How does a researcher know if their treatment has an effect or not? In this lesson, we’ll look at two-group experimental designs, contrast control and treatment groups, and examine random assignment and matched groups.
Rory is a psychologist, and he is interested in the effect of watching a popular science fiction show. He wants to know if watching the show will cause people to believe more in aliens than if they don’t watch the show.
Experimental design is the process by which a researcher decides how to run a study. For example, Rory might decide to get a bunch of subjects and divide them into two groups. He presents the show to one group and doesn’t present it to the other group.
Afterward, he asks whether or not they believe in aliens. If the group who watched the show answers ‘yes’ more often than the group that didn’t, he knows that watching the show will increase belief in aliens.Rory has chosen a two-group design, which is when an experiment is done on two groups of subjects and the results are then compared.
For example, Rory is going to compare the belief in aliens of two groups: those who watched the show and those who didn’t. Let’s look closer at the decisions Rory has to make in order to make his two-group design work.
A two-group design almost always involves a control group and an experimental, or treatment, group. The control group does not get the treatment, while the treatment group does get the treatment (hence its name). In Rory’s case, the treatment is the television show, so the treatment group is the group that watches the show.
The control group is the group that does not watch the show.In a simple, two-group design with a control group, the researcher wants to know whether the treatment has an effect or not. Rory, for example, wants to know if watching the television show will have an effect on belief in aliens.
His alien-belief survey will give him an idea of how much each group believes in aliens; and if the group who watched the show believe more, he can draw conclusions about the effect of the show.
But Rory still has to decide how to assign his subjects to each group. Most researchers use random assignment, which means that they put participants in groups using a random method. Maybe Rory flips a coin to see if someone will be in the control or treatment group. Maybe he draws names out of a hat.
There are also online randomizers that can help Rory assign people to groups.Random assignment works well if Rory has a large number of subjects because chances are that the groups will end up about the same. But what if Rory only has a small number of subjects? If he randomly assigns, he might end up with groups that are not equivalent.For example, what if he only has a few subjects and he randomly assigns them, but somehow all of the subjects in the treatment group turn out to be science fiction fans? They might be more likely to believe in aliens to begin with, so Rory can’t know if it’s the TV show he’s showing them that’s having an effect or if they were already believers.That’s why some researchers choose to create matched groups, which means they put participants in groups based on a common variable.
In Rory’s case, he might match up all of the science fiction fans, grouping them in pairs. For each pair, he can assign one to the control group and one to the treatment group. When he’s done, he has two groups with equal numbers of science fiction fans.
A two-group design is when a researcher divides his or her subjects into two groups and then compares the results. The two groups usually consist of a control group, who does not get the treatment, and a treatment or experimental group, who does get the treatment. In order to assign subjects to groups, researchers have two options.
Random assignment is often used with a large number of subjects, while matched groups is a good option for a smaller number of subjects.
You should be able to examine random assignment and matched groups in two-group design experiments upon completion of the lesson.