A quasi-experimental design is used when experimentation is not possible. The researcher selects two groups, one which has been exposed to the independent variable, and the other which has not. The researcher then compares the two groups to determine the impact of the independent variable on the dependent variable.