Posted on: Monday 27th of June 2011
I was speaking about choice tools at a conference recently when someone asked what’s becoming a standard question: “you talk about all the different ways accessing and using information can help make better decisions, but people don’t make decisions like that. When push comes to shove they make their decisions on gut feeling: do I like/feel comfortable with this decision or not? If so, isn’t all this stuff about better uses of information pretty irrelevant?”.
I fluffed my answer then and I’ve fluffed it many times before – it’s a very simple question requiring quite a complex answer … “well, yes and no”. So let me try to outline a better answer.
The bottom line is this: Gut feelings are an indispensable element in human decision-making. Choice tools and other decision support services are and should not be an attempt to replace gut feelings. Choice tools should be designed to do two things:
a) help our gut feelings work better as decision-making guides
b) help us avoid the pitfalls of relying on gut feelings when these feelings are leading us astray.
Let’s investigate further.
‘Gut feelings are good’
The philosopher Alfred North Whitehead once remarked that “Civilization advances by extending the number of important operations which we can perform without thinking about them”. Human beings have evolved the same way. If we had to consciously direct the workings of our digestion, the pumping of our heart, our breathing and so on we wouldn’t have the time or energy to do anything else. They do their business automatically.
Gut feelings – automatic responses to environmental situations – work the same way. Some are purely physical kneejerk reactions: better to jump the instant you feel the prick of a drawing pin as you sit down rather than stop to think about it.
The human brain processes conscious, articulate thoughts at a speed of around 40 bits per second. In that same second, however, our eyes process around 10 million bits of information, our skin about a million bits, our ears and sense of smell about 100,000 bits each, and our taste buds around a thousand bits.
So, crudely speaking, the ratio of conscious to unconscious is about 1/300,000. That’s pretty much how it should be. To test this out, stop reading this for a moment and become aware of everything that’s in your peripheral vision, the feel of your clothes on your skin, any background sounds and smells, and so on. Your senses are constantly scanning your environment non-stop to alert you to any signs of danger – under the radar of consciousness.
Much of this environment scanning is emotional as well as physical. It’s about liking (opportunities for food, sex, pleasure) and disliking (danger, discomfort and so on). It extends to moral and reciprocal instincts, empathy, and the ability to pick up and interpret all manner of subtle social cues (eye movement, posture, tone of voice etc) that you get when interacting with other people. These create feelings of trust, liking and so on that we find hard to explain but are nevertheless indelibly important in all decisions involving judgements about people.
Gut feelings and emotions
Human beings are not ‘rational’ in the ways described by 20th century economists’ description of rational economic man. We do not coolly assemble, sift and judge all available facts to arrive at some objectively defined best possible decision. What’s ‘best’ is defined by what we want, and what we want is primarily emotional. Without these emotional drives decision-making becomes impossible – literally. Patients with brain damage disconnecting the ‘emotion centres’ from the ‘reasoning centres’ of their brains find it almost impossible to make the simplest of decisions.
Gut feelings and learning
Alfred North Whitehead’s point about not having to think about things applies to individual lives as well as evolution. When you are learning to drive a car, every move you make requires careful planning and consideration. You have to think a lot. But then, after a while, you just get in the car and drive, going through all the motions without a moment’s further thought. You have internalised the learning so that your responses become automatic.
This is true of intellectual skills as well as manual skills. Skilled, experienced doctors ‘just know’ what to do when they come across certain patients, while junior doctors fuss and fret as they try to remember all the things they learned in medical school. The experienced doctor’s decision-making seems mysterious, almost magical. But it’s just the same process of embedding learnings into the circuitry to make them automatic.
‘Gut feelings are bad’
These are all good reasons to trust our gut feelings. We wouldn’t be able to function without them. And they work the way they do for very good reasons.
But they are not perfect. Our instinctive reactions to events and situations sometimes prompt us to take actions which we regret in hindsight. The emotionally-driven nature of our decision-making means we tend to look for information that fits our desires and goals – and to ignore information which ‘gets in the way’. This makes us highly selective in the information we pay attention to – which can be very dangerous.
That’s probably why we evolved the ability to ‘stop and think’ – to imagine alternative scenarios, to sketch them out in our heads and think through the consequences; to gather and analyse data to check to see if what our gut feelings are telling us really is supported by the evidence. Speaking evolutionarily we evolved ‘reason’ both to support emotionally-driven, automatic decision-making processes – to think up better ways of achieving what we want to achieve – and to help us correct the errors that pure instinct sometimes makes.
Even so, we still struggle: the rapidly growing field of behavioural economics charts the myriad different ways that human beings display ‘predictable irrationalities’; where our gut feelings lead even our reasoning astray. So while we can’t live without ‘gut feelings’ and shouldn’t even want to, we do need better ways of identifying when they are working well and when they are not.
That’s what 21st century ‘choice tools’ and decision-support services will do. They won’t even try to offer individuals ‘perfect’ decisions as described by 20th century economists. They will use our understanding of human decision-making processes, including gut feelings and ‘predictable irrationalities’ to help us a) ‘stop to think’ when our gut feelings might be leading us astray while b) also helping us improve the speed and efficiency with which we make our decisions.
Here are few simple examples:
- A retailer may cleverly use all manner of presentational tricks to make his product seem good value – by, for example, offering it at a ‘discount price’ in the context of lots of other much more expensive products. All these tricks are designed to give us the feeling of good value. But a quick price comparison check via a mobile phone could reveal that it’s actually being sold a lower price elsewhere.
- Conmen succeed by giving us the feeling that they are likeable and trustworthy. A simple reputation test – getting the reviews of other people who have used this person – could quickly dispel the mirage.
- Often we have a very poor idea as to how we actually behave: our memories, perceptions and feelings trick us. Hard data showing our actual behaviour patterns – how much energy we use, calls we make, money we spend on alcohol – can force us to see the reality of what we do.
21st century choice tools will be designed to help us check, monitor and resort to these extra support services so that we can be confident that we are making better decisions without having to give up all reliance on gut feelings. Looking forward, these support services are likely to become quite sophisticated – taking the form of carefully researched and highly designed ‘smart’ decisions trees.
Gut feelings and heuristics
Not all ‘gut feeling’ decisions are embedded learning, however. Despite the theories of 20th century economists, real human beings are ‘cognitive optimizers’ and ‘cognitive misers’. Far from wanting to assemble all possible facts, to sit and judge them all according to their merits, they prefer to find a short cut.
There are two possible ways of catching a ball, for example. One way is to make detailed measurements of the ball’s weight speed and angle of the balls trajectory to calculate its trajectory, all the while factoring other conditions such as air resistance. By the time you would have completed your calculation, the ball would have passed you by.
Or else you can take a different approach: fix your eyes on the ball and run, adjusting your running speed so that the angle of your gaze remains constant. By focusing on just this one piece of information – the angle of your gaze – and reacting to that, it will lead you to exactly where you need to be to catch the ball. That is how sportsmen and women actually catch balls. And most of them haven’t a clue that they are doing it. It’s just a gut feeling – but this time it’s taking the form of a simple ‘if-then’ rule; an algorithm.
With a little bit of careful investigation, it’s possible to unearth these algorithms and to turn them into easy-to-follow decision trees. Take the example of coronary care unit in a Michigan hospital which was trying to improve physicians’ performance in diagnosing problems. The problem was that their accuracy in diagnoses was about equal to chance – so the doctors compensated by sending every patient for tests even when most of them didn’t need one.
They then tested an expert system that captured all the possible symptoms and diagnostic techniques. It worked very well, except for one thing: it was so long and complicated nobody could be bothered to use it. However, there were some experienced physicians who were very good at diagnosing problems. On closer inspection, it turned out that all they did was ask themselves three simple ‘yes/no’ questions that got to the heart of the matter in no time.
The German psychologist Gerd Gigerenzer calls this ‘fast and frugal’ decision-making. Human beings evolved in an environment when they had to make make decisions when both time and information was scarce, he points out. In such situations, seeking out more information is bad, not good (as in the ball-catching example). The fact that people often ignore information isn’t proof of their ‘irrationality’ – it’s proof that they work better using by using ‘fast and frugal’ or ‘smart’ heuristics such as ‘fix the angle of your gaze and move so that your angle of gaze always remains the same’.
It’s taken a while to get there, but I think this issue of gut feelings and emotions in human decision-making is very important.
Person-centric decision support services – services that provide individuals with choice tools that they can trust – will never succeed by trying to impose ‘rational’ decisions on people. They will succeed by working with the grain of how real people behave.
They will study how different groups and types of people make different types of decision in different circumstances and design ways to help them streamline the process when its working well, and avoid pitfalls and errors when it is not – via smart decision trees for example. They will do this by combining the strengths of the human brain with the strengths of the computer.
Computers are very good at doing some things human beings can’t do, but no good at doing other things. For example, humans find it very hard to remember more than seven things at a time, while computers find it easy remember millions of things. They’re good at crunching structured data – doing price comparisons, collating peer reviews, identifying and analysing trends and patterns. We are not. But we’re good at picking up emotional cues, doing things such as recognising faces and understanding contexts – which they find very difficult indeed.
The purpose of choice tools is to create the best of both worlds. Creating powerful, effective choice tools will be a massive research and design challenge. But then, that’s what intelligent human beings are good at.