A four-year-old girl sits at a table in a featureless room. A friendly researcher places a marshmallow in front of her and tells her that if she can resist eating it for 15 minutes, he will be back with another one and she can then eat both. He leaves, and what she does next will predict her success and mental health for the rest of her life. Such is the power of the now classic marshmallow study, long thought to be a measure of self-control.

The initial research began in the late 1960s, and follow-up work has suggested that the length of time a child waits before eating the marshmallow is a better predictor than intelligence of success as an adult. A new study published last October in Cognition, however, indicates that children's behavior in such situations may not always reflect only their innate self-control. A child may also be making a rational decision on whether to trust that the second marshmallow is indeed coming soon.

Celeste Kidd, a doctoral candidate in brain and cognitive sciences at the University of Rochester and lead author of the new study, suspected there might be a common misconception about the classic marshmallow study—namely, that waiting is always the right choice. While volunteering years ago at a homeless shelter for families in Santa Ana, Calif., she realized that all the kids around her would eat their marshmallows straight away, living as they did in an environment where anything they had could be taken away at any time. “Delaying gratification is only the rational choice if the child believes a second marshmallow is likely to be delivered,” Kidd says.

Although previous marshmallow-type studies have acknowledged that external factors might influence kids’ ability to wait for the bigger reward, none had directly tested for those factors’ effects. So Kidd and her colleagues ran a study in which they manipulated the reliability of their young participants’ environment. A researcher gave children with an average age of four years some poor-quality art materials and told them if they could wait, she would return with better supplies. In a “reliable” condition, she did exactly that, but in an “unreliable” condition, she returned to explain she did not have any better materials after all. A marshmallow test followed. Those in the reliable condition lasted an average of 12 minutes, whereas those in the unreliable condition lasted only three.

With that in mind, the findings of the many decades of follow-ups to the marshmallow study [see timeline at right] are cast in a different light. The studies invariably point to a strong association between how long a child was able to wait before eating the marshmallow and various measures of mental health, competence and success in later life. A recent imaging study of the kids in the original study, now in their 40s, even found differences in the activity of key brain areas between those who could and could not resist temptation as children.

If Kidd is right, these differences may be the result of more than just innate self-control, such as socioeconomic status, parenting quality and other environmental factors that influence decision making. “It's incorrect to presume lack of willpower is the only relevant factor in determining children's wait times and, subsequently, the primary driver of children's successes later in life,” she says.

1972: Walter Mischel's classic “marshmallow study” is published. Preschoolers were given a treat and told that waiting 15 minutes to eat it would earn them a second marshmallow. They waited, on average, six minutes. Children who hid the marshmallow from view or who distracted themselves were able to delay gratification much longer.

1981: An experiment in pigeons produces similar results—distractions and visual obstruction helped the birds delay gratification.

9 Comments

1. 1. TTLG 10:58 AM 3/11/13

   I think belief in one's own ability to control others may also play a part. If the child has faith in their ability to get the adult to deliver up what is promised, then they will also be more likely to wait for the promised reward. Since believing in one's ability to control the external world is a key part in trying to change things, it seems reasonable that this would have a big influence on how successful one is in life.

   Reply | Report Abuse | Link to this

2. 2. NeuroRu 04:14 PM 3/11/13

   I just want to ask, how in world did they do this with pigeons? Did they have a pigeon-whisperer or something?

   Reply | Report Abuse | Link to this

3. 3. nicholasjh1 05:06 PM 3/11/13

   Not to mention, low ability to trust a wait for another marshmallow could = bad schools, bad parenting, poor nutrition, and a number of other factors. Meaning will power may have no direct causation for success. It was simply showing that these students were poorly taught, malnourished and facing hard odds.
   

   Reply | Report Abuse | Link to this

4. 4. phalaris 06:55 AM 3/12/13

   Here we can see the left at work. From a single study we have the tentative opinion expressed :
   
   "these differences may be the result of more than just innate self-control"
   
   and immediately we have it from nicholasjh1 that it's nothing to do with innate characteristics.
   
   Even the conclusion that non-innate factors made the difference is not necessarily justified: it may be that the kids just had the (innate?) ability to intuit that they were in an unpredictable environment, where practicing self-control is unlikely to produce results. I think TTLG is making a similar point.

   Reply | Report Abuse | Link to this

5. 5. nicholasjh1 01:25 PM 3/12/13

   But Phalaris, you are willfully overlooking the data given that those who chose not to wait did poorly in life. Those were just examples of what may have been the reasons for success/failure. The real point I was trying to make though is how easily correlation can be construed as causation when scientist start making assumptions like this and spinning stories about why they got a specific correlation. There was a strong correlation in this test and now psychologist have been using it for years to try an prove something that may not exist because they never controlled for confounding factors such as education / nutrition / and balance in home life. Poor science is my point. I would agree that TTLG's idea is another possibility, and in fact having grown up with many blacks and whites in Detroit (I am white) I found that I was constantly receiving the message that I would succeed and having people agree to my requests and assume that I would do well on the assignments. This was not automatically true of my peers who were not white. It has a real effect on people.

   Reply | Report Abuse | Link to this

6. 6. DonaldS 05:05 PM 3/12/13

   Did the authors of the original study actually claim that the observed impulse control was innate?

   Reply | Report Abuse | Link to this

7. 7. DonaldS 05:09 PM 3/12/13

   I would appreciate a reference for the pigeon study.

   Reply | Report Abuse | Link to this

8. 8. bucketofsquid 03:47 PM 3/27/13

   Due to the massive bigotry inherent in almost everything done in the 1960s I'd have to say that the study is meaningless and a new zero sum starting point study should be done where no assumptions are made and everything is considered possible as a reason. Then gather real empirical evidence and follow where it points.

   Reply | Report Abuse | Link to this

9. 9. WizeHowl 05:45 AM 4/1/13

   The first question I would ask is "was there already any relationship between the child and the adult?" If there was none then the child had no reason to trust that the adult would give them the second marshmallow.
   
   Secondly, small children have no concept of time, to them 15 minutes may as well be 3 minutes, so telling them to wait for 15 minutes before they get to eat the lolly is like telling someone in a dark room with no time piece that they can be released in two days.
   
   This whole test was done so that only the richest, and therefore the best educated kids were going to succeed. It was done so that some rich scientist could get richer on more government funding to prove nothing useful.
   
   
   
   

   Reply | Report Abuse | Link to this