Concept Learning: Similarity and Number of Exemplars Category Number Information, and Description of Prototypical Features

Abstract

College students classified the sex of day-old chicks from drawings of their genital eminences. Employing geometrie forms as one category of training stimuli increased errors on test stimuli. This between-category similarity outcome contrasts with others to some extent, and it also argues against the sufficiency of a template model of conceptual functioning. Training with only a single pair of exemplars also impaired transfer. Other research has confounded exemplar number with amount of training Informing subjects of the number of categories the drawings belong to failed to facilitate transfer significantly. A description of the prototypical features differentiating the two sex categories enhanced transfer. Several aspects of the communication may have been influential.