Learning from “The Book of Why” about Casual Science
As I started reading “The Book of Why” by Judea Pearl I was reminded how often analysts trot out “correlation is not causation“. It’s a well known warning. Indeed, I often encourage those learning data visualisation to ensure their designs don’t imply causation. But this book helped me think much deeper about this.
In this book review I hope to bring to life that this is not a light read but it is an important one. As Judea Pearl makes so clear in his case studies, causation is central to how we think as human beings. Most non-statisticians are not really interested in patterns & correlations in the data. They want to know what causes an effect, so they can take appropriate action.
Sadly this book also chronicles how for decades the formal study of Statistics has not helped that quest. Leading figures (who are critiqued in this book) declared it unknowable or that the only answers lay in more data. As well as the veneration of Randomised Control Trails (RCTs) this mindset played into the Big Data revolution. But lacking any documented causal model of the real world, this can lead to misleading & false results. When you only have a hammer, you start to see nails everywhere.
A ladder to guide our understanding of causation
But I am getting ahead of myself. Let me share with you first an overview of what to expect in this book. Then I will go on to explain some of the helpful frameworks & approaches Judea outlines. After a helpful scene setter (on why this subject matters & his background), he shares a ladder as the chief framework. It is called thew Ladder of Causation & makes clear the difference between three levels of understanding causation:
- Seeing = Association (e.g. correlation, What if I see…? What does a survey tell us about election results?)
- Doing = Intervention (e.g. experiments, What if I do…? What if we ban cigarettes?)
- Imagining = Counterfactuals (e.g. understanding possibilities, What if I had done…? Why? What if I had never smoked?)
Building on that fundamental structure, Judea goes on to share both the history of progress & set backs. New heroes & villains emerge as away from the Statistics mainstream progress is made. Including the important contribution of Bayes & Bayesian Networks. Along the way & beyond the history lesson, he explains what is needed for each rung of the ladder. Judea explains the importance of confounding variables, paradoxes, designing interventions & modelling counterfactuals.
He finishes this critical textbook (for understanding Casual Science) with considering the relevance of this progress to AI & Big Data. As I considered when reviewing “Rebooting AI” we are still a long way from more helpful general intelligence. From robots understanding why they do & perceive what they do. He brings to life why the answers lie more in Casual Science than in more Big Data.
Models to help analysts think deeper about the real world
As I mentioned before this is not an easy read. That is not because it is badly written. Not at all. Judea has the warm human writing style of his Jewish heritage. Rather, it’s because this is not an easy topic. This book will challenge you to think. It’s a chance to go ‘back to school‘ and learn a science that should be as commonly taught as other branches of Maths & Statistics. This review cannot do all that material justice, but let me reference a few tools and approaches that I recommend analysts explore & learn.
Early on, the author introduces us to Causal Models. Diagrams with arrows that capture out understanding of known or possible casual relationships. Like the power of analysis being informed by real world domain knowledge, this makes all the difference. Time & again he shows how such a visual graph can inform approaches & the mathematics needed. Building on that, he teaches the reader (with sufficient maths background) his Do Calculus. A mathematical way to capture casual models & resolve these expressions of what happens is you take certain actions.
There are two further topics I would recommend analysts & statisticians learn from this book. All the above are powerful enough by themselves. Even more transformation of current approaches can be made by climbing the ladder of causation to the top. Together with being aware what can go wrong. Statisticians will find the sections on Counterfactuals informative & at times philosophical. It’s amazing to reflect how easily a human child can imagine what might have happened if they had acted differently, yet how elusive has been the maths to represent it. Lastly, I would direct all analysts to read his explanations & mitigations for Simpsons Paradox, Monty Hall Paradox & Berkson’s Paradox.
Could you benefit from reading “The Book of Why?”
Despite my warnings that this book will require you to think & flex your maths muscles, I recommend it. Indeed, this book makes a good case for all Data Scientists needing to study Causal Science. It also shows why this level of understanding will help analysts. I hope you take up my recommendation and can’t wait to see a lot more causal diagrams. If nothing else it will give you a chance to sit at the feet of a true master of this science. Judea Pearl has left the world a rich legacy from his pioneering work in this field.
What would you have analysed differently if you knew how to answer the simple question: Why?