EV-0029
The famous marshmallow finding shrank when it was retested
Watts, T. W., Duncan, G. J., & Quan, H. (2018). Psychological Science, 29(7), 1159–1177 · Source
What this entitles us to say
A larger conceptual replication of the marshmallow test finding that early delay of gratification still predicted adolescent achievement, but at roughly half the original correlation — and that the association shrank by about two thirds once family background, early cognitive ability and home environment were controlled.
What it found
A larger, more demographically varied conceptual replication of Shoda, Mischel and Peake (1990). Concentrating on children whose mothers had not completed college:
- An additional minute of waiting at age 4 predicted roughly one tenth of a standard deviation gain in achievement at age 15. - That bivariate correlation was about half the size of the original studies'. - It was reduced by about two thirds once family background, early cognitive ability and home environment were controlled.
Why it matters here
The marshmallow test is the single most likely piece of pop psychology to be raised in a room where this program is being sold — usually by someone supporting us. That is exactly why it needs to be in the file. Citing a finding that a well-read buyer knows has been substantially walked back would cost more credibility than the story could ever buy.
It also carries a real lesson for the trait rather than merely a warning. Much of what looked like early self-discipline predicting later success turned out to reflect circumstances — a stable home, resources, what a child had learned to expect from adults. Discipline is not evenly available. A leader looking at a low score on this trait in someone's report is not looking at a moral failing, and the module should say so plainly.
That connects directly to the program's first principle, and to EV-0027: if discipline is largely built in advance out of routines and environment, then it is unsurprising that environment predicts it — and encouraging, because environments can be changed.
The limits
This does not show delay of gratification is irrelevant — a measurable association survived the controls. It is a conceptual rather than direct replication, and the analysis concentrated on children whose mothers had not completed college, which shapes what it can say. Debate continued afterwards, including a direct comparison paper in 2019 arguing the two studies are more compatible than the headlines suggested. Verified to abstract level.
What we do not say
"Delay of gratification doesn't matter." An association survived the controls. This is a correction of an overclaim, not a refutation — and the follow-up literature argues the two studies agree more than the headlines suggested.