What I’m Reading: If correlation doesn’t imply causation, then what does?

Michael Nielsen, author of Reinventing discovery: The new era of networked science wrote a great piece explaining, in part, the theory of causal calculus, introduced by Judea Pearl. Nielsen describes the three important ideas of causal calculus: causal models, causal conditional probabilities, and directional separation (or d-separation). The ‘clever imaginative leap’ that causal calculus takes, as Nielsen explains it, is that the method establishes a set of rules that help us infer causation, even when we’re unable to run a randomized, controlled experiment. Potentially powerful stuff!

Want a little practice with these ideas? Nielsen includes optional exercises for the reader (I glossed over them, admittedly), and tackles problems that he has with the model. This has lead to a very rich comments section – make sure you check it out if you want to dig deeper into this topic, and see what other great minds think of Nielsen’s work!