Experiments vs. Quasi-Experiments

In study designs where there are multiple groups of participants (e.g. the Two Groups Pre-Test/Post-Test design from the previous sub-module) the method used by the researcher to allocate the participants to the groups determines whether the study is an Experiment or a Quasi-Experiment. The primary difference is that in an experiment the researcher uses randomization or control to create the groups. While in a quasi-experiment the researcher does not have this level of control.

In an experiment the researcher

  • Manipulates the Independent Variable to create groups with different levels of the variable
  • Compares the group based upon the measured Dependent Variable(s)
  • Keeps all other variables as constant as possible through randomization or control.

In the case of randomization, the researcher assigns participants randomly to the groups. If the sample size is large enough, then randomization should help balance the effects of any uncontrolled variables across the groups. In the case of control, if the researcher has a belief that there is a key uncontrolled variable that may impact the results, the researcher can use that variable to balance the groups. For example, in an educational setting, a researcher could use student performance in prior courses as a measure of ability and balance the groups based on this measure.

Internal Validity

The overall goal in study design is to create a study such that the only factor that can impact the Dependent Variable is the Independent Variable under study. The more the researcher can reduce the potential effects of other Uncontrolled Variables, the higher confidence they can have in the results.

Previous submodule:
Next submodule: