Return Home

Mischel’s Marshmallows

Monday, March 09, 2009 - 01:04 AM


How are your New Year's resolutions holding out? This might at least help you feel better about them.

Psychologist Walter Mischel explains how one little test involving a marshmallow might tell you a frightening amount about what kind of person you are. And Radiolab favorite Jonah Lehrer helps us make sense of the results. This one's all about our will power (or lack thereof).


Correction: An earlier version of this piece incorrectly stated that the kids who performed better on the marshmallow test had higher GPAs in high school and went to better colleges. Those elements were not a part of Mischel’s original study. The audio has been adjusted to reflect this fact.


More in:

Comments [59]


I remember going to a seminar given by Philip Zombardo (the guy responsible for the famous prison experiment). In the seminar we watched a film showing himself and an associate performing the same "marshmallow experiment." But he used the terms "past, present and future oriented." Is this the same study? I'm confused.

Sep. 18 2013 12:44 AM

The question I wonder about:

How often do the kids get treats at home? A child who doesn't get sugar often might eat the treat right away versus one for whom treats are a daily occurance.

Aug. 25 2012 09:31 PM

Tom and Betsy (and all other attachment parenting fans out there) -

There is absolutely not simple "pick up the crying baby" or "don't pick up the crying baby" rule for every baby, every family, and every situation. I'm a nanny, and I've seen just about every approach and every personality type in a baby (and all babies are NOT created equal when it comes to crying). The only consistent rule of thumb I have ever been able to develop is that you can't have a rule of thumb. You have to pay really, really close attention to the child in question, and to the emotional needs of the adult in question (I've been shocked by how much self-righteousness often seems to be involved in attachment parenting, and suspect that many parents stick to it out of a need to prove themselves right).

I should perhaps point out that I was raised by a mother who absolutely did not come to me when I cried, and I have ended up leaning way over to the other end, where 99% of the time I will go to a crying child. I should also mention that I have the best willpower of anyone I've ever met. The only predictor I can think of that seems plausible in light of this study is whether the parents follow through on promises. If a parent says, "You can have a piece of chocolate if you eat your broccoli," then not only do they need to give the child the promised piece of chocolate, but it MUST come after the child finishes their broccoli. I know some folks don't like the use of bribes with kids, but it's not about bribing; it's about follow-through.

Aug. 19 2012 11:53 AM
Emely from Puerto Rico

I wonder if they've tracked the economic standing of the parents of the children, because I have a feeling that this is another test that shows difference in how children are raised in different social and economic standings.

Mar. 21 2012 02:39 PM

really interesting blog post about this topic (not my blog)

Jan. 06 2011 10:55 AM
leland from Planet Earth

As it happens, it was discovered that kid # 1 refused to eat them because they were "w"s.

Oct. 09 2010 12:11 AM

someone sent me this from youtube... the marshmellow test...

Apr. 19 2010 11:13 PM

Have studies looked for genetic, protein, or metabolic biomarker differances in the kids who had more impulse control VS the kids who had less? I understand about 1 of 3 kids will "Pass" the test.

Mar. 24 2010 01:42 AM

Before I watched the video I just read the post and I was thinging like how in the world can a marshmallow tell what kind of person you are

Feb. 22 2010 11:31 AM

This is my first time ever watching a marshmallow video radio lab or even hearing about a radio lab. It is very interesting to see and hear new things and to know what happend so many years ago and to compare it to now.

Feb. 22 2010 11:25 AM

This is my first time ever watching a marshmallow video radio lab or even hearing about a radio lab. It is very interesting to see and hear new things and to know what happend so many years ago and to compare it to now.

My self-control and discipline comes in waves. RadioLab had done another show (which one I don’t remember… stress maybe?) that illustrated how it’s more difficult to make healthy choices when you are under stress and pressure. I have noticed that when I’m in the middle of exams I want a lot of candy, but over the summer when I don’t have class I eat like a health nut. I wonder if this factor affects the kid’s choices.

Feb. 22 2010 11:24 AM

Thinking off from the previous post, I now have to question the validity of this experiment in gauging a person's will power. The experiment is based on the continual sacrifice of the child's will each moment renewing all for the sake of what we think is the child's increased benefit. We assume the ones who can't wait DO in fact think it is worth waiting for, but that they simply can't. The development of our sense of humanity, of our empathy towards one another, is (at least for me) the most significant force of our will power. This isn't included here at all. Its simply testing a person's selfish willpower. Maybe the failures have a stronger will power through empathetic and selfless values, while the ones who wait tend to be the selfish, self-interested type.

May. 17 2009 08:26 PM

What if the kids who "couldn't resist" are actually the smarter ones? What if they subconsciously calculated that the suffering wasn't worth the extra payoff. We know about the economic law of diminishing returns. The second one isn't as good as the first one, nor is the third as good as the second. The law is relatively new, but not the centuries-old consciousness it describes. Maybe the smartest kids realize its better to enjoy the one treat without any suffering.

May. 17 2009 08:12 PM

SAT tests measure academic learning - not intelligence. It would have been more useful to compare learning capability/intelligence (IQ tests) with study participants to see how much innate intelligence may have had to do with later SAT and life success and earlier will power/coping strategies. If there were any good instruments to measure creativity, that would have been useful as well.

May. 01 2009 03:59 PM

What if the kids just ate right before the study? Better yet, what if they just had something sweet right before the study.

Clearly, that would increase their willpower so that they would be more easily able to resist indulging on the marshmallows immediately.

Apr. 22 2009 09:53 PM

Several folks have expressed, what I believe, is most logical explanation in that the parents role is the determining factor. Specifically, CJ’s comments regarding discipline and Joe’s regarding strict vs. less strict. It is nature (especially in children) to develop cooping mechanisms to deal with limits/expectations/goals/uncontrollable events (whatever they may be). The key, in my opinion, is in setting and enforcing those limits during early development (childhood) to develop “naturally” these coping mechanisms that are used for a lifetime from everything from being inpatient waiting for a stop light to dealing with an aging and ill parent to all the trials of marriage and work.
So for me, to answer the questions of Lorenzo, I’d say.
1. Use this information as proof to the positive impact of limits that are enforced when raising children.
2. You can’t teach a child a distraction for each and every unique challenge of life. Instead stick with #1 and copping mechanism will “inherently” be found when needed for those challenges.
3. Depends on which method you use. If you go about teaching for every obstacle, it seems to me end the end that would be very confusing (let’s see what page of the 5000 covers making a move to another state or which one covered how to deal with my obnoxious boss??).
4. If teachers had children that were less interested in mayhem and knew how to cope with being someplace that wasn’t, in fact, a fun factory, then yes.
The marshmallow test is just one more example of the positive impacts of some type of discipline during childhood so for me, study done.

Apr. 20 2009 10:42 PM
Lorenzo Ortiz

If there is truth to the conclusions reached about the marshmallow test, the question is, "How do we use this information?" Will teaching kids to distract themselves with thoughts or activities? Will it help them in some way? Would teachers be given time to teach real-life skills? I know that in Texas, teachers wouldn't have time because they are preparing kids to pass a standardized test! Study that.

Apr. 19 2009 09:24 AM

There are too many unaccounted for variables, and the definition of success is limited. It does seem more to do with rule following. Also the idea of a marshmallow holding the same value between individuals is weak, while it seems that they all value the marshmallow, 2 marshmallows is not always "better" and it might take some a few minutes to weigh this out...
This seems to uphold a "more is better" model of success (more money=better job ect)but this model is likely what caused the subjects to delay gratification.
It does seem that weighing economic or social class into the equation might be beneficial, but also asking the adult guinea pigs how they define success.

Apr. 18 2009 11:41 PM

I agree with Tom, babies that are responded to,nurtured and loved learn how to trust. In the experiment, more m & m's were promised if they could wait. Maybe a child that could not "delay gratification", really did not "trust" the adult to deliver their promise, because they did not "learn" trust from their own parent. Many kids that live in povery, know that if they don't get it now, they may not get it later. I think SES and trust may have more to do with this experiment.

Apr. 18 2009 10:57 PM

Tom: Interesting. I was just thinking that I might have eaten the marshmallow out of a lack of trust that the tester would follow through with the reward.

Apr. 17 2009 12:06 PM
Kathy in Florida

why do people insist on taking the information so seriously? Can't you just listen for the entertainment value? Quit taking yourselves so seriously.

Apr. 15 2009 12:19 PM

Where's the video that you guys tout as "some of the greatest video every shot!" I'm a month behind in my podcasts, so shoot me?!

What a tease, what a come on - I'm so disappointed! Please relink!

Apr. 11 2009 09:12 PM

On 3/11/09, Mark commented...
"parents should not respond immediately when a child cries during the night, so that the baby learns that the parents will not respond to every noise he or she makes. Is it possible that parents who knowingly or unknowingly adopt this strategy are teaching delayed gratification to their small children, the results of which are seen in the marshmallow experiment?"

Mark (and others) do not follow that advice. Studies show that babies who's cries are ignored, do not learn delayed fact it's the opposite. Babies that are responded to and nurtured, are well fed, loved and develop strong attachments that bolster feelings of confidence and safety. Later, these traits manifest themselves within the child and give them the ability to delay gratification...their needs have have been met--Maslow, psych 101. Babies who's cries are denied, grow up with a world view that their needs will not be met, and graduate into a "feast of famine" response. These kids will gobble up the first marshmallow because they don't trust/know if the other treats will materialize; they don't know when they'll get to eat again.

ALWAYS pay attention to your baby when it cries.* Yes you may not get to sleep as well in the short term, but in the long term, you will raise a far better adjusted child.

*Safety note: Your child also may be crying because it's arm may be caught in a painful or uncomfortable position.

Mar. 30 2009 11:52 AM

I think this test raises larger questions about what we value in an education.

We seem to lean towards control and have traditionally educated and tested along those lines. Students who underperform may do so more due to their inability to fit that mold (the one valuing marshmallow restraint).

An impulsive learning style might well be good, provided the structure of society were different -- witness great minds who cannot fit in, and our means to measure success were more suited towards those with different learning styles.

Mar. 24 2009 04:15 PM

I'll just add to the commentary about the environment that the children were raised in. I believe that children who were raised in a disciplined environment would have done better in the marshmallow experiment. Further, this atmosphere of parental discipline carries forward as the children move through the educational system, and it should be no surprise that they perform better on their SATs and beyond.

I see the correlation. I would challenge the implication that it is a result of being hard-wired. Although there is certainly a set of children who are naturally more self-disciplined, I think that most children raised in an environment of goals and discipline will take that forward throughout school and beyond.

Mar. 22 2009 10:54 AM

Wasn't the video mentioned in this podcast up on the radiolab youtube account? I remember watching it a while back and laughing at the little kids and what they did, as well as the one that cheated and started licking the orea. But i was pretty lost as to what was going on. Now it all makes sense. Radio lab should put that video back up!

Mar. 20 2009 03:28 AM

When are you coming back? I need new episodes. What's with all these breaks?

Mar. 18 2009 09:33 PM

I would like to argue in defense of Jad. A lot of the posters are quick to point out what he did wrong and how his conclusions are flawed. In most of those cases, the listeners may have just not been listening closely enough. First, he clearly states that the information from this study points more in favor of nurture rather than nature being the primary factor in one's ability to exert willpower.

Also, they never claimed that the ability to exert willpower and delay gratification is causative of success later in life; they simply said that such ability is "diagnostic for predicting aspects of cognitive and social competence later in life." (And that quote comes directly from the abstract of Mischel's 1989 Science paper on 'Delay of gratification in children.') Let me say that again: this does not CAUSE better performance, it is merely a PREDICTOR of better performance (in an admittedly post-hoc analysis of the data).

Mar. 16 2009 10:55 PM

I think a few people on this board are onto something else that this piece doesn't mention entirely. Is this experiment testing just will-power? Perhaps some children don't wait for long because the reward isn't very gratifying. Maybe it has to do more with comprehension and following rules. Perhaps you'd get a different response depending on how who is giving the rules. (i.e. a parent or a complete stranger)
I think this is very interesting and would like to hear more about the variations of this experiment.

Mar. 16 2009 04:26 PM

I think that we all are a little more "hard-wired" than we like to think. Some people in neuroscience (Christof Koch) think that most of our behaviors and even most of our *thoughts* aren't really consciously controlled. When we do change, that's very interesting... and maybe even weird. You guys brought this out with by talking about your New Year resolutions. Good job!

Laura's idea about an episode on stats is interesting. Maybe exploring the scientific process itself could make for an interesting episode(s). There's lots of things that go wrong in this process and many more factors that go into research than just a desire to study ants, for example...

I'm also thinking the history/philosophy of science might be interesting for you to tackle. I'd love to hear Robert and Jad argue about Popper, Kuhn, and others.

It's a little different than what you're doing now. You seem to work your way through scientists' stories of their processes to get to the big ideas. This would be the big idea of scientific process. But I think you could do it!

Mar. 15 2009 10:30 AM

Technican issue: The download has the last 45 seconds truncated but the stream has the whole thing. Its a bit annoying because one might think there's a long discussion on it.

Mar. 15 2009 09:09 AM
Mabel Adams

Is it possible that the kids who were destined to do better on SAT tests had minds that were better at coming up with the bag of tricks? The conclusion here is that having the skills to resist temptation led to better test scores, but maybe it is that whatever wiring and training by age 4 dictates better test scores also provided a better "bag of tricks."

Mar. 14 2009 06:13 PM

Where's the video? This ABC one clearly isn't the one mentioned in the podcast.

Mar. 14 2009 04:50 PM
Richard Champalbert

What I want to know (and this was never made explicitly clear in the story) is were the kids ever given the promised reward, or were they all doomed to "fail" eventually?

Do any of the kids realize that the whole thing is a set up, and there is no second marshmallow (or cookie)?

As for the long term effects of will power, what seems apparent to me from the story is that once a child is headed down the path of academic success or failure, continuing down that path appears to be what comes naturally.

I would be curious to know how effective measures would be to changing that path.

Mar. 14 2009 03:43 PM
Benedict Schurwanz

and regarding the "marshmallow temptation extended version" video from YouTube, they're obviously just using that experiment for its religious implication instead of psychology/behavioral study. Same stuff, seen from a different perspective.

Mar. 14 2009 11:26 AM
Benedict Schurwanz

This story raises some dramatic implications in my mind. The idea that impulse control can be correlated to performance in various aspects of life makes a lot of sense to me. How many aspects of success as an adult relate to doing something you may not want to do because it has to be done, or for another reason than what you want right now?

But the most important part, I think, is that impulse control CAN BE TAUGHT. It is NOT hard-wired. I would guess that it's easier to teach to a four-year-old, and the applications of impulse control at age four are much simpler. Obviously some kids are naturally better at coming up with strategies, and some kids have more trouble.

The next step, that's missing here, is studying what effect teaching impulse control in a thorough way to four year olds can help them through the rest of their life, and in what aspects of life. And studying what sort of factors in a child's environment had an effect on their impulse control ability.

I think this idea could have profound implications for the education system.

So who's doing that study?

Mar. 14 2009 11:14 AM

"Statistically" is such a problematic way of looking at things. And to say that statistics show that kids are "screwed" (or set for life) at age 4 is an awfully dangerous suggestion.

We rarely (if ever) remember the constructedness of statistics, and the fact that they only pretend to predict things in many cases--and that part of their predictive power may well be not prediction, but suggestion--self-fulfilling prophecy.

I would LOVE to see Radiolab do a show on this history and development of statistics. It would be fascinating--the development and implications of bell-curve thinking, the way sociology used statistics to make itself seem like a "hard science" and how that's linked to determinism and social darwinism...seriously, guys, after this sketchy bit of not-nearly-skeptical enough reportage, I think you owe it to us.

Mar. 13 2009 11:24 PM
Dale Keiger

Sorry, I see that now. I was fooled because my podcast cuts off in mid-sentence. Apparently, the middle of one of the last sentences.

Mar. 13 2009 02:16 PM

The comment that this example of self control is either "hardwired" or a "bag of tricks" that some kids just were good at leaves out completely the possibility that this relates to the kids upbringing. With my young daughters I've noticed a huge difference in self-control when my wife has been watching the kids (she's strict) vs. when by mother in law has been watching the kids (not so strict). Surely discipline has as much to do with an experiment in self control as innate hardwiring.

Also, my 2 year old can delay gratification on things like treats, so I'm not so sure about 4 yrs being a magic age for that ability to develop.

Mar. 13 2009 09:53 AM

It is only a 15 minute podcast.

Mar. 13 2009 04:37 AM
Dale Keiger

No matter what I do, I cannot get more than 15 minutes of this episode to download. I have unsubscribed, resubscribed, done all I can think of. Is anyone else having this problem?

Mar. 12 2009 10:30 PM

As I understand it, the problem with a Pavlovian approach to trying to teach of a broader skill like patience and delayed gratification is that the specific training elements often don't transfer over into a general practice.

You quite easily can teach kids to be remarkably good at delaying gratification with an oreo in that one specific situation, but so far we just don't seem to have a good sense of how to turn that into a general personality trait of the sort we'd hope would lead to success in all areas of life and study.

Mar. 12 2009 02:58 PM
Sargun Dhillon

Hm, it would be somewhat interesting to use some sort of therapy on the kids during this experiment. For example, make it so the M&Ms change taste over time. You can do this by putting something nasty on them that breaks down easily by UV light... Teach the kids to get delayed reactions. Then introduce them to another food item, and see if the instinct carries over. Of course you'd have to do this at the age level of 2-3 years old, before they're fully aware that they can walk away from the situation.

Mar. 12 2009 07:34 AM
Senor Calvo

I wondered about the role of social class in all this. Being able to hold of the oreos for longer as a child is strongly correlated with educational and job success, being slimmer and healthier adulthood, all things we know are also strongly correlated with social class and parental income and education. So is it that bringing your children up to defer their gratifications (something that is certainly seen here in the UK as a middle class characteristic) is simply another thing to be correlated with social class rather than having any causitive link with success in later life? I'll have to find the original research and see if the predicitive value of deferred gratification holds true across social class.

Mar. 12 2009 05:46 AM

Does anyone know where to find a video of the original Marhsmallow test? The podcast said there'd be a link, but I only see the M&M reenactment.

Mar. 11 2009 06:21 PM

So where can we watch the video of the kids and the cookies?

Mar. 11 2009 04:31 PM

Can you please tell us the music you used?


Mar. 11 2009 02:00 PM

I'm going to disagree with Jad's comment in the podcast that the experiment says that willpower may be "hardwired". I think those that are picking up on the parents probably have it right.

This has caused me to think about parenting advice I've heard for getting babies to sleep through the night. I've been told that parents should not respond immediately when a child cries during the night, so that the baby learns that the parents will not respond to every noise he or she makes. Is it possible that parents who knowingly or unknowingly adopt this strategy are teaching delayed gratification to their small children, the results of which are seen in the marshmallow experiment?

Mar. 11 2009 10:43 AM

to Bill - LOL

Mar. 11 2009 10:17 AM

This skips over the most obvious point: The children which are given attention from their parents.

In schools, the " underachievers", as a group, are in a community where the parents simply do not make the students do there homework. Similarly, a four year old who is not given positive reinforcement for following directions will not follow directions.

I did the experiment with my four year old, and I must admit, I did not think she would do well. She reacted similarly to her reactions to the directions such as" hot stove" or " stop pushing" and waited patiently.

All children have their strengths, and some need to work on things harder than others. If you want a real test of self-control, try asking them not to pick their nose. Oddly, once you mention that, even to adults, they can not stop picking their nose.

Mar. 11 2009 09:33 AM

where are the promised oreos experiments?

Mar. 11 2009 08:56 AM

I like RadioLab, but again this is not good work. Like some posts above I agree the conclusions drawn in the show are lacking in many ways. The explanation gave on the show is one of many, but still the show presents it as absolute. Part of being scientific is being skeptic, pointing to weaknesses in conclusions, and acknowledging there are other possible explanations. In this regard RadioLab does harm to science.

Mar. 11 2009 02:10 AM


I fully agree with Levi's post. Additional data would be great. What kind of models is this study using?

Mar. 10 2009 11:45 PM

My self-control and discipline comes in waves. RadioLab had done another show (which one I don't remember... stress maybe?) that illustrated how it's more difficult to make healthy choices when you are under stress and pressure. I have noticed that when I'm in the middle of exams I want a lot of candy, but over the summer when I don't have class I eat like a health nut. I wonder if this factor affects the kid's choices.

Mar. 10 2009 11:39 PM
Jacob Kramer

What was really being tested in those four year-olds was the ability to think logically and creatively under emotional duress. The children who could keep their goal in mind (logical faculty) and distract themselves (creative faculty) while the treats were tantalizing them were able to hold out.
Willpower, defined as the ability to retain higher brain function under emotional duress, would certainly help anyone succeed in life. It wouldn't be the only factor -- take Robert Krulwich, for example. He claims not to have much willpower, but his logical and creative faculties are obviously powerful enough (at least while he isn't stressed out) to create a body of work that gets him job satisfaction and security.
As a side note, it may be Krulwich's very "lack of willpower" that makes his radio pieces so engaging. "Lack of willpower," understood as "very strong emotional response," is a hallmark of his work. Krulwich's outsized emotional reactions to science help the rest of us think about science as tantalizing.

Mar. 10 2009 05:28 PM
Dan Sack

Bal, I think that preacher's message was terrible. The ten commandments come *after* Israel is "saved" from Egypt. The message is more like: "hey I've helped you (because I love you), so here is how you should respond" rather than: "you have to follow these rules to get to heaven." Anyway, that's all I'll say about religion.

Like Levi said above, It would be interesting to see info on the parents using positive/negative reinforcement. Also concerning parents, what are their eating habits? Some kids may think cookies are worse to eat than others.

Thanks for posting this!

Mar. 10 2009 10:00 AM
Levi Wallach

An interesting show, but the one thing I found missing was how different variables might cause one kid to have more self control where another has less. The thesis seemed to be mainly that the kids who had more self-control had this due to genes and that you MIGHT be able to effect this by behavior "tricks."

But were other things controlled for, specifically parents and parenting styles? What was the willpower, educational levels, etc. of the parents? What kind of parents were they - did they use negative reinforcement, positive, a combination? Bribing and threatening have both been shown in recent studies not to be effective in motivating kids beyond the immediate issue at hand. IE, it doesn't make them truly want them to something because they think it's right, but rather to do it just to get an award (or avoid a punishment) and thus they will go ahead and do it once they think they won't be caught. I'd love to see some additional data around these things...

Mar. 10 2009 07:55 AM

Even as a staunch agnostic, I have to admit the preacher in the "extended " marshmallow video has stumbled on an wonderfully effective metaphor for the pleasure of sin versus the promise of heaven. Unfortunately, his promise of heaven is just double the sin, with a chocolate center! 72 virgins anyone?

Mar. 10 2009 04:39 AM

What's with the "extended" version of the marshmallow temptation video? It's a preacher guy at the end making a religious correlation... weird.

Mar. 10 2009 12:45 AM

Leave a Comment

Register for your own account so you can vote on comments, save your favorites, and more. Learn more.
Please stay on topic, be civil, and be brief.
Email addresses are never displayed, but they are required to confirm your comments. Names are displayed with all comments. We reserve the right to edit any comments posted on this site. Please read the Comment Guidelines before posting. By leaving a comment, you agree to New York Public Radio's Privacy Policy and Terms Of Use.

Supported by