ABSTRACT Personal Construct Theory (Kelly, 1955) provided a framework in which the congruence of social perceptions was examined Triads of close acquaintances engaged in personality description procedures presented at two levels of abstraction (global and specific) Using a modification of Kelly's (1955) Role Construct Repertory Test (REP) grid methodology, target persons in each triad generated personally relevant self-constructs, contexts in which such self-constructs might be displayed, and actions which might be expressive of those self-constructs This resulted m an idiographic data base on which peers attempted to construe the target's personality in a variety of contexts The relationships between several self-construct system properties (such as Mean-ingfulness and Prototypicality) and congruence (shared construing) were assessed Patterns of correlational analyses indicated that the most meaningful self-constructs (i e, those manifested in extreme form), as well as self-constructs exhibiting the least variation across contexts, produced greater agreement among triad members about the target's personality It was suggested that the study of self–peer congruence is feasible using idiographic, self-identified personality themes, explicit contextual information, and a systems perspective