Integrative Approach To Decision Making

Approaches to decision making

A fundamental distinction has to be made regarding the approaches to decision making. Normative theories involve the so-called decision-science, the development of formal models about how people should think and take decisions. Such theories are strongly based on the concepts of “rational” people, perfect information and optimisation, and are therefore highly formalised. Descriptive theories study and suggest models about how people in real context really think and take decisions, therefore addressing the main critics to the first kind of theories. Finally, the integrative approach blends insight from both theories, striving to develop formal models and principles, but taking into consideration the complexity of

It suggests that people infringe the conditions posed by the rational approach, as they have limited information, limited ability to process it and often take satisfactory rather than optimal decision. The perspective is based on the research on heuristics and biases. Heuristics are “decision shortcuts”, quick methods such as rules-of-thumb and reliance on past experience often used in practice. Although they are useful can potentially lead to the same solution as more pondered approaches, under certain conditions they can

Group and organisational perspectives see decision making as an organisational process, as such, it is influenced by organisational policies, procedures and politics. Supporters of this school of thought identify risks (e.g. groupthink) and suggest best practices to manage decision making in groups.

Finally, the naturalistic perspective argue that decision making cannot be described in general terms, thus it cannot be formalised in models and procedures. Rather, managers should “reflect in action”. Particular stress is placed upon intuition, which depends on experience and allows experts to recognise and quickly react, almost automatically, to situations. The shortcoming of such view is the inability to address in the same way novel situations.

Models and tools for decision

In this case, non-probabilistic approaches are needed (such as maxi-max, maxi-min and mini-max).

Decision tree

The decision tree allows to decompose a decision situation in an orderly temporal sequence of decisions, environmental states and consequences, representing it graphically. While any decision table can be translated in a decision tree, the opposite is not true. In a single tree, more than one, temporally subsequent, table can be illustrated. Moreover, with trees it is possible to handle non-symmetric situations. While I will not explain the tree’s rollback solution method (Goodwin, Wright, 1999), the lying criterion is the simple: choose the alternative with the highest expected value.

Influence diagram

A graphical representation of a decision similar to the decision tree, rather than a temporal sequence of events the influence diagram emphasises the relationship of causal dependence or independence of the various components of a model. Therefore, elements that are not connected in a tree can be connected in the diagram because of a causal dependence. As such, a diagram is usually more complex than a tree. For a practical comparison see (Clemen, Reilly, 1996,