#47: The Shape of Probability – Harry Crane
Harry Crane is a professor of statistics at Rutgers University.
In this episode, Joe catches up with Harry in New Jersey to discuss the history of probability, how experts routinely misuse statistics, “The First Rule of Probability”, the replication crisis in science, and how Harry thinks about probability in shapes rather than numbers.
Show notes
Selected links
Connect with Harry Crane: Website | Twitter
Harry’s book, Probabilistic Foundations of Statistical Network Analysis
Harry’s project, Researchers.one: Website | Twitter
Nassim Taleb, Foundations of Probability Seminar
Daniel Kahneman, Foundations of Probability Seminar
Harry Crane, Foundations of Probability Seminar
The Signal and the Noise, by Nate Silver
The modern-day snake oil salesman, Harry’s medium article on Nate Silver
A technical companion to the above medium article
Benjamin et al’s paper on redefining statistical significance
Harry’s paper on the replication crisis, a reply to Benjamin et al
Skin in the Game, by Nassim Taleb
Daniel Kahneman’s open letter on priming
Thinking, Fast and Slow, by Daniel Kahneman
The ‘Linda problem’ (original paper)
The ‘base rate fallacy’ (original paper)
Topics discussed
- Harry’s definitions of probability and statistics. [14:20]
- Misuses of statistics, including why Nate Silver is always right (and therefore wrong). [23:12]
- Harry introduces “The First Rule of Probability” – an application of “skin in the game”. [28:14]
- The replication crisis in science. [33:15]
- Harry’s radical proposal to address the replication crisis. [59:27]
- Joe and Harry discuss Kahneman and Tversky’s famous “Linda problem” and whether people being “bad intuitive statisticians” is necessarily a bad thing. [01:08:48]
- Gerd Gigerenzer’s riposte to Kahneman and Tversky. [01:19:39]
- Kahneman and Tversky’s “blue cab problem” and the base rate fallacy – and how they are predicated on the supposed rightness of Bayesian probability. [01:21:23]
- How Harry thinks about probability in shapes rather than numbers. [01:31:33]
- Harry’s new book Probabilistic Foundations of Statistical Network Analysis and the main idea it advances. [01:35:37]
- Harry’s project Researchers.one and how he’s hoping to improve academia. [01:39:00]