Differentiation and Sociality in Terms of Elicited and Provided Constructs

Previously unacquainted undergraduates (15 men, 15 women) interacted briefly in randomly assigned pairs, and subsequently attempted to infer the self-evaluations of their partners in terms of both personal constructs elicited from the latter and a set of supplied constructs. Inferences based on elicited constructs were significantly more accurate (p < .005) than were those based on supplied ones. Also, the self-evaluations of participants with relatively less differentiated personal-construct systems, as assessed by the repertory-grid procedure, were inferred by their partners significantly more accurately (p < .01) in terms of elicited constructs than were those of participants with relatively more differentiated personal-construct systems. These findings are interpreted within the framework of personal-construct theory.