her participants into two groups. In her experimental group, or group of subjects that do get the treatment, the preschoolers will watch the television show and then count for her. In the control group, Cynthia will just ask the preschoolers to count for her without watching the television show. After both groups have counted for her, Cynthia will compare the two groups and see which group, on average, could count higher.
If there's no difference in the groups, then the television show makes no difference on counting abilities. But if the participants in the experimental group are able to count higher than those in the control group, Cynthia knows that there's a good chance that watching the television show will make preschoolers better counters. This is a basic control group/experimental group design in research. There are two groups; one gets the treatment and the other doesn't, and then the researcher compares the results to see if the treatment worked.