I can't be trusted around marshmallows

A pet peeve of mine is when news articles make it sound like parents can completely control their children's behaviors by simply saying the right thing. One gift from our children is the lesson that we cannot control everything -  even if we do everything "right." This confounding behavior comes in part from children's low self-control. Even if we tell them not to do something it is very hard for them to stop themselves from doing it anyway.

However children differ in the their abilities to delay gratification, their abilities to fight temptation. Classic research by Mischel and colleagues in the 1960s and 70s outlined a protocol that is often referred to as "The Marshmallow Test." In these studies preschool children were tested individually by a familiar adult experimenter. The child chose a preferred treat (along with marshmallows, animal cookies, pretzels, and other treats were choices) and then the experimenter said that he or she could eat that one piece of treat now, OR wait until the experimenter returned in 15 minutes to get two pieces. The children differed in how long they were able to wait: some only lasted a few moments and gobbled down the single treat; while others distracted themselves or literally sat on their hands until the 15 minutes was up to claim the double treat. Follow-up studies demonstrated that the children who waited the longest grew up to have better grades in school and show other markers of success. This work is still cited today to show the importance of self-control and is sometimes compared to the modern concept of grit.

Through the years some have raised questions about The Marshmallow Test. For example, Duckworth, Tsukayama, and Kirby (2013) asked if it really tests for self-control or if there is a third variable, another quality, that drives both the longer waiting time and higher grades. Based on past research, it could be that children who are higher in intelligence prefer to wait for a larger reward and tend to get higher grades than children who are lower in intelligence. Likewise, children's behavior in The Marshmallow Test and their achievement in school may be reflect how much they are tempted by reward: some children are more tempted by treats while others have little temptation. It may be that children who have low temptation do not have to work very hard to control their impulses - which allows them to wait and to do better in school.

To clarify this issue Duckworth et al. conducted two studies. Study 1 included 56 local 5th graders (average age 10) who participated in a modified version of The Marshmallow Test: the waiting time was extended to 30 minutes and the students completed a survey about what they like eating before the delay of gratification test began. On a separate occasion the children participated in two tests of intelligence, and answered questions related to temptation. The questions addressed: how excited they get when a reward is offered; how actively they seek out the reward; and the extent that getting in trouble impacts them. Teachers rated each child on self-control (as it relates to schoolwork) and reported end of year grades.

The results demonstrated that the children who waited longer were also rated as having more self-control. Although grades were marginally related to waiting time, intelligence and temptation by reward were not related to waiting time in The Marshmallow Test. However, Duckworth et al. note that the children did not differ very much from each other - especially on the variable of intelligence.

Because the first study involved a small sample lacking in diversity, Study 2 analyzed a data set from the NICHD Study of Early Child Care and Youth Development that covered a broader sample of 966 children. These participants had completed The Marshmallow Test at age four (in this version, the experimenter would return after a maximum of seven minutes) and were tested using two age-appropriate measures of intelligence. At that time, their parents and preschool teachers rated them on two measures of self-control: their abilities to concentrate and their abilities to control their behaviors. As well their mothers rated them on their temptation by reward: how excited they became when offered a reward.

When these children were in eighth and ninth grade, the students' grades were included in the data set. In ninth grade the participants also indicated how often they engaged in "risky behaviors" as a measure of impulsiveness.

The results demonstrated that the children who waited longer were higher in intelligence and higher in self control at age four: they received higher ratings on their abilities to concentrate and to control their behaviors. To a lesser extent, children who waited longer were also less tempted by rewards. The authors then used sophisticated statistical methods to separate out the influences of self-control, intelligence, and temptation by reward.

In line with past research, the children who waited longer at age four were more likely to become adolescents with higher grades and fewer reported risky behaviors. Analyses revealed that level of self-control beat intelligence as a better predictor of the associated high grades. However, unlike high self-control, being less tempted by reward did not predict grades or behaviors in adolescence.

Taken together, these studies support the long-assumed notion that The Marshmallow Test is a good test of self-control. Self-control was the most reliable predictor of the higher grades associated with longer waiting times in The Marshmallow Test; intelligence and temptation by reward were less effective or failed as predictors.

We can only assume that Mischel would be pleased with this confirmation that his famous protocol is a valid test of self-control. Considering that he had to wait more than four decades for this news, I would say that he has passed the "ultimate" marshmallow test (too bad he doesn't like the taste of marshmallows)!

Further Reading:

A pdf of the Duckworth et al. (2013) article can be read online courtesy of the University of Pennsylvania.

For an update on Duckworth's work, read "Measuring Students' Self-Control: A 'Marshmallow Test' for the Digital Age", an article by Ingfei Chen for Mind/Shift. Move over marshmallows! Can today's students delay gratification when it comes to video games?

Watch Walter Mischel explain delay of gratification to Stephen Colbert on the comedy news show, The Colbert Report. Very entertaining!

BONUS: If kids totally lose their cool waiting for an extra marshmallow, imagine their reactions to enormous portions of their favorite foods! Watch what happens in this BuzzFeed video: