Exemplar Prototypes and Clinical Diagnosis: Toward a Cognitive Economy

Abstract

Diagnostic categories are analogous to fuzzy concepts. As a result, the current classification system of mental disorders (DSM-III) describes diagnostic categories as prototypes with sets of “correlated features” rather than selected “necessary and sufficient” criteria. The correlated-features assumption implies that category membership is a matter of degree: Cases can vary in typicality on a continuum of goodness of fit from the typical instances to the borderline, atypical ones. The studies reported in this paper tested the utility of different representations of diagnostic information in aiding the detection of this continuum and increasing confidence in diagnosing typical cases. Results replicated the typicality effect (i.e., reliability of diagnostic judgments is greater for typical than atypical cases; Cantor, Smith, French, & Mezzich, 1980) and suggested that summary prototypes served untrained diagnosticians especially well in recognizing variations in typicality and enhancing confidence, whereas exemplar prototypes were more helpful for trained diagnosticians.