Wisdomly❦
Psychology

Noise: A Flaw in Human Judgment

Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein · 2021 · 9 ideas · 9 min

The authors argue that unwanted, unpredictable variability — noise — in human judgment causes as much real-world harm as bias does, yet organizations rarely measure or address it because it's harder to see.

Start reading♡ Save

Why this book

The authors' central argument distinguishes two distinct kinds of error in human judgment: bias, a systematic deviation in a consistent direction, and noise, an inconsistent, unpredictable scatter of judgments around the same case that has no consistent direction at all. They demonstrate through studies of judges, underwriters, doctors, and forecasters that noise, meaning the same decision-maker or different decision-makers reaching wildly different conclusions on materially identical cases, is often a bigger and more overlooked source of error than bias, precisely because noise is invisible without deliberate measurement, whereas bias at least leaves a directional signature.

The book matters because it reframes how organizations should think about improving judgment quality: rather than focusing solely on removing bias, they should audit and reduce noise through structured decision processes, algorithms, and what the authors call "decision hygiene." This has direct implications for fairness and consistency in fields like criminal sentencing, insurance, medicine, and hiring, where unnoticed noise can produce arbitrary and unjust variation in outcomes.

Who should read it

Managers, professionals in judgment-heavy fields like law, medicine, or underwriting, and anyone interested in decision science will find concrete, applicable value here. It rewards readers willing to engage with some statistical framing, though the authors keep the math accessible.

About the author

Daniel Kahneman was a Nobel laureate psychologist known for pioneering behavioral economics; Olivier Sibony is a strategy professor and consultant; Cass R. Sunstein is a legal scholar known for work on behavioral law and regulation.

The ideas

01Noise and bias are separate, independently measurable errors02System noise splits into level noise, pattern noise, and occasion noise03Judges reach different sentences for identical cases far more than expected04Occasion noise means the same judgment can shift based on irrelevant context05The mere presence of noise erodes trust even without proof of specific unfairness06Simple rules and algorithms often outperform expert intuition specifically by removing noise07Structured decision processes reduce noise through independent judgment aggregation08Some noise reduction is resisted because judgment feels like professional identity09Not all variability is bad — some judgment tasks value it
decision-makingbiasjudgmentorganizational-behaviorstatistics
About this summary. Wisdomly re-expresses a book's ideas, arguments, and structure in our own words — nothing here is the author's text. Summaries are a map, not the territory: if the ideas land, the full book is worth your money and your evenings.
Noise: A Flaw in Human Judgment by Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein — summary & key ideas — Wisdomly