Re: Marshmallow redux

1

I don't think the marshmallow test was so publicized because of it was cute and pithy. I think it was so publicized because it let rich people say, "I'm successful because I'm better and people who aren't successful have failed because they lack self-restraint."

Good luck with the needle.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:32 AM
horizontal rule
2

You're saying the marshmallow study was soft science?


Posted by: Mossy Character | Link to this comment | 05-30-18 7:33 AM
horizontal rule
3

People keep posting really long articles. If only my parents hadn't given me so many marshmallows, I could read them.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:36 AM
horizontal rule
4

1: Can't it be both?


Posted by: heebie-geebie | Link to this comment | 05-30-18 7:42 AM
horizontal rule
5

If you want.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:43 AM
horizontal rule
6

I saw someone quip something like "The real marshmallow test was seeing if social scientists were disciplined enough over a 30 year period to refrain from treating as real a juicy result that had not been shown to be replicable. They failed."

I don't know enough about social science to know if that's actually funny.


Posted by: Robert Halford | Link to this comment | 05-30-18 7:53 AM
horizontal rule
7

Can I also just say that the "kittens" in the OP 100% definitely sounds like a euphemism.

I'm not advising replacing the word "kitten" with the word "penis" in the OP but if you do the results are interesting.


Posted by: Robert Halford | Link to this comment | 05-30-18 7:54 AM
horizontal rule
8

Sweater kittens.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:55 AM
horizontal rule
9

I don't know enough about social science to know if that's actually funny.

A little bit.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:56 AM
horizontal rule
10

I have a joke that I think works better when I pretend I don't get it, although it's been explained to me over the course of presenting it this way. I like to tell it as though I'm just the messenger, repeating garbled science that other scientists seem amused by.

What does the H. stand for in Jesus H. Christ?
Haploid.


Posted by: heebie-geebie | Link to this comment | 05-30-18 7:57 AM
horizontal rule
11

That's a classic.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:58 AM
horizontal rule
12

Yeah! Yeah! Let's go scribble kittens all over every page of our high school textbooks!


Posted by: heebie-geebie | Link to this comment | 05-30-18 7:59 AM
horizontal rule
13

Have you considered taking Ambien for the pain?


Posted by: Moby Hick | Link to this comment | 05-30-18 8:01 AM
horizontal rule
14

I can't read the paper right now but that doesn't sound like a replication failure if "rarely statistically significant" refers only to the effect after inserting a bunch of statistical controls that weren't in the initial study. I guess it depends on what you decide as your threshold for "replication failure" but a same-direction effect size reduced by half doesn't obviously sound like one to me. That a correlation gets smaller when you control for a bunch of stuff that would likely be related to both variables is just to be expected. If anyone ever argued that ability to delay gratification was some causal factor that existed entirely independently of cognitive ability, the home environment, etc. then I suppose it disproves that, but I'm not aware of anyone doing so.


Posted by: Tia | Link to this comment | 05-30-18 8:03 AM
horizontal rule
15

They used the marshmallow test as a hook to attempt to teach self control skills--

"Renowned psychologist Walter Mischel, designer of the famous Marshmallow Test, explains what self-control is and how to master it. "

that the actual results can get controlled away is pretty relevant to the whole deal

https://www.amazon.com/Marshmallow-Test-Self-Control-Engine-Success/dp/0316230863/ref=tmm_pap_swatch_0?_encoding=UTF8&qid=&sr=&dpID=51fY-krMLCL&preST=_SY344_BO1,204,203,200_QL70_&dpSrc=detail


Posted by: lemmy caution | Link to this comment | 05-30-18 8:09 AM
horizontal rule
16

14: But everyone knows that a set quantity of grit is attached to the soul at conception.


Posted by: peep | Link to this comment | 05-30-18 8:10 AM
horizontal rule
17

The reason that some kids don't eat the marshmallow is that gritty marshmallows are gross.


Posted by: heebie-geebie | Link to this comment | 05-30-18 8:11 AM
horizontal rule
18

I agree with 15.

But it does sound like the study was a re-analysis, not a replication. Which is a good thing to do because models get misspecified and analysts need money to buy land for cob cabins.


Posted by: Moby Hick | Link to this comment | 05-30-18 8:11 AM
horizontal rule
19

17: The organic marshmallows made from real marsh are both gritty and phenomenally good.


Posted by: Moby Hick | Link to this comment | 05-30-18 8:12 AM
horizontal rule
20

Kitties hooray!


Posted by: Barry Freed | Link to this comment | 05-30-18 8:16 AM
horizontal rule
21

I notice from watching very silly French sci-fi show Missions that the marshmallow experiment (or rather, a massively dumbed down version of it) is now media shorthand for establishing that your protagonist is a psychologist.


Posted by: Ginger Yellow | Link to this comment | 05-30-18 8:31 AM
horizontal rule
22


Meh, self-control can be a thing that's valuable, and can be taught, and is highly related to and/or determined by things like cognitive skill and home environment. I mean, if you showed that scores on math tests at age 8 predicted a bunch of life success factors at age 20, but that this effect was greatly reduced when controlling for home environment and cognitive skills (which it certainly would be!), it wouldn't mean that math skills weren't independently valuable and teachable and didn't have a causal role in good outcomes.

These pop theoretical books by psychologists are usually some mix of "here are my research findings" and "here are some beliefs I have about the world". I definitely would like people to treat all this stuff with slightly more skepticism. Maybe people will have less appetite for them now, which is fine. It remains that "replication failure" ought to have a specific meaning, and it shouldn't be "correlations become lower when you control for a bunch of stuff." The first says "this finding may have always been spurious." The second says "there is more than one way to understand this finding."

18.2 makes sense; I was a little surprised there was a decade plus replication going on.


Posted by: Tia | Link to this comment | 05-30-18 8:37 AM
horizontal rule
23

So, I looked at the actual article. They did use a different dataset and are thus replicating.


Posted by: Moby Hick | Link to this comment | 05-30-18 9:15 AM
horizontal rule
24

Obviously self-control is a thing. Laura Branigan couldn't have had hers taken if it weren't. I'm sure it can be taught in the sense of "this is a time in which you can profit by delaying gratification." But I have no idea if it can be taught in the sense of "I know I would be better off delaying gratification here, but I just don't have it in me."


Posted by: Moby Hick | Link to this comment | 05-30-18 9:19 AM
horizontal rule
25

YAY KITTIES!


Posted by: Anonymous Troll | Link to this comment | 05-30-18 9:32 AM
horizontal rule
26

24: Well, little creative tricks can certainly be taught - "Why don't you try distracting yourself instead of staring at it with your mouth watering?" - which is how schools find themselves in the knot of pretending to teach resilience.


Posted by: heebie-geebie | Link to this comment | 05-30-18 9:58 AM
horizontal rule
27

Kids today are going to have less self control anyway because of the violent video games and all the trespassing on my lawn.


Posted by: Spike | Link to this comment | 05-30-18 10:03 AM
horizontal rule
28

So, I looked at the actual article.

Isn't that cheating? Sounds to me like the behavior of a non-marshmallow eater.


Posted by: politicalfootball | Link to this comment | 05-30-18 10:05 AM
horizontal rule
29

27: Don't forget all the porn they're getting at grocery stores.


Posted by: Minivet | Link to this comment | 05-30-18 10:57 AM
horizontal rule
30

The list said 'quorn".


Posted by: Moby Hick | Link to this comment | 05-30-18 12:09 PM
horizontal rule
31

Which grocery stores? Asking for a friend.


Posted by: Spike | Link to this comment | 05-30-18 12:31 PM
horizontal rule
32

QuornHub


Posted by: Barry Freed | Link to this comment | 05-30-18 12:34 PM
horizontal rule
33

Did your friend lose his phone, computer, Kindle and library card?


Posted by: Moby Hick | Link to this comment | 05-30-18 12:41 PM
horizontal rule
34

What is happening in the libraries? We have librarians here, you know.


Posted by: Robert Halford | Link to this comment | 05-30-18 12:43 PM
horizontal rule
35

I am biased so as to believe everything can be taught, but that doesn't mean it can be taught so effectively in an artificial intervention that it overwhelms all the other influences in the person's life. Although I ought to hope that it can because as of tomorrow I'm about to resume working as a therapist for a subclass of addictions, which is more or less exactly that. I suppose what I actually mean is that teaching children to visualize marshmallows as puffy clouds, or even more generally teaching them one method of regulating their response to a hot, tempting stimulus, is not likely to be the single transformative intervention that changes the course of their life.


Posted by: Tia | Link to this comment | 05-30-18 3:07 PM
horizontal rule
36

I suppose what I actually mean is that teaching children to visualize marshmallows as puffy clouds, or even more generally teaching them one method of regulating their response to a hot, tempting stimulus, is not likely to be the single transformative intervention that changes the course of their life

But I suppose I should add that I'm unaware that anyone ever made a claim so dramatic, or near it.


Posted by: Tia | Link to this comment | 05-30-18 3:09 PM
horizontal rule
37

Tia!


Posted by: teofilo | Link to this comment | 05-30-18 4:38 PM
horizontal rule
38

hot, tempting stimulus

IYKWIM


Posted by: Kreskin | Link to this comment | 05-30-18 5:06 PM
horizontal rule
39

35: I have no doubt most things can be taught. I have enormous doubts about what you can learn about the future by bothering 3-year-olds.


Posted by: Moby Hick | Link to this comment | 05-30-18 5:17 PM
horizontal rule
40

You don't learn much about the future, but you do IME learn a lot about Transformers: Rescue Bots.


Posted by: Robert Halford | Link to this comment | 05-30-18 5:29 PM
horizontal rule
41

That one sucked.


Posted by: Moby Hick | Link to this comment | 05-30-18 5:30 PM
horizontal rule
42

||

Bay Area peoples. On June 16th, I will be one of the storytellers at TEACH's annual fundraiser in SF. The tickets are pretty pricey ($100 each), so I would never expect anyone to pay that or go to the show, but maybe you wanted to spend a lot of money supporting abortion access anyway.

In case you did, and it would make a difference if you internet-knew someone in the show, here's your chance.

|>


Posted by: Megan | Link to this comment | 05-30-18 5:41 PM
horizontal rule
43

Ribs do seem like they would be a painful spot.


Posted by: apostropher | Link to this comment | 05-30-18 5:55 PM
horizontal rule
44

Teo!

39: Sure. I mean, that's just a subset of the general observation that psychology is a field with weak theory and small effects. Which, fair. But maintaining 50% of the effect size in a conceptual replication (I don't know what they did and I'm not just being lazy -- I would love any reason to procrastinate, even including actually reading this article; I don't have journal access right now) doesn't make this finding seem worthless to me. I think there's some kind of disconnect between the level of predictive power that people from other disciplines expect and what research psychologists are willing to find interesting because they know their theories are weak and their effects are small and they'll take what they can get. Reading the abstract more carefully, I see the claim is more like, academic effects replicate, but effect size is smaller, but social-behavioral effects don't. It remains striking to me that you can get a relationship at all between a single test, not straightforwardly academic, administered at age 4 and achievement measures at 15 -- precisely because there are so many influences on a kid's life.


Posted by: Tia | Link to this comment | 05-30-18 6:08 PM
horizontal rule
45

Until you measure them at three, switch them to different parents for ten years, and then measure them again, I'm going to assume it is spurious.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:01 PM
horizontal rule
46

Randomly-selected different parents.


Posted by: Kreskin | Link to this comment | 05-30-18 7:09 PM
horizontal rule
47

You put the parents in a big, wire basket. Spin it around a few times and then grab a set.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:11 PM
horizontal rule
48

Also, give the kids Ambien and see if they turn out racist.


Posted by: Moby Hick | Link to this comment | 05-30-18 7:40 PM
horizontal rule
49

Give the kids different doses of Ambien and see how racist they turn out.


Posted by: foolishmortal | Link to this comment | 05-30-18 9:56 PM
horizontal rule
50

||

Further to our discussion of overdrafts a while back, the UK FCA today has proposed "for discussion" to ban fixed overdraft fees and to require banks to warn people by text/app notification that they'e about to go overdrawn or that they have had a transaction refused for lack of funds.

Taking this into account we intend to model a package of measures which would aim simplify prices and place a degree of control over unarranged overdraft prices. These are:
• A ban on all fixed fees including daily, monthly and allowed payment fees, for arranged and unarranged overdrafts. This would not include refused payment fees.
• Arranged overdrafts to be charged using a single interest rate on each individual account. This could vary for different account types, or even different customers holding the same account, but could not have different tiers within a single account.
• Introduction of a rule to require firms to provide a representative APR advertising of arranged overdrafts, as currently required for other forms of consumer credit.
• Alignment of arranged and unarranged prices. Unarranged overdrafts are also to be priced using a single interest rate, no higher than a fixed percentage uplift of the interest rate for arranged overdrafts. We will carry out further work to determine what this uplift should be or whether unarranged should be no more expensive than arranged.

|>


Posted by: Ginger Yellow | Link to this comment | 05-31-18 2:20 AM
horizontal rule
51

There is no minimum safe dosage of Ambien established not to cause racism.


Posted by: Walt Someguy | Link to this comment | 05-31-18 4:21 AM
horizontal rule
52

OT: If you attended an institution of higher learning with a single-minded devotion to including and capitalizing the definite article in the name of the university, should you do that in your résumé?

Asking for a friend.


Posted by: Moby Hick | Link to this comment | 05-31-18 6:55 AM
horizontal rule
53

heebie:

please be careful with tramadol.

they gave it to me after surgery, but the nurses said it's never used for extended pain relief because after a couple days it starts to kill your kidneys.


Posted by: joel hanes | Link to this comment | 05-31-18 7:01 AM
horizontal rule
54

She has four kids. No shortage of kidneys out there.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:06 AM
horizontal rule
55

I don't think those eight kidneys are ripe yet.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:09 AM
horizontal rule
56

52 - I would only do that if I were sending a resume to The state of The University's origin and thus had a decent shot at a positive "people this will please/people this will vaguely annoy" ratio.

If its a video resume that has you turning into the camera and announcing your college of origin, you have only one choice.


Posted by: Robert Halford | Link to this comment | 05-31-18 7:11 AM
horizontal rule
57

Say "The" Ohio State University while making air quotes?


Posted by: Moby Hick | Link to this comment | 05-31-18 7:12 AM
horizontal rule
58

Is there another, different, Article-less University in the same State?


Posted by: Mossy Character | Link to this comment | 05-31-18 7:13 AM
horizontal rule
59

Speaking of video resumes, I remembered recently the "what is the definition of success" video resume guy from like 2007 and it turned out he died a few years ago. It's tragic because he could have had a cabinet position in the Trump administration.


Posted by: Robert Halford | Link to this comment | 05-31-18 7:15 AM
horizontal rule
60

58: Nobody knows because it's hard to recall anything between Zanesville and Wheeling.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:17 AM
horizontal rule
61

58: There are many universities in that fair state, but only one with its own special article.


Posted by: peep | Link to this comment | 05-31-18 7:18 AM
horizontal rule
62

I wonder how many people apply to Miami University thinking that it's in a somewhat warmer state.


Posted by: peep | Link to this comment | 05-31-18 7:20 AM
horizontal rule
63

Miami was a university before Florida was a real estate investment scam.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:22 AM
horizontal rule
64

There are many universities in that fair state, but only one with its own special article.

"From 1995 to 1998 I attended Some University in Ohio".


Posted by: ajay | Link to this comment | 05-31-18 7:26 AM
horizontal rule
65

61: wrong! There's also The College Of Wooster


Posted by: Cryptic ned | Link to this comment | 05-31-18 7:36 AM
horizontal rule
66

Not to be confused with The Sauce.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:37 AM
horizontal rule
67

Oh wait, that's not a university.


Posted by: Cryptic ned | Link to this comment | 05-31-18 7:39 AM
horizontal rule
68

Or Jeeves Seminary, where one studies Butlerian Jihad.


Posted by: Ginger Yellow | Link to this comment | 05-31-18 7:40 AM
horizontal rule
69

So, while googling on this, I learned that "The University of Toledo" is a thing. Ohio is a leader in definite articles.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:41 AM
horizontal rule
70

And of course the college that would really benefit from a "The" in front of the name doesn't have one.


Posted by: Cryptic ned | Link to this comment | 05-31-18 7:43 AM
horizontal rule
71

"The University of [PLACE]" is pretty standard.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:44 AM
horizontal rule
72

Except at the beginning of a sentence isn't here.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:47 AM
horizontal rule
73

70: Nah. People would think it's a trade school.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:47 AM
horizontal rule
74

70 is where the aliens are.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:49 AM
horizontal rule
75

"The University of Chicago sucks donkey balls." No problem.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:49 AM
horizontal rule
76

@#7: I too perplexedly thought about what kittens could be an euphemism for finally actually visualizing heebie putting kittens (actual ones) on various parts of the body for some some potentially deviant purpose (I moved past that picture quickly) before remembering the discussion of tattoos (which is still kind of scary).


Posted by: zb | Link to this comment | 05-31-18 7:49 AM
horizontal rule
77

74 is super fucking racist.


Posted by: Mossy Character | Link to this comment | 05-31-18 7:51 AM
horizontal rule
78

Sorry. Extraterrestrials.


Posted by: Moby Hick | Link to this comment | 05-31-18 7:52 AM
horizontal rule
79

The marshmallow test has clearly been overhyped as a simple intervention (as have many other versions of simple, cheap interventions that are sold to have huge effects). Wouldn't it be fabulous if we could change the course of a child's life (and, society's development) by teaching them to resist a marshmallow, praising effort not ability, having them stand in power poses?

I would also, though, like to distinguish among the different results that get cast as "non-replication." There's actual fraud, sloppy science and statistics (i.e. p-hacking, dismissing outliers, exaggerating results), discovering potential hidden moderators (and not just presuming them or using potentially problematic statistical corrections), effects that decrease in size on replication. Not differentiating results in a dismissal of data on all the hard problems, and the data is sometimes relevant, though conclusive.


Posted by: zb | Link to this comment | 05-31-18 7:55 AM
horizontal rule
80

I have to say I hadn't come across the marshmallow test as an intervention before, as opposed to something with predictive power, but that's probably just because of the circles I move/read in. I don't read a whole lot about early childhood interventions in general.


Posted by: Ginger Yellow | Link to this comment | 05-31-18 8:11 AM
horizontal rule
81

I don't get why aliens would be at Wright State University.


Posted by: ajay | Link to this comment | 05-31-18 8:16 AM
horizontal rule
82

When a kid comes to the door on Halloween, I always tell them they can have one fun-size Snickers now or come back in a half hour for two.


Posted by: Moby Hick | Link to this comment | 05-31-18 8:16 AM
horizontal rule
83

And when they come back you say, "Tricked you!" and slam the door in their faces.


Posted by: Mossy Character | Link to this comment | 05-31-18 8:18 AM
horizontal rule
84

81: You're right. They're at Wright-Patterson Air Force Base. I wasn't thinking clearly. They're a couple of miles apart.


Posted by: Moby Hick | Link to this comment | 05-31-18 8:21 AM
horizontal rule
85

Not so fast
https://jasoncollins.blog/2018/05/31/the-marshmallow-test-held-up-ok/


Posted by: wotton | Link to this comment | 05-31-18 4:48 PM
horizontal rule
86

There was a good take on the marshmallow test back in 2012. It wasn't so much about replication, but about determining the causality. Did kids who waited longer for two marshmallows do so because they had learned that patience would be rewarded or did kids who waited longer for two marshmallows do so for intrinsic reasons that would serve them well in the real world?

The experiment involved promising the kids better stickers or better art supplies if they waited a bit. Some kids were actually given the better stuff while others were told that would have to make do with the inferior stickers or art supplies on hand. Needless to say, the kids who had been rewarded for waiting by getting better art supplies were more likely to wait for a second marshmallow. The kids who had gotten shafted settled for a single marshmallow.

It actually says more about our society than about individual character.

http://europepmc.org/backend/ptpmcrender.fcgi?accid=PMC3730121&blobtype=pdf


Posted by: Kaleberg | Link to this comment | 05-31-18 8:33 PM
horizontal rule