The goal of the proposed research is to study how an untutored reasoner comes to know that one thing causes another. The process of causal discovery concerns the mechanism by which an intelligent system learns to differentiate sequences on which the system bases explanation and control (i.e., causal relations) from the indefinitely many that are incidental or merely covariational. Previous psychological research reveals that natural causal discovery is remarkably rational, in fact, more so than the current explicit statistical methods that scientists and lawyers use to help them infer causality. An analysis of the two previous dominant approaches to the psychological process of causal discovery -- the pure causal-power view and the pure covariation view -- shows that they respectively specify an incorrect input and an incorrect output. The latter approach is that adopted by standard statistics. Cheng (1997) proposed an integration of these approaches that overcomes their individual problems. Her theory concerns the discovery of relations involving causes and effects that each can be represented as a single binary variable. Novick and Cheng (accepted) have extended it to apply to conjunctive causes. ? ? The proposed research (1) tests the assumptions underlying the derivation of simple and conjunctive causal powers, (2) tests the novel predictions made by the two components of the theory, (3) examines whether there is a dissociation between implicit and explicit causal discovery, (4) examines why some previous findings deviated from the predictions according to causal power, and (5) tests the implications of this theory for causal attribution in the law. The method involves presenting adult human participants with information regarding the presence and absence of candidate causes and the presence and absence of effects, and asking them to make causal judgments. The project directly aims at gaining a deeper understanding of the process of causal discovery; it indirectly aims at providing a basis for the development of an effective program for teaching statistics and scientific methodology.
| Liljeholm, Mimi; Cheng, Patricia W (2009) The influence of virtual sample size on confidence and causal-strength judgments. J Exp Psychol Learn Mem Cogn 35:157-72 |
| Lu, Hongjing; Yuille, Alan L; Liljeholm, Mimi et al. (2008) Bayesian generic priors for causal learning. Psychol Rev 115:955-84 |
| Liljeholm, Mimi; Cheng, Patricia W (2007) When is a cause the ""same""? Coherent generalization across contexts. Psychol Sci 18:1014-21 |
| Novick, Laura R; Cheng, Patricia W (2004) Assessing interactive causal influence. Psychol Rev 111:455-85 |
| Buehner, Marc J; Cheng, Patricia W; Clifford, Deborah (2003) From covariation to causation: a test of the assumption of causal power. J Exp Psychol Learn Mem Cogn 29:1119-40 |
