When you think about errors of judgement, you probably think about bias - the type of mistakes that bend us one way rather than another. But there’s another kind of mistake that is often missed: noise, which scatters judgement all over the place, depending on the person, the situation, or even the time of day.
In Noise, by Daniel Kahneman, Olivier Sibony, and Cass R Sunstein, shows us that noise is everywhere and is seriously disruptive in just the kinds of places you’d expect us to be consistent.
In 1973, a famous judge, Marvin Frankel, drew attention to the problem in the justice system with a series of powerful anecdotes. Here’s one: two men, neither of whom had a criminal record, were convicted for cashing fake cheques for similar amounts. One was sentenced to fifteen years, the other for thirty days, for what were essentially identical crimes. The judges involved had with similar training and experience simply made different decisions.
A larger study, in 1981, involved 208 federal judges who were exposed to the same sixteen hypothetical cases. In only three of the 16 cases was there a unanimous agreement to impose a prison term. In one case the mean prison term decided on was 1.1 years, but the longest recommended was 15 years.
Such variability doesn’t matter when professionals are in direct competition with each other - the best decisions win out. It does matter when professionals - like lawyers, doctors, and those running university admissions - are supposed to agree.
This problem creeps into medicine, public health, economic forecasting, forensic science, creative strategy, performance review, hiring and child protection. Yet it is commonly ignored. After years of experience, professionals typically have high opinions of their own judgement, and of that of their colleagues - so the assumption is they will agree. Faced with the disparity, they are astonished.
What’s the solution? One idea is imposing strict guidelines everywhere - giving people such as judges many more rules to follow - but this meets with resistance. People in positions of authority do not like having their discretion taken away - it makes them feel like robots, they say. Then too, judgement by algorithm lowers the possibility of compassion - giving someone a second chance who does not, technically, deserve it.
The authors comes up with a plan - decision hygiene. There are six principles for organisations or individuals to take on if they want to minimise noise. First is to accept that decisions are about accuracy, not individual expression. The second is to think statistically, and take an outside view of the case. (The default mode of thinking, the authors say, is to focus on the case at hand and embed it in a causal story).
The third is to structure judgement into independent tasks - this prevents the problem of excessive coherence, where people distort information that doesn’t fit into an emerging story. Fourth, decision makers should resist premature intuitions. Fifth, they should take independent judgements from multiple judges and factor them in, and sixth, they should favour relative judgements, which tend to be less noisy.
The book is a satisfying journey through a big but not, the authors suggest, unsolvable problem, with plenty of fascinating case studies along the way.
Humans are often bad at making decisions. But we can get better.
Noise: A Flaw in Human Judgement by Daniel Kahneman, Olivier Sibony and Cass R Sunstein (HarperCollins, £25)