The efficacy of the resource-scarcity and outcome-conflict views in explaining dual-task interference was examined. A discrete-continuous task pair was purposely chosen to allow fine-grained analysis of time-shared performance. The relative priority of the dual task was manipulated by a secondary task technique to test for performance tradeoff that would be indicative of resource allocation. The temporal predictability of the discrete stimuli was manipulated to examine possible strategic avoidance of interference. The moment-by-moment data did not reveal any evidence for a switching strategy. It was concluded that the intricate interference patterns could be more easily interpreted within the resource framework than within the outcome-conflict framework.Time-shared performance typically degrades from single-task performance level. The present study examines the viability of two explanations for this degradation: resource scarcity and outcome conflict.First, it is important to make clear what is meant by timesharing performance. Time-sharing in the computer domain refers to several terminals accessing a central processor at different times. In other words, one central processor serves several terminals by rapidly switching between them. Time-sharing in this sense is sometimes used in the psychological literature. For example, Hirst and Kalmar (1987) used time-sharing to refer to rapid switching between the processing oftwo tasks. In contrast, time-sharing in the present paper is used along the tradition followed by Gopher (1993), Friedman, Polson, andDafoe (1988), Moray (1967), andWickens (1990). Here, time-sharing refers to the simultaneous processing of two or more tasks. Resource TheoryAccording to resource theories (e.g., Gopher, 1986;Kahneman, 1973), time-shared performance will compete for limited resources if certain conditions are met.

Interactive effects of age, expertise, and structural similarity on time-sharing efficiency were examined. Half of 90 subjects who ranged from age 20 to 80 years were nonpilots. The other half were pilots who were considered to have expertise in time-sharing. Five dual tasks were selected to represent various cognitive aspects of flight performance and to represent various degrees of structural similarity defined by Wickens' multiple resource model. Several main findings were of note. One, time-sharing efficiency increased as structural similarity decreased. Two, time-sharing efficiency decreased with increased age. Three, pilots had higher level of time-sharing efficiency than nonpilots. Four, expertise in time-sharing appeared to be able to moderate some of the deleterious age effects. Theoretical and practical implications of these findings were considered.

The secondary task technique was used to test two alternative explanations of dual task decrement: outcome conflict and resource allocation. Subjects time-shared a continuous tracking task and a discrete Sternberg memory task. The memory probes were presented under three temporal predictability conditions. Dual task performance decrements in both the tracking and memory tasks suggested that the two tasks competed for some common resources, processes, or mechanisms. Although performance decrements were consistent with both the outcome conflict and resource allocation explanations, the two explanations propose different mechanisms by which the primary task could be protected from interference from the concurrent secondary task. The primary task performance could be protected by resource allocation or by strategic sequencing of the processing of the two tasks in order to avoid outcome conflict. In addition to examining the global trial means, moment-by-moment tracking error time-locked to the memory probe was also analyzed. There was little indication that the primary task was protected by resequencing of the processing of the two tasks. This together with the suggestion that predictable memory probes led to better protected primary task performance than less predictable memory probes lend support for the resource explanation.