DESIGNS THAT FIT - Human-Computer Interaction and Management Information Systems: Foundations - page 192

This limited view of physical fit and its impact on well-being may very well change with the
advent of new input-output devices. In particular, virtual reality systems in the workplace will
sense human behavior by means of physical devices, such as displays that track eye movement
and gloves that track hand motion. The computerized system (i.e., the virtual environment) will
be designed to react appropriately and communicate back to the user with haptic feedback, advanced
graphics, and auditory cues. Thus, the physical aspects of human interaction will increase dra-
matically the diversity and complexity of the physical design, and, most likely, will produce a cor-
responding demand for a broader view of physical fit.
In sum, physical fit ensures minimal physical effort to accomplish the task and consideration
for the user's overall well-being. The reliance on limited devices to enable input (primarily motor)
and output (primarily visual and some auditory) has characterized computers in organizational
settings. The general tendency to design hardware that fits all sizes and all tasks may, however,
change with the advent of more advanced interactive technologies such as virtual reality systems
that build on multiple senses. With it may come a stronger interest in physical fit and in the impact
of fit on performance at the task level.
Cognitive Fit
In the IS literature, the most influential case of fit in HCI design is cognitive fit. In summarizing
the design implications of cognitive fit, Vessey and Galletta (1991) state that “for most effective
and efficient problem solving to occur, the problem representation and any tools or aids employed
should all support the strategies required to perform that task.” While the previous discussion
emphasized the fit with user capabilities and implicitly assumed a set of tasks to be performed,
cognitive fit emphasizes the fit with the user's model of the task demands. Moreover, cognitive
fit proposes a theory that explains the link between fit and performance. Assuming the user has a
mental model (representation) of the task demands, the computer representation displayed to
the user can either match or not match the user's model. A match ensures consistency between the
problem-solving processes that are appropriate for both representations, reducing the propensity
for error and reducing the effort and time required. On the other hand, a mismatch between
the two representations requires a transformation of one representation to suit the processes that
fit the other representation, reducing accuracy and increasing time. Hence, fit is positively related
to performance.
Cognitive fit theory was developed for the case of spatial and symbolic tasks that rely on per-
ceptual and analytic problem-solving processes: Graphic displays fit spatial tasks, encouraging
perceptual processes, and tabular displays fit symbolic tasks, encouraging analytic processes. Vessey
(this volume) explains how the theory applies to simple tasks (Vessey, 1991) as well as complex
tasks (Vessey, 1994). In complex tasks, the user chooses between alternative problem-solving
processes on the basis of a cost-benefit analysis, in which a user balances the effort required to
execute the process with the expected performance benefits. At some point, users judge the effort
required to execute a more effortful problem-solving process such as analytic processes to be too
high relative to the expected benefit and, therefore, shift to the easier perceptual processes. The
theory cannot, however, predict the point at which a user will shift from one process to another.
Nevertheless, the theory can predict the direction in which external influences, e.g., motivation to
be accurate rather than fast, will influence the choice of problem-solving processes. Furthermore,
fit is binary. The fit between representations is defined indirectly as positive when a graphic dis-
play is used for a spatial task or when a tabular display is used for a symbolic task. Other combi-
nations result in no fit. This precludes intermediate levels of fit.