Causality and causal inference deal with expressing and reasoning about relationships of cause and effects, and structural causal model provide a rigorous formalism to assess causality.
Within this framework, confounders constitute an important concept and tool to understand data and give it a causal interpretation. Recognizing and controlling confounders allow us to reach consistent conclusion, and provides a solid base for decision-making.
In this notebook we experimentally review several standard scenarios of confounding, ranging from greedy casinos to the evaluation of new medicines. We implement these scenarios, we play around with them, we compare them, and we analyze how confounder affect our results and how our conclusion should be properly qualified. In conclusion, these study cases will provide an illustration for the different ways in which confounder may affect our models.