A Study on System Dialog Message Design by Age Considering Interruption in Task-Switching

Abstract

Due to the wide use of smart devices, users are often exposed to system dialog messages during interaction. These messages are mainly interruptive, since what they say is not relevant to the intention of the user and to the existing task. The interruption requires extra work to be done in a different context from the existing task, which needs additional cognitive effort. In particular, this cognitive task-switching process may have a negative impact on the performance of the old users, whose cognitive abilities decline. In this study, the interrupting task was designed to handle several smartphone system dialog messages in order to infer which terminological properties make them interruptive by age. 24 subjects of two age groups (twenties and sixties) followed text reading tasks which contain dialog messages with two terminological attributes (the number of technical jargons, the length of message), observing the timeline of three task switching process (main task, interruption task, resumption) as subtasks. Considering these conditions, task performance time, comprehension rate, and subjective workload were measured. As a result of the experiment, it was shown that the task performance time of sixties was significantly longer than those of twenties. Correlation analysis for the partial time of the task switching revealed that for old users, the time spent for the resumption was positively correlated with the time cost for interruption. In addition, the old subjects experienced higher subjective workloads by the messages with many jargons, and this load tended to continue in the following process, the resumption. While previous message studies only considered the complexity of the dialog messages themselves, this study tried to consider them in the total work condition discussing behavioral characteristics of processing the interruption by age during the task-switching simulation. In addition to the design guideline for the practical message construction, we also suggest cognitive considerations to maintain the performance of the old user at the appropriate level in such a task-switching condition.