Abstract
In an experiment testing the extension of the von Restorff/isolation effect to feature recognition, observers viewed a study set of 20 faces with 0, 1, 2, or 4 distinctive stimuli on each face. In Experiments 1 and 2, the stimuli were black spots (simple, distinctive stimuli), and in Experiments 3 and 4 they were inverted es (complex, distinctive stimuli). Observers then viewed a test set containing 20 study-set and 20 distractor faces. In 2 of the experiments (1 & 3), the test sets contained the same distinctive stimuli as the corresponding study set; in the other 2 experiments (2 & 4), the test sets had no distinctive stimuli. In Experiments 1 and 2, faces with 1 spot on them in the study/learning phase were recognized significantly more frequently than those that had 0, 2, or 4 spots. That this one effect was still present when there were no spots on the faces in the test set demonstrates that it is an encoding phenomenon rather than a recall phenomenon. As predicted, that result was not replicated when inverted es were used (Experiments 3 and 4).