The "Marshmellow" test

I remember reading somewhere about the "marshmellow test" – the finding that: if a child could wait longer for a marshmellow (or other snack they liked), the child would be more likely to do well later in life.

Found the original 1990 experiment!

Interestingly, this 2018 study replicated the 1990 experiment.

It basically found the link (between waiting for the marshmellow and later outcomes) to be smaller than the 1990 original – especially after controlling for other factors (like family background, early cognitive ability, etc).

What does this mean?

It doesn't "debunk" the marshmellow test entirely, it just says that other factors matter more.

It seems to me the "marshmellow test" is still an interesting "proxy" test for later outcomes – but unsurprisingly, the full picture matters more.