NOISE: Deciphering the New Flagship Book in Behavioral Sciences

Daniel Kahneman, Olivier Sibony and Cass Sunstein, three lovers of cognitive functioning who need no introduction, are therefore the co-authors of the new book: Noise - A Flaw in Human Judgment, published this year in English (to be published soon in French).

In this book, they discuss with pedagogy the importance of noise in explaining the inconsistencies in our judgments and raise the associated ethical or credibility problems. I suggest summarizing three key ideas here: first, cognitive biases produce statistical biases; second, noise is inevitable in our judgments; and finally, the interesting and useful idea that it is possible to reduce noise.

Idea 1: Cognitive biases produce statistical biases

Although they are fascinating, I will refrain from presenting to you yet again what are Cognitive biases. I Will Only Highlight Their Character Systematic. Indeed, decades of research in behavioral sciences have shown that, under certain conditions, this type of inconsistency in our reasoning almost inevitably occurs. By following the Experimental method, these studies were able to demonstrate that we all have the same cognitive deviations (where rational reasoning would be expected). What does that mean statistically?

Daniel Kahneman, Olivier Sibony and Cass Sunstein schematize this phenomenon in Figure C where a statistical bias is observed. In fact, we notice that all the shots from the target in question (i) are not in the center of the target, (ii) but rather grouped together in the lower right quarter. One could therefore assume that the rifle used to fire has a defect that causes a certain behavior - i.e. in all shooters. In this context, the error (Mistake, relative to the center of the target) is shared, making the error (Error, in the statistical sense of the term) directional. Because of their systematicity, biases are relatively Foreseeable, it is then possible to anticipate them in order to correct the error and thus gain in precision, using the method of Nudge.

Idea 2: 'noise' is unavoidable when making a judgment

Before addressing this point, I suggest that you illustrate the concept of noise with an example that speaks to everyone. Every morning, I plan to wake up at 7:00am to start my day. Sometimes I manage to get out of bed at 6:45am while other days I hang out until 7:10. Recently it was 6:57am and 7:08am the next day. This Variability, it's noise: the number of minutes between the actual sunrise time and the planned sunrise time does not really follow a trend (-15 min, +10 min, -3min, +8min). It is indeed an unpredictable fluctuation. Statistically, we can note an error compared to what is expected (sunrise at 7:00). This error is represented schematically by the three co-authors in FIG. B. Shots can be seen on the whole target: it is therefore not an error caused by a bias.

The catch with statistical noise is that it's much less visible than bias. Indeed, if you average all my alarm data, you might think that I am perfectly fulfilling my objective (since on average, I get out of bed at 7:00am). What should be kept in mind is that in a data set, one should expect to continuously observe noise (or even noise AND bias, as shown in Figure D).

However, noise can be a problem in our Judgments. Judging is measuring information that presents a certain Uncertainty. In their book, Daniel Kahneman, Olivier Sibony and Cass Sunstein are specifically interested in so-called “professional” judgments, for which a certain consensus is envisaged (i.e. where variability is not desired). For example, it is expected that when faced with the same case, two doctors will make the same diagnosis; in the same way, it is expected that a judge's verdict -for a given case- will not vary according to the time of the decision. And yet...! The authors present various studies that highlight the importance of noise, regardless of the sector (medicine, justice, insurance, recruitment...). Although it is not very visible, noise is everywhere.

‍

Idea 3: It is possible to reduce noise

As humans, we judge a situation based on our Singularity (taking into account -often unconsciously- our history, our own experiences, and also our own cognitive biases at work at the time of judgment). This singularity creates noise. This is what the authors of Noise call the “noise pattern,” which can be seen as the signature of our uniqueness. In other words, we are unique and that certainly causes noise in our judgments. The problem is that very often, we don't see it... Worse: several large-scale data analyses have revealed our tendency to minimize it (we expect reasonable disagreement between different judgments, which can, in fact, be five times greater than we imagine). For this reason, the authors suggest that organizations carry out a” Noise Audit ”, i.e. an evaluation at the organizational level of the importance of noise in judgments.

More generally, the authors propose to practice a Decision hygiene. It is the idea that it is possible to make our judgments more robust, i.e. more accurate, by creating favorable conditions for decision-making. For example, cutting up the mass of information available to us to avoid feeling like we're judging the whole thing. In reality, it is a skill we are incapable of: we will naturally tend to focus on information that supports us and confirms our hypotheses. However, our sense is unique to us, by definition, and is therefore very sensitive to noise!). Also, aggregating several judgments collected independently (to avoid the influence of one, the other or even the group on the judgment) is a method for preventing noise. For example, instead of asking your employees about a process during a meeting, instead ask for their opinion individually.

___

One thing is certain, it is impossible to eliminate all the error of (professional) judgments. Yes, it's human nature. We are not machines that make their “judgments” by strictly following an algorithm and where error is technically impossible (we can safely think that an Excel formula will never be wrong, so we expect zero variability). Let's accept that our uniqueness makes us poor judges and prevent noise by adopting good decision hygiene!