Causal judgement: use of different types of contingency information as confirmatory and disconfirmatory

In three experiments, participants made causal judgements from summary presentations of information about occurrences and non-occurrences of an effect in the presence and absence of possible causes. Participants' judgements could best be accounted for by the hypothesis that some were tending to use cell information in idiosyncratic ways: types of contingency information normatively regarded as confirmatory were sometimes treated as disconfirmatory and vice versa. In all three experiments, participants' judgements were better predicted by a model based on this evidence than by the probabilistic contrast model (Cheng and Novick, 1992) or the Power PC theory (Cheng, 1997)