Join 3,438 readers in helping fund MetaFilter (Hide)


Marshmallow Test. Again.
March 1, 2010 12:18 PM   Subscribe

Previously on Metafilter, we discussed the Marshmallow Test. But does the study hold up to scrutiny?

Bronson and Merryman recently published this book.
posted by wittgenstein (61 comments total) 5 users marked this as a favorite

 
But she’s unequivocal about whether the marshmallow test should be used to determine entrance to private schools.

The marshmallow test is used as an entrance metric for private schools? Seriously?
posted by fremen at 12:28 PM on March 1, 2010 [4 favorites]


The article does include some trenchant criticisms of the notion that the marshmallow test should be used to determine private school admissions.

On the other hand, who the hell ever thought the marshmallow test should be used to determine private school admissions?
posted by grobstein at 12:29 PM on March 1, 2010


Private school admissions should be determined by whether the subject thinks using the Marshmallow test to determine private school admissions is a good idea.
posted by qvantamon at 12:36 PM on March 1, 2010 [9 favorites]


It also turns out that whether the child waited thirty seconds or fifteen minutes, there was no difference in long-term outcome for the marshmallow.
posted by lore at 12:42 PM on March 1, 2010 [38 favorites]


Seriously, the whole point of the original study was that that behavior was learned and could be overcome and wasn't a predictor of much of anything but just a correlate with other high-impulse low-self-control behaviors, which also tend to correlate with less academic success.
posted by Scattercat at 12:42 PM on March 1, 2010 [1 favorite]


What does the test say if someone just eats the first marshmallow and then left because they have a bag of marshmallows at home?
posted by kcalder at 12:57 PM on March 1, 2010


Yeah... Using the marshmallow test as an actual predictor would be insane. The original test used only offered the kids a second marshmallow to hold back. But if the parents told the kid it was super important not to eat the marshmallow, plus made sure they were full before hand it wouldn't even be an issue.
posted by delmoi at 12:59 PM on March 1, 2010


I thought the test was to see whether or not you'd choke to death on the marshmallow. I thought I aced that damn test.
posted by mccarty.tim at 1:04 PM on March 1, 2010 [1 favorite]


On the other hand, if certain private schools want to signal their dubious abilities to evaluate children early on, they'll save me a lot of effort in my selection process.
posted by akash at 1:06 PM on March 1, 2010 [4 favorites]


In my lab we felt that this could be a potentially useful class of testing for self-control. A rational individual would find that either 1) eating the marshmallow now is worth more than eating two in fifteen minutes or 2) eating two in fifteen minutes was worth more than eating one now. Either of these options make sense depending on the temporal discounting rate of that particular individual.

What doesn't make sense (in classical economics anyway) is if you have a big block of people who hold out for 10 minutes and then quit. People often try to argue that this implies the existence of some self-control mechanism that is finite and expendable. The problem is that most of the 'aborts' can be accounted for by simply adding a constant probability that the person will switch from not eating the marshmallow to eating the marshmallow. If there was a strange hump around 10 minutes, that would be one thing, but there's usually just a constant trickle from waiting to not waiting which could be explained by having the value of waiting and not waiting being essentially equal and then adding a little noise to the system.

In short, self control probably exists in some form, but it's damn hard to properly demonstrate it in an experiment.

Does this mean that the study didn't hold up to me?
posted by scrutiny at 1:14 PM on March 1, 2010 [4 favorites]


I did the marshmallow test when I was really young, about three or so, and I think I messed up some well-meaning psychologist's data. (I think the proffered treat may have been Rice Krispie treats instead of marshmallows, but I can't recall for sure.) I spent the waiting period sitting quietly, not touching the marshmallow/Rice Krispie treat. When the grown-up came back in and offered me the second treat, I replied "oh, no thank you" and that was that.

After all, I've never really liked marshmallows or Rice Krispie treats. Too sticky and puffy.
posted by Metroid Baby at 1:19 PM on March 1, 2010 [24 favorites]


The marshmallow test used for fraternity admissions is really different.
posted by GuyZero at 1:20 PM on March 1, 2010 [7 favorites]


Scattercat has it.
posted by mrgrimm at 1:24 PM on March 1, 2010


A lot of the New Yorker articles on business topics (including the one where this study on marshmallows was reported) are fizzy, frothy trivialities that get taken waaaay too serious, like Malcolm Gladwell's hair.
posted by KokuRyu at 1:29 PM on March 1, 2010 [2 favorites]


Scattercat has it.

He has? Gimme the marshmallow, Scattercat!

*scuffle scuffle scuffle*
posted by davejay at 1:30 PM on March 1, 2010 [3 favorites]


My reaction to this kind of test has always been "fuck you". Seriously, since I was five.

"If I eat the marshmallow now can we go do something else less manipulative?" would probably be a more complete articulation of my feelings.

We even did something very similar as some kind of demonstration on the nature of sin when I was at bible camp. (part of what made me the atheist I am today.) I went ahead and ate the candy immediately and publicly. When the teacher asked me why I said "because this doesn't seem like a very fun game".
posted by poe at 1:39 PM on March 1, 2010 [12 favorites]


I use this test all the time in my hiring interviews.

You'd be surprised how many adults fail this. (I've failed it many times myself in my twenties, but I'm wiser for it in my thirties)
posted by blahblah at 1:44 PM on March 1, 2010 [1 favorite]


You'd be surprised how many adults fail this.

I'm really surprised because it's not a pass/fail test.
posted by GuyZero at 1:49 PM on March 1, 2010 [9 favorites]


I think I would fail the test out of distrust rather than the inability to delay gratification. You're always told not to take candy or cookies or whatever from strangers, and then you have this stranger offering you marshmallows or cookies, and THEN he says, "You can have double if you wait for me."

I'd be sitting there thinking, "What if he comes back with a big axe and chops off my head?"

I'd probably run screaming out the door.

Hopefully, I'd have the sense to grab the bowl of cookies first.*

*That is, if I didn't think they'd been poisoned. Told you I have a problem with trust.
posted by misha at 1:51 PM on March 1, 2010


Do you only hire people who eat the marshmallow straight away, because the ones who wait are obviously greedy and therefore untrustworthy, and also possibly gluttonous and sinful?
posted by dng at 1:52 PM on March 1, 2010 [2 favorites]


I use this test all the time in my hiring interviews. You'd be surprised how many adults fail this.

what the fuck i would rather eat my own hands than go through the "marshmallow test" for a job

fucking wal-mart has more respect for their employees than that
posted by Optimus Chyme at 1:57 PM on March 1, 2010 [17 favorites]


Chubby Bunny
posted by Kabanos at 2:09 PM on March 1, 2010 [4 favorites]


Rice Krispie treats are made with marshmallows. There's a point in there somewhere.
posted by longsleeves at 2:12 PM on March 1, 2010


I don't like marshmallows. So yeah, offer me two if I don't eat one? I'll take none, thanks, and what does that say about impulse control?
posted by sandraregina at 2:17 PM on March 1, 2010


Does this mean diabetics get automatic admission to all private schools?
posted by blue_beetle at 2:42 PM on March 1, 2010


Qu’elle surprise! The behaviourists are wrong again, and (again) resort to statistical tricks and misinformation to sell their populist science-lite books. Those darn humans are /always/ messing up the pretty graphs!
posted by clvrmnky at 2:42 PM on March 1, 2010 [1 favorite]



>Scattercat has it.

He has? Gimme the marshmallow, Scattercat!

*scuffle scuffle scuffle*


Nooo! Mine! Mine! *thump thump crash*

...my precious, yes. We eats it now, precious...
posted by Scattercat at 3:07 PM on March 1, 2010 [3 favorites]


I still trying to figure out how you use the marshmallow test in a job interview. Well and not just have every semi-qualified applicant just go home in disgust during those 15 minutes.
posted by aspo at 3:12 PM on March 1, 2010


I still trying to figure out how you use the marshmallow test in a job interview. Well and not just have every semi-qualified applicant just go home in disgust during those 15 minutes.
posted by aspo at 6:12 PM on March 1 [+] [!]


Well, you wouldn't use marshmallows, of course. What you'd do is you'd tell them to show up at like 10, but then you'd have the interviewer running behind so they'd have to have enough self control to wait for fifteen minutes, and -- wait a second ...
posted by Comrade_robot at 3:17 PM on March 1, 2010


Who was the genius that picked marshmallows instead of something like... candy? I bet half the kids who held out didn't really like to eat marshmallows anyway.
posted by digsrus at 3:21 PM on March 1, 2010


Really? I mean I know that work places don't run like clockwork, and I've waited for plenty of interviews, but if I even caught a hint that someone is playing stupid power games like that when interviewing me I'd only take the job if I desperately needed the money. Oh, and I'd be looking for a new job from day one.
posted by aspo at 3:25 PM on March 1, 2010 [2 favorites]


Oh, I don't know really, but if people were to do it, that's one way it could look.
posted by Comrade_robot at 3:34 PM on March 1, 2010


Wow, there's a lot of pathetic posturing going on in this thread. Bravo, you strong-minded individualists.
posted by waxbanks at 3:37 PM on March 1, 2010


clvrmnky: "Qu’elle surprise! The behaviourists are wrong again, and (again) resort to statistical tricks and misinformation to sell their populist science-lite books."

Yea, those silly behavioral scientists, repeating lies and selling books. Oh wait, it's the scientists who are publishing the data used to refute them in Predicting Cognitive Control From Preschool to Late Adolescence and Young Adulthood, and Bronson & Merryman the ones selling a book. And on actually reading the study, it's not clear that Eigsti's work refutes any previous findings unless you assume SAT scores equal IQ scores. Rather, the study is interested in whether your strategy for not eating the marshmallow (direct your attention away from it or not) predicts response time on a drag race style 'go/no-go' test.

I have no clue who is promoting a marshmallow test of consequence, but I've seen no such thing from behavioral scientists.
posted by pwnguin at 3:42 PM on March 1, 2010


aspo: I still trying to figure out how you use the marshmallow test in a job interview. Well and not just have every semi-qualified applicant just go home in disgust during those 15 minutes.

Obviously not with marshmallows. Still, it seems like a bad idea; amateur psychological experimentalists are likely to have their subjects pick up on being gamed, and having the more dignified / ethical ones walk off in disgust, where the more amoral / antisocial ones will simply game you back.

Oh, and if it really does just involve making people wait for you as Comrade_robot suggested, you're a lot less likely to find out who has self-control and much more likely to find out who scheduled something really important after the interview.
posted by Mitrovarr at 3:44 PM on March 1, 2010 [1 favorite]


Who the fuck wants one *or* two marshmallows? That's not a fucking meal. It's a test of conformity and eagerness to please authority.
posted by drjimmy11 at 3:48 PM on March 1, 2010


I guess if you want someone who will grovel before authority, fail to take advantage of opportunities, and blindly believe vague promises, this is a good way to get your man though.
posted by drjimmy11 at 3:51 PM on March 1, 2010


IF ONLY IT WERE A MEAL
posted by grobstein at 3:52 PM on March 1, 2010


This is the dumbest thing since the supposed Microsoft test of "what color is the wall behind you" and if you look you fail. That one rewards a) people who notice inane useless details and b) people who take wild guesses and ignore the easy, foolproof way to find out the correct answer.
posted by drjimmy11 at 3:53 PM on March 1, 2010


Oh, it's the marsmallow test again, is it?

*breaks out camp stove, graham crakers, chocolate bar*

Yeah, the last guy didn't want a creative solution to this either...
posted by InfidelZombie at 3:54 PM on March 1, 2010


And finally, private school admissions should continue to be based on the same criteria they have been since time immemorial: how much money the kid's parents have.
posted by drjimmy11 at 3:55 PM on March 1, 2010


[The school psychiatrist] gives me a chocolate Easter bunny. And this shows how tricky those guys are. I eat the chocolate and I think, wait a second... this isn't around Easter. "Was this a test?" He said, "Yes." "And what does it mean?" He said, "Well, had you eaten the ears first you would have been normal; had you eaten the feet first you would have had an inferiority complex; had you eaten the tail first you would have had latent homosexual tendencies; and had you eaten the breasts first you would have had a latent oedipal complex."

I said, "Well, go on. What does it mean when you bite out the eyes and scream, 'Stop staring at me!'?'" He says, "It shows you've a tendency towards self-destruction." I said, "What do you recommend?" He says, "Go for it!"


--Emo Phillips
posted by Horace Rumpole at 3:56 PM on March 1, 2010 [3 favorites]


Who the fuck wants one *or* two marshmallows? That's not a fucking meal. It's a test of conformity and eagerness to please authority.

If you were testing adults, or if the experimenter stayed in the room with the kids, then I'd argue that yes, you're probably right. However, with kids that young, experimenter-driven demand effects virtually disappear if the experimenter is not present.
posted by scrutiny at 4:00 PM on March 1, 2010


With regards to hiring someone, this could be seen as an general "rule-of thumb" test of temporal discounting which would make it appear valuable.

The problem is that doing any sort of test like this in an interview will fail miserably because people are not acting and choosing how they normally would. They'll pick the things that they think would make you want to hire them. Obviously the test wouldn't be about marshmallows, it'd be about some investment opportunity for the company, yada yada yada, but even there the problem is that there is no right answer. If you only hire people who give one sort of answer - assuming that everyone is being totally honest and not padding their answers at all - you'd quickly have a work environment that was full of the same sorts of people and it probably wouldn't be very effective.
posted by scrutiny at 4:07 PM on March 1, 2010


I was never fond of plain marshmallows as a kid (still not), so I imagine I'd have sat there and stared at it a while before picking it up and playing with it like clay, which is what I did with most marshmallows given to me as a child. Now put it between some graham crackers with chocolate, and I'm all over it, but plain marshmallows? Yuck.

Seems a ridiculous way to determine entrance to a school too. Too many variables would go unaccounted for. Maybe the kid doesn't like marshmallows and doesn't want one or two of them. Maybe the kid is full. Maybe the kid's parents don't allow sugary foods for the kid, and he's not going to eat it only to make parents happy. What if the family is vegan? Or the household has diabetics in it, so there aren't sweets. So forth and so on. Sure, interesting to see how a kid would react, but I don't think it's going to tell anyone too terribly much about how a small child will turn out later in life.

And any employer who I even suspected was playing games with me would be an employer I wouldn't want to work for.
posted by Orb at 4:09 PM on March 1, 2010


I just bought and ate a bag of marshmallows, and I don't even like them very much.

Will people please posting on this subject?
posted by sebastienbailard at 4:10 PM on March 1, 2010 [1 favorite]


"We need a test that will help us weed out kids who are distrustful of authority, and kids who can't just sit still for long periods of time, doing nothing. Bonus points if the test can claim to test something else, like achievement or intelligence. Double bonus points if the test can help me use up this big ass bag of marshmallows."
posted by 23skidoo at 4:21 PM on March 1, 2010 [3 favorites]


If anyone ever tries this on me, I'm going to take out my cell phone, put it on my head, and then take notes in my Moleskine on their response. I'll then furrow my brow, and say, "I'm sorry. I just don't think it's going to work out. You're clearly a sociopath."
posted by mccarty.tim at 4:21 PM on March 1, 2010


Nooo! Mine! Mine! *thump thump crash*

You know, if you'd let me have it, I would have given you two back fifteen minutes later...

---

By the way, my kids go to a preschool, and we just found out they sometimes bribe the kids with marshmallows for good behavior. So I sat down with them, and explained that marshmallows are fake food, so they can have one, or maybe two, but seven might be too many -- so they should decide what's more important, another marshmallow or getting sick from too many marshmallows. I also told them to say "no thank you" if they decide to turn down a marshmallow, but how many they eat and how many they turn down is up to them.

The next day, upon being picked up, their teacher told my wife "Oh, your kids are so smart! When we tried to give them marshmallows, they said 'no thank you' and wouldn't eat them!"

I suppose if they were given this test, they'd probably accept the first marshmallow right away and tell the tester that they weren't interested in a second marshmallow, thanks. I wonder what the test results would say.
posted by davejay at 4:50 PM on March 1, 2010 [1 favorite]


Man, I really want marshmallows, now. Not the store brand, either, or just any bag of marshmallows. I'm talking, big, puffy, fresh marshmallows that are so soft they're like silk in your mouth. Fresh from the factory.

But if I buy a bag, do I fail? What if I make my own? Or marshmallows' delicious cousin, divinity? Or Pavlolvas?

According to the behavioralists, I should be obese.
posted by mccarty.tim at 4:55 PM on March 1, 2010


Davejay, turning down future marshmallows means that you are an underachiever. You lack the ambition to get what you deserve. Greed is good, and all that nonsense. Certainly won't get you a job in Wall Street.
posted by mccarty.tim at 4:58 PM on March 1, 2010


Isn't the twin test of private schools basically the bank accounts of the parents and the blind obedience of the students? (usually the former more than the latter?)

When I worked HR, we didn't bother with the marshmallow test, we had much stricter requirements - like whether you bothered to put your name and contact information on your resume, whether you were the type of person to call HR every 3 days screaming about why don't you have a job yet, or to mail us naked photos of yourself with handscrawled letters about being the reincarnated King of England.

The nice thing about working for the state was that people tend to self select. The not so great part was you still had to officially shuffle them through the system as if they were legit...
posted by yeloson at 5:07 PM on March 1, 2010


I was kidding folks - I think it's just be weird to put a marshmallow in front of an job interviewee in walk out of the room.
posted by blahblah at 5:13 PM on March 1, 2010


"I was kidding folks - I think it's just be weird to put a marshmallow in front of an job interviewee in walk out of the room."

Yeah, but after that thread, don't you want to try it now? Just once?
posted by verb at 5:21 PM on March 1, 2010


Seems a ridiculous way to determine entrance to a school too. [. . .]

GUYS! The "school entrance" thing is almost certainly FUD. No one is using or considering using this test for a school entrance exam. Any suggestion thereof is either a joke or a ridiculous strawman.
posted by grobstein at 5:50 PM on March 1, 2010


In my experience, most schools are run by ridiculous strawmen.
posted by Faint of Butt at 6:13 PM on March 1, 2010


I love we're all so fucking self-absorbed and exhibitionistic that this has turned into the "MeFites demonstrating their particular quirks about marshmallows, psych testing, and self-control" storytelling hour.

BUT I HAVE A MARSHMALLOW IN MY PANTS!
BUT REFINED SUGER IS BAD FOR GROWING CHILDREN!
BUT MY MOMMY TOLD ME NOT TO TAKE CANDY FROM BEHAVIORISTS!
BUT I DON'T LIKE MARSHMALLOWS SO I FOOLEDED THEM!
posted by Joseph Gurl at 6:23 PM on March 1, 2010 [2 favorites]


That's some pretty weak scrutiny.

"First, it’s the easiest test in the world to fool."

First, fooling the test isn't a criticism because fooling it at the time wasn't an issue. It's an issue for its application, yes, but not for the underlying results (i.e. delayed gratification/impulse control) predicting success. Using it for screening purposes is ludicrous anyway.

"Second, reportage of the marshmallow study has obfuscated just how few kids were included in Mischel’s analysis."

Second, I think you misunderstand statistics and the concept of power/significance.

"So all this hype about the Marshmallow Task’s incredible ability to determine SAT scores comes from a handful of tots who were hungry"

They were all hungry kids? Is that some kind of argument for sampling bias?
posted by tybeet at 6:43 PM on March 1, 2010 [1 favorite]


The idea to use this test to screen private school applicants comes from this article. It's not any particular school doing it, just one columnist who thinks it would be better than the current policy: using IQ tests.
posted by Danila at 7:56 PM on March 1, 2010


Maybe I'm just an old fuddy-duddy...but whatever happened to reading someone's resume to determine whether their indicated skills and experience were appropriate to the job in question, interviewing the applicant to get a sense of (a) how well they might fit in to your business environment and (b) whether they're BS-ing on their stated skils and experience, then deciding to hire them or not based on those grounds?

P.S. - I'm thinking of hiring some kids to STAY OFF MY LAWN. If it works I'll pay 'em in marshmallows.
posted by Greg_Ace at 8:52 PM on March 1, 2010


While reading this I couldn't stop thinking about the cupcake dog revelation.
posted by Demogorgon at 9:35 PM on March 1, 2010


« Older Richard Dawkins has been tweeting on the controver...  |  Jump On Contact is a blog abou... Newer »


This thread has been archived and is closed to new comments