Marshmallow Study and class

You've no doubt heard of Walter Mischel's Marshmallow Test and its followup study, which examined the relationship between delayed gratification (the ability to resist the temptation to eat a marshmallow right away with the promise of more if you succeed) and overall life success. Celeste Kidd, a U Rochester doctoral candidate, has published a paper in Cognition challenging Mischel's findings, arguing that children from more unpredicatable circumstances may choose the single marshmallow because they have a rational basis for suspecting that the experimenter is lying to them about the additional marshmallows that await them if they follow instructions.

The Marshmallow Test is sometimes used to suggest that people are poor because they have low self-control; Kidd's paper implies that poor people behave wisely when they grab opportunities as they present themselves, because they are often lied to when it comes to promises of greater rewards down the road.

Celeste Kidd adds:

The video discusses a study we recently did at the University of Rochester that revisits the original 'marshmallow task' experiments from Stanford in the 1960's. Our results suggest children's waiting during the marshmallow task might actually result from a rational decision-making process--not just a deficiency in self-control.

In the Stanford experiments, most children--75% of 3- to 5-year-olds in one study--appeared unable to resist the temptation of an immediate low-value reward (one marshmallow now) over a future high-value one (two marshmallows after 15 minutes). There's a popular misconception about these studies, though, which is that waiting for the second marshmallow is always the right thing to do. In fact, there are a lot of situations in which waiting is a bad idea. If you're skeptical that a second marshmallow will ever become available--or you believe there's a risk that your first marshmallow might be taken away--you should enjoy the smaller reward right away.

In our study, we preceded marshmallow-task testing with evidence that the experimenter running the study was either reliable or unreliable. Children who believed the experimenter was reliable then waited about four times longer before eating the marshmallow than those who thought she was unreliable (12 minutes vs. 3 minutes). These results suggest that children engage in very sensible decision-making that considers environmental reliability. They may also provide an alternative explanation for why marshmallow wait-times correlate with later life success--successful people grow up in reliable situations. Broadly, the study illustrates that children build a model of the reliability of others' behavior--and use this model to inform their decisions.

The Marshmallow Study Revisited

Rational snacking: Young children’s decision-making on the marshmallow task is moderated by beliefs about environmental reliability (Cognition), PDF

(Thanks, Celeste!)


  1. This is also why you should always just make one wish as soon as you catch the leprechaun rather than listen to any of his deceptive talk about getting more wishes down the line.

    And if it’s that “Lucky Charms” guy, I wouldn’t trust his promises on either front.

  2. They should study whether it works with punishments in the same way it does with rewards.  Seems like it should. Environments that are unreliable with rewards would also be unreliable with consequences. I’m just wondering if that could explain how a person can make the decision to steal a Pepsi even when he knows he’s on his third strike. Doesn’t seem rational, but in the context of an unreliable environment, maybe it is.

    1.  Well, something like this has been done with dogs.  Wikipedia has a good description of it, better than I can do with my time-addled memory:

      So learned helplessness, the idea that no matter what you do, you’re gonna get screwed, could explain why that guy steals a Pepsi.  He’s learned that no matter how well you do, you’re still going to be treated like dirt, so you might as well get what little enjoyment you can out of it.  That, or he’s a kleptomaniac.

  3. Have they studied how the kids do later in life?  If you can delay gratification you can write a novel or learn an instrument, etc.

    1. That was the purpose of the follow-up to the original study. The take-away was much as you might guess: kids who can delay gratification for greater reward tend to be more successful adults. This new study seems to show that it’s actually a little bit more complicated than that.

  4.  What if the kid knows that marshmallows are really bad for your teeth, and doesn’t want two marshmallows?  Eh, Mr. Smarty-pants psychologist?

  5. Excellent follow-up, thank you for pointing it out. Now that you mention it: a friend’s dad spent ages 14 to 19 in a Siberian prisoner-of-war camp (after being captured in Hitler’s last defense of Berlin) where hunger was the worst enemy. He said you learned to eat anything immediately; the best place to store food was inside you. Otherwise it could get stolen from you.
    (N.B.: Other interesting stuff I remember: Anyone who wasn’t willing to steal food died early. Prussian officers weren’t as willing to steal food. Tall people starved faster. There wasn’t much of a fence–the guards relied on the remote isolation of the place–so the prisoners organized a rota to go out at night and collect turnips or some kind of root-like pig fodder being grown on the peasants’ fields. If you volunteered for absolutely all proposed tasks, even things you had no idea how to do, that helped your chances of survival. Plus when they put you in charge of a plow horse you could repurpose its oats.)

  6. The implications ar actually huge. The right loves to claim that impoverished people are in their predicament due to , as Ryan said recently, lack of discipline and character.

    This indicates that the poverty in which they live–a very unpredictable and unrewarding atmosphere–may indeed be the problem.

    That means we need to ameliorate (end) the effects of poverty, especially on young children.

    1. My Russian friends tell me the Soviet Union fell because you can’t run large groups of humans on a basis of “we reward slackers the same as hard workers”.  They claim that somebody has to be visibly poor, compared to the average man, or else nobody will bother to plant the seed corn or fix the leaks in the roof and everything falls apart and we all starve.

      But even the hardliners (like Jesus – see Matthew 26:11) would probably agree with you as long as you take the comma and the word “especially” out of your last sentence.

      1. I think there’s more an argument to be made about making sure the base margin of reward reduces struggling to a minimum rather than capping the rewards of the excellent.

        For a given value of excellent, mind you.

  7. There is ample evidence that “extremely successful” people are those who always take the shortcut to instant gratification.  Wall Street.  CEOs taking short-term profits.  Etc.  They get successful exactly by taking the short-term gains that lead to long-term pain.  They typically are able to shift that pain away from themselves and onto others by leveraging the short-term gains.

    It’s not so much “success by delayed gratification”, it’s just success by whatever means, often by pushing others down.  “Investment” of these short-term gains is really just another form of instant gratification – generally pushing yourself up and others down.  I’ve never met a really (financially) successful person who did not have a dog-eat-dog mentality.  They often distance themselves from the carnage though – again, leveraging the short-term success to do so.  In this way, they often can appear quite genteel.

  8. Importantly the observation in this study doesn’t actually preclude the previous interpretation of the experiment from being correct. Both interpretations can be correct and complimentary. Growing up in an unreliable environment will shape your psychological development afterall, neuroplasticity and all that.

    Reading the full press release they state in it that having selected 28 children “The study results were so strong that a larger sample group was not required to ensure statistical accuracy” but this would only be true if the kids were genuinely selected at random from the whole population. And given how participants for small prospective studies are usually chosen at universities I’d be very sceptical of that being the case.

    1. I took a course in psychological statistics in college. It was taught by an excellent professor and has turned out to be one of the more useful classes I took because it gave me some basis to evaluate all these scientific studies that are reported on. This professor used to do a lot of consulting on studies to help people determine how big the sample size needed to be and the metrics they would use to evaluate the results. One thing he told us was that a large sample size was not needed to get significant results. He said that a study with as little as 6 participants could yield valid results.

      1. I’m not objecting to the size of the sample. I’m questioning the way in which the sample was chosen, which is often far from random in small prospective studies.

        Small studies can certainly yield significant results but that doesn’t mean that they yield true results. The likelihood of a significant result also being true is connected to both the prior probability of the result being true and the inherent bias in the experimental set up. It’s well worth checking out John Ioannidis’s excellent paper on this subject.

        Sstatistical significance is not directly commutative to truth

      2. Significant results does not necessarily equal Valid results.   If significance was the only thing that mattered, there would be no need for sampling theory. 

  9. My sister is a grade school teacher and when she was working on her Masters she had to read a book that I wish to God I knew what the name of it was, because it sounds SOOO interesting. Maybe someone here has read it and will give the name. The book was written by a woman whose husband grew up poor. She grew up middle class. She had a hard time understanding, for example, how her husband would just spend money the instant he got it, so she studied poor people in a kind of anthropological way to try to understand these behaviors and wrote this book about it. From what my sister told me, it was a very similar conclusion that that writer came to as to the authors of this study, that in general people were not reliable and that it drove all kinds of behaviors around money and choices.

Comments are closed.