| Abstract: | Regulation and monitoring of successful performance in complex tasks requires decisions regarding which processes to engage in and in what order. Coordination of these activities appears to be a separable aspect of the demands imposed on working memory by cascaded serial processes (Carlson & Lundy, 1992). In the experiments presented here, subjects indicated whether symbolically suggested transformation rules (rotation, reflection) correctly described the difference between two patterns of filledin squares within a 3x3 matrix. Two experiments were carried out with different levels of practice and common transfer tasks. Our aim was to provide evidence for an executive coordinating mechanism on the basis of differential practice and transfer effects. The following results were obtained: (1) Coordination complexity (Mayr & Kliegl, 1993) has a nonadditive effect on transformation time at all levels of practice, suggesting that a substantial deal of time is spent on processes other than executing transformations. (2) Performance on the transfer task indicated that the practised coordination skill was generalised under the conditions of the experiment. Level of practice influenced the transfer pattern, showing more specificity in transfer after extended practice. Taken together, the nonadditive effect as well as the transfer effect suggest that coordination cost reflects separable processing demand. The results constrain the type of interaction between the visual working memory system and the central control component of multicomponent working memory models (Baddeley, 1989). |
