Marshmallow Test Revisited (2020) Michaelson


Preschoolers who have families where they learn to delay gratification tend to have later life success.


“Thus, a child’s ability to wait might be less important than the social and environmental circumstances influencing their willingness to wait.”

Michaelson, L. E., & Munakata, Y. (2020). Same Data Set, Different Conclusions: Preschool Delay of Gratification Predicts Later Behavioral Outcomes in a Preregistered Study. Psy- chological Science, 31(2), 193–201.

Research Story

Two researchers from American Institutes for Research and UC Davis looked at new data from a long study that included the famous marshmallow preschool test. They pre-registered their plan to see if the marshmallow test predicted good life outcomes and to see if the differences later in life could be attributed to factors other than self-control.

They took data from 1,364 participants followed from birth to age 15. They found that the marshmallow test does predict later life outcomes in 3 of the 5 original findings. They found that kids who delayed gratification on the marshmallow test did do better academically, had fewer problem behaviors, and had better social skills.

So the marshmallow test does capture something about early childhood that is important for life success. But what? Is it self-control or social support? Did kids who waited do so because they had individual self-control or be- cause in their homes they delay gratification in other ways and homes where parents have their children delay gratification might tend to do other things that predict better outcomes later in life. For example, those kids might have more trust that things will go like they expect. They may have more practice with decisions involving the future.

Delay of gratification did predict life success, but social support explained the relationship better than self-control. So in essence – the marshmallow test might be a better measure of social support (the kind of family a pre- schooler has) rather than individual self-control.

Another possibility is that the marshmallow test doesn’t predict the life outcomes, but instead predicts academic and social outcomes in preschool and those stay generally stable during deveopment.

Another set of 3 researchers led by Tyler Watts did a similar study with
the data set, but found that the marshmallow test doesn’t predict later life outcomes. The two researches for this study made a dozen different research decisions and came up with different conclusions. Thes authors of this paper made a number of arguments about whey their study was better.

The marshmallow preschool experiment does predict life outcomes and now there’s a piece of evidence why. Rather than self-control, it could be the family environment. It might be that rather than being better able to wait, they are more willing to wait.

So What – Application

If the marshmallow test really measures social support in preschoolers rather than self-control, then the test could be used by schools and others working with young kids to identify at-risk kids.

If it’s really the social support aspect of delayed gratification that kids get at home that’s making the difference, then one possible application would be to do activities with our kids that help them to practice delaying gratification and looking to the future.