Causal Inference Lecture Notes: Covariate Adjustments in Randomized Experiments

In our previous discussion of classical approaches to randomized experiments, we did not talk about the situations in which we know more about experimental units. Indeed, typically such covariate information is available to experimenters. For example, in social science experiments, the demographic information about subjects may be known to researchers. Here, we consider how we may take an advantage of such information in randomized experiments.