Noise
Daniel Kahneman, Olivier Sibony & Cass R. Sunstein · 2021
Noise (2021) is a book by Daniel Kahneman, Olivier Sibony & Cass R. Sunstein on Decision-Making & Judgment. Kahneman, Sibony, and Sunstein argue that unwanted variability in professional judgment, not bias, is a huge and mostly invisible source of error in courts, hospitals, and companies, and show how to reduce it.
Two equally expert professionals can look at the identical case, a claim, a sentence, a resume, and land on wildly different verdicts, and almost no organization ever checks.
Two of the ideas you’ll keep
The Clif summary of Noise pulls out 7 insights. Here are the first two.
- Total judgment error splits into two independent parts: bias, the average deviation in one direction, and noise, the unwanted scatter around it. Because organizations obsess over bias and rarely measure noise, cutting noise is often the cheaper, larger lever on accuracy.
- A noise audit gives many professionals identical case files to judge independently and blind, then measures the spread. At one insurance company, executives guessed underwriters would differ by about 10 percent pricing the same risk; the actual median difference was 55 percent.
The other 5 — and the full 11-minute summary — are in the app.
What’s inside: 9 chapters
The whole summary runs about 11 minutes, read or listened to.
- Two Kinds of Error: Bias and Noise
- Judgments Are Measurements
- The Noise Audit: What the Insurance Study Found
- Level Noise, Pattern Noise, and Occasion Noise
- Why We Don't Notice Noise: The Illusion of Agreement
- The Psychology of Judgment: Matching and Excessive Coherence
- Noise in the Wild: Forensics, Hiring, and Forecasting
- Decision Hygiene: Reducing Noise Without Knowing the Answer
- The Limits of Noise Reduction: Rules, Standards, and Costs
Finish Noise tonight.
Read or listen to the full summary in about 11 minutes, then keep the ideas with a minute of review a day. Start with the 2-minute quiz — no card required.