Review of Noise: A Flaw in Human Judgement


Another Kahneman book felt needed in my life so I picked up Noise. My previous experience with Kahneman was from his Thinking Fast and Slow which I thoroughly enjoyed. The whole idea of humans consistent lack of consistence is a fun idea that is surprisingly little talked about in my psychology classes. In general I've been very disappointed with my academic pursuits of the field of psychology. I guess I am just surprised at what problems people are fixated on. Some are still trying to argue that psychology isn't a pseudo-science which feels almost like these psychologists want to be martyrs? Focused on the past and the absolute basics, psychology in university at the moment feels strongly incorrect. No wonder people thought this was a pseudo science if the stuff taught by Kahneman and Cialdini (and Cialdini still feels a bit aged to me). 
    Anyways, onto the book. The main idea of noise is to shine a large spotlight on a massive source of error in judgment- which is noise. You've probably heard of the accuracy versus precision explanation in some lab class of yours but I always forget so I'll explain it real quick with the classic dartboard analogy. An accurate and precise thrower would land all darts within a small area- like an inch, around the bullseye. A precise but inaccurate shooter would still land all darts within a small area but not around the bullseye. This is either called systematic error or bias and is both obvious to spot and obvious how to fix usually (if you're consistently landing 2 inches above the bullseye, aim for 2 inches below it). This is not what the book is here to address. Instead it focuses on the imprecise shooters (accuracy can be good or poor). Who's shots are not consistent and seem to have lots of random error aka- noise. Noise is less attractive and less obvious than bias and not as obvious to fix. This book sets about methods and tools to fix noise in judgment which has serious implications, from two prisoners committed of the same crime getting 30 years difference of prison based on something out of their control (who the judge is or their mood or other apparently random and uncontrollable factors). The main guidelines and implementations of the book are more directed at organizations than individuals but they are still applicable. I personally am a big fan of algorithims in judgement protocols in medicine, although have a few reservations about particular ones. 
    Similar to how Thinking Fast and Slow is most effective (for me) in its application of theme rather than rigorous study (which could also be very useful, I am just not willing to put in the time)- I think the general theme and motto of this book "Wherever there is judgement, there is noise" is the message the authors wanted to get across the most and I think it does carry an effect in me. In general the healthiest thing you can do when making judgments consistently (or even once off) is to fully analyze the different factors. Because if you can work out how to become precise or less noisy, then fixing the bias or system error is much easier and usually carries significant rewards. 
    A large theme in this book is the superiority of judgment by algorithms and machine learning, and the impossibility of predicting the future. Reducing confidence and adding weight to multidimensional aspects in our judgements, and turning them into simple algorithm. If possible, implement machine learning. This will reduce time and significantly increase ability to manipulate goals and rewards. This could run into resistance in the medical community but results speak for themselves and I for one am excited for our new robot overlords.
    Okay re-saying what the book talked about is boring. I love Khaneman and I like this book. I think keeping books like in my frequent reading helps a lot in keeping my mindset more logical (or at least suspicious of over-coherance and overconfidence). I might not implement the strategies that much in my life but I suspect the writers realized the largest benefit would be to instill the reader with noise-senses at best so they can have a slightly better ability of identifying a particularly noisy scenario. If so- job well done!


Popular Posts