Noise

A Flaw in Human Judgment

Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein · 2021 · Psychology

14 min · 12 chaptersNarrated in your voice

Original summary · AI-drafted, human-published · added by Library

What it's about

Kahneman, Sibony, and Sunstein argue that unwanted variability in professional judgment — two doctors, judges, or underwriters reaching different conclusions from the identical case — is a source of error as costly as bias, yet almost invisible because organizations rarely compare judgments side by side. Drawing on studies of sentencing, insurance, hiring, and forecasting, the book proposes measurable audits and structured 'decision hygiene' practices to reduce this scatter, reshaping how leaders should think about consistency, fairness, and expertise.

Reading plan

Pick a finish date and Genius lays out the days — the plan shows today's target and keeps you honest.

Read together

Start a circle and share the code — everyone sees everyone's honest place in the book. Accountability, not leaderboards.

Who it's for

- Executives and managers who rely on employees to make consequential case-by-case judgments - Professionals in law, medicine, insurance, or hiring who want to understand why their own verdicts vary - Readers of behavioral economics curious about a lesser-known companion problem to cognitive bias

What’s inside

1
Two Kinds of Error: Bias and Noise
Organizations that focus only on correcting bias are ignoring a second, often larger source of error: unwanted scatter among judgments that should be identical.
1 min
2
The Insurance Company That Didn't Know It Had a Problem
A properly designed noise audit — giving identical cases to many judges independently — reveals scatter that no one in an organization believes exists until they see the numbers.
1 min
3
Level Noise and Pattern Noise
Most disagreement among judges comes not from some being generally harsher or more lenient, but from each responding idiosyncratically to different features of a case — a source of error that simple recalibration cannot fix.
1 min
4
Occasion Noise: When the Same Person Disagrees with Themselves
Reliability cannot be secured by hiring better people, because even a single skilled judge is not a stable instrument — their own verdict shifts with mood, fatigue, and the order in which information arrives.
1 min
5
Why Groups Don't Cancel Noise, They Amplify It
Putting several qualified people in a room to discuss a case does not average out their individual noise; it typically produces a falsely confident, sometimes more extreme, shared verdict.
1 min
6
The Illusion of Agreement
Noise persists because judgment never feels arbitrary from the inside, so each professional wrongly assumes a colleague seeing the same facts would reach essentially the same conclusion.
1 min
7
Decision Hygiene: Structuring Judgment
Breaking a judgment into separate, independently scored sub-assessments before forming an overall verdict removes much of the noise caused by first impressions and order effects, without eliminating human expertise.
1 min
8
Algorithms, Rules, and the Resistance to Replacing Judgment
Even crude statistical formulas often outperform expert human judgment because they never have a bad day, yet people distrust and punish algorithmic errors more than identical human ones.
1 min
9
What Noise Reduction Costs
Consistency is not the same as fairness, and driving noise out of a system leaves any underlying bias intact while potentially sacrificing the individualized judgment that gives some decisions their legitimacy.
1 min
10
Where the Argument Is Thin
1 min
11
Final summary
1 min
12
Actionable advice
1 min

About Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein

Daniel Kahneman, a Nobel laureate psychologist and author of Thinking, Fast and Slow, pioneered research on judgment under uncertainty. Olivier Sibony is a strategy professor and former McKinsey partner focused on decision-making in organizations. Cass R. Sunstein, a Harvard legal scholar and co-author of Nudge, has advised governments on behavioral policy and regulation.