Sure you made the right choice?

How does decision-making confidence change as we age and gain more experience? Is it always a good thing to be confident in your choices? Katie Haylor spoke to social neuroscientist Dan Bang from UCL...

Dan - So when we talk about decision making and in particular confidence - which is a feature of cognitive control process, you might say - then we often talk about something called metacognition, broadly defined as thinking about thinking. And more specifically it's the set of processes that are involved when you reflect on, evaluate, and control your on-going thought and behaviour. And metacognition has different features which are going to be relevant for understanding how confidence might change over and how decision making might change over the lifespan. So on one hand we have what we call sensitivity or resolution, which is how good you are telling your correct beliefs from your wrong beliefs. And then on the other hand we have something we call calibrational bias, which is a general tendency for over or under confidence if you like. The former is more related to development, the maturation of neural infrastructure, if you like. Whereas the latter is more tied in with social context and the different social roles that confidence plays.

Katie - I guess both of those things change as we grow up. Right?

Dan - Exactly. So if you think about the sensitivity of our confidence, that's likely to depend on different forms of cognitive function that we know are slow to develop, such as cognitive control or executive function. And then on the other hand, this bias side of metacognition is more tied into the current social context if you like. And as we know, the social context changes dramatically across the lifespan. So the different kinds of functions that confidence might fulfill also going to change. And that might lead to what we think of as stereotype behaviour in different periods of time.

Katie - Can I put something to you? Because I've often heard the metaphor of growing up and confidence being a bit like being a fish in a pond. And so when you're a kid, the pond or your area of influence, the area in which you're sure about things is relatively small, but as you grow up, the pond becomes a lot bigger. So you might feel a bit less confident in your decisions. But that's because your context is so much broader. Does that metaphor have relevance to what you're saying?

Dan - I think it does. So in decision neuroscience, we often think of confidence as a mathematical construct. So when we talk about confidence, we talk about some internal estimate of the probability that a belief or a decision is correct. And in order to compute that quantity, that estimate accurately, you need to have access to all the kinds of information that bear on that hypothesis, belief or decision. And naturally as you world sort of grows, as you gain more knowledge about different domains, you'll realise that there are more and more pieces of knowledge or evidence that should bear on your hypothesis or belief, which naturally I think will lead you to feel less confident about that hypothesis because you realise that you can't have full confidence in it in a sense, I mean obviously in broad terms.

Katie - Are there factors that we know tend to influence how confident we might be at making a decision?

Dan - As the decision grows in complexity, the more factors are going to be at play. So if you have simple decisions, your confidence is likely to be based on all the relevant information there because you have that available to you. So you can compute a more accurate sense of confidence. But as you get to more complex things, other factors are going to be in play. And then we stop relying on what we call heuristics or rules of thumb. We have something called confirmation bias. So if you already believe something, we're more likely to accept evidence or supporting that hypothesis rather than another hypothesis. You also have things like ease of processing. So if a hypothesis seems sort of intuitive to us or if it's easy to understand, we're more likely to think that that's correct. So these are sort of rules of thumb we generally hold, but then when you get to more complex problems can probably lead you into trouble.

Katie - I see. So it sounds like actually making a decision objectively can be pretty difficult.

Dan - Exactly. And I think decision neuroscience went through a phase where we thought of humans as quite silly, making silly decisions. But then when we start recognising that we make decisions under huge resource constraints, we don't have infinite computation available, we don't have infinite time available when you make decisions, then you can start thinking about these different biases or heuristics as quite efficient solutions to what's actually a very difficult problem.

Katie - So how do scientists like yourself study confidence in decision making?

Dan - So again, I think it takes different forms. So what I do in my own work is use simple games based on simple perceptual tasks. So you see a cloud of dots on your screen and you have to say "are they moving to the left or to the right?"Then we manipulate the amount of noise, that sensory stimulus, and then we check whether your confidence sort of tracks the accuracy of your choices about that sensory stimulus. So we try to come up with constraints, scenarios where we ask you questions over and over again, and we can sort of quantify how good are your decisions overall, and then how good are you at evaluating your decisions? So you can imagine you might have people that are very bad at making decisions, but they know they're bad. Or you might have people who are very good at making decisions, but they're not able to recognise the times they make mistakes.

Katie - Does that distinction relate in any way to someone making the right decision? I'm wondering if being more confident in your decision making makes you more or less likely to actually make the right choice?

Dan - Again, confidence has many different functions, but I think for you as an individual, the danger of having too high confidence is that you're not going to seek dis-confirming evidence. You're not going to go out and test your hypothesis about the world extensively, because you already believe it to be right. So there can be a sense in which over time, having a miscalibrated sense of confidence is gonna lead you to have more extreme beliefs if you like.

Katie - Do you think there's any sort of interesting social reflections there, because we seem to favour confident people, right? Whether it's in the media or perhaps in politics or something.

Dan - It's true for example, in politics. You can say that the general voter tends to prefer confident people. We want to note that those people who are in charge know what they're doing, but that sort of introduces a different incentive scheme, if you like, for confidence. Such that if you're a politician, you might realise that even though you have a doubt about a certain policy or a certain plan for government, then it's going to be in your interest to have overconfidence in that policy or in that plan. Because you know that that's going to turn into some political currency for you. It just really highlights the social aspect of confidence when you get to this larger political scale.

Katie - Do you think there's anything interesting to say about confidence in decision making and the scientific method?

Dan - Absolutely, because I think confidence in decision making, it's about evaluating the strength of your belief about some hypothesis that you base your decisions on. And what we do in science is that we have hypotheses. We go out to do experiments to test the strength of our belief in those hypotheses. Now you could say that the scientific method is implicitly informed by what we know about biases in confidence in decision making. So we know we have this confirmation bias. We know that there are other factors such as the appealingness, the simplicity of our hypothesis that might sway us in that direction. So we think of experiments, we think of statistical tests that are going to allow us to control for those biases that might drive confidence in decision making and everyday life.