We propose that the concept of cognitive effort in memory is both useful and important. Cognitive effort is defined as the engaged proportion of limitedcapacity central processing. It Was hypothesized that this variable might have important memorial consequences and might also be a potential confounding factor in levels-of-processing paradigms. The first experiment tested this possibility using two types of incidental-learning tasks factorially combined with two degrees of effort. It was found that high effort led to better recall than low effort, but that level-of-processing effects were nonsignificant. A second experiment clearly demonstrated the feasibility of using performance on a secondary task as an independent criterion for measuring effort, and two further experiments ruled out alternative accounts of effort effects. A reliable levels-of-processing effect was obtained in the fourth experiment in which the incidental-learning tasks were blocked. Implications and possible future applications of the cognitive effort concept are discussed.

Three experiments examined the nature of individual differences and the role of advance information in reading comprehension. Subjects read short passages, in some cases preceded by a given type of advance organizer, then recalled the information therein, and finally sorted ideas from the passage into groups of similar ideas. Parameter estimates for the Kintsch and van Dijk (1978) model, together with a derived measure for the idea-sorting task, showed that good readers were better at recalling propositions and organizing ideas than poorer readers. When the effects of different types of advance organizers were considered, it was found that good readers usually showed greater recall of detail when given either type of advance organizer, whereas poorer readers displayed enhanced recall of detail only for a particular type of advance organizer.

During the development of a system, software modules can be viewed in terms of their commitments: the constraints imposed by their own structure and behavior, and by their relationships with other modules in terms of resource consumption, data requirements, etc.. The Comet system uses explicit representation and reasoning with commitments to aid the software design and development process in particular, to lead software developers to make decisions that result in reuse. Developers can examine the commitments that must be met in order to include an existing module, and can explore how commitments change when modules are modi ed. Comet has been applied to the domain of sensor-based tracker software.

This paper describes the architecture and prototype of a system which dynamically determines how to present information to a user. The system utilizes a rule based approach to select one or more modalities for presenting information.Next the system determines one or more techniques to present the information within each of the previously selected modalities.This system also adapts to individual users providing flexibility not found in traditional presentation systems. Finally, models are used for storing knowledge about the user resulting in a system which can be easily enhanced as new data is obtained and can adapt to the needs of its users.

No abstract