Differences between Observation and Intervention in Causal Learning

Previous studies have suggested that learning is improved when people actively intervene rather than when they passively observe in causal structure learning tasks. Two experiments were conducted to investigate whether a facilitative effect will occur in the judgment of causal strength. In Experiment 1, participants were asked to learn causal strength in a situation where the target cause and context independently produced the effect. The intervention group could manipulate the state of the cause, which was later presented to the observation group (i.e., yoked-control procedure). The results demonstrated that participants made similar evaluations for the target cause, but not for the context. Experiment 2 was designed to examine whether different estimations were because of facilitation or bias in which participants undervalue other causes. The results provide support for a facilitative effect, but suggest that the improvement with intervention may be limited to the estimation of weak causal strength.