Character  &  Context

The Science of Who We Are and How We Relate
Editors: Mark Leary, Shira Gabriel, Brett Pelham
Jul 15, 2019

If Life Is Uncertain, Should We Eat Dessert First?

by Brett Pelham
Young boy looking at plate of cookies

Many decades ago, social psychologist Walter Mischel developed a clever test of delay of gratification in kids.  In both kids and adults, delay of gratification is the ability to refrain from doing something you’d like to do right now with the promise that you’ll get a bigger reward if you wait.  For reasons that will soon become obvious, Mischel’s delay of gratification test has become known as the “marshmallow test.” 

In Mischel’s original work on this test, kids at a Stanford University nursery school were seated at a table populated by nothing but a delicious marshmallow. Children were told by an experimenter that they were free to eat the marshmallow.  However, before they could pop the marshmallow in their mouths, the kids learned that the experimenter had to run an errand.  If kids could sit tight for a while without eating the marshmallow, the experimenter would return with a bonus marshmallow. 

But that’s just the set-up.  Here’s the punchline.  About 11 years after Mischel’s lab studies of the marshmallow test were completed, Mischel and two colleagues, Yuichi Shoda and Philip Peake, tracked down 200 teenagers who had taken part in the original marshmallow tests as preschoolers. For these teenagers, Shoda, Michel, and Peake got both teachers and parents to report how much self-control the kids had and how good the kids were at planning ahead and dealing with stress. For some of these teenagers, the researchers even got access to their SAT scores. 

Analyzing these new data, the researchers found that the longer these kids had waited as preschoolers before caving in and eating the marshmallow, the better they were doing 11 years later in high school!  Their ability to resist eating the marshmallow as kids was related to how much self-control the teenagers seemed to have, how well they coped with stress, and how well they did on the SAT.  These findings suggest that the ability to delay gratification early in life is highly stable across time – and that it is related to important adult life outcomes.             

But this clever study proved to be only part of the story. Tyler Watts, Greg Duncan, and Haonan Quan recently asked what would happen if one were to repeat this study using a more diverse set of kids, rather than the generally White and financially comfortable kids Mischel and his colleagues had studied. This question is important because, as they noted, in the wake of the study by Shoda, Mischel, and Peake, many experts had designed educational interventions for struggling school children based on this single study.  If we could just teach preschoolers to delay gratification, maybe they’d all get into Stanford.

But very few kids in the Stanford preschool, where the original study took place, were struggling in any way. Thus, Watts and his colleagues conducted their study with much more diverse and representative samples of four-and- a-half-year-olds.  They gave these kids the marshmallow test, along with many other psychological measures. Then, they waited for these kids to make it to high school – where the researchers assessed the same outcomes Shoda and his colleagues had assessed.  

In their follow-up study, Watts and his colleagues also found that kids who waited longer for a highly desirable treat to double did better academically more than a decade later.  But Watts and his colleagues also found that, in their studies, preschoolers from poorer and less educated families had substantially less ability to delay gratification than kids from wealthier and more highly educated families.  They also observed as association between early delay of gratification and high school SAT scores that was real but much weaker than the effect observed in Shoda, Mischel, and Peake’s earlier study.   

In other words, as in the original study, preschoolers who waited longer than average for a marshmallow did do a bit better on the SAT when they were in high school. But this tendency was much weaker among poor kids than the comparable tendency observed among more financially comfortable kids. 

Watts and his colleagues also found that preschoolers who scored high on the marshmallow test later did better on the SAT, in part, because they already had excellent intellectual skills as preschoolers – not simply because they were good at delaying gratification in and of itself.  So it looks like delay of gratification measured early in life does have a modest but impressive ability to predict how kids will be doing in school more than a decade later. 

But social context matters, too.  A recent study by Celeste Kidd, Holly Palmeri, and Richard Aslin showed that it’s easy to change how long preschoolers are willing to wait for a marshmallow – by changing how reliable their social worlds are.  Here’s how they came to this conclusion. Kids began this new “marshmallow study” by interacting with a friendly female experimenter who told them they’d be doing an art project.  Kids further learned that if they wouldn’t mind waiting a bit to start their art projects, the experimenter would step out and return with a set of art materials that was much better than the small set of heavily-used crayons the kids could see in a glass jar. All the kids agreed to wait for the better set of art supplies.

For half of the kids, the experimenter returned after a couple of minutes, bringing with her a truly awesome set of art supplies. For the other half of these kids, the experimenter returned with a sincere apology rather than a sweet set of supplies.  She explained that she had made a mistake and that the awesome art supplies were not available.  She quickly and cheerfully reminded the kids in this “unreliable” condition that they could still use the pitiful, worn-out crayons she had shown them earlier.                  

Once the preschoolers finished their art projects, the experimenter returned and gave them Mischel’s marshmallow test.  The preschoolers who had recently experienced a “reliable” world – a world in which the experimenter followed through on her promise of better art supplies -- waited four times as long (about 12 minutes) for the bonus marshmallow than did the kids who had experienced an unreliable world in which the experimenter did not follow through on her promise (about 3 minutes).   

Kidd and her colleagues seem to have approximated in the lab the very conditions that poverty and instability often create in the worlds of kids who are not lucky enough to attend elite preschools. Taken together with the classic and recent longitudinal studies, this clever experiment suggests that the relative inability to delay gratification may result, in part, when people perceive their worlds as uncertain and unreliable. When life is uncertain, it may not be a bad idea to eat dessert first.         

For Further Reading

Kidd, C., Palmeri, H., & Aslin, R.B. (2013). Rational snacking: Young children’s decision-making on the marshmallow task is moderated by beliefs about environmental reliability. Cognition, 126 (1), 109-114.                                        

Shoda, Y., Mischel, W., & Peake, P. K. (1990). Predicting adolescent cognitive and self-regulatory competencies from preschool delay of gratification: Identifying diagnostic conditions. Developmental Psychology, 26(6), 978- 986.

Watts, T. W., Duncan, G. J., & Quan, H. (2018). Revisiting the marshmallow test: A conceptual replication investigating links between early delay of gratification and later outcomes. Psychological Science, 29(7), 11591177.


Brett Pelham is a social psychologist who studies gender, the self-concept, social inequality, and other social psychological issues. This blog is adapted from his upcoming book with David Boninger – which is tentatively entitled Psychology in Modules: Understanding Our Heads, Hearts, and Hands. Brett is also an associate editor at Character and Context.      

About our Blog

Why is this blog called Character & Context?

Everything that people think, feel, and do is affected by some combination of their personal characteristics and features of the social context they are in at the time. Character & Context explores the latest insights about human behavior from research in personality and social psychology, the scientific field that studies the causes of everyday behaviors.  

Search the Blog

Get Email Updates from the Blog