Total Pageviews

Wednesday, February 10, 2016

Statistics and Psychoanalysis I – Why does the story matter so much?



Among the most powerful relationships in health is the relationship between smoking and lung cancer.  The studies vary, but there is about a .40 magnitude correlation between smoking and lung cancer.  If that number were 1.0, every smoker would get lung cancer.  If that number were 0, someone who smoked would be no more likely to get lung cancer than someone who did not. A correlation of .40 is a huge number.  It is big enough to win gigantic lawsuits and to require warnings to be put on every pack of cigarettes.  It is big enough to scare an entire society to change its ways, so that we no longer smoke in restaurants and in college classrooms – indeed most buildings that we enter are smoke free and the number of smokers has plummeted.

What does a number of that magnitude really mean, though?  Well, there is a trick that statisticians use to interpret correlations.  They square them.  That’s right, they multiply them by themselves in order to determine the “percentage of variance accounted for in the second variable by the first variable."  So, if we square .40, (4 X 4 = sixteen and move the decimal to the right place) we get the number .16.  So, 16 per cent of the variance in lung cancer is accounted for by smoking.  That means that eighty four percent of the variance in lung cancer is accounted for by something else.  What else?  No single factor.  Smoking is the best predictor of lung cancer – so many, many factors – hundreds or even thousands of factors are related to lung cancer in addition to smoking

On some level this makes sense.  We’ve all heard stories of someone who has lived to 100 years of age and, when asked the secret of their longevity, they’ve said that they smoke a pack of Lucky Strikes every day.  And we’ve heard stories of people who have never smoked in their lives and yet they have contracted lung cancer.  The first person likely had many, many of the environmental or constitutional factors that prevent cancer on their side, and the second person apparently had other factors, none of them as strong as smoking, working in concert to bring about their unfortunate situation.

What does this have to do with psychoanalysis?  Well, Freud introduced the concept of multiple determinism over a hundred years ago.  This is a fancy way of saying what the cigarette smoking example illustrates – our thoughts, feelings, dreams, and actions do not result from a single cause, but from multiple sources acting in concert and as a result of weird compromises that emerge between conflicting sources.  So the cigarette smoking example should be something that psychoanalysts understand and live by, right?

Well, it’s not as easy as that, whether you are a psychoanalyst or the woman or man on the street.  We get overwhelmed when things get too complicated.  We want things to be simple.  When I hit the billiard ball, it rolls in the direction that I hit it every time.  That is a correlation of 1.0 and we want correlations of 1.0 – or very close to 1.0 – all the time.  And remember, in health, .40 (which we have discovered – despite its power – is actually a very small number) is about as big as numbers get in the world of predicting some distant outcome from some set of behaviors.  But this means that we are only able to predict what will happen some of the time for some of the people when a lot of other things line up in important but hard to infer ways.  It’s pretty remarkable that we have any kind of behavioral science at all, actually.  

So Freud, who was a keen observer and a brilliant theoretician, was able to see through this maze and recognize those big predictors – the smoking cigarettes, if you will – and bring them to our attention.  That said, he was frequently wrong, or saw the wrong aspect, but even when he was wrong, he was often able to provide an explanation that made narrative sense – meaning that it held together the way that a good story does – so we gave him the benefit of the doubt.  So, for instance, as part of a grand theory related to sexual development, Freud proposed that the way we are toilet trained will predict aspects of our later personality – specifically that if we are toilet trained in a lackadaisical way, we will be messy in our approach to the world, but if we are given rigid direction, we will be neat, orderly, even obsessive – and thus the term anal has come to describe, in common parlance, those traits. 

If toilet training actually is related to obsessional symptoms, which it probably is, it is way down the list of things that predict how obsessional an adult will be – so it has a really low number – and the things at the head of the list – the ones that have the most power to predict - don’t have a very big number.  We don’t have a great handle on what causes obsessive functioning – genetics is part of it, and certainly the ways in which people interact with caregivers during their childhood are part of it – but Freud’s explanation was one that was so clear and had such a compelling narrative arc that we were willing to take it as a valid explanation.

We were also willing to do that because we want to reduce complicated situations to more simple ones.  Freud was actually pretty humble about the reach of many of his theories.  He would say – look, this is based on a small sample and I don’t know that it is relevant to other populations or in other situations.  He did this at the beginning of Mourning and Melancholia, where he noted that he was basing his formulation on very few cases, but the distilled idea – which does violence to the very complex and subtle description that Freud offered – that depression is caused by anger turned towards the self – became first a platitude that was used to describe all depression and then a symbol of how psychoanalysis over reaches – because it simply isn’t the case that all depression is related to either the boiled down version or the more extended version.  Some aspects of some depressive experiences are very well explained by the principles set forth in Mourning and Melancholia.  It is a very useful way to understand some situations (just as cigarette smoking is the critical causal agent in many cases of lung cancer), but it is not always the determining factor – and many times is not even one of them.

But we are drawn by elegant descriptions and prefer simple explanations, and so we boil down complex situations and people into very simple equations and apply them beyond their useful limits.  But that doesn’t mean that narratives don’t serve very useful functions.  We use narrative descriptions to capture some important aspect or element of the lived life of the people that we interact with.  Psychoanalysts do this when they offer an interpretation to a client.  Novelists and poets and screenwriters do this when they construct a story in a particular way.  I do it when I interpret a book, play, or movie in a certain way – I reduce it to a particular dimension.

I am almost always dissatisfied when I re-read a post about just about anything.  Despite being generally satisfied - I usually feel like I have portrayed some important element or story line of the piece I am describing, I also feel that I have not gotten all of it.  So, I have come to accept that I can’t do that within the length parameters that I give myself, but frankly I need to come to grips with the idea that a complete explanation can’t be offered.  Art – whether in the form of a good book, play, movie, or lived experience (a dream or an interaction with another person) – hints at the complex universe that we live in, but it is not a complete representation of it.  The author uses his or her conscious and unconscious mind to construct a narrative that reflects what could have led to a series of events that could have occurred.  We, as the audience, use our conscious and unconscious minds to fill in the gaps in the narrative – to make it into a plausible story that is both reducible to something understandable and part of something more complex than we can hope to grasp.  And this process, which I try to engage in with these posts, captures (hopefully) some small but powerfully useful per cent of the variance of something that is very complex – something that goes beyond the story and merges into the vast murkiness of the millions of lives that people actually live – lives that are never fully understood by anyone.  The story doesn’t – and can’t - get anywhere near all of that complexity, but having a handle on even a corner of it gives us some power to appreciate it, to know it, and to be able to communicate about it.


The number .40 can be both a very large and a very small number – and our ability to make sense of the world necessarily falls short of accounting for all of the variance in it, but it can provide a useful handle that can help us account for more than we otherwise might.   A good friend of mine is one of the statistical gurus in our department.  It is his position that a good research paper tells a good story.  The statistics need to be used to create a narrative arc.  The statistics, just like a good analysis of a story, and just like a story that is about some event – real or imagined – is a reduction.  And a psychoanalytic interpretation is a reduction – one that draws us back, away from what we are engaged with, but also equips us to re-engage with it, to wrestle with it, but to do that with a new tool in our hands, one that gives us a little more leverage to move a little closer to flipping the thing we are wrestling with on its back so that we can, at least for a moment, pin it down.  That is just fine as long as we realize, whether we are statistician, psychoanalyst, or artist, that we have only pinned down a small corner of a vast network.


 To access a narrative description of other posts on this site, link here.     For a subject based index, link here.

To subscribe to posts (which occur 2-3 times per month), if you are on a computer, hit the X button on the upper right of this screen and, on the subsequent screen, hover your cursor over the black line in the upper right area and choose the pop out box that says subscribe and then enter the information.  I'm sorry but I don't currently know how you can subscribe from a mobile device - hopefully you have a computer as well...




No comments:

Post a Comment

The Covenant of Water: Is it a Great Book?

 Covenant of Water, Abraham Verghese, Psychoanalysis, Psychology, Diversity, Quality Is The Covenant of Water a Great Book?   Abraham Vergh...