Noise: A Flaw in Human Judgment

Noise Review

Noise is Daniel Kahneman’s return to the psychology of judgment, co-authored with Olivier Sibony and Cass Sunstein. Where Thinking, Fast and Slow focused on bias — the systematic, predictable ways that human cognition errs — Noise focuses on a different and less studied problem: random variability in human judgment that is neither systematic nor predictable.

The core distinction is this: a scale that consistently reads two kilograms too heavy is biased. A scale that gives a different reading every time is noisy. Human professional judgment, Kahneman and his co-authors argue, is both biased and noisy — and while bias has received enormous attention since Kahneman’s earlier work, noise is almost completely ignored, despite being equally costly.

The examples are striking. Studies of medical diagnosis find that two doctors examining the same imaging will often reach different conclusions. Studies of sentencing find that the same crime committed before the same judge will receive sentences that vary enormously based on factors that have nothing to do with the crime or the defendant — including, famously, whether the judge’s local sports team won or lost that morning. Insurance underwriters price the same risk differently depending on mood, time of day, and ambient temperature. All of this variability is costly and almost none of it is necessary.

The practical proposals — decision hygiene, structured assessments, statistical base rates — are the most prescriptive section of the book and the most useful for organizations seeking to improve their decision-making. Noise is drier than Thinking, Fast and Slow and less narratively unified, but it is doing genuinely new intellectual work and deserves to be read alongside its more celebrated predecessor.

The Authors and the Shadow of Thinking, Fast and Slow

Noise carries the weight of expectation created by Daniel Kahneman’s Thinking, Fast and Slow, the 2011 synthesis of decades of his work with Amos Tversky that became one of the most influential popular psychology books of the century and helped earn Kahneman, a psychologist, the 2002 Nobel Memorial Prize in Economic Sciences. Here he writes not alone but with two formidable collaborators: Olivier Sibony, a strategy and decision-making specialist with a background in management consulting, and Cass Sunstein, the legal scholar and behavioural-policy thinker who co-wrote Nudge with Richard Thaler. The pairing is purposeful — Kahneman supplies the cognitive science, Sibony the organisational application, Sunstein the legal and regulatory dimension — and it explains both the book’s breadth and the occasional flattening of voice that some readers miss from Kahneman’s solo work.

The collaboration also signals the book’s ambition. Noise is not aimed only at the general reader curious about how the mind works; it is addressed substantially to the people who run institutions — hospitals, courts, insurers, hiring committees — whose decisions accumulate the costs that noise imposes. That practical orientation is the book’s most distinctive feature and the reason it reads less like a popular-science narrative than like a manual for reform.

Decision Hygiene and the Practical Payoff

The book’s most original contribution may be the concept the authors call “decision hygiene” — a deliberately unglamorous term, chosen by analogy with handwashing, because reducing noise means following dull procedural disciplines whose benefits are invisible precisely when they work. The authors lay out concrete techniques: breaking complex judgments into independent components, aggregating multiple independent assessments before discussion, using structured interviews and relative rather than absolute scales, and appointing a “decision observer” to flag the conditions under which noise creeps in. The argument that noise is in some ways a more tractable problem than bias — because you can detect and measure it without knowing the correct answer, simply by observing variability — is genuinely useful, and it gives organisations a way to improve judgments without first solving the much harder problem of identifying every cognitive bias at work.

Strengths, Limits, and Who It Is For

What the book does best is establish that random variability in expert judgment is pervasive, expensive, and largely unrecognised, and it backs that claim with disquieting evidence from medicine, law, forensic science, and corporate decision-making. Where it strains is in length and pacing: at 464 pages the core insight is stretched across more elaboration than it strictly needs, and the back half can feel like a series of applied case studies rather than a building argument. Readers who want only the central idea can come away satisfied from the first third; readers responsible for designing real decision processes will find the later chapters worth the investment. It pairs naturally with Thinking, Fast and Slow for the cognitive grounding and with Philip Tetlock’s Superforecasting for the complementary project of improving prediction, and it is best read by anyone who makes consequential judgments at scale and suspects, rightly, that their colleagues do not always reach the same conclusions they do.

Limitations

At 464 pages, it overstays its welcome — the core argument could be made in 200 pages without losing force. Co-authored books sometimes lack the distinctive voice that made Kahneman’s solo work so compelling. These are worth knowing before starting, though they are unlikely to diminish the experience for the readers the book is written for. The most efficient way to read it may be selectively: absorb the conceptual chapters that establish the bias/noise distinction in full, then treat the applied case studies as a reference to dip into according to one’s own professional concerns, rather than a continuous narrative demanding cover-to-cover attention.

Final Verdict

Our rating: 3.9/5 — A necessary companion to Thinking, Fast and Slow — Kahneman and colleagues identify noise as an equally important but far less studied source of human judgment error, with profound practical implications.