ABSTRACT
In this chapter we introduce the fundamental concepts of experimental design. We show how experimental studies can be used to establish cause-effect relationships and why this is typically not possible using observational data. We study the main ingredients of an experimental study: The different treatments a researcher wants to compare, the experimental units (the objects on which the treatment is applied) and the response (the output that is being measured on each experimental unit). In addition, we illustrate why the random allocation of treatments to experimental units (randomization) is critical to drawing the correct conclusions and how confounding variables can be a problem with observational data.
