Two studies were designed to investigate the effects of prior information about plausible causes on subsequent attributions. In the Experiment 1, prior information was given about an internal and an external cause. It was predicted and found that the stronger an expectancy for a behavior, the more the behavior would be attributed to the cause that formed the basis of the expectancy. It was also predicted that (1) for facilitative causes, the stronger the expectancy based on a given cause, the less the behavior would be attributed to other causes (discounting), and (2) for inhibitory causes, the stronger the expectancy, the more the behavior would be attributed to other causes (augmenting). These predictions were not supported. It was suggested that discounting and augmenting did not occur because subjects had been given information about both causes which "locked in" their attributions to each cause. To test this explanation, a second study was undertaken in which observers were given information about only one cause. The results of Experiment 2 indicated that discounting and augmenting may only occur for attributions to causes about which no prior information is available.According to Kelley (1972), causal schemata are employed by attributors when two or more causal factors are perceived to interact in causing a given effect. In situations where there is more than one possible cause for an event, and when each cause alone could have caused the event, Kelley suggests that attributors employ a schema for multiple sufficient causes. In these situations, Kelley suggests that the discounting principle (Kelley, 1971) can be used to predict the attributions that will be made for the event. This principle states that the role of a given cause in producing an effect is discounted to the degree that other causes are present. Experiments conducted by Jones and others (Calder, 1974; Jones, Davis, & Gergen, 1961; Jones, Worehel, Goethals, & Grumet, 1971;Messick & Reeder, 1972) to test the theory of correspondent inferences (Jones & Davis, 1965) provide strong support for the discounting principle. In these experiments, subjects were given information regarding an effect (the behavior of an actor) and the state of a plausible external cause for it. They were then asked to judge the magnitude of a second possible cause. If the external cause led subjects to expect the behavior, they attributed the behavior less to a second plausible cause (dispositional characteristics of the actor) than if the external cause did not lead them to expect the behavior.. Although all of the experiments just reviewed
97provided subjects with prior information on only one external cause, the discounting principle should also apply to situations in which the attributor has information on an internal cause or on more than one cause. In the latter situation, the degree to which a behavior is attributed to any given cause will depend upon the strength of all of the other causes.Strength in this context refers to the expectancy that a particular beha...