Strategies to intervene on causal systems are adaptively selected

摘要: How do people choose interventions to learn about causal systems? Here, we considered two possibilities. First, test an information sampling model, gain, which values that can discriminate between a learner’s hypotheses (i.e. possible structures). We compare this discriminatory model positive testing strategy instead aims confirm individual hypotheses. Experiment 1 shows behavior is described best by mixture of these alternatives. In 2 find are able adaptively alter their and adopt the more often after experiencing confirmatory leads subjective performance decrement. 3, time pressure opposite effect inducing change towards simpler strategy. These findings suggest there no single describes how intervention decisions made. Instead, select strategies in adaptive fashion trades off expected cognitive effort.