Integration of features in comparing multifeature stimuli

For a series of trials, the subject indicated whether or not two face-like stimuli were identical with respect to the features they contained (a positive or negative response, respectively); reaction time (RT) was recorded. The stimuli were separated by .5 or 4.0 sec, and each contained either two or three features. On some trials, the features of the first stimulus were represented within a single head outline; on others, they were distributed over several spatially separated outlines (components). Positive-response RT increased with the number of stimulus features and with the number of first-stimulus components (c); these two factors were additive. Negative-response RT increased with c, decreased with the number of differing features, and these two factors interacted. These effects did not vary with interstimulus interval. The results indicate that the subject did not integrate the first-stimulus components, but instead compared the second stimulus to each component, holistically (for positive responses) or analytically (for negative responses).