Two-Group Pretest–Posttest Design

Design Setup

Experiments are a style of research that attempt to establish causality (the idea that changes in independent variable A are the cause of changes in dependent variable B). In order to establish causality, researchers must create designs with internal validity, or assurance that no other explanation exists for the relationship between variables A and B. In addition, researchers strive to achieve external validity, or likelihood that their results will be similar to those found in reality rather than a laboratory.

There are many different forms an experiment can take. Depending on the number of groups, exposure to manipulations, and the number and timing of measurements, these types of experiments take on different names and have various advantages and disadvantages. The setup of a two-group pretest–posttest design has two essential components that contribute to its unique research advantages: measurement of the dependent variable and the groups. In this particular design,[Page 1798] measurement of the dependent variable occurs before and after the intervention. The pretest refers to a measurement made prior to the intervention, which serves as a baseline to compare against a measurement taken after the intervention, the posttest. In a medical drug trial, researchers may compare a patient’s blood pressure before and after administering a drug to see if the drug had any effect on that patient.

However, these effects may have been caused by any number of things other than the drug itself; for example, the doctor may have been soothing, or the patient may have changed his or her diet between the two tests. In order to come closer to establishing the drug as the reason for any positive health changes, researchers must demonstrate its effects, specifically, and rule out other possible causes. This is solved by allowing for a comparison between groups. In this design, the “two groups” are typically referred to as the treatment group and the control group. The treatment group is a subset of the sample that receives an experimental manipulation in which the researchers are interested. The control group is ideally identical to the treatment group in every way, but does not receive the experimental manipulation, and serves as a way to measure what might have happened without any intervention.

For example, in the previous case of a medical drug test, individuals in the control group would be given a placebo (a pill which has no known effects), while those in the treatment group would receive the actual blood pressure drug. Researchers would then test to see if there were significant changes in the blood pressure of patients in the treatment group as compared to the control group. By comparing two groups who have been otherwise exposed to identical conditions (e.g., the same doctor and diet), researchers can be more certain that changes were due to the drug and begin to rule out other possible causes.

...