DECISIONAL GUIDANCE - Human-Computer Interaction and Management Information Systems: Foundations - page 81

(2004) found that guidance contributed positively to the “integrity” component of trust. Jiang and
Klein (2000) found that decisional guidance affected the strategies users employed.
Targets
The nine studies were fairly evenly spread over the two targets of guidance (Figure 6.1), with three
targeting the structuring of the decision-making process (selecting a forecasting method and select-
ing an appropriate data display), three focusing on the execution (assigning jobs to machines,
answering questions posed by an e-commerce recommendation agent, and designing databases),
and two doing some of each (Table 6.3). This dimension of the typology seemed fairly straightfor-
ward, with the various studies easily placed in categories. Nonetheless, a potential ambiguity arises
in a study not framed in terms of decisional guidance that could be recast as such. The paper
(Singh, 1998) studied two cognitive aids that support the “strategy execution process” by helping
users select the proper next step in an already structured decision-making process. Since the
process was already formulated, referring to the invoking of each step as “executing the process”
seems entirely appropriate. The typology, however, would characterize this activity as structuring
the process rather than executing it, since structuring the process refers to the selection of func-
tional capabilities and executing it refers to controlling or interacting with those capabilities once
invoked. Given the conflict between the formal definitions and the intuitive usage in this study, the
formal definitions should be revisited.
Forms of Guidance: Informative Versus Suggestive
Three of the empirical studies compared the effects of informative guidance with those of sug-
gestive guidance. Parikh et al. (2001), in the most extensive study of decisional guidance thus far,
constructed a DSS for a forecasting task, finding that suggestive guidance outperformed inform-
ative guidance in terms of decision quality and decision time (not total time), but informative
guidance led to greater user learning. The task required users to choose a forecasting method for
a time series. In predefined mode, informative guidance for a given forecasting method provided
information such as the following:
Least-square regression model
When you plot the data with time on x-axis and a variable
(sales, profit, cost, etc.) on y-axis, if you see a linear trend,
this model should be used. A linear trend gives same slope any
point on the trend line. Least-square regression technique
identifies the trend and gives us slope and intercept.
Suggestive guidance in this mode included the same information plus such additional content
as this:
Your data has . . .
A linear trend
Low fluctuation
No seasonal change
Under the current typology, the study's suggestive guidance does not fit comfortably in either cat-
egory. It does not strictly qualify as suggestive guidance because it does not explicitly recommend a