What the marshmallow test really tells us

Photo by Flickr user Tequiua. Photo by Flickr user Tequiua.

Editor’s Note from Paul Solman: One of the most exciting developments in economics in recent years has been its conjunction with psychology. What do we really want? How might we behave in what’s truly our own best interest?

These are questions we’ve explored on Making Sen$e with, among others, Dan Ariely of Duke, Jerome Kagan of Harvard, Jeremy Bailenson of Stanford University’s Virtual Reality Lab, and Grover of Sesame St., to whom we administered the fabled “Marshmallow Test”: could he hold off eating just one marshmallow long enough to earn a second as well?

Now comes an essential book on the subject of gratification delay by the father of the Marshmallow Test, Columbia University psychologist Walter Mischel: “The Marshmallow Test: Mastering Self Control.” Our interview with him, posted as part 1 today and part 2 tomorrow, is — how to put this emphatically enough? — well worth delaying other gratifications to read.

The Marshmallow TestPS: Let’s start with some of the basics. I read the interview that the woman at The Atlantic did with you, and I was so struck by the fact that what she was mainly concerned about was that her child had, and I use the term in quotes, “failed the marshmallow test.”

WM: She is representative of so many parents.

PS: So what’s going on there?

WM: The unfortunate interpretation that’s been made of the research, which I must say the media have helped to create, is that your future and your destiny are in a marshmallow, which in turn translates into the widespread belief, I think, in the genes. And I think both of those are really deep misunderstandings that have very serious negative consequences for how we think about self-control.

From my point of view, the marshmallow studies over all these years have shown of course genes are important, of course the DNA is important, but what gets activated and what doesn’t get activated in this library-like genome that we’ve got depends enormously on the environment. And what we as individuals do and think and experience, and the stress levels we encounter, the stuff we smoke, the toxins we inhale, and the things we do and feel — the way we manage our emotions, the way we regulate our lives — enormously influences how the DNA plays out.

The good news in this is really that human beings potentially have much better potential for regulating how their lives play out than has been typically recognized in the old traditional trait series that willpower is some generalized trait that you’ve either got or you don’t and that there’s very little you can do about it. The research shows there’s a great deal you can do about it; there’s a great deal that is being done about it in many kinds of — not only experiments, but school programs, pre-school programs, and so on. I think that the evidence that self-control skills are highly protective is, to me, much more interesting that the evidence that extreme differences in high self-control versus low self-control play out in different kinds of minds in different degrees of efficacy and success.

PS: But the New Zealand study, for example, which is not subject to the criticisms sometimes leveled at your studies, which is that your sample is too small (because they’re talking about 10,000 people or more followed longitudinally where you had fewer than 100 that you followed for 30 years) …

WM: Actually, by now, it’s over the course of 40 years … and it actually is a bit over 100. And it, of course, depends. There’s no question that the sample becomes increasingly selective. That sample in itself, I think, is open to lots of loose interpretation because, to me, Paul, the amazing thing is that they found any long-term differences in a sample that began with such enormous homogeneity. All of those kids were essentially white kids from an elite university — either the children of Stanford faculty or the children of Stanford graduate students in which the conversation scene in kindergarten between kids was about things like, “What area did your father get his Nobel prize in?”

PS: But doesn’t that imply your results, and the much larger sample results from New Zealand, that there is a significant genetic factor? I’ve corresponded with psychologist and behavioral economist George Ainslie about your work and the New Zealand study, and he, for example, thinks it’s entirely plausible — not demonstrated — but plausible that there is a self-control trait (not to say “gene,” but “trait”) that, all else equal, is predictive of, among other things, and of particular interest to me, the ability to save and plan and prosper financially in the future.

WM: I have several comments on that. I don’t think there’s any question that genetics are enormously important. But I think that what the research, for me, over the years has shown is that whether we call it willpower or whether we call it the ability to delay gratification, what’s involved is really a set of cognitive skills for which the current label is “executive control” or “executive function.”

And what executive control fundamentally involves is the activation of the areas in the pre-frontal cortex (the attention control areas) that allow you to do really three things: to keep a goal in mind (I want those two marshmallows or two cookies), to inhibit interfering responses (so I have to suppress “hot” responses, for example, thinking about how yummy and chewy and delicious the marshmallow is going to be), and have to instead do the third thing, which is to use those attention-regulating areas in the prefrontal cortex to both monitor my progress toward that delayed goal, and to use my imagination and my attention control skills to do whatever it takes to make that journey easier, which we can see illustrated beautifully in any video that I can show you of how the kids really manage to transform the situation from one that is unbearably effortful to one that’s quite easy. They throw off their sandals and turn their toes into piano keys in their imagination and play them and sing little songs and give themselves self-instruction, so that they’re doing psychological distancing to push the stuff that’s fun (the treats and the temptations) as far from themselves as they can.

To me, the interesting thing about the marshmallow study is not so much the long-term correlation as is what we discover when we look at what those kids are doing and what the parallels are that we can do when dealing with retirement planning or with giving up tobacco and so on.

PS: So even Ainslie’s argument about hyperbolic discounting and that you have multiple selves battling against one another — even that involves the executive function, if you will, some role for the prefrontal cortex that then inculcates habits, or strategies that can become habits, like the playing of your toes, that will affect your behavior regardless of your predisposition to wait.

WM: I think that’s putting it very well, yes. As you know, the point of the marshmallow studies is, after you’ve made the choice, and you’re in the restaurant and you’re facing the dessert tray that the waiter is flashing in front of you, and you’ve gone into the restaurant with the resolution “no dessert tonight,” what happens when you actually see the stuff?

I’m right now in the midst of a very interesting collaboration with David Laibson, the economist at Harvard, where our teams are working on that Stanford sample doing a very rigorous, and very well designed and very well controlled study to see what the economic outcomes are for the consistently high-delay versus the consistently low-delay group.

PS: So explain what it is exactly you’re doing with Laibson’s team?

WM: Well, what we’ve done is used very complete and rigorous measures that David’s team came up with of the wealth, of the credit card debt, of the endless stuff that economists love about their financial situations.

I keep reminding myself of the extraordinary nature of finding differences in this sample, where, when we’re talking about educational level, for like 500 kids (which is a large sample in psychology), in that whole bunch of kids, we found, I think, three who didn’t complete college, and they probably went on to start Microsoft or something! So when we’re talking about educational outcomes, we’re talking about how many advanced degrees they got.

That’s why I have been both fascinated by getting any long-term results here, and why I moved from Stanford to Columbia, in New York City, where I’m sitting on the edge of the South Bronx. I came, originally, with the idea of doing studies in the South Bronx – not in Riverdale – but in some of the most impoverished and stressed areas, where we find very interesting parallel results. And to me, the most interesting thing in the Bronx studies – and we’ve had them repeated now in areas of Oakland, California – what’s much more interesting than the predictive effects of the correlations of these relatively small samples is the protective effects, by which I mean that kids, for example, who are severely predisposed to aggression and to violence and to acting out, if they have self-control skills — that is, if they wait longer for more m&m’s later rather than just a few now — the level of aggression that they have is much less. They’re still aggressive, but they don’t hit the counselor over the head with a flashlight and give her a concussion.

PS: So to you, what that says is not that there’s this genetic endowment – people are stuck with it and there’s nothing you can do – it’s just the opposite.

WM: Exactly right. It means that no matter what the DNA lottery has dealt them, people have a hell of a lot more choice and freedom if we can reduce their stress levels and if we can give them access to the kinds of skills and the kind of mental transformations that let them think differently about delayed and immediate outcomes, their temptations, their own dispositions and so on. That’s why I think both the philosophical and the policy implications are profound. I’m meeting this month with people from the British cabinet in London who worry about this kind of stuff. The problem here is that we’ve got economic advisers in the White House, but we don’t have psychology advisers.

Editor’s Note: Find the continuation of Paul’s conversation with Walter on Making Sen$e Thursday.