A friend shared this fantastic resource about [statistical] Causal Inference, about which I know very little. This book (and website) looks like a great place to start, and it’s very well made. The combination of theory, demonstration / application and exercises for doing the actual real thing it’s talking about on real data looks great, and his definition of competency just about sums up what I’d like to understand in a world of data and statistics. Recommended.
So how would I define causal inference? Causal inference is the leveraging of theory and deep knowledge of institutional details to estimate the impact of events and choices on a given outcome of interest. It is not a new field; humans have been obsessing over causality since antiquity. But what is new is the progress we believe we’ve made in estimating causal effects both inside and outside the laboratory. Some date the beginning of this new, modern causal inference to Fisher (1935), Haavelmo (1943), or Rubin (1974). Some connect it to the work of early pioneers like [pioneering researcher into the cause of Cholera] John Snow…
But however you date its emergence, causal inference has now matured into a distinct field, and not surprisingly, you’re starting to see more and more treatments of it as such. It’s sometimes reviewed in a lengthy chapter on “program evaluation” in econometrics textbooks (Wooldridge 2010), or even given entire book-length treatments… The market is quietly adding books and articles about identifying causal effects with data all the time.
So why does Causal Inference: The Mixtape exist? Well, to put it bluntly, a readable introductory book with programming examples, data, and detailed exposition didn’t exist until this one. My book is an effort to fill that hole, because I believe what researchers really need is a guide that takes them from knowing almost nothing about causal inference to a place of competency. Competency in the sense that they are conversant and literate about what designs can and cannot do. Competency in the sense that they can take data, write code and, using theoretical and contextual knowledge, implement a reasonable design in one of their own projects. If this book helps someone do that, then this book will have had value, and that is all I can and should hope for.
Scott Cunningham – Causal Inference: The Mixtape
Note: it gets pretty technical pretty fast.