2018 Replication Halves Marshmallow Test Effect in 918 Children as Controls Cut It by Two-Thirds
Updated
Updated · spacedaily.com · Jun 15
2018 Replication Halves Marshmallow Test Effect in 918 Children as Controls Cut It by Two-Thirds
3 articles · Updated · spacedaily.com · Jun 15
Summary
A 2018 replication found preschool delay of gratification predicted far less teenage success than the marshmallow test’s popular story suggests, with only about a tenth of a standard deviation gain in achievement at age 15 per extra minute waited.
In 552 children whose mothers had not finished college, that link was already roughly half the size reported in the 1990 study and shrank by about two-thirds after adjusting for family background, early cognitive ability and home environment.
The larger study also found most predictive value came from clearing about 20 seconds of waiting, while longer delays added little, undermining the idea that heroic self-control at age 4 strongly maps onto later outcomes.
Behavioral outcomes weakened most: associations between waiting time and age-15 behavior were small and rarely statistically significant, especially compared with the original study’s stronger claims.
Using a broader NICHD sample of 918 children rather than the small Stanford-linked original, the replication suggests the test may capture family advantage and circumstance as much as stable willpower.