The Effect of Communication Style on Task Performance and Mental Workload Using Wearable Computers

This thesis measured the mental workload associated with operating a voice activated software application run on wearable computer under five different communication styles (buttons, command line, icon buttons, icon text menus, and text menus). The goal of this thesis was to determine which communication style would be best allow wearable computer users to simultaneously perform other non-computer tasks. Thirty subjects were randomly assigned to using one of five software versions (n = 6), each of which utilized a unique communication style. The mental workload associated with operating each version was assessed by monitoring the performance of secondary tasks. Secondary tasks consisted of completing a block assembly, digit subtraction, and walking along a marked pathway. Each secondary task was performed twice by itself and once while operating one of the software versions, creating a total of nine trials per subject. Block assembly task performance measures included average assembly time, percentage correct blocks, and percentage correct blocks attempted. Digit subtraction measures included percentage of correct digits. And path walking measures included average walking speed. Subjective estimates of mental workload were also collected for those trials in which subjects operated the wearable computer and performed physical tasks using the NASA Task Load Index (TLX). Finally, usability information was collected for each software version via a questionnaire form.

Each of the five versions of the experimental software application was operationally identical to the others, but utilized a separate communication style. The button version displayed available functions via sets of labeled buttons in the control screen. The icon button version replaced the appearance of these buttons with labeled icons. The text menu version displayed available functions textually via a pull down main menu. The icon text version displayed appended icons to the left of each main menu item. Finally, the command line version displayed no labels, buttons, menus, or icons for any functions. The experimental software was designed as a day planner/scheduling application used to set reminder dates on a calendar, edit task lists, and edit phone listings.

Under the multiple resource view of mental workload, it was hypothesized that the different versions and secondary tasks would demand distinct types of mental resource and, consequently, that mental workload would be observed as lowest when the version and secondary task demanded different types of mental resources. In contrast, it was also hypothesized that mental workload would be observed as highest when the version and secondary task demanded the same type of mental resources. Although separate one way ANOVAs performed on all secondary task measures failed to indicate statistically significant differences in mental workload across the versions, secondary task performance was consistently observed as best for subjects using the icon button version. Analysis of NASA TLX subscale data indicated that the block assembly task was rated as requiring less effort and the digit task rated as requiring less mental demand when the icon button version was used. These results generally support using an icon button communication style for wearable computer software applications.

Results of this study are applicable to the design of the user interface of wearable computers. These results not only report subjective and objective measures for assessing the amount of mental effort associated with operating a wearable computer and performing various physical tasks simultaneously, but also provide estimates for determining the amount of physical task performance decrement to expect when wearable computer are also operated. Such data may be used to determine human factors guidelines for matching wearable computer interfaces to physical tasks so that interference between the two is minimal.