Without a nod to context and power there is just noise
Reviewed in the United States on June 5, 2021
We all know that people, including the experts, are poor at predicting the future and that every judgment is potentially full of error and bias. This book attempts to explain why and what can be done about it.
“Noise is the unwanted variability of judgments, and there is too much of it.” I personally think “static” would be the better descriptor but terminology is clearly an author’s prerogative. And there is no question that there is a lot of error in the judicial system, medical diagnosis, insurance underwriting, the education system, the recruitment, evaluation, and advancement of employees, forensic science, and just about everything to do with business and political decision-making.
But to think anyone can wrap all of that up with one neat bow and in doing so provide a potential roadmap to address them all is attempting to map the intricacies of the universe on the back of a cocktail napkin. You inevitably end up with so much clinical jargon (e.g., decision hygiene, mediating assessments protocol) as to make it all, well, outrageously noisy—and unnecessarily, if not painfully, long.
These three individuals are clearly brilliant and well schooled on the topic. Their academic peers may well give the book rave reviews. That’s okay. That’s the way judgment works.
As I am not a peer, however, I took exception with some of the assumptions. Judges are notoriously inconsistent in setting prison sentences. But setting a prison sentence is not the same as diagnosing breast cancer. One is clearly a judgment. The other seems better characterized as analytical.
Perhaps more importantly, setting prison sentences and medical diagnosis are largely judgments made by individuals. When judgments are made within an organization, however, as in deciding who gets promoted in a corporation, the individual is dwarfed by the organizational power structure that the authors never fully address. It is often power, not noise, which compromises the judgment of most organizations.
I believe, to be fair, that they would argue that power and noise are interrelated. Fair enough. It is the power, however, that defines the culture, the processes, and the judgments of most organizations. And if the power structure is faulty, which it often is, the reduction of noise won’t matter much and probably won’t get addressed anyway.
There is, to be fair, a lot of good common sense here. It is true that “correlation does not imply causation” and the “wisdom of crowds”, probabilistically anyway, is pretty well established. And I could not agree more that “even the most enthusiastic proponents of AI agree that algorithms are not, and will not soon be, a universal substitute for human judgment.”
In the last third of the book or so the authors acknowledge that some believe that more noise (i.e. fewer rules, guidelines, etc.), not less, would lead to more equitable and fair outcomes in many social systems (e.g., assigning prison sentences). They acknowledge that it ultimately depends on the process but they generally come down on the side of less noise.
If you put the issue in a broader human context, however, I don’t believe the case is convincing. If we lived in a society defined by collective identity and a sense of mutual obligation I would probably agree. But we don’t and in that sense the idea of reducing noise may be counter-intuitively untimely.
Sometimes the reduction of noise through rules or standards can actually increase inequity and reduce fairness. The US Tax Code is a case in point. The noise has, over decades, been reduced to a dull hum. We have the lengthiest and most complex tax code on the planet. But is it fairer as a result?
No. By reducing the noise we have given a distinct advantage to those with the wealth to hire armies of lawyers, accountants, and lobbyists to find and exploit the noise that’s left behind because one byproduct of rules and guidelines is loopholes.
The procedures for qualifying to vote and certifying elections are another example. More rules may provide further protection against fraud. But there is a very good chance it will also prohibit people with the legitimate right to vote from doing so or even compromising the very ideal of democracy. Which is the worse outcome?
Most social systems, in other words, exist in a larger context. If there exists a strong sense of collective social identity and mutual obligation reducing the noise in most social systems might make sense. That, however, is not the world we currently live in. And until that context changes, more rules and guidelines are as likely to result in less equity rather than more.
23 people found this helpful