DECISIONAL GUIDANCE - Human-Computer Interaction and Management Information Systems: Foundations - page 93

“timing” dimension. Long-range guidance is now a form of prospective guidance, and short-range
guidance is now concurrent guidance.
The topics of feedforward (Bjorkman, 1972) and feedback—especially, cognitive feedback
(Te'eni, 1991)—have been receiving increasing attention in the information systems literature.
How do they relate to the various types of guidance? Feedforward, which provides information
about a task before the person begins to perform it, would most likely be prospective guidance or
possibly concurrent guidance, depending on how it was implemented. Feedback, defined as
“information about the decision-making process and its outcome” (Te'eni, 1991, p. 644) is a bit
trickier. If the feedback is given for the reasons suggested already—to allow the user to correct an
entry or to guide the user in an iterative task—the feedback would qualify as retrospective guid-
ance. But much feedback is intended not to influence present system use at all, rather to enable
user learning for the future. Such feedback would lie outside the realm of guidance, which
informs, sways, or directs the user in exercising discretion within the system. So not all feedback
is decisional guidance. But research in cognitive feedback could be a good source of mechanisms
for retrospective guidance.
Guidance for Groups
Decisional guidance for groups seems to be an especially fertile area for research. Although there
can be some crossover—for instance, some of the MCDM guidance included in Limayem and
DeSanctis's (2000) system drew upon individual techniques—single-user guidance and group
guidance are likely to be largely independent bodies of knowledge. A variant of this typology
with added dimensions may be needed for group research and practice, since much of GDSS
guidance focuses on the special needs of group interaction. For example, most of the guidance
provided by Limayem and DeSanctis (2000) was targeted at three types of “group breakpoints.”
Similarly, Wheeler and Valacich (1996) distinguish between active and passive appropriation
mediators—here, we can think of these as guidance features—noting that the multiple channels
of group interaction may weaken the power of the computer-based system to guide. A prominent
topic in GDSS research is the role of the human facilitator. Research on guidance for groups may
help understand if and how a human facilitator can be replaced by computer-based features. Other
topics of interest would likely include the role of guidance when groups are not co-located or
when they meet asynchronously.
RESEARCH AGENDA FOR DELIBERATE DECISIONAL GUIDANCE
With the revised typology in place, we return to the question with which the paper began:
• How does decisional guidance affect user behavior?
Although Decisional Guidance is focused on user discretionary behavior, fully understanding
guidance and its effects requires also studying user perceptions and user learning. People's per-
ceptions may influence if and how they use a system and its guidance. Improved or degraded user
learning may be a side effect of guidance that is of concern to designers and users. All told, three
sets of consequences are of interest: effects on (1) behavior, (2) perceptions, and (3) learning.
My purpose here is not to create an exhaustive set of issues to be studied but to pare the possible
issues down to a manageable size, allowing us to focus on the most important. Even so, the list is
substantial.