The impact of predictability on dual‑task performance and implications for resource‑sharing accounts

The aim of this study was to examine the impact of predictability on dual-task performance by systematically manipulating
predictability in either one of two tasks, as well as between tasks. According to capacity-sharing accounts of
multitasking, assuming a general pool of resources two tasks can draw upon, predictability should reduce the need
for resources and allow more resources to be used by the other task. However, it is currently not well understood what
drives resource-allocation policy in dual tasks and which resource allocation policies participants pursue. We used a
continuous tracking task together with an audiomotor task and manipulated advance visual information about the
tracking path in the first experiment and a sound sequence in the second experiments (2a/b). Results show that performance
predominantly improved in the predictable task but not in the unpredictable task, suggesting that participants
did not invest more resources into the unpredictable task. One possible explanation was that the re-investment
of resources into another task requires some relationship between the tasks. Therefore, in the third experiment, we
covaried the two tasks by having sounds 250 ms before turning points in the tracking curve. This enabled participants
to improve performance in both tasks, suggesting that resources were shared better between tasks.