It's looking more like wealth leads to increased patience, not the other way around. Intuitively this makes sense to me, as if you're food-insecure waiting for a better deal that probably won't come is irrational.
It's likely a self-reinforcing feedback loop. Knowing that you'll always have life's basic necessities provided for gives your brain time to focus on the long-term and not get distracted by imminent survival needs, which in turn lets you take advantage of opportunities that will only pay off in the long-term. The immediate pathway into the middle class is usually getting a college education; into the upper middle class is getting an advanced professional degree; into the wealthy is founding a successful business or participating in a successful financial or real estate deal. All of these require working hard and delaying gratification for years for a payoff that you can't necessarily see when you start. If your mindset is "Well, I've got a lifetime to spare; what else am I going to do with my time?" that's not a huge loss; if your mindset is "Gotta take what's in front of me now because it won't be there tomorrow" it becomes a lot harder.
Er, the paper cited shows that intelligence (and proxies thereof, like family background, as intelligence/SES are heritable) is prior to the marshmallow test and also predicts patience/success in the marshmallow test. Kids aren't wealthy or not, so that can't possibly be the causal mechanism...
Family income shows different tends in the marshmallow test. Kids from rich families often do much better. To my knowledge, no one thinks this is because rich kids come from better genetic stock (you'd be amazed how many experiments and studies change when you control for income levels). This supports the gp's hypothesis of the causality stack being flipped upside-down. Having more money leads to less stress, because you have less want. If you're a child of wealth you may not directly have the money, but you sure have the benefits of it.
> To my knowledge, no one thinks this is because rich kids come from better genetic stock (you'd be amazed how many experiments and studies change when you control for income levels).
They do, they're just polite about it because no one wants to put it that way. However, this is just Herrnstein's syllogism in action: if intelligence/education are heritable, education leads to higher incomes, then higher income families will have higher genetic potential which is passed onto kids (especially with the existing high levels of assortative mating on education). And this is precisely what we see with polygenic scores: there certainly are class differences in means.
I'm not surprised if it's true, but is there data showing that kids from rich families do better on the marshmallow test? I can't find it, and I'd be pretty interested to see the size of the effect.
The 2018 replication just added a single control for a whole battery of environmental factors. Family income was included (and correlates like parent's education), but it also lumped in biological factors (e.g. birth weight) and child-psychological factors (e.g. cognitive performance one year earlier). Given that, it's really hard to understand how much impact each specific element had.
Here's a study [0]. From my understanding every behavioral economist would not be surprised by the results. You'll notice the mention of the 2/3 due to "the presence of controls for family background, early cognitive ability, and the home environment." That's family income.
That's the same study I was referring to, though. They didn't provide the impact of any individual control variable, just an aggregate of about eight different factors that are both social and biological. Also, they didn't show that their control corresponds to marshmallow wait time specifically, but instead that it's at least as good a predictor of academic outcomes. (Which means the control and wait time must be broadly correlated, but not necessarily that their 'errors' are aligned.)
What I'm hoping for is something that specifically breaks out the relationship of family income to "marshmallow score", even without a long-term followup. I certainly believe that they correlate, but I'd like to see how strongly, and how it varies across what range of incomes. Even among people who agree that wealth drives marshmallow patience, explanations range from "wealthier families can consistently fulfill long-term commitments" to "food insecurity makes kids take food now". An obvious way to differentiate those theories is to see whether a wealth/time correlation remains strong past the point of food insecurity. (Since the original study was limited to relatively wealthy families, I expect it does.)
I agree that behavioral economists wouldn't be surprised by this result, but I'm not sure they're sold on "it's just wealth", either. While news stories are calling this a failed replication, I'm seeing economists argue that this was a successful replication of the test's predictive power, and only indicts some specific casual theories proposed to explain it. In particular, several of the control variables in the replication are measures of child academic performance, so without an individual-factor breakdown, there's a risk that this is effectively controlling for the variable being tested. I think the "patience makes you successful in life" narratives are laughably wrong, but I wish we had some clearer data on what the actual correlations are.
The issue is that it's patience at age 4, so kids have had maybe two years to learn even very basic expectation-and-outcome patterns. The effect people often describe where adult poverty teaches you to spend now before you lose your money isn't going to have kicked in yet.
That doesn't rule out food insecurity or other unfulfilled commitments teaching kids this result, but it has to be something earlier and more subtle than the established patterns of wealth in adults.
"Pick up on their parents' patterns" isn't the same thing, though. At the very least, it's an extra remove away, and probably modified by other control variables in the study like 'parental responsiveness'.
I'm not claiming that family wealth doesn't affect patience, but I see lots of people giving the specific rationale that being poor teaches people to spend their money immediately because it won't be available later. That doesn't make much sense to me, partly because people learn that pattern first-hand in late childhood or later, and more importantly because the original marshmallow test study stayed predictive within an entirely above-middle-class population.
In general, I don't think the current data actually offers much evidence for any specific narrative about how wealth causes this effect. Assessing that would call for information like the effect of relative versus absolute wealth, or at the very least a breakdown of what income levels this correlation happens over.
I'm only one data point but I disagree. I think their premise makes a lot of sense. You can be patient and still be very poor. I've seen a lot of people be that way. But growing up those people are very likely not to stay that way.
Wealthy parent, I would guess know that patience is a valuable skill and can be taught. Poor parents don't know this. But the poor kids that just so happen to be patient for whatever reason from my experience excelled at socio-economic mobility the way I didn't see less patient people. Every person I know including myself who made it out are very very long term thinkers.
The Atlantic's account of the marshmallow test is pretty overzealous.
In particular, Calarco describes it as a "failed replication", but in reality the new study (with 10x the participants and more diversity) also found a significant correlation between childhood wait time and teen test performance. I think calling this a failure to replicate is simply wrong.
The new result is that adding two large sets of controls makes the effect disappear. One of those control sets was environmental, including family income - but it also included traits about the individual child up through age three. The other was the child's score on a battery of other psych tests alongside the marshmallow test. Data for individual control variables isn't available, only these two aggregate scores. And the environmental control alone wasn't sufficient to erase the effect; it weakened below significance for some but not all groups studied, but only with both controls did the effect drop to near-zero for all groups. So "wealth causes patience" is one explanation, but "controlled for the variable" is another. And the age-15 test assesses metrics like reading comprehension, math skill, and general knowledge. The age-4 battery included metrics like "memory for sentences", "applied problems", and "complete words". Is it really surprising that controlling for vocabulary at age 4 makes it hard to find differences in vocabulary at age 15?
Even if this isn't controlling for the variable, we might have controlled away the cause instead. Wealth causing lower discounting is one obvious candidate, but we haven't ruled out the possibility that bad circumstances directly lower patience, which then causes bad outcomes. This study controlled for infant temperature, birth weight, and cognitive functioning at age two; if any of those things are a cause of or close correlate with lowered patience, the study loses all of its sensitivity.
And even if that isn't the case, I think the 'wealth causes patience' explanation is just permitted by the data, rather than demonstrated. I'm not sure exactly how you'd test for reverse causation on wealth/patience, but it would probably involve assessing patience at least twice to see whether kids gained/lost it corresponding to wealth. But as far as I can tell, no one has even checked it a second time without controls; wouldn't it be fascinating if age 4 patience correlated with SAT scores better than age 15 patience?
(For the record, I think the real answer is probably that wealth does increase patience in teens/adults, but the early-childhood 'patience' with marshmallows is a secondary symptom of any of numerous factors actually driving these results.)
https://www.theatlantic.com/family/archive/2018/06/marshmall...
The paper referenced in the article: https://journals.sagepub.com/doi/abs/10.1177/095679761876166...