Information Technology Reference
In-Depth Information
overall. Studying guidance more generally seems reasonable before taking on the contingencies.
But the contingencies may be the keys to understanding guidance effects, which might wash out
at higher levels of analysis. One would expect individual, task, and environmental differences to
be fertile research areas in the future.
Montazemi et al. (1996) compared the effects of guidance for tasks with high and low com-
plexity. Their surprising result was that with low task complexity suggestive guidance led to the
best performance, but with high task complexity suggestive guidance was inferior to informative
guidance and no better than solving the problem without a DSS. They attribute the finding to dif-
ferences in the type of suggestive guidance provided, since the complexity of the more complex
task prevented inclusion of feedback in the form of “task information.” Analyzing the study raises
two significant matters. First, the study suggests that finding a means of supplying effective sug-
gestive guidance may be challenging for highly complex tasks. Second, since different suggestive
guidance mechanisms were employed for the low and high complexity tasks, one might ask if the
study is truly comparing the same form of guidance across two task types. Some might argue that
the findings are as much a statement about differences in the guidance mechanisms as about dif-
ferences in the tasks. Others might claim that the differences in task types drive the differences in
guidance mechanisms so the comparison and conclusions are appropriate. Either way, one won-
ders how often the issue of different guidance mechanisms will arise when trying to compare
guidance effects across different task types. And it presages the difficulties yet to be encountered
when trying to generalize findings from individual studies of guidance.
Mahoney et al. (2003) compared the effects of decisional guidance for selecting displays
between field independent subjects and field dependents, finding that with guidance the two
groups performed equally well. Because field independents outperformed dependents when dis-
plays and tasks were mismatched, they see decisional guidance as a means of improving the per-
formance of the field dependents. More generally, this study's finding suggests that individual
differences may be important for understanding the effects of decisional guidance.
Restrictiveness
System restrictiveness and decisional guidance are two different system features that affect user
behavior. But while they are distinct system attributes, they have an important relationship: They
are alternative means to an end. Systems intended to direct user behavior can do so by restricting or
by guiding. Several of the studies have implications for understanding the restrictiveness-guidance
relationship.
Antony et al. (2005) implemented two versions of their knowledge-based system for database
design. One was restrictive, forcing users to comply with the rules, and the other was guiding,
facilitating and encouraging, but not requiring, rule compliance. The two approaches were
equally effective in contributing to user performance. But the restrictive system was perceived as
easier to use. Although this study is just one data point, it raises the possibility that—at least in
some situations—restricting may be more desirable than guiding as a means of directing users.
This possibility is especially important given the emphasis some DSS commentators place on
flexibility (non-restrictiveness).
Wilson and Zigurs (1999), as well as Mahoney et al. (2003), used the theory of cognitive fit
(Vessey and Galletta, 1991) to direct decision makers to data displays fitting the tasks they were
performing. The findings of both studies support the theory of cognitive fit. Wilson and Zigurs
(1999) found that subjects given only displays matching the task were more accurate than those
who chose their own displays. Mahoney et al. (2003) found that subjects given only displays
Search WWH ::
Custom Search