Using Causal Bayes Networks as a normative theory for statistics and causal discovery, we have developed an extensive simulation environment for teaching the theory through hands-on, interactive exploration. I will present and demo the "Causality Lab" - as well as experiments we have conducted to explore how students learn the theory using the lab. I will also present evidence showing that students who use causal diagrams to represent social policy debates are better at analyzing these debates than those who do not, even with only 10 minutes of training.