 |
PRACTICAL PHILOSOPHYTHE JOURNAL OF THE SOCIETY FOR PHILOSOPHY IN PRACTICEwww.practical-philosophy.org.uk www.society-for-philosophy-in-practice.org |
|
|
|
PRACTICAL PHILOSOPHYTHE JOURNAL OF THE SOCIETY FOR PHILOSOPHY IN PRACTICEwww.practical-philosophy.org.uk www.society-for-philosophy-in-practice.org |
|
Rational Choice in an Uncertain World: the psychology of judgment and decision making
Reid Hastie and Robyn M. Dawes
2001 London: Sage Publications,
pp. 372.ISBN 0-7619-2275-X (pb) £35.00
Practical Philosophy (Book Reviews) Autumn 2002 Volume 4.2
Reviewed by: David Arnaud
This book, written by two prominent decision researchers, is very much in the mainstream of academic psychological research into decision-making. The reader will find in here plenty of clear explanations of probability theory, normative and descriptive decision theory, decision trees and the role of heuristics in forming judgements. As is common within this mainstream tradition, they believe we are often poor at making wise choices in an uncertain world, resisting the recent movement by the naturalistic decision-making lobby to seeing humans as naturally good decision-makers. An unlikely claim, they believe, as evolution only equips its success stories with being fitter, rather than fittest. Moreover, even if evolution instilled our ancestors with cognitive mechanisms that enabled them to make wise decisions in our environment of evolutionary adaptedness, we now face the kind of choices that were not faced by our evolutionary forebears who didn’t have to choose one college/course/car/profession/house rather than another, or between working for another ten years and retiring, or to decide if children or a career would make us happier, or if a hip replacement is a good or bad idea. Decision-making is a skill that can, and should, be learned, the authors claim. It’s not an evolutionary product: we are born with it and we need it.
So, does this book help us to acquire it? Yes, partly, is the answer, I think, and I will try to illustrate some of what it captures and some if what it misses. Part of decision-making in an uncertain world is judging our beliefs about unknown aspects of the world for their accuracy. This means, for the authors, assessing the probability of the belief being true. This, they suggest, can be understood within the framework of the lens model developed by Egon Brunswick. Suppose you see a man and wish to judge his age. The acceptability of this judgement depends upon how well we use the ‘cues’ that are available to us such as the colour of his hair, the wrinkles on his face, his gait and so on. Ideal judgement, they claim, is based upon summing up how well each of the cues, weighted through regression analysis, allows us to probabilistically predict the man’s age. The fallibility of human judgement can be tested by seeing how well humans live up to this ideal. This method of finding out about the world seems to introduce a rather odd kind of epistemological scepticism as we can have justified claims about the world only in cases where there is enough regularity already in the world to build up a statistical picture of how well cues predict aspects of the world. If we accept this epistemological model, we are not going to be able to make any justified belief claim, however provisional and probabilistic, about the many uncertain aspects of our lives for which there is not a relevant history of frequencies. Whether or not we should accept this picture in total (and it must be noted that it provides no room for critically assessing arguments or identifying likely causal mechanisms based upon our general understanding of relevant aspects of the world), the authors illustrate how our failure to take into account a sufficient number of cues and their probabilistic weighting leads to many mistakes in judgement. In one example, of many presented, experienced clinical psychologists, asked to judge how likely newly admitted psychiatric patients were to be violent, only managed to produce a correlation of +.12 to the patients’ actual behaviour. A statistical model, given the same cues, managed a correlation of +.82.
Why are we prone to these mistakes? The answer, the authors argue, can be identified by analysing our reliance upon our ‘toolbox’ of cognitive heuristics. These heuristics are thinking strategies we bring into play when we make judgements. Perhaps the most researched of these is the availability heuristic. In our environment of evolutionary adaptedness we would have had to judge the frequency of important events, such as the number of predators at a water hole, or the occurrence of types of plants in a particular area. Evolution has equipped us with the availability heuristic for making these kinds of judgements. We judge likelihood, so the theory goes, by judging how easily instances come to mind; if we can easily think only of a few cases we judge it unlikely; if many, likely. While this is in some cases an adequate, get by, mechanism (fitter rather than fittest) it can also mislead. For example, how is a therapist to judge the seriousness of a client’s threat to commit suicide? If the availability heuristic is used, the therapist searches memory for similar clients who have made similar threats – however those cases likely to come to mind are those in which the client did attempt suicide (these are more easily remembered), so the therapist will judge the likelihood of suicide to be much higher than it actually is.
The other heuristics that the authors identify are anchoring and under-adjustment, judging by similarity and using scenarios. Why do businesses fail to underestimate the completion times of projects? Their initial estimates provide an anchor against which they fail to sufficiently adjust when new evidence becomes known. Why do we think that a college student who writes poetry and wins prizes in calligraphy is more likely to be an art student than a psychology student? Because this person fits our image of what an art student is like, even though, given that vastly more students study psychology than art, many psychology students are likely to have artistic interests (perhaps even more than the total number of arts students). Why do successful trial lawyers seek to tell a good story to the jurors? Because we store much of our knowledge in story form and judge truth by the overall coherence of a story, rejecting or accepting evidence depending upon whether it fits our story.
The authors do not discuss critical thinking but it is clear that both this literature and the analysis of heuristics are concerned with the errors we make in our judgement. One difference is that the heuristics literature seeks to examine the psychological mechanisms that cause errors while the critical thinking literature seeks to categorise errors based on their formal structure. For example, the similarity heuristic might be the causal mechanism underpinning the fallacies of poor analogical reasoning and the neglect of base rates, and the availability heuristic the fallacies of overconfidence and the failure to consider counter-evidence.
This raises an interesting speculation. The fallacies arm of the critical thinking literature recommends learning to identify, and inoculate ourselves against, a wide range of mistaken patterns of reasoning. For the practical philosopher, is a quick route to improving reasoning instead to help clients identify the heuristics behind these fallacies? To get a reasonable answer to this question we would have to see how many fallacies can be accounted for in terms of these heuristics (many would not, I suspect), and, empirically, whether teaching people about the heuristics enables people to spot and correct these fallacies. A difficulty is that the heuristics might lead to well-justified judgement when, for instance, the ease with which we can think of cases really does map onto the actual frequency of cases, our first thoughts really are correct, or the similarity between two objects in the attributes we are aware of really is truly indicative of similarity of attributes of which we are unaware. This suggests that making people aware of the heuristics they use cannot be a replacement for an assessment of why fallacies are fallacious. Being aware of these heuristics might though be itself a useful heuristic. To see how this works, imagine we are making a judgement about the frequency of the occurrence of an event, such as whether car dealers are swindlers or smoking is likely to lead to cancer. We need to remember that we might be tempted to judge this on the basis of ease of recall, and ask ourselves if this case is one where ease of recall is likely to lead to probability errors, and whether these probability errors would have serious implications. If the answer to both these questions is yes, we need to use strategies more likely to produce a realistic judgement. On this note, new to me, was the valuable advice about how to understand and illuminate statistical information using Venn Diagrams and frequency trees.
Besides gaining a reasonably accurate, probabilistic, understanding of what the world is like, decision-making requires working out what we value. Three examples of common reasoning failures about our happiness offered are that we tend to think our desires in the future will be more varied than they turn out to be, that when it comes to predicting our own future happiness (by which, more or less, the authors mean pleasure) we are poor judges as we tend to predict big changes to life’s events whereas in fact we fluctuate around a hedonic ‘set-point’ (for example, junior faculty members incorrectly thought they would be elated or devastated when they found out whether they would receive a tenured position), and we overestimate our ability to be self-controlled in the face of temptation. The advice they drew from the illusion of ‘hedonic certainty’ was that in our decision-making we should be less concerned with focusing on global assessments of how happy something will make us feel, but rather focus on more specific attributes such as our health, variety of experience, productivity, opportunities to help others and so on. Nothing was offered though on how we might generate or assess the desirability of these kinds of values – an omission made less surprising when one reflects that this is a philosophical rather than a psychological question.
Finally, the authors reflect upon what philosophical attitudes we should take given the uncertainty of the world. ‘People abhor uncertainty,’ they write (p. 313), and we turn to religion, astrology, or tarot cards to try and make sense of life’s underlying uncertainties, to convince ourselves there is a hidden causal order that we can control, and to explain how the world (against the run of the evidence) is really just, rather than morally random. While the chapter is entitled, ‘In Praise of Uncertainty’ a main concern is rather praising the acceptance of uncertainty. Without this we are easily led down morally suspect roads. ‘Why was there a plague?’, ‘The Jews are poisoning our wells.’, ‘Why am I a success?’, ‘I must be an exceptional person to deserve it.’, ‘Why did I suffer great misfortune?’, ‘I did something terrible to bring it about.’ Counsellors, influenced by research into locus of control (people vary in believing whether events are within or outside their control) falsely assume that believing we have control over our fate is better for our mental health, whereas a shrug of the shoulders and an acceptance of chance might well be better for us. Instead of assuming false certainty, we should strive, using probability theory, to make the best choices we can in an uncertain world and remember the wisdom in Ecclesiastes (9:11, quoted on p. 313): ‘The race is not to the swift, nor the battle to the strong, neither yet bread to the wise, nor yet riches to men of understanding, nor yet favour to men of skill; but time and chance happeneth to them all.’
|
|
|