Deriving decision heuristics

Assume that the decision-maker will define a set of threshold values and apply decision heuristics, which are logically consistent with his/her preference structures. Because for different decision-makers or in different contexts preference structures may differ in terms of the pattern of the sets of accepted and rejected overall values, the cognitive process model automatically generates heterogeneous decision heuristics. One extreme is the strictest preference structure in the sense that no overall value survives the overall threshold,


That means that regardless of the attribute states, the object under consideration will be rejected. In this case, information search is not needed (or the heuristic of 'no action').

Relaxing a little could lead to the preference structure where only the overall value of attribute states exceeding the highest respective threshold values,


This preference structure implies a conjunctive rule in the sense that the object will be accepted only when all attributes are in their highest states. During the decision process, any single attribute being unsatisfactory will cause the decision process to stop, regardless of the states of other attributes.

At the opposite end is the most relaxed preference structure, representing the case that all overall values will be accepted.


This preference structure implies the other 'no action' heuristic since attributes being in whatever state will lead to the object being accepted.

A little larger λ may result in a preference structure where all but the overall values of non-activated attribute states are accepted,


Disjunctive heuristics can be inferred from this preference structure since any attribute being satisfactory (activated) will cause the decision process to stop and accept the object, regardless of the states of other attributes.

Within the spectrum, various other preference structures and heuristics can be inferred. For example, the lexicographic heuristic is implied in such a preference structure,


According to this preference structure, at least one attribute j exists. When some states of this attribute are not activated, the decision process will stop and the object will be rejected. When some states are activated, the decision process will stop at accepting the object. In-between are those states based on which neither accepting nor rejecting the object can be determined and further consideration of other attributes is needed.

By varying the value of the overall threshold, many other decision heuristics can be exactly inferred based on this framework. The location of λ determines how a decision problem will be treated (a serious problem with very high standard or a minor problem with very low standard) and the involvement of the decision-maker (extensive or limited information search), which, of course, varies from person to person, from context to context.

< Prev   CONTENTS   Next >