Causal Models and Cognitive Representations in Multiple Cue Judgment

Recent studies suggest that humans can infer the underlying causal model from observing the distribution of variables. In a multiple-cue experiment we investigated if people can infer the causal structure from mere observation, and if different causal models invite different cognitive processes. Participants performed 220 training trials in two judgment tasks with different underlying causal structure. The result shows a poor ability to discriminate between causal models, and poor manipulation insight, but a correlation between causal models and cognitive processes. This study suggests that people do not represent multiple-cue judgment tasks in terms of causal models, but that common effect causal models invite reliance on processes of explicit cue abstraction