The Influence of the Serial Order of Visual and Verbal Presentation on the Verbal Overshadowing Effect of Dynamic Scenes

This article investigates the influence of verbalization processes on the visual recognition performance of dynamic scenes. In the recognition phase different distractor items were used: event model incompatible and event model compatible. It was hypothesized that source confusion, which is said to be responsible for inducing the verbal overshadowing effect, is reduced with event model incompatibility. In two experiments people viewed a dynamic scene and read a verbal summary related with the scene. The first experiment showed a verbal overshadowing effect when the verbal summary was presented after the film. In the second experiment – the verbal summary was shown before the film – recognition performance improved with event model incompatibility. Source confusion could be eliminated by changing the order of visual and verbal presentation. Verbal Overshadowing The question whether verbally describing a visual stimulus fosters or hinders its subsequent recognition has got a long tradition in cognitive psychology. Read (1979) showed that face recognition performance was improved using verbalization. More recently verbal overshadowing has come in the focus of research. Schooler and Engstler-Schooler (1990) examined the role of verbalization processes. In the verbal overshadowing paradigm, subjects first see a visual stimulus (e.g. a face) and then have to describe this stimulus verbally. Typically, in a subsequent recognition test, their recognition accuracy is lower (Schooler, 2002). In a meta analysis Meissner and Brigham (2001) analyzed 29 studies examining the verbal overshadowing phenomenon and found a negative effect of verbalization processes on the recognition performance. While active verbalization induces a transfer inappropriate processing shift, that dampens individuals’ ability to apply certain non-verbal operations, Dodson, Johnson, and Schooler (1997) evinced that a verbal overshadowing effect (VOE) even appears if participants read a verbal description passively. As recognition performance did not show a VOE when participants were asked to ignore their verbal representation, source confusion about the validity of which mental representation should be the basis for the recognition was made accountable to this kind of VOE. Although verbal overshadowing is a robust effect it could be shown that it disappeared under certain circumstances. Kitagami, Sato, and Yoshikawa (2002) found a VOE when distractor items were designed highly similar to the target items, while no verbal overshadowing effect appeared at low similarity. In a former study Bartlett, Till, and Levy (1980) found, using realistic photos as stimulus material, that verbalization led to higher recognition performance, when it enabled the participants to distinguish between target and distractor items. These findings are in accordance with the assumption that viewers develop a kind of model which specifies the characteristic and relevant features of the visual stimulus. This model enables the viewers to distinguish between target and distractor items in a recognition paradigm. On the one hand, model incompatible distractor items are easier to identify and on the other hand, model compatible distractor items are less likely identified. Additionally, it can be assumed that models directly derived from the visual stimulus contain more detailed information than models derived from the verbal description of the visual stimulus.