How the recent past shapes our decision-making

A recent study highlights the extent to which we magnify events from the recent past when making decisions about similar situations

Imagine a science writer who has been up all night chatting with friends, listening to music, and then decides to walk to the beach with them to see the sunrise. There’s a deadline for an article such as this one in a few hours, and the writer was well aware of this, long before the café owner closed the shutters. But he decides to gamble a little, there was a deadline last week too, and the writer completed it after staying up all night – it worked out, there was a bunch of likes on Facebook. A new study by Professor Wolfram Schultz and Dr Martin Vestergaard from Cambridge University’s Department of Physiology, Development and Neuroscience suggests the writer’s decision to do this again could be based on ignoring more considered evidence from a longer period of time. The writer’s decision, their research suggests, may have been the result of what they called the ‘Banker’s Fallacy’ – when we are over-optimistic based on being overly impressed by some immediate gain. The accompanying notion of ‘temporal markdown’ is when we are overly discouraged by recent decline. The further back the experience we have of a situation, the less-weight it carries in our decision-making, and it isn’t factored into a broader-based and measured view.

Dr Martin Vestergaard explains that the research is exploring such responses based around the effect of a ‘happy ending’, as it is popularly called.  Dr Vestergaard’s work on decision-making in general is inspired by Psychologist and Behavioural Economist Daniel Kahneman’s work on cognitive biases. Vestergaard explains that ‘Kahneman and his co-workers identified a number of – some would say – types of irrational behaviour. But they go under the term “cognitive biases” which is an umbrella term for peculiarities in decision-making where it seems that people are making irrational choices, where a compelling case could be made in favour of the unchosen options.’

Easier mental computation

In the paper for the Royal Society Proceedings B (the Society’s biology research journal) Professor Schultz and Dr Vestergaard explained this decision-making process by making an analogy of a three-course meal on a menu. We might imagine that our memory of a three-course meal we had at a restaurant, and in principle our judgement of the overall value of that meal, would be built from the value of each course. For example, a dinner with an OK starter, good main course and excellent dessert, all things being equal (that I’m not just someone with a sweet tooth!), should have the same evaluation as a dinner with an excellent starter, good main course, and OK dessert. What happens though is that our decision-making is overly influenced by the most recent experience, whether good or bad. The meal with the ‘happy ending’ is given greater weight in our decision.

Why do we do this? ‘Instead of the very computationally taxing task of calculating the actual summary evaluation all the time, of all the parallel experiences we have in our lives,’ says Dr Vestergaard, ‘we look at what happened just now in relation to what we have experienced in total so far. You can think of that as a contrast in the experience – are things getting better or are things getting worse? Then we don’t need such complex computational machinery to constantly update our internal log book, instead we just look at if there is difference upwards or downwards.’ 

Propensity for temporal markdown?

So for example, is this the process we see in sport when players or teams become overconfident? ‘If these psychological attitudes are driven by contrast guided decision making it might be’, says Dr Vestergaard. Equally ,driven by the same contrast guided decisions making, teams who were recently badly beaten but overall have a good record might overestimate the value of the latest most immediate poor result, disregarding the positives in the past.  

In the study just under a quarter of the participants (9 out of 41) were actually able to evaluate accurately their experience in the experiment – there was no temporal markdown. What makes these people better able to make effective decisions?  And as Vestergaard points out, if you look at the remaining participants whose evaluations did include a temporal markdown, ‘there is still a big variation in how well they do. I am interested in what it is that determines the extent to which people are guided by contrasts or recency and it is something I am looking at right now.’ The next step in the research is looking at other factors which make people more likely to engage in temporal markdown, such as education, socioeconomic background, profession or career.