I have just read a complementary copy of the Brafman brothers' (Ori and Rom) book titled Sway: The Irresistible Pull of Irrational Behavior. One of the review's cited on Amazon.com said:
"If you think you know how you think, you'd better think again!"
The human brain evolved, and the product of that evolution is truly wonderful. One of the wonders is consciousness, and we can consciously analyze not only the causes of our decisions, but also our decision making processes. Freud revealed to us that there is also an subconscious, and that not all aspects of our decision making are available to our conscious analysis.
Indeed, we are evolved to be social animals, capable not only of advancing our own individual interests but also of altruism. We live socially, and our big brains have allowed us to create and elaborate cultures which in turn affect the ways in which we think and make decisions. Our brains evolved in a dog eat dog (or rather predator eats man) world, in which sometimes quick response to a threat is a better survival tactic than careful analysis of the threat and its appropriate response. Moreover, evolution leaves appendices which seem purposeless,
Many years ago, when I was a student, there was a team at UCLA that was trying to figure out what the SWAC computer did when a sequence of data was fed into its control register that did not correspond to any of the commands which it had been designed to implement. Evolution too can leave the organism with responses that were never positively or negatively selected, due to the fact that their triggers had never occurred naturally. In the millennia of prehistoric human development there were few chances to select for organisms that were more or less successful when living a sedentary life with an abundance of rich foods.
Modern psychology is revealing that there are many subconscious biases that exist in human thinking that affect the decisions we make. The Brofmans have done a great job in providing a very readable, short introduction to this research.
So what?
Danial Kahneman, a Nobel Prize winner for his efforts to inform Economics with Psychology, is supposed to have said that the most important defect in decision making is arrogance. People think that they think better than they do in fact think.
The Brafman book presents many examples of people suffering real damage because they fell victim to the failures of their own decision processes. They cite an airliner crash due to such a lapse by an experienced pilot who was responsible for safety training for his airline, as well as a major financial loss made by an investor who had become a multimillionaire due to wise investments, and (of course) the White House decisions that lead to the horrors of Iraq.
If the book does nothing else, it will perform a great service in helping people to realize that their important decisions may not be made as rationally as they appear to the decision makers, and in cautioning people to reexamine the bases for their risky decisions.
The book fortunately does more by identifying a number of specific biases that are known to color decision making and by suggesting strategies to protect against such biases.
Bayes Theorem
I have one specific complaint about the analysis. The brothers note that the frequency of diagnosis of bipolar disorder has increased, while doubting that the incidence or prevalence of the disorder has gone up. The ascribe this increase to the change in diagnostic criteria for the disease.
Classification is fundamental to science. Unfortunately, it is not always possible to make useful classifications, at least in the early stages of science. In medicine it is nice to have utilitarian classification schemes. If a person has this disease, then he will with high probability get better with this treatment. As we enter the time of personalized medicine, physicians will have more accurate classifications for this purpose. Thus in the future, physicians will increasingly use rules such as if a person with this genetic profile has this condition, then he will benefit from this treatment regime. But the fact that we will have better classifications in the future does not mean that the current classifications are not useful now. We use the best we have, as long as they are better than nothing.
Similarly, the human animal uses classifications whether or not they are diagnotically instrumental. A farmer thinks of Bossie as a cow, and of cows as animals whether or not that classification helps get Bossie to give more milk. He thinks of the north 40 as planted in corn, which he recognizes as a crop, even though those classifications may not help him increase his crop's yield. So too, it may be useful for practitioners to think of people as having bipolar disorder, even were that classification not to improve the expectation of outcomes of their therapy. Sometimes you just need shorthand to be able to discuss something.
I would also note that Bayes theorem says the a posteriori probability of a diagnosis is equal to the a priori probability of the diagnosis times an odds multiplier (determined by the symptomology). Thus a patient presenting with fever in a hypermalarial zone is much more likely to be diagnosed with malaria than would be a patient with the same degree of fever in the Arctic. It is not necessarily an error in decision making to give different diagnoses with the same symptoms if the a priori probabilities of disease are in fact different.
More to the point, one might expect that as a higher proportion of the cases of a disease came to be recognized in a population, then the frequency of diagnosis of that disease in the population should also increase. The trend should stabilize when the frequency of diagnosis actually approximates the incidence of the disease. As any disease is discovered and characterized by the medical profession, its apparent prevalence increases.
The Bottom Line
I was in the midst of reading another couple of books when Sway arrived, and I found myself putting the others aside and reading Sway avidly. The discussion of the use of Crew Resource Management (CRM) techniques for improving decision making within teams dealing with emergency situations would alone be worth the effort of reading the book. Indeed, I am thinking of going through the book again soon in order to make more careful notes on the specific decision making biases that are identified.
My thanks to the authors and editors!
No comments:
Post a Comment