Ever heard of Mischel’s marshmallow experiment?
In Walter Mischel’s original marshmallow test, a series of children were given a marshmallow or another sweet and then offered a choice: they could either eat the marshmallow now, or wait a while and then receive a second marshmallow. The purpose of the experiment was to study a child’s ability to delay gratification and to examine whether this ability could predict adult behavior.
That self-control at a young age predicted SAT scores, career success and other bits of life after 20 years. The New Yorker featured an article and then, more recently, linked to a video version of kids going through the motions of self-control.
We’re in the midst of measuring self-control before a series of behavioral and economic interventions in Africa. I’m interested in the link not just to impulsive spending, or lack of savings, but violence as well.
We’ve been trying to develop an adult version of the Mischel test, with only part success so far. The survey, cognitive tests and behavioral games take 2-3 hours in total. The whole, blisteringly hot time (it is Liberia, after all) there’s a cold, glistening can of soda sitting in front of the respondent. If they wait, they get more, or maybe even cash. We time how long before they take it.
I don’t know whether to expect results, but the fun is in the trying. Any other suggestions?
Hat tip to Angeli.
12 Responses
The fun for whom is in the trying?
I somewhat draw into question the use of a cold Coke as the object of desire. It’s my observation in Sudan, at least, that a lot of people have grown to prefer their soda (and beer) warm… Sorry I can’t at the moment think of a good alternative, although some interesting ones have been posted by colleagues above. Very cool and prospectively instructive experiment to be considering, though.
This idea of trust is a crucial one in what you are looking at in adults. It brings to mind an experience I had recently here in northern Uganda: A community group had their election of a new chair, vice and treasurer and I was invited as their observer. At the end I gave a gift of about $10, which is enough to open a group bank account –it is also enough to buy a crate of sodas. A heated debate erupted about who should be entrusted to open the account and who would be the signatories. In the end they bought a crate of soda. It’s hard to say how much of their choice was based on limited resistance to the temptation of instant gratification or a lack of trust among the members. I’ve also observed this in ‘pay as you earn’ savings schemes run by Ugandan NGOs. The ‘aversion’ to saving, I think, is not that at all. Rather, it has much more to do with how much the employees trust the management to actually give them the money in the future.
I’d keep in mind that trust can be a big role. They may take the Coke now because they can see if and trust that you will give it to them, even if they’re willing to wait. You might want to show all of the possible prizes to them at the start and say, “you can have this one now, or these later”, although that might change what you are testing.
– Ryan
Hello,
Along the same line of thinking, we’re conducting a survey in Senegal and are planning to measure self-control using bank notes (real money).
The idea is the following: surveyed people receive 1000 CFA from
a lottery on risk aversion and discounting preferences elicitation during the first visit of our enumerator. (we keep track of the number of the bank note.)
If they’re able to keep this bank note until the second visit of our enumerator (1 or 2 weeks later), they get an extra 500 CFA.
Since we’ll control for their discounting factor for these same period and amount of money, we should be able to disentangle between discounting preferences and self control issues.
What do you think of this idea?
Thanks a lot in advance for your comments,
Best,
Olivier
Hi Chris. I had some ideas for extensions to the marshmallow experiment not so long ago, I wrote them on my blog:
http://cdelcarpio.blogspot.com/2009/06/extensions-to-marshmallow-experiment.html
They were inspired by a very funny video by Joachim de Posada at a TED talk last year about the same topic. Saludos,
I say try a small portion of “dry rice” and if they wait, larger portion of rice and soup – something everyone loves like cassava leaf with some heavenly smelling meat! Oh – makes me want to try with kids in LIB……like a piece of fruit and if you wait – ICE CREAM!!!
It seems less cruel and possibly a better idea for methodological reasons to offer something that’s not directly tied to the specific environment in which you’re testing people. Maybe offering a $1 mobile phone minutes card immediately but a $2 after three hours?
It has to be immediately consumable and tempting, which a phone card wouldn’t do. They wouldn’t want to use it right away.
What’s cruel about telling someone they can have a second drink if they wait?
Culture/environment might twist the results of your study if you’re using a cold coke as the “candy.” Granted everyone gets thirsty, but my experience with most Africans is that they are able to hold out for a drink a lot longer than us whities can not necessarily because they have better self control but simply because they are more acclimatized and/or culturally adapted to do without such “candy.”
Bear in mind some criticisms of the original study:
http://www.thedailybeast.com/blogs-and-stories/2010-02-19/just-let-them-eat-the-marshmallow/
If you are combining all of your other tests with this one–that may cause problems. How might I have done on the SATs with some sort of temptation in front of me? Maybe people that are better at delaying gratification are not affected, while those that are do worse on the test. In that case, all your other results will just pick up the Marshmallow effect, not what you’re testing.
One thing to try would be to offer a differentiated payment structure (I assume you’re paying respondents). Offer them X today, or they can some back in a week for 2X.
cigarettes?