A COMPONENTIAL INVESTIGATION OF THE RELATION BETWEEN STRUCTURAL MODELING AND COGNITIVE ACCOUNTS OF HUMAN JUDGMENT

Structural modelling and cognitive process approaches have developed r ather different accounts of human judgment and decision making. Two hy potheses to explain these differences were evaluated in the context of a judgment task, and formulated in terms of predictions concerning me asurement of attribute importance. First, following suggestions made b y Billings and Marcus (1983), it was argued that measures of judgment behaviour based on structural modelling reflect cognitive activity lat e in the judgment process, whereas measures derived from cognitive pro cess approaches reflect cognitive activity early in the process. A new componential judgment task was developed which not only provided esti mates of attribute importance based on structural modelling, but also two sets of cognitive process measures based on cognitive components a ssumed to occur early and late in the judgement process. A greater deg ree of convergence between approaches was predicted when the cognitive approach was based on activity in the component occurring later in th e judgement process. Second, it was argued that in previous research s ubjects have had unlimited time to make their judgements, reducing the need for attribute importance to provide the dominant basis for deter mining processing strategy. The present experiment introduced a time p ressure condition and, on the basis of previous research, predicted th at this would increase the amount of information processing based on a ttribute importance, thereby increasing the convergence between estima tes of attribute importance derived from the two approaches. The first , but not the second hypothesis was supported and the results were dis cussed in terms of their implications for understanding previous diffe rences between the two approaches to human judgement.