Abstract
This article examines four distinct relationships between covariation patterns and attributional dimensions (consensus -> locus; consistency -> stability; distinctiveness -> globality: action-contingency -> control) as specified in the Joint Model by Van Overwalle and Heylighen (1995), using an adapted stimulus-response compatibility paradigm. Subjects were first told that a person was rejected after an application interview and then given additional covariation information. Next, they had to judge whether a particular cause was the “correct” explanation for this rejection, on the basis of a causal rule given before. This rule was either compatible or incompatible with the joint model’s predictions. Assuming that this model accurately captures subjects’ natural causal judgments, it was predicted that incompatible causal rules should be more difficult, leading to more errors against the imposed rule and longer response times, whereas rules that are compatible should be easier, leading to less errors and shorter response times. Two experiments confirmed the predicted covariation-attribution relationships, except for the contingency-control relationship which was confirmed only by the accuracy data.