Brian Cadena - The Marshmallow test

Brian Cadena, Associate Professor, Department of Economics at University of Colorado Boulder discusses the marshmallow test and its role when it comes to measuring impatience.


The idea of the marshmallow test was that young children were given a single marshmallow in a room and they were asked, or told that they could either eat it now, or if they were willing to wait five minutes they would receive a second and then they could eat both of them. What those original researchers found was that the kids who were able to wait for that second marshmallow then had much better life experiences, much better success in life going forward.

The recent criticism is that if you control for a bunch of other characteristics, a bunch of other ways that the children who were able to wait were different from the children who were unable to wait, a lot of those things, like family background, family income, the way that the children were raised, and a variety of other dimensions, can go a long way toward explaining those differences and outcomes. So it’s less clear whether it was actually that particular piece of impatience that was leading that difference in outcomes. I don’t know if that’s a great way necessarily of measuring impatience, and so economists have used other things, other measures of impatience that are more robust to those kinds of control.

One surprising one is actually interviewer ratings in survey. When people take these in-depth surveys like the National Longitudinal Survey of Youth, the interviewers at the end are asked how the respondent seemed to them during the interview, and one of the choices is impatient. And it turns out that respondents who were coded as impatient by their interviewers are incredibly like to engage in a variety of behaviors that are associated with self-control problems.

They’re more likely to smoke, more likely to drink, less likely to save, all those sorts of things and those relationships continue to hold once you control for a variety of background factors that were also measured relatively early in those respondents lives. So things like parental income, parental education, the home environment, even things like whether the parents were investing in their kids by getting them a library card, whether they had subscriptions to magazines, all these kind of things that reflect sort of how parents are investing in their kids.

Even once you adjusted for all of that, these interview ratings are surprisingly effective at predicting the kinds of impatient behaviors that economists are interested in.