If one the founding fathers of behavioural psychology can be guilty of the instinctive bias of drawing conclusions from a small amount of data, then anybody can
Should you harbour ambitions that your next child will be a computer genius, then try and make sure they are born between 22 December and 19 January. That would of course make them a Capricorn and, from 1966 to 2013, no fewer than 10 of the 61 winners of the Turing Awards – given annually to those deemed to have made contributions of lasting and major technical importance in computing – were Capricorns.
Now, 10 out of 61 is clearly twice as many as you might expect – and, as it happens, five times the number of Scorpios (23 October to 21 November) to have won. Equally obviously it is nonsense. As this AnalyticBridge blog points out, having carried out 100,000 random simulations of a group of 61 people, 10 or more of them end up sharing the same birth sign around a third of the time.
What we have here is the problem of limited amounts of data or small ‘sample sizes’ – or, to put it another way, of how the human brain has evolved over the years to be predisposed to generalise and jump to conclusions. It is why people see things happen two or three times and conclude that constitutes a reliable pattern that can inform future decisions. It is, as you might imagine, deeply unhelpful in investing.
Nevertheless, it is something everyone is prone to do and indeed, given it is essentially a naturally ‘hardwired’ response, it is very hard not to do it. Still, if you had asked us to pick one person who would not fall into this trap, here on The Value Perspective, we would probably have bet all our cash on one of our heroes, the Nobel Prize-winning behavioural psychologist Daniel Kahneman. As it turns out, we would have lost all our cash.
Even experts fall into behavioural traps
In his book Thinking Fast and Slow, there is a chapter on ‘priming’, which is the idea exposure to words or ideas can make us act in a way we might not expect. A famous example of this is ‘the Florida effect’ – after an experiment where people were timed walking down a corridor and those who had just been looking at various words and ideas associated with elderly people walked more slowly than those who had not.
Last month, another blog – Reconstruction of a train wreck – decided to review the various studies cited by Kahneman in his book and, because the sample sizes were so small and thus the studies were not replicable, it awarded them a combined “Grade F for Fail”. And, rather brilliantly, among the various comments on the blog is one from Kahneman offering a measured admission he had indeed been guilty of a behavioural sin.
Or as he notes at one stage: “What the blog gets absolutely right is I placed too much faith in underpowered studies. As pointed out in the blog, there is a special irony in my mistake because the first paper that Amos Tversky and I published was about the belief in the ‘law of small numbers’, which allows researchers to trust the results of underpowered studies with unreasonably small samples.”
If one the founding fathers of behavioural psychology is not immune to millennia of evolutionary conditioning, then nobody is.
Just the sort of grown-up response we would expect from one of our heroes then – but of course the wider point is that if one the founding fathers of behavioural psychology is not immune to millennia of evolutionary conditioning, then nobody is. Investors therefore need to be especially vigilant – and always questioning a sample size before building grand theories around it would be a step in the right direction.
Or, as we noted in On the bounce, in the context of basketball’s so-called ‘hot hand effect’: “Whether one is looking at basketball or fund management, performance is more random than most people believe and is not a reliable indicator of future performance. As such, investors need to guard against mistaking short-run underperformance or outperformance for a change in a long-term trend.”