A new take on the marshmallow test

Most people — even the non-psychologists among us — have at some point heard of the legendary marshmallow test, which measures the ability of preschool children to wait for a sweet treat.

Researchers have found that the amount of time children are willing to wait for their marshmallow is surprisingly predictive of various life outcomes, such as educational attainment during adolescence, as well as social competence and resilience to stress throughout development.

A recent fMRI brain scan study even found that people's performance as kids is related to their ability to suppress their impulses, and is reflected in neurological signatures of cognitive control, 40 years later.

The test is clearly tapping into something crucial that shapes children's futures to a considerable degree. But what exactly is it? Does the test  capture an ability that is akin to intelligence or intrinsic cognitive control, or might performance be a marker of some other underlying factor — such as the privilege of living in a supportive home where children can develop the trust capacity that enables them to wait for a reward?

The list of potential explanations is long — and now it has received a surprising new addition from a study recently published in Psychological Science. Fengling Ma from Zhejiang Sci-Tech University and colleagues have discovered that children can radically improve their performance on the marshmallow test if they believe their social reputation might be at stake — an effect that begins to emerge as early as three years of age.

In the original experiments, conducted at Stanford University in the 1970s, preschool-age children were left in a room with a highly desired treat (e.g. a cookie or marshmallow), and a less appealing one — oftentimes a pretzel. They received the following offer: if they could wait for fifteen minutes on their own, they could eat their favourite treat. If the wait became unbearably long, they could always bring back the experimenter. But then they could only have the less exciting snack.

Waiting the full fifteen minutes is a challenge, especially if the treats are tauntingly on full display. Children participating in the original experiment waited an average of just over three minutes under these conditions. And the recent study found that waiting was just as difficult for more than 270 three- and four-year-olds from preschools in one Chinese city.

In these experiments, children were asked to choose between a single instantaneous reward (a colourful sticker or a cookie), and double the reward if they waited quarter of an hour. Under baseline conditions, fewer than 25% of children had the patience to make it to the end.

But this figure dramatically changed in two other conditions, where children were made aware of the possibility that their performance might affect their reputation. Here, the experimenter told each child that either their teacher or friend would find out how long they ended up waiting for their reward. This had a radical impact on behaviour: 40% of children waited till the end when they thought their performance would be communicated to a peer, while almost 70% did so when they believed their teacher would find out about their wait.

These results are interesting given that the experimenter did not explicitly mention to the children that waiting longer was a desirable behaviour. Instead, kids seemed to understand that patience is a socially valued attribute, and adjusted their behaviour in the hopes of being perceived favourably by other people.

The fact that children waited longer when their teacher, as opposed to peer, was their theoretical spectator is intriguing. It implies that children as young as three can make sophisticated judgements about the value of their audience and strategically alter their behaviour to impress them.

The fact that a desire for social recognition radically improved children's performance offers us much food for thought. For years, many researchers have assumed that the key to the marshmallow test's ability to predict children's future success lay in its ability to measure their cognitive control. There is, after all, clear value in being able to suppress impulses in favour of more deliberative behaviours.

Alternatively, some researchers have argued, the test taps into some other cognitive skill that might have been shaped by a child's environment. In this debate, largely centred on cognitive abilities, "soft" social skills have perhaps taken a backseat.

And yet the new study points at a new, and previously unsuspected, ingredient of good marshmallow test performance: a child's awareness of the behaviours people value, such as patience, might be just as important. Perhaps it is the marshmallow test's ability to tap into social perceptiveness that holds the key to this simple test's ability to predict future success, time and time again.

Further reading

– Delay of Gratification as Reputation Management

About the author

Sofia Deleniv holds a degree in Experimental Psychology and has just completed her PhD in Neuroscience at the University of Oxford, where she investigated sensory processing using a mix of electrophysiology and computer modelling.

In 2015, she decided to try her hand at science writing by starting her blog 'The Neurosphere'. Since then, her work has appeared in magazines such as the New Scientist and Discover. You can visit her Twitter feed for updates on her written work and other exciting bits of science.