Testing the Marshmallow Test

A few months ago, moms from all over the world flooded TikTok with videos of their toddlers being put to the “marshmallow test,” in which a marshmallow was placed in front of a child with the promise that they would receive two treats if they did not eat it while the parent left the room. The claim was that children who displayed enough self-control for a greater reward would have the self-discipline to become successful as adults. 

Leah S. Richmond-Rakerd, assistant professor of psychology at the University of Michigan, and Terrie Moffitt, professor in the Department of Psychology and Neuroscience at Duke University, tested this theory in their recent paper. Richmond-Rakerd explained that the results of this study could help people age more healthily by better understanding the effects of childhood self-control on life decisions made later in life. “We were interested in whether people with better self-control also age more slowly and are better prepared to manage their health, financial, and social demands of later life,” Richmond-Rakerd said. This life skill may be more important now than ever before. 

Richmond-Rakerd and Moffitt conducted their longitudinal study using data collected from the Dunedin Study, a prospective study of a birth cohort of over one thousand babies followed from birth to age forty-five. The study members’ self-control was measured at ages three, five, seven, nine, and eleven using a multi-occasion and multi-informant strategy in which reports were collected from parents, teachers, and even the children themselves. This differs from most approaches: other studies of self-control use behavioral tasks, but the accuracy of the findings from these less holistic experimental models are highly contested in how well they actually predict behavior in the real world. The reporting system instead identifies behavior in the children’s day-to-day lives, such as how well they are able to wait their turn to play a game or how frustrated they get when something doesn’t go their way. In the study participants’ adulthood, aging was measured, biologically and socially—how quickly they were doing so across different organ systems, in addition to their financial, health, and social skills. Did the participants have sufficient financial knowledge? How strong was their social network? 

The study found surprising results. Although childhood self-control has long-lasting implications, there is still an opportunity to prepare ourselves for aging even when we are forty. Moreover, for the participants, self-control was not confounded by IQ, education, or cognitive skills as expected. This is an optimistic finding: it tells us that social skills such as financial literacy are teachable. “Early beginnings matter, but adulthood matters too,” Richmond-Rakert said. “We found that adults who exercise better self-control developed more health, financial, and social reserves for old age—even if they didn’t have so much self-control in early life. This is encouraging because it opens up middle age as a potential intervention window. A lot of research has focused on intervening in childhood, and our results indicate that the early years are certainly important, but middle age may also be a good time to revisit the opportunity to get better prepared for later life.”

The study results may also have social implications on our social security system. The results shed light on a concerning pattern: people are beginning to struggle with their physical fitness and health at an earlier age. “Those rules for when you can retire and when you can get support for your retirement were developed years and years ago when in the U.S., the median age of death was sixty-five. What we have now is that people are living longer but they’re also falling apart younger,” Moffitt said. This opens up an essential conversation of the importance of chronological age compared to that of biological age. 

To conclude, Moffitt gave a more concrete example. “There are factory workers who have had really intense, heavy-duty, physical labor all their lives. By the time they’re fifty-five, they are pretty worn out and they should be able to retire. Whereas someone like me, a college professor who has sat in a comfy office with air conditioning and worked on a computer, I could really work until I’m seventy-five,” Moffitt said. 


Richmond-Rakerd, L. S., Caspi, A., Ambler, A., D’Arbeloff, T., De Bruine, M., Elliott, M., . . . Moffitt, T. E. (2021). Childhood self-control forecasts the pace of midlife aging and preparedness for old age. Proceedings of the National Academy of Sciences, 118(3). doi:10.1073/pnas.2010211118