Skip to content

Three-territory gold-mining reinforcement learning task with hidden agents

Design · All designs
Evidence · 1 claim (1 for) · 1 study (1 causal), q3 · 0 of 1 report an effect size · 1 claim rests on one study

Description

A developmental adaptation of the Dorfman et al. task in which participants choose between two mines with fixed reward probabilities (80% vs 20% gold) across three 50-trial blocks. Each block occurs in a territory visited by a hidden agent that intervenes on 30% of trials: a Millionaire who adds gold, a Robber who replaces gold with rocks, or a Sheriff who intervenes randomly. After each outcome participants report whether the agent caused it. The task measures learning via optimal choice and causal attribution via forced judgments, supporting computational model comparison across ages.

Design Implications

Context

Requirements

  • Participants must understand task instructions; piloting indicated age 7 was the youngest age at which children reliably understood the instructions, and experimenter corrections occurred only during practice trials.

Constraints

  • The exact intervention probability (30%) was not told to participants; intervention was invisible and ambiguous on each trial.

Target Learners

  • children ages 7-12
  • adolescents ages 13-17
  • adults ages 18-25

Learning Goals

  • reinforcement learning from positive and negative feedback
  • causal attribution of outcomes to external agents

Claims

Examples

-

Key Sources

  • Alexandra O. Cohen, Kate Nussenbaum, Hayley M. Dorfman, Samuel J. Gershman and Catherine A. Hartley. (2020). The rational use of causal inference to guide reinforcement learning strengthens with age. npj Science of Learning 5:16. https://doi.org/10.1038/s41539-020-00075-3
Page metadata
Field Value
type design
id gold-mining-hidden-agent-rl-task
title Three-territory gold-mining reinforcement learning task with hidden agents
description A developmental adaptation of the Dorfman et al.
status review
generated by process:settle-candidates on 2026-10-27
trust unverified — no one has confirmed this page's content
sources 1 — alexandra-o-cohen-2020

Frontmatter is the page's machine-readable record — see the Schema & Guide.