Verbal overshadowing: a sound theory in voice recognition?

Abstract

SUMMARY Verbal overshadowing is the impairment of a person’s recognition ability as a result of generating a verbal description. Two experiments involving 169 participants examined the effects of verbal overshadowing, race of voice (own/other) and cognitive style (holistic/analytic) on voice recognition. In Experiment 1, participants heard a recorded voice (own- or other-race) saying a short phrase. After completing a cognitive style analysis and 15-minute filler task, the verbalisation group gave a written description of the voice while the control group did a filler task. Participants then attempted to identify the voice from a 6-voice lineup tape. Experiment 2 manipulated the similarity of the own-race voices by using telephone recordings of the voices, and the encoding-test similarities of the stimuli by using different phrases at encoding and test. Results showed a strong own-race bias with superior own-race voice recognition and no verbal overshadowing in Experiment 1, and a strong verbal overshadowing effect in Experiment 2. Cognitive style was predictive of voice identification in both experiments. Results are discussed with reference to the own-race bias, cognitive style and encoding-test similarity of the stimulus in verbal overshadowing. Copyright # 2005 John Wiley & Sons, Ltd. While much research into human memory has suggested that verbal rehearsal and elaboration of information improves memory performance, a series of experiments by Schooler and Engstler-Schooler in 1990 raised doubt that this is always the case. In their experiments, Schooler and Engstler-Schooler asked participants to watch a 30-second videotape of a staged bank robbery. After a 20-minute delay, during which participants did an unrelated filler task, they were asked either to describe each feature of the robber’s face in detail, or to continue with the filler task. Finally, participants were asked to identify the robber from a line-up of eight verbally similar faces. The results showed that participants who had described the target face were significantly less likely to identify the face. As Schooler and Engstler-Schooler (1990) observed, much of the previous research into memory performance used verbal stimuli, and such stimuli were well-suited to verbal rehearsal and elaboration. In contrast, the modalities of stimulus and task in their own experiments were disparate (visual and verbal), with faces not lending themselves easily to verbal description. They hypothesised that verbal rehearsal or elaboration would be