Control Groups

Random Assignment

The most basic control groups are those that are assigned randomly. That is, a participant sample is randomly divided up into an experimental group, which receives the experimental manipulation, and the control group, which does not. As far as the two groups, they are the same in all characteristics save the experimental manipulation; any differences in the DV between the two groups can be attributed to the experimental manipulation.

Nonequivalent Groups

With pre-experimental designs, experimental groups and control groups are not randomly assigned. Instead, these groups preexist, and researchers often choose them out of convenience. The selected groups are as similar as possible, but without random assignment, researchers cannot know for sure that they are equivalent. Thus, these designs are referred to as nonequivalent groups. For example, if a researcher is interested in two different classes and assigns one class to the experimental manipulation and one class to the control group, the researcher can compare the two. The researcher is limited, however, in the conclusions that can be drawn.

Placebo

Placebo groups are designed to prevent any potential change in outcomes due to participant awareness of membership in an experimental group or control group. Placebo designs are common in medical research. The experimental group receives medical treatment and the control group receives a placebo (a substance or treatment that has no therapeutic effect). For communication research (in which medical treatment is rare), [Page 253]placebo designs are often used by introducing a task or other type of experimental manipulation that has no effect on the DV without disclosing to participants that the control activity is unrelated to the research design. For example, if the experimental manipulation is a training protocol to reduce communication anxiety, individuals in the placebo control group might instead be asked to play a game or complete a different task that is unrelated to communication anxiety.

Matched

With a matched control group approach, participants are first grouped on preexisting characteristics such as gender or level of experience, considered to be potential influences on the relationship between the IVs and the DVs. From these matched groups, half of the participants are assigned to the experimental group and half are assigned to the control group. This approach can better control for extraneous variables so long as researchers are fairly confident as to which characteristics might be relevant, but it reduces the number of eligible participants, as only those members of the sample who fit the matching categories are included in the study.

...