Is there an audiobook of Noise: A Flaw in Human Judgment?

Yes. Noise: A Flaw in Human Judgment has an Audible audiobook edition, narrated by Jonathan Todd Ross.

How long is the Noise: A Flaw in Human Judgment audiobook?

The Noise: A Flaw in Human Judgment audiobook runs 13 hr 28 min.

Noise: A Flaw in Human Judgment by Daniel Kahneman book cover for the Book of the Day psychology book review

Our score

4.3/ 5

Excellent

Listen to Noise: A Flaw in Human Judgment free with an Audible trial Narrator, runtime & our audio review

Buy on Amazon

Kindle$12⁹⁹

As an Amazon Associate we earn from qualifying purchases. Prices may vary.

EDITION PUBLISHEDMay 2021

Genres

Non-Fiction Psychology Business

Themes

Bias against random scatter as sources of errorWhy simple rules outperform expert intuitionDecision hygiene as organizational procedureWhat standardizing human judgment costs

Noise: A Flaw in Human Judgment

Name: Noise: A Flaw in Human Judgment
Price: 12.99 USD
Availability: InStock
Author: Daniel Kahneman

Daniel Kahneman

Daniel Kahneman, Olivier Sibony, and Cass Sunstein argue that organizations bleed accuracy through random scatter in judgment, not only through systematic bias. Two underwriters priced the same policy 55 percent apart. Noise names that problem, shows how to measure it, and hands managers a working set of procedures for cutting it down.

The Review

Our take

Bias is everyone pulling to the same side of the target. Noise is everyone landing somewhere different. Kahneman, Sibony, and Sunstein spend this book on the second kind of error, the scatter that shows up when two doctors read the same file, two judges sentence the same crime, or the same underwriter quotes the same policy on a Monday and a Wednesday. Their claim is arithmetic rather than rhetorical. Bias and noise contribute to error in equal measure, and almost no organization measures the second one at all.

The evidence is good and mostly comes from inside real companies. A noise audit at an insurance firm asked executives how far apart two underwriters would price an identical case. The executives guessed around 10 percent. The median gap was 55 percent. From there the authors split the problem into parts that can be attacked separately: level noise, where one judge is simply harsher than another; pattern noise, where a judge reacts idiosyncratically to a particular kind of case; and occasion noise, where the same person differs from themselves depending on mood, hunger, or the time of day.

The remedies come under the label decision hygiene, and they are procedural rather than inspirational. Break a complex judgment into separately scored components and hold the overall verdict until the end. Have people write independent estimates before the meeting, because the first person to speak contaminates everyone after them. Average multiple independent judgments. Judge candidates against each other rather than against an abstract scale. Hand repeatable calls to a simple rule, since even a crude linear model beats expert intuition on the same inputs. Be honest about the price of all this. It is an organizational program, not a personal habit. A real noise audit needs a sample of cases, staff hours, and a manager willing to hear that their team is inconsistent. A solo reader gets the ideas and very little to run.

Why you should read

Great if Thinking, Fast and Slow left you wanting the sequel argument
Great if you run hiring, underwriting, or performance reviews
Great if you would rather have statistics than anecdotes
Great if you want procedures instead of mindset advice

What to expect

Roughly 450 pages with a slow statistical middle
The core distinction repeated across many professional fields
Concrete tools collected in three closing appendices
Textbook register more than popular science storytelling

It is also long, and the middle stretch on measurement is dry going. Anyone arriving from Thinking, Fast and Slow expecting the same steady supply of surprising experiments will notice the difference in tone within fifty pages. What holds the book up is that it takes the objections seriously. There is a full section on what noise reduction costs in dignity, accountability, and the ability of a rule to entrench a bias permanently. The most practical pages are at the very back: an appendix laying out how to run a noise audit, and a checklist for the person assigned to watch a decision while other people make it.

Reviewed by Jordan · How our reviews are written

Our verdict

4.3/ 5

Excellent

Listen to Noise: A Flaw in Human Judgment free with an Audible trial Narrator, runtime & our audio review

Buy on Amazon

Kindle$12⁹⁹

As an Amazon Associate we earn from qualifying purchases. Prices may vary.

Where to read or listen

Is there an audiobook of Noise: A Flaw in Human Judgment?: Yes. Noise: A Flaw in Human Judgment has an Audible audiobook edition, narrated by Jonathan Todd Ross.
Listen to Noise: A Flaw in Human Judgment free with an Audible trial
We earn from qualifying Amazon purchases.
How long is the Noise: A Flaw in Human Judgment audiobook?: The Noise: A Flaw in Human Judgment audiobook runs 13 hr 28 min.
How much does Noise: A Flaw in Human Judgment cost?: On Amazon: Kindle $12.99. Prices as of August 26, 2026 at 3:46 AM EDT and may change.

Filed underPsychology Business Browse the archive →

More by Daniel KahnemanSee every Daniel Kahneman book we've reviewed, with our scores.All Daniel Kahneman books →

Common questions

What is Noise: A Flaw in Human Judgment about?: Bias is everyone pulling to the same side of the target. Noise is everyone landing somewhere different. Kahneman, Sibony, and Sunstein spend this book on the second kind of error, the scatter that shows up when two doctors read the same file, two judges sentence the same crime, or the same…
Is Noise: A Flaw in Human Judgment worth reading?: We scored Noise: A Flaw in Human Judgment 4.3 out of 5. Daniel Kahneman, Olivier Sibony, and Cass Sunstein argue that organizations bleed accuracy through random scatter in judgment, not only through systematic bias. Two underwriters priced the same policy 55 percent apart. Noise names that problem, shows how to measure it, and…