Review of Thinking Fast and Slow

 

Another recommendation to me by my friend Jake who's taste I had come to greatly appreciate. This is the first psychology book I had ever read and it served me greatly for many books to come afterword (such as Buddhism is True). I was also unaware when reading it (about a year ago now) what a big name Daniel Kahneman is. The central idea of this book and is that all judgement and calculations done by our brain- most "thinking" that is, is executed by two wholly different systems. System 1 is fast thinking and is like a circus performer able to juggle an almost unlimited number of tasks with the catch that these tasks must be small and familiar. System 2 is the slow and methodical decision maker who can take many different inputs of information to make a judgment but it can handle pretty much only one task at a time. Driving on a familiar road is system 1. Doing calculus is system 2 (for most people). They both come assigned with their own flaws. The errors that arise from heuristics are often due to the flaws in either these systems or in our assignment of a problem to each system. An evolutionary psychologist that these heuristics make sense and are not flaws or a blindspot in our thinking but intended through evolution. The person who makes a snappier judgement thats right 80% of the time will do better than their friend who takes triple the time to make a judgement that is right 85% of the time. The book will use this fast and slow thinking as a lens to examine the other two segments which deal with heuristics then sources of overconfidence in what the mind thinks it know. The heuristics are interesting and there's tools to get around them excellently provided in the book. However the hard part seems to be to be able to spot them. The third part of the book though, I was enthralled. Daniel Kahneman many times through-out this book implemented little "tests" for the reader to take to exemplify a heuristic or bias he was explaining. I "failed" (or maybe succeeded?) very many of these tests. The discussions of loss-aversion, experienced well being, and two-selves all fueled Kahneman's theme of "What you see is all there is." The largest sources of error consistently in this book come from overvaluing recent events. In general over the course of this book and some others of psychology that I've read I cannot help but come to the feeling of failure in our ability to think statistically. It's hard. It's really REALLY hard to think of situations purely from a stance of statistical probabilities yet we've designed a world that will functions and can easily be improved by such thinking. I catch myself now a year later thinking if I had come to a conclusion only because I am overlooking a crucial self bias or simplifying the situation in order to make it more digestible. If you go on any app store there are a plethora of algorithms made for providers to calculate any score. These algorithms have proven time and time again to be better than human judgement. These algorithms themselves have problems. As a CNA I get a lot of patient complaints about bed and chair alarms. We stick these things on everybody. I have done what now in retrospect seems too much reading on fall statistics looking for a simple solution to the problem. There have been many implementations over the course of my 3 years as a CNA at the same hospital of fall score algorithms which assesses who is high and who is low risk for falls. I even became a part of the fall team and attached reducing patient falls to my goals that I would achieve for my yearly 3% (less than 25 cents) raise. I placed bed alarms on all my high risk patients- nailed my hourly documentation, made sure everyone was aware what patients were high risk for falls when I went to lunch. I was a machine. At the next evaluation for my raise, my total falls that year had tripled from my first year. I had 0 falls happen while I was with the patient. Every single one of my falls was unwitnessed by me but also they were uncorrelated to staffing ratios. The patients mostly all had fall alarms. They were all high fall risk but then because our algorithm is designed to have 25 different factors, 90% of patients are high fall risks. It is not only impossible for  staff to walk every single patient any time they need to do anything out of bed- but that is the single possible reduction of falls I could find. If I wasn't there- they fell. I have come to the conclusion that falls only happen during one possible time: when patients are walking. If patients were all 100% reliant on staff to walk though, hospitals would cease to function at a nursing level. After that year I didn't know what to think and for the last 11 months I have been incredibly more relaxed in how I allow my patients to walk. I have had 0 falls this year is the most baffling thing. I do not know if because I am more relaxed- more patients walk and become more steady or just luck or what. But the important lesson this book has taught me is to what statistics and who to listen to and how to ignore. It taught me how to think and to know that luck plays in most things- an absolutely massive role. We cannot know everything, so if you ever think you know enough to make a prediction with over 90% accuracy- you are making a mistake somewhere. However we must make decisions anyways without all the knowledge, without a great ability to predict. This is the core fear of medicine I believe.