THE INFLUENCE OF MODELS ON THE EVALUATION OF ABILITY

A hypothesis derived from social comparison theory was tested through 3 experiments. Ss were required to predict their performance on a novel task after a good or poor model had actually completed it. The experimental situation was varied in the respect of Ss' experience on that task (Exp. I), of the information about Ss' ability (Exp. II), or of the degree of the information enhances or threatens Ss' self-esteem (Exp. III). With no experience or information, Ss predicted higher performance when the model was good than he was poor. But when Ss had experienced the task or received information about abilities, model had no influence. While Ss were influenced only by the good model when the information enhanced their self-esteem, the poor model had strong influence when their self-esteem was threatened.