Matching of facial features: Continuous processing, improper filtering, and holistic comparison

Abstract

The aim of the present study was to investigate the effect of irrelevant features on reaction time (RT) in a matching task, where the relevant and irrelevant features are parts of a well established unit, that is, the human face. In Experiment 1, subject had to decide whether successively presented pairs of photographed faces were the same or different with respect to relevant features (eyes or hair). As a function of the number of irrelevant different features (from zero to three) the "same" RT increased. The rate of increase was larger when the relevant feature was eyes rather than hair. RT was shorter in the hair stimulus condition than in the eye stimulus condition. "Same" RT was longer than "different" RT. In Experiment 2, the facial features were presented alone. RT for the hair was shorter than for eyes. In the hair stimulus condition, the "same" RT appeared to be shorter than the "different" RT, whereas the opposite result was obtained for the eyes. In Experiment 3, the main results of Experiment 1 were replicated in a procedure in which a "standard" face was continuously presented during an experimental run. Although no single model constructed for the matching tasks can explain all the data, the response competition approaches, supplemented by the possibility of stimulus set selection, and the contribution of holistic facial representation might be interpreted as accounting for the results. © 1985 Psychonomic Society, Inc.