Tag: bias

Dealing with Life’s Decisions – (1) Blinded by Science

This piece is the first in two on the question of how to make decisions in life.

Some decisions we simply take too hastily, using cognitive shortcuts pre-programmed in our brains, and we would benefit from slowing down, thinking them through, applying a structured decision framework to them, and so on. A good example of this class of decision problem is investment. Provided you can frame the problem narrowly enough – i.e. that logically prior issues have been solved – investment decisions will be improved by thinking them through, because they will be freed from several cognitive errors which typically characterize them. This type of problem is simple and much has been written on it, so I will put it aside.

I also want to put temporarily aside the “big decisions” of life – whether and whom to marry, whether to have children and how many, and so on. These decisions will have consequences which, obviously, you cannot compute when taking them. I will reincorporate this type of decision in the second piece in this series.

In this article I want to look at decisions which are more everyday, which obviously may also have uncomputable effects (you sign up for the art class, at which you meet someone who changes your life) but which are usually designed to have more prosaic ones. For example: which classes to take at school; which sports to practice; dietary regime and so on. These are decisions in relation to which a certain amount of “evidence” exists, but where this evidence is not conclusive. Those are most decisions in life, and there is a reason why we have evolved cognitive short cuts to deal with them. It is not my intention to argue that there is nothing to be gained from applying more structured thinking to this type of decision (or other decision heuristics which go beyond “gut feeling”). What I do wish to do is to show that, almost inevitably, we think about this sort of decision in the wrong way. In short, we are so conditioned to acknowledge the supremacy of “rational” reasoning over our instincts that in fact we allow ourselves to be swayed by arguments which have the appearance of rationality but suffer from shortcomings which are so pervasive and fundamental that we would almost always do better to ignore these arguments altogether.

I am going to take an entirely typical example, of the kind we encounter many times on a daily basis, at least if we try to keep ourselves abreast of the news. Let us say we read a journalistic article, purported itself to be based on a scientific article, reporting on certain alleged health benefits of yoga. Those benefits speak to some issues or concerns we have with our own health, and so the idea has been put in our minds of giving yoga a go. Should we?

Please note that this example is just that. The media disseminates claims like this all the time. For example, we might read that playing a musical instrument is associated with higher intelligence. Or that bilingualism is negatively correlated with Alzheimer’s. Or that a diet rich in proteins results in more durable weight loss. And so on, and so on (I made all those examples up just to give the flavour of the type of truth claims we are dealing with and the problem which they pose).

Now, let us suppose that the underlying scientific study is at least correctly carried out and that the journalist has not entirely misrepresented its conclusions. Those are already two hefty assumptions which may or may not apply, but the context may give us an indication as to the confidence we can have that they indeed hold (for example, this is, ceteris paribus, more likely to be true of an article on the BBC than in the Daily Mail). What errors may we still make if, on this basis, we allow the article to modify our behavior?

A whole host.

Changing behavior has costs. There are the obvious direct costs, which may be greater or less depending on the case: in the yoga example they are likely to be fairly limited (yoga subscription, transit to the class, kit….). But then, there are also the sizeable opportunity costs. Yes, this may be a good use of my time, but is it the best use? Do I need to pre-commit resources up front?

This question cannot be answered unless you know what your priorities are: those outcomes which will make the biggest difference in your life. Ideally, that would be a pre-existing exercise. But even if you know you need to address a particular issue – say high blood pressure – and the evidence presented in the article actually shows some efficacy for the course of action in question (yoga), you can still go very wrong. By plumping for yoga, you go with the availability heuristic, which privileges the course of action you just heard of over what you might need to do more work to identify. By taking action, you lessen cognitive dissonance, and therefore the nagging feeling inside which might have prompted you to do more serious research or thinking customized to your own situation. Yoga will work on some cellular pathways, but those are certainly not the only factors involved in giving rise to your condition. There may be much more important ones, but ones which you are much more resistant to addressing – say your work, your relationships or where you live.

Even if the information is accurate, it has neither been produced, nor has it reached you, by chance. Someone decided to test a particular yoga program (which may have nothing to do with what is on offer in your locality). They did so because they have a predisposition to finding a favourable effect from yoga. But the same favourable effect might be produced in any of a number of other, unresearched ways – a problem which is particularly acute if the mechanism of action is not elucidated or hypothesized subject to a great degree of speculation. So there is a selection bias. This cannot be ruled out on the part of the media either, and if you got the article second hand, say through Facebook, your friend has also selected it in preference to others – with what reason?

In addition, the study may very well be partly or entirely attributable to the placebo effect (which is a great effect, but could be produced in other ways), with remaining variation explained by factors which yoga shares with other forms of exercise and/or other spiritual practices. The participants in the study may have self-selected, and therefore share attributes which differ from those of the population in general, and perhaps also from you. For example, imagine that those who do yoga are twice as likely to be vegetarian. Correlation is not causation: it could be their diet that explains all or part of the variation observed. You, in any case, are not Ms or Mr Average – you are older or younger, fitter or less fit. Yoga may be a fit for your other activities, or it may duplicate the benefit of them.

Now, I am certainly not saying you shouldn’t do yoga, nor that it doesn’t have benefits. I am saying that it’s almost worthless to read the article, and it may be worse than worthless to produce or distribute it. This article has in all likelihood not given you any new information at all. All it has done is make an incremental contribution to the “brand” of yoga as perceived by you. This, by itself, is not the core of the problem, however. The core is the idea you have that science should be your main tool to solve the problem you started out with. Although you have this idea, you have not in fact been scientific at all. You could have been more scientific – for example, read a book that discusses a series of approaches to your problem. That would probably have been a good idea (the article you read was not some kind of breaking news, so no need to be afraid that the book would be out of date). But even if you had done this, the problem would still have been orders of magnitude too complex for you to decide it on the basis of science alone. You not only will decide it on the basis of factors which you cannot really rationalize. But this is the only way to decide it. All the research you can possibly do is merely preparation, hopefully valuable preparation to make a better decision (there is, of course, a trade-off with the time you invested), but it will never provide an algorithm which decides on your behalf. Some people will choose to view this irreducible subjectivity as a lamentable concession to human nature. But, as I will explore in part two, it seems to me that all of the alchemy which turns research into outcomes is there, in the giddying sovereignty of the moment of decision.

Science, just like the mind, is a tool; something else – you – must be in the driving seat. Positivism is unscientific. Science makes a contribution, and yet if you have the belief that your decisions should be guided by science, it is very likely that, in combination with cognitive and selective biases, you in fact are led into decisions which are worse than those you would have made had you not had this belief at all.

Between the idea
And the reality
Between the motion
And the act…
Between the conception
And the creation
Between the emotion
And the response…

Between the desire
And the spasm
Between the potency
And the existence
Between the essence
And the descent
Falls the Shadow

T.S. Eliot, The Hollow Men

The twin errors of “Sex at Dusk”

For those who do not know it, “Sex at Dusk” is a book by Lynn Saxon which purports to “debunk” Sex at Dawn. Everyone seems to be agreed that Saxon has a lot of science at her fingertips and that this unreadable book nevertheless makes a number of valid points.

What people are missing in this debate is, however, fundamental. It seems to me that Saxon commits an error of method, epistemological in nature, and an error of genre.

The question in which the readers of Sex at Dawn are interested is what is the most plausible account of evolved human sexuality given the balance of the evidence. This question is not answered by pointing out errors and misinterpretations in the book. Even if couched in scientific terms, this comes pretty close to an ad hominem attack. On the contrary, given all the obvious societal interests vested in the “standard narrative”, it is the scientific underpinning of that narrative regarding which we should be particularly attentive and skeptical. This is clear from Kuhn, Popper, Bourdieu, and any number of other philosophers and sociologists of science. If you have to choose a null hypothesis, it would be better to go with S@D and not with the standard narrative because this would at least counterbalance to some extent almost everyone’s internalized biases. This would be good and correct scientific method, just as it is the church that should have had to prove the terracentric view of the universe and not Galileo the heliocentric one, once 51% of the evidence was on Galileo’s side. What evidence is there that S@D’s conclusions are wrong, not just that the authors made some errors getting there? The conclusions can only be wrong (in the normal sense of this word within the social sciences) if there are other conclusions which are more plausible. What are these conclusions? A vision of a more “polyamorous” ancestral environment is not scientifically suspect just because the Church Fathers have indoctrinated us with the idea that it is morally suspect. This is an epistemological error which I personally find inexcusable on the part of a contemporary social scientist.

This brings me to the error of genre. S@D is a work of popular science. The intention of the authors is demonstrably to affect the terms of the broader, and hugely important, social debates which sexuality feeds into. They do this by presenting science, but they are allowed, and even required, to be selective given how biased much of the “evidence” is. They are even allowed (thank you, Seneca) to make mistakes. It is a book with an agenda (as all books have an agenda, for, as Derrida famously observed, “there is no text without context”). That is why Saxon’s book can only be an ad hominem attack. Her choice of method condemns her to this. If Saxon believes the balance of evidence points towards monogamy then that is the book she should write. This would, however, be surprising as I think one of the criticisms one might make of S@D is that the “standard narrative” is not actually a standard narrative from a scientific perspective. What it is, is a socially standard narrative (something very different) which is a hidden bias in much scientific writing, especially the more distant the theme of that writing is from actually investigation into evolved human sexuality. Saxon’s book, less excusably given its subject matter, inadvertently proves the point. Barash and Lipton (The Myth of Monogamy, 2001) do the same when they show that we are not naturally monogamous and then claim that we “should” be anyway.

The point is that what we “should” be is up for grabs. We no longer have to take Plato’s word on it. This emancipation from the patriarchal bias in classical moral thought around sexuality, so brilliantly analyzed by Foucault (The Use of Pleasure, 1984), is what S@D sought to achieve, and what it has achieved. S@D has been justly successful in reaching its goals because it is engaging, humanistic, humorous, optimistic, and entertaining. This is how you change the world, if you are courageous enough not merely to analyze it, and particularly in ways that have an unrecognized bias towards the status quo. Wanting to change the world is not illegitimate and the fact that so many in the scientific community seem to think it is shows, I think, something of the power relations between vested social interests and the scientific establishment.

I am not saying that S@D is the last word on the subject, or even that it is a Copernican moment (and the authors are very quick to disown such an idea, as witnessed by Chris Ryan’s comment on one of my earlier articles). There are elements in the conclusions which I myself have argued are incomplete. In a way, in places it’s an engaging caricature. Perhaps this is a moral failing on my part, but I find it hard to be appalled by that. I think what it nevertheless is, is a brilliant popularization of the relevant science combined with true wisdom and compassion for the human condition. This makes it, as I think its short history has shown, a defining moment in the Kuhnian process (which is to be interpreted in a post-structuralist sense given the nature of “truth” in the social sciences(*)) by which one scientific paradigm is replaced by another. The authors achieve this by undermining the forces which maintain the status quo. They manage to dissipate some of the fear inculcated in us by established social discourses according to which we have to hang on for dear life to the disintegrating institution of monogamy because of the imagined catastrophic social consequences of giving it up; rather, we can trust our biology and imagine better ways of ordering our affairs than those which served Roman and later European militaristic expansion so well, and therefore survived that “evolutionary” race, but perhaps are not relevant to life on the planet in the 21st century.

Note

* Cf V. Romania (2013), Pragmatist Epistemology and the Post-Structural Turn of the Social Sciences, in Philosophy Today, Summer 2013 (link).

PS: Before anyone is tempted to conclude anything from the ratings of Saxon’s book on Amazon, do recall selection bias and confirmation bias…

Tricks of the mind

I have just finished reading Daniel Kahneman‘s book Thinking, Fast and Slow. Kahneman is one of the fathers of behavioral economics, having won the Nobel prize in 2002. Thinking, Fast and Slow summarizes much of his work and the state of knowledge on cognitive biases in decision making. There’s a short summary of the book over on Wikipedia. But I encourage you to read the whole thing, which is a treasure trove of insight into how the mind works.

In recent years, Kahneman has been particularly interested in hedonics, i.e. the study of happiness. Experiments show that traditional utility theory cannot be squared with our evaluation of pleasure and pain. Objectively greater levels of pain may be evaluated and recalled as less painful if either the peak of pain is lower or the final parts of the experience are less painful than the experience as a whole. Conversely, the pleasure in an experience can be ruined by less pleasurable moments at the end of it; nonetheless, objectively the pleasure has been had. Kahneman talks of two selves: the experiencing self versus the remembering self. Yet even if the remembering self makes expensive errors, Kahneman cannot dismiss its judgment entirely. Memories are the afterlife of experience, and they matter to happiness also.

Kahneman’s work is of tremendous practical importance, but it raises some theoretical issues which he does not discuss in the book. Specifically, he appears to simply assume that the basis for all the behavior he describes is biological, and ipso facto it is universal. This is doubtless true in part, but it is unlikely that it is true of everything that he discusses. Indeed, the mind can be trained, showing that there is a sociocultural dimension to observed biases and that they are path dependent.

While some errors can be given a plausible evolutionary etiology (but Kahneman warns us to avoid the seductiveness of stories), others, especially the misevaluation of happiness, raise questions as to how the mind has been endemically conditioned by society. Does the remembering self simply get it wrong, or is the experiencing self not really present to its experiences? Our learnt attitudes to experience seem to get in the way.

Ultimately, Kahneman provides a lot of evidence that accidents of life have no long term impact on happiness. Happiness is not obviously generated by what we experience in (adult) life at all, but rather by how we experience it – which has largely been settled by the time we get there. If we experienced life in fullness, we would not be prone to imagining that environmental changes – in job, place, relationship – would generate massive shifts in our well-being. This illusion is a direct result of the lack of well-being we experience because of the bound state of our libido. Given this fact, if we are aiming at gross national happiness, avoidance of disturbances in childhood to our ability to abandon to the natural flow of life should be an overarching priority of public policy.