Revisiting the Marshmallow Test: A Conceptual Replication Investigating Links Between Early Delay of Gratification and Later Outcomes.
Level 3 - non-randomized controlled study
Prospective longitudinal cohort study (by design analogy, non-clinical)
PubMed 29799765 · doi:10.1177/0956797618761661
What was done
Researchers conducted a conceptual replication and extension of Shoda, Mischel, and Peake's (1990) marshmallow test. They assessed children's ability to delay gratification at age 4, focusing on children whose mothers had not completed college, and evaluated academic achievement and behavioral outcomes at age 15 with controls for family background, early cognitive ability, and the home environment.
What was found
In bivariate analysis, an additional minute waited at age 4 predicted a gain of approximately 0.10 standard deviations in achievement at age 15, which was half the size of previously reported correlations. Controlling for family background, early cognitive ability, and the home environment reduced this correlation by two thirds. Most variation in adolescent achievement came from waiting at least 20 seconds. Associations between delay time and behavioral outcomes at age 15 were much smaller and rarely statistically significant.
Why it matters
This replication shows that the apparent long-term benefits of early delay of gratification on later achievement are largely explained by family background and early cognitive environment.
Limits
The abstract does not report the total sample size. The focus on children whose mothers had not completed college restricts generalizability. As an observational cohort study, residual confounding remains possible.
Cited by
- contradicts Longitudinal follow-ups of Walter Mischel's marshmallow test found that longer delay times in children correlated with higher academic achievement, higher income, better health, and lower rates of incarceration and behavioral problems.
- supports A large replication study of the marshmallow test found that after controlling for socioeconomic status and other covariates, delay time at age 4 no longer predicted adolescent life outcomes.