DECISIONAL GUIDANCE - Human-Computer Interaction and Management Information Systems: Foundations - page 72

C HAPTER 6
DECISIONAL GUIDANCE
Broadening the Scope
M ARK S. S ILVER
Abstract: Decisional guidance was introduced as a design feature of decision support systems
(DSS) more than a decade ago. It was originally defined as how a DSS enlightens or sways its users
as they choose among and use the system's functional capabilities. Unlike mechanical guidance, the
interface feature that helps users with the technicalities of invoking and using functionality, deci-
sional guidance provides more substantive support for exercising discretion when choosing which
functions to employ and while employing them. During the years since its introduction, decisional
guidance has been used to study not only DSS but also a variety of other information systems,
including group support systems, executive information systems, and CASE tools. This paper begins
by summarizing the basic concepts of decisional guidance and reviewing how it has been studied
empirically over the years. The paper revises the original definition in light of the review and broad-
ens it to apply to information systems more generally. The typology of guidance is also updated. The
paper concludes by presenting a more focused agenda for research based on the refined typology.
Keywords: Decisional Guidance, System Restrictiveness, System Design, System Effects, System
Features, System Flexibility
Decisional guidance refers to the features of an information system that affect the choices people
make when interacting with that system. Users of an interactive system typically confront numer-
ous choices, such as what to do next, which option to select, or what input value to provide.
System features can guide these choices either deliberately, when guidance is intentionally built
into a system, or inadvertently, as an unplanned consequence of the system's design. Even guid-
ance that is deliberate is not necessarily directive; decisional guidance may sometimes attempt to
influence the user's choice but other times may only enable a more informed choice.
Consider a few examples:
• A user of a statistical package is about to run a regression analysis. The software asks him if
he wants to transform the data first. The user does not know why he would want to do this,
but when he presses the Help button the system explains the implications of autocorrelation.
• An electronic spreadsheet supports ten different chart types, each with ten subtypes. A user
is about to plot time series data and must select one of the one hundred possible chart types.
The system recognizes that the data represent a time series and recommends the chart type
and subtype the user used for time series data in the past.
• An architect is designing a building with CAD software. After he makes a series of design
choices, the system warns him that the building is structurally unsound.
90