Noise – A Flaw in Human Judgement by Daniel Kahneman, Oliver Sibony, and Cass Sunstein, is long, so it sat on my bookshelf for awhile. But it’s worth the time. Noise is the variation in judgements that shouldn’t vary. Judges should give similar sentences in similar cases, underwriters should find the same expected risks from identical cases, and fingerprint matches shouldn’t vary between experts. But they do – a lot.
Noise is different than bias — predictable poor judgements that don’t vary in direction, such as discrimination. Bias gets a lot of attention – which it deserves – but noise may be the bigger problem for society.
Noise is exceptionally high in healthcare. Price variation is well-documented, but healthcare delivery is also very noisy. Disagreement among radiologists reading the same image can lead to massive over and under-treatment. Doctors seeing the same care, should agree on a diagnosis and treatment plan more often than they do. Not only can noise harm patients, it also wastes a lot of money.
The book describes the many causes of noise – for instance, the time of day and weather conditions shouldn’t affect judgements, but they do. Artificial intelligence (AI) can help, but it’s not going to fix everything. The authors provide concrete options to lower noise and improve judgements. An appendix includes a toolkit to create a “noise audit” to identify the problem in your organization or system, if is exists. And the authors give very specific advice to reduce noise in your work.
Definitely worth adding to your summer stack of books.