Points of Intersection between Randomized Experiments and Quasi-Experiments

There are several contexts in which randomization in research design is not possible. Researchers may be interested in the effects of large-scale events or policy initiatives, or of complex causal processes that cannot be replicated. Threats to external validity in randomized experiments will lead to greater dividends for some research questions in a quasi-experimental design. Quasi-experiments can offer opportunities for testing causal theories, if they are carefully constructed to counter threats to internal validity. Furthermore, randomization is not a simple guarantee against threats to internal validity where few units of observation are used. An appreciation of the intersection of these approaches—between stronger quasi-experiments and weaker randomized experiments—can help guide our choice of experimental method. One can make the same argument about the benefits of pursuing good opportunities with quasi-experiments over scant opportunities with randomized experiments as is made more often in reverse.