Why We Needed to Retake the Marshmallow Test

We might have to change what we think about the marshmallow test.

Fifty years ago (and many times after that), a child and a single marshmallow or some other treat were left in a room. If the child resisted temptation and ate nothing, she could have two when the adult returned. Years later, researchers discovered that the SAT scores of children who held out for 15 or 20 minutes were 210 points higher than those who lasted only 30 seconds. Further confirming the significance of delayed gratification, they found that the kids who waited wound up with better jobs and more emotional resilience.

Now a group from NYU and UC Irvine is again looking at the significance of the marshmallow study. Their conclusion? We got it somewhat wrong.

The New Marshmallow Test

We’ve always thought that the marshmallow experiments were about children who could wait. But they might not have been. We could have been looking at the wrong set of facts.

In the new study, the researchers changed the group’s composition. Whereas the originals were based on a Stanford University preschool with an affluent and educated constituency, the contemporary replication changed the number and socio-economic group. The children were more varied, the follow-up numbers were larger, and the experiment itself was more consistent. They even looked at how many books were in a child’s home and whether their mothers were responsive. Later, analyzing the data, they made sure to control for income and other variables that might skew the results.

What a difference.

The replication told researchers that family background, the home, and cognitive ability were the facts on which to focus. No longer was self-control the key variable. Instead, the ability to hold out is an offshoot of affluence and education. For kids whose mothers had comparable educational background, those who could and could not delay gratification had similar standardized test scores. And for kids whose lives made them believe they might not get a second treat, it made sense to eat one now.

Our Bottom Line: Replication

Perhaps the big lesson here is replication. Researchers believed that a classic study needed a second look. But even the study’s originator, Walter Mischel has a different perspective. Rather than emphasizing the predictive value of delayed gratification, he recently said he was more interested in self-control. Looking back on his work at age 84 (in 2014), he said he was most concerned with how we increase our will power .

For smiles, you might want to see the children in this marshmallow study:

https://www.youtube.com/watch?v=0mWc1Y2dpmY

My sources and more: I do recommend this Atlantic article and Quartz for the up-to-date summary of the new marshmallow experiment. Then, if you want more, the paper on which the articles were based has all of the details. But if you are just looking for a good read on the topic, do go to this New Yorker piece on Dr. Mischel.

Please note that my SAT facts come from this econlife.