Y11W20WR What the marshmallow test really predicted

Evidence Mapping
The writing prompt

Map what the marshmallow-test tradition actually established and what it didn’t, given the 2018 replication work.

1Retrieval check

Q1.What did Watts et al.’s 2018 replication of the marshmallow test find?

  • AThe predictive effect grew larger
  • BAfter controlling for family income and maternal education, the predictive effect largely disappeared
  • CThe original findings were confirmed exactly
  • DWillpower is the sole predictor of success

Q2.What does the article suggest the original marshmallow test probably measured?

  • APure willpower strength
  • BChildren who trusted that adults would deliver on promises — correlated with family stability and resources
  • CGeneral intelligence
  • DAppetite differences
Show answer key

Q1 → B. After controlling for family income and maternal education, the predictive effect largely disappeared.The original study may have measured children’s trust in promise-keeping adults — correlated with family stability and resources — more than pure self-control.

Q2 → B. Children who trusted that adults would deliver on promises — correlated with family stability and resources.A reasonable child in an unreliable environment eats the marshmallow now — this isn’t weak self-control; it’s accurate forecasting.

2Prompt deconstruction

Command verb
MAP — categorise claims, don’t argue
You pick
specific claims to place in robust / oversold / alternative / built-on-weak categories
Goal
distinguish what held, what’s been walked back, and what would settle the remaining interpretations
Must reference
Mischel’s original AND Watts’s 2018 replication AND the trust-in-environment interpretation

3Pick nudge

Which parts of the marshmallow-test evidence need to be separated?

Robust
The original phenomenon (variation in delay)
Oversold
The willpower-predicts-success story
Alternative
Trust-in-environment interpretation
Built on weak evidence
Educational programs citing the original

4Planner — categorise the claims

Robustly supported
The original phenomenon of variation — does that still hold?
Walked back by replication
Which specific popular claim has been weakened?
Supported but not definitive
Which alternative interpretation does the 2018 work point to?
Built on weak evidence
Which educational applications rest on the earlier claim?
A study that would settle it
How would you test trust-in-environment directly against residual-willpower?

5Sentence stems

  • The claim that ___ is robustly supported, because ___.
  • The claim that ___ replicates only partially — specifically, when ___.
  • The popular version of ___ has been walked back; the careful version is ___.
  • The genuinely open question is ___.
  • A study that would resolve this would ___.
  • On the weight of evidence, the article’s own position is ___.

6Exemplar paragraph (not about this article)

(1) Robust: children do show real variation in delay of gratification — the original phenomenon survives. (2) Walked back: the claim that this variation predicts later academic and life outcomes, which Watts et al. found largely disappeared after controlling for family income and maternal education. (3) The trust-in-environment interpretation — that waiting reflects reasonable confidence in promise-keeping adults — is supported but not definitively established. (4) Educational programs built on the original willpower framing rest on weak evidence. (5) A settling study would randomise children to a reliable-environment cue before testing delay, then track outcomes separately from family resources.

What this paragraph does, move by move

  1. Robust phenomenon named.
  2. Specific popular claim placed as walked back.
  3. Alternative interpretation placed with its status.
  4. Weak-evidence applications identified.
  5. Proposes a design that would settle it.