Structure and strength in causal induction q

We present a framework for the rational analysis of elemental causal induction—learning about the existence of a relationship between a single cause and effect—based upon causal graphical models. This framework makes precise the distinction between causal structure and causal strength: the difference between asking whether a causal relationship exists and asking how strong that causal relationship might be. We show that two leading rational models of elemental causal induction, DP and causal power, both estimate causal strength, and we introduce a new rational model, causal support, that assesses causal structure. Causal support predicts several key phenomena of causal induction that cannot be accounted for by other rational models, which we explore through a series of experiments. These phenomena include the complex interaction between DP and the base-rate probability of the effect in the absence of the cause, sample size effects, inferences from incomplete contingency tables, and causal 0010-0285/$ see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.cogpsych.2005.05.004 q We thank Russ Burnett, David Lagnado, Tania Lombrozo, Brad Love, Doug Medin, Kevin Murphy, David Shanks, Steven Sloman, and Sean Stromsten for helpful comments on previous drafts of this paper, and Liz Baraff, Onny Chatterjee, Danny Oppenheimer, and Davie Yoon for their assistance in data collection. Klaus Melcher and David Shanks generously provided their data for our analyses. Initial results from Experiment 1 were presented at the Neural Information Processing Systems conference, December 2000. TLG was supported by a Hackett Studentship and a Stanford Graduate Fellowship. JBT was supported by grants from NTT Communication Science Laboratories, Mitsubishi Electric Research Laboratories, and the Paul E. Newton chair. * Corresponding author. Present address: Department of Cognitive and Linguistic Sciences, Brown University, Box 1978, Providence RI 02912, USA. E-mail address: [email protected] (T.L. Griffiths). T.L. Griffiths, J.B. Tenenbaum / Cognitive Psychology 51 (2005) 334–384 335 learning from rates. Causal support also provides a better account of a number of existing datasets than either DP or causal power. 2005 Elsevier Inc. All rights reserved.