I’m currently an instructor at the Asian Spring Program on Rationality, a camp for mathsy teenagers in rural Taoyuan, Taiwan.
I chose to teach a class about reasoning through cause and effect in the social sciences. You can view the slides of my presentation here.
The basic idea was to introduce students to the fundamentals of causal inference. It didn’t make it into the slides, but I was also drawing out the scenarios I was describing using Judea Pearl’s directed acyclical graph notation. I hope I got some of them at least somewhat interested in this topic, although one of my main conclusions was that, no matter how wonderful your students are, you are facing an uphill battle in any room containing bean bags.
Later that day, one of our students patiently walked us through the proof of the Gibbard-Satterthwaite theorem. The kids are going to be alright.
Earlier in the camp, I taught a class about anthropic reasoning, for which I used these slides. That was an updated and extended version of an earlier talk I gave about the sleeping beauty paradox.
These are the books and papers I discussed in the session. I recommend the ones with a star:
Judea Pearl, Causality: Models, Reasoning, and Inference, 2nd edition. ★
Scott Cunningham, Causal Inference: The Mixtape. ★
Josh Angrist, Jörn-Steffen Pischke, Mostly Harmless Econometrics: An Empiricist's Companion. ★
Jeffrey Wooldridge, Introductory Econometrics: A Modern Approach, 7th edition. ★
David Card, Alan Krueger, Minimum Wages and Employment: A Case Study of the Fast Food Industry in New Jersey and Pennsylvania.
Josh Angrist, Victor Lavy, Using Maimonides’ Rule to Estimate the Effect of Class Size on Scholastic Achievement.
Joshua Angrist, Lifetime Earnings and the Vietnam Era Draft Lottery: Evidence from Administrative Records.
Scott Alexander, Two Dark Side Statistics Papers. ★
John Donohue, Steve Levitt, The Impact of Legalised Abortion on Crime.
Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined. ★
In the case of the journal articles above, there have been important later criticisms of the methodology – but I was trying to explain the basic causal reasoning behind the original paper. The actual conclusion is somewhat besides the point.
The slides should be readable without having seen the presentation. Given the theme of the challenges of causal reasoning, I thought it would appropriate to open with M.C. Escher. The only slide you won’t understand without context is a picture of the draft lottery balls being turned, which is a reference to the allegation from some statisticians that the lottery system for being drafted into the Vietnam War wasn’t truly random (presumably because the guy turning the machine didn’t crank the handle hard enough). In fact, the part of the class that the students were most surprised by was the fact that America once selected soldiers to be drafted on the basis of a televised lottery…
Many thanks to the organisers for having me!
Judea's other shorter/more friendly primer is also really good: https://www.amazon.com/Causal-Inference-Statistics-Judea-Pearl/dp/1119186846 :). That's the main thing I've read for causal inference, will look through your slides 🫡 it's so cool you taught this at ASPR! :)
Wait, the student led the class through the proof of GS? Without assistance?!
Kids these days! There’s no point in even trying to keep up.
The Pearl book, iirc, requires some background in pretty advanced probability theory. Eg, he uses Markov chains. Or am I mistaking the book you mentioned for a different book?