A Test of Two Implicit Assumptions of Algebraic Modeling: Salience and Familiarity Effects in Affective Evaluations of National Leaders

Abstract

Within the framework of a hierarchical model of social cognition, algebraic models of attitudinal evaluations such as Fishbein's (Fishbein & Ajzen, 1975) rely on two undefended implicit assumptions. When predicting the evaluation of category subordinates (i.e., specific national leaders such as Reagan and Thatcher), Fishbein often uses attributes relevant and salient to the category superordinate (the “Ideal National Leader”) and assumes that the attributes are (a) familiar and (b) salient to participants' views of the subordinates. Undergraduates (N = 50) were asked to make probability, familiarity, favorability, and importance ratings of 10 national leaders on attributes either salient or nonsalient to the Ideal National Leader. Results supported Fishbein by showing that attributes most salient to the superordinate are acceptable for predicting the evaluation of subordinates. Other results cast doubt on the superiority of Fishbein's summative model over other algebraic models and on his method for collecting favorability ratings of attributes.