An Activation-Based Account of Belief Bias in Relational Reasoning: The Effect of Concurrent Working Memory Load

Abstract

Working memory plays an important role in explaining reasoning performance. A model of belief bias in relational reasoning is presented here which explains the typical belief bias effects through the activation level of conclusions in memory. This account is formalized in an ACT-R model. The model predicts that a concurrent working memory load will increase the effect of belief on responses to reasoning problems, specifically on indeterminately invalid problems. This prediction is confirmed experimentally.


Back to Saturday Posters