THE IMPLICATIONS OF SUCCESS FOR HIERARCHICALLY AND PARTIALLY RESTRICTIVE ABILITY DIMENSIONS

Recent research (e.g., Trafimow & Schneider, 1994) suggests that succe sses lead to more correspondent ability attributions than do failures. However, this research leaves open the possibility that the different ial importance of successes versus failures is greater for some abilit y dimensions than for others. Three experiments were conducted to test a hypothesis that what distinguishes these two types of ability dimen sions is the expected probability that a person without the ability co uld perform a relevant behavior. In Experiment 1, participants' abilit y judgments following a target person's successful performance of rele vant behaviors correlated with judgments of how likely a person withou t the ability would have been to successfully perform the behaviors. H owever, there was no correlation between ability judgments and judgmen ts of how likely a person with the ability would have been to fail to perform the behaviors. Experiment 2 resulted in similar findings, but with a different set of behaviors and a different experimental paradig m. In Experiment 3, participants were told about an ability with which they were unfamiliar, and led to believe that someone without the abi lity could either never or sometimes perform the relevant behavior. Th ey then made ability judgments about someone who either successfully o r unsuccessfully performed the behavior. Consistent with the hypothesi s, the strongest correspondent inferences were made following a succes sful performance when participants had been led to believe that someon e without the ability could never perform the behavior.