Information Technology Reference
In-Depth Information
covering a broad range of features. And generalizing is likely to be made even more difficult as
differences in tasks, people, and environments introduce contingencies that matter. So few find-
ings are likely to be universal.
As a starting point, researchers should carefully position the guidance mechanisms they study
within the cells of the typology so that we can compare similarly focused studies to draw broader
conclusions about a given kind of guidance. Doing so should also facilitate generalizing or mak-
ing useful comparisons across cells of the typology. Studies that compare guidance mechanisms
that differ only on a single dimension are especially useful as we look to generalize beyond indi-
vidual mechanisms. The empirical studies reviewed here were limited to those explicitly defined
in terms of decisional guidance. Many other existing studies of system features could have been
framed as studies of guidance. Had they been expressed in terms of guidance and positioned
within the typology, those studies could have contributed not only to our understanding of the
given features but also to our cumulative understanding of decisional guidance.
Studies whose guidance is theory-based and whose effects are explained by the theory are
especially valuable and likely sources of generalization. Using DSS research again as an analogy,
Todd and Benbasat (1999) conducted a program of research focusing on essentially the same set
of decision aids for one class of problems—multi-attribute decision making. Their designs and
their findings draw on behavioral decision theory pertaining to the way decision makers trade off
effort and accuracy. Positioning their work this way makes their findings applicable to a much
broader range of systems and problems than just this specific set of aids or class of problems.
Their findings about how people trade off effort and accuracy in the context of computer-based
decision aids can be applied to studying other decision-making problems and implementing other
support systems. Similarly, Wilson and Zigurs (1996) and Mahoney et al. (2003) both based their
guidance on the theory of cognitive fit (Vessey and Galletta, 1991). One can therefore investigate
the applicability of their findings to tasks and displays other than the ones they studied.
Studying the Typology
Before concluding, there is value in identifying the most promising issues to study for each
dimension of the typology:
Targets
The targets dimension distinguishes two very different kinds of guidance: guidance for choosing and
guidance for using the system's functional capabilities. These are not alternative types of guidance
for a given discretionary opportunity. Rather they correspond to two different points of interaction
with the system: points at which one chooses what to do and points along the way of doing it.
Comparing the two would not make much sense and one should take care in applying findings con-
cerning one target to the other. It may be interesting, however, to see if such factors as automatic ver-
sus on-demand guidance produce the same findings for both targets.
Directivity
Informative guidance seems fundamentally different from the two directive forms of guidance
(quasi-suggestive and suggestive), since the motivations are different: Informative guidance reflects
a desire to enlighten without influencing, whereas directive guidance points users in a given direc-
tion. Given the difference in motivations, it would seem that designers do not need to choose one
Search WWH ::




Custom Search