Stimulus Pretest

Experimental Design

There are several different types of experimental designs put forward by Donald Campbell and Julian Stanley that use a stimulus pretest. However, it is important to note first that a pretest can be used in both a quasi-experiment and a full experiment. The main difference between the two terms is the usage of random assignment. If participants are randomly assigned to their cells, it is a full experiment. If they are not randomly assigned, it is a quasi-experiment. For a full experiment, the researcher could first use random selection to place participants in cells and then run the pretest, again gathering evidence that the cells are not significantly different before the influence of the manipulation. This is called the pretest-posttest equivalent groups design. In a quasi-experiment, the researcher could run the pretest before placing participants in groups. This process creates quasi-equivalent groups instead of true random assignment, which still allows the researcher to [Page 1678]see the influence of the manipulated independent variable across the length of the experiment. This is called the pretest-posttest quasi-equivalent groups design.

Three other specific designs are of interest to this discussion. First, there is the one-group pretest-posttest design. This is a pre-experimental design highlighted by Campbell and Stanley as a flawed usage of a pretest. This design contains only an experimental group that fills out a pretest, is exposed to a manipulation, and then fills out a posttest. Due to the absence of a control group, there are several validity concerns, namely history (there might be other factors occurring between tests that would explain for the difference between the pretest and posttest) and maturation (differences in the participants themselves between the two tests). This design is included here to note that the presence of a pretest is not enough, in and of itself, to make an experimental design valid. Second, there is the pretest-posttest control group design. This design contains both an experimental group and a control group. Both groups are given the stimulus pretest and posttest, but only the experimental group is exposed to the manipulation. Here the pretest serves the functions of assessing group equivalence before the manipulation and as the first time point in assessing group difference following the manipulation. Finally, there is the Solomon four-group design. Two of the four groups in this design are the same as the previous design. However, this design adds another experimental group and another control group, both of which do not complete the stimulus pretest. This design was created to address one of the main weaknesses of the stimulus pretest, which will be discussed in the next section.

...