Mischel's Marshmallows

Jonah Lehrer, in "How We Decide" writes about the famous Mischel Marshmallow experiment. There is also an entertaining presentation about it on RadioLab.
Lehrer recently published an article in the New Yorker going into much more depth about the subject and the people involved.

This is some psych research done at Stanford by Walter Mischel in the late 60's or early 70's which was set up as a test of the self-control of four-year-olds. Take a four-year-old into a fairly boring room, sit the kid at a table. Put a marshmallow (or other treat) in front of the kid and give them a choice. The adult researcher is going to leave them alone in the room. They can eat the treat whenever they want, but if they can wait until the researcher gets back, they get a second treat in addition to the first. Of course the kids varied in their ability to wait. Some ate the thing right away, about thirty percent held out for fifteen minutes and got the second treat. The really interesting part is the follow up. Mischel hadn't initially planned to followup, but the kids were in the Stanford area and many of them went to school with his kids. He noticed a pattern which caused him to follow up the research when the kids were all teenagers. The results were quite interesting, among other things:

The child who could wait fifteen minutes had an S.A.T. score that was, on average, two hundred and ten points higher than that of the kid who could wait only thirty seconds.

He kept following them into adulthood and found things such as:

... low-delaying adults have a significantly higher body-mass index and are more likely to have had problems with drugs ....

How long you can wait for a marshmallow at four is a useful predictor of your SAT score and your adult body-mass index. A simple test ends up having less than simple ramifications.

“What we’re really measuring with the marshmallows isn’t will power or self-control,” Mischel says. “It’s much more important than that. This task forces kids to find a way to make the situation work for them. They want the second marshmallow, but how can they get it? We can’t control the world, but we can control how we think about it.”

Of course, one has to keep in mind that this is all statistical tendencies over the hundreds of kids who took the test. This is not some inevitable destiny. The had examples of being able to teach some of the kids various strategies for getting the second marshmallow, such as imagining a frame around to marshmallow and pretending it's a picture. The main key seems to be an ability to move ones mental focus from a "hot" stimulus, like a tempting marshmallow, to something else. They didn't set up the study to compare the kids who were taught strategies with ones who weren't, but it's encouraging to note that some kids were able to turn it around themselves somehow:

Mischel is particularly excited by the example of the substantial subset of people who failed the marshmallow task as four-year-olds but ended up becoming high-delaying adults. “This is the group I’m most interested in,” he says. “They have substantially improved their lives.”

The New Yorker article goes on about some of the current research on the subject to see if the ability to delay gratification can be taught.

One really useful addition to the New Yorker article is Jonah Lehrer answering reader questions about the article.

In one of the answers, he makes a statement which would be so nice to hear from other popular writers:

These statistical correlations can’t be reliably applied to the individual.

For example, I've heard Malcolm Gladwell speak of such statistical correlations as if they were inevitable. In this RadioLab, Gladwell refers to this research by Ap Dijksterhuis where people primed with the idea of "professors" performed statistcally better at Trivial Pursuit questions than people primed with the idea of "soccer hooligans". When Gladwell mentions it in the RadioLab, rather than showing Lehrer's restraint, he says that if you're primed with the hooligans "you basically lose". Gladwell is a really good writer, but he's a bit reckless with his data. (Robert A. Burton in his book "On Being Certain" also takes Gladwell to task, with good cause.)

I find Jonah Lehrer a welcome voice popularizing psychology and neuroscience. He shows respect for the subject.

One of the bits of the New Yorker article I would like to hear more about is Mischel's attitudes toward personality science back in the 1950's and 60's.

Mischel noticed that academic theories had limited application, and he was struck by the futility of most personality science. He still flinches at the naïveté of graduate students who based their diagnoses on a battery of meaningless tests.

He was hired to try to evaluate the personalities of Peace Corps applicants and:

Volunteers were tested for standard personality traits, and Mischel compared the results with ratings of how well the volunteers performed in the field. He found no correlation; the time-consuming tests predicted nothing. At this point, Mischel realized that the problem wasn’t the tests—it was their premise. Psychologists had spent decades searching for traits that exist independently of circumstance, but what if personality can’t be separated from context?

I guess this really caught my attention, because I've recently had people point me at personality tests like Myers-Briggs and Enneagram. Some times the results of such tests seem to ring true, but mostly in a very similar way to the way horoscopes ring true. I like Mischel's statement that personality can't be separated from context, I think that is a very important point. Static tests just aren't going to capture what your personality or your character, is really like. Behavior is inseparable from situations. People always want to find the one static thing which determines something else, but it is rarely that simple. It's the same way with the perennial nature/nurture debate; it's a false dichotomy. Just because we have two different words for them does not mean that the two can be separated from each other. As Mischel says:

“In general, trying to separate nature and nurture makes about as much sense as trying to separate personality and situation,” he says. “The two influences are completely interrelated.”