Decision Making Models

Richard W. Scholl, Professor of Management, University of Rhode Island
Revised: October 2, 1999

Decision Making Models Summary
We discussed the three components of every decision:

Criteria- The standards by which decision makers evaluate alternatives. These are what Fisher and Ury term "Interests"

Alternatives- Specific courses of action or options, being considered. "Positions"

Cause and Effect Beliefs- Cause/effect belief are cognitions linking specific alternatives to specific criteria. These are often referred to as models, theories, assumptions, beliefs, or alternative attributes. Our attitude discussion, covered the relationship among the four attitude components namely; Cognitions, Affect, Evaluations and Behavioral Intentions. These C/E beliefs are cognitions or individual theories.

We will work from the deductive premise that ineffective decisions result from either:

Problems with the development of criteria such as incorrect criteria, improper weighing of criteria, or not all relevant criteria considered;

Search for alternatives- Did not identify alternative that would satisfy criteria set.

Use of incorrect cause and effect information.

What is a decision-making construct?
A decision-making construct is a model or plan which determines the process by which a decision is made. There are three major elements to a decision-making construct:

How are criteria determined? Which criteria should be used in making a decision? How does one deal with multiple, conflicting criteria?
How are alternatives generated?
How are alternatives evaluated against criteria, that is, how are cause/effect relationships established and how are cause/effect conflicts resolved?

These questions involve differences between rational and political models of decision-making, differences between group and individual decision-making, and relate to the manner in which problems are identified, diagnosed, and ultimately solved.

The effectiveness of the decision making construct may be judged by the following:
Does the construct allow/encourage the decision maker to include all relevant criteria?
Does the construct allow/encourage the consideration of all alternatives?
Does the construct allow/encourage the decision maker to have access to the most accurate cause/effect information?

Conflict from a Decision Making Perspective. Additionally, from a decision making perspective, we can view conflict as a breakdown in the decision making process. As such the following assumptions apply:

Conflict manifests itself as a disagreement over alternatives or positions. When there is agreement as to what option to pursue, what plan to implement, or what course of action to take no substantive conflict exists and thus, a decision is made. This does not mean that there is no emotive conflict or that parties to the decision share either the same values (criteria) or theories (cause/effect beliefs).

When conflict does exist (an impasse as to what option to choose), its source is one of the following:

Value Based Conflict- Disagreement over Criteria or Interests. This is a broad category which not only includes individuals attempting to accomplish mutually exclusive goals, but also includes differences in importance or weights placed on multiple criteria. Also included in this category are hidden agenda. Conflicts generated from this source are the type dealt with by the Fisher Getting to Yes strategies.

Theory Based Conflict- Disagreement over cause/effect beliefs or theories. In this case, individuals share the same goal or interests, but disagree about the best way to accomplishment this goal. The underlying cause is that they do not share a common theories regarding the relationships between alternatives and criteria.
Criteria or Goals Criteria or Goals
Agree Disagree
Cause/Effect Beliefs Agree No Substantive Conflict Valued Conflict
Cause/Effect Beliefs Disagree Theory Conflict Both Value and Theory Based Conflict


Dealing with Multiple Conflicting Criteria

The simplistic Auto Choice (in some classes we used a Network Printer choice decision) case required the use of a strategy for dealing with a set of multiple, conflicting criteria, that is, a set of criteria where a different alternative is selected when each criterion is used. The three strategies that were identified were:

Singling out a criterion to maximize- Using this strategy the decision maker simply eliminates the conflict by selecting what he/she believes to be the most important criteria and eliminating all others until the conflict is resolved.

Maximizing/Minimizing- This strategy requires the development of a utility index function. The decision maker then maximizes or minimizes this index and chooses the alternative with either the highest or lowest index. To compute an index each alternative must be scored on each criteria, a weight determined for each criteria, and an index computed by added the product of the criteria score and the weight for each alternative.

Satisficing- Using this strategy the decision makers develop a minimum acceptable level for each criteria. They then evaluate each alternative against this criteria set. Those alternatives not meeting all criteria minima are eliminated. Sub-strategies for two conditions were discussed.

Condition 1: No Alternatives meeting the criteria set
Eliminate criteria in order of importance until an acceptable alternative emerges.
Change minimum acceptable levels until an acceptable alternative emerges.
Continue search for an alternative that meets the entire criteria set (Innovation).

Condition 2: More than one alternative meets entire original criteria set.
Choose one criterion to maximize or minimize, or a small subset of criteria and computer utility index.
Add additional criteria that differentiate remaining alternatives.
We also made the important distinction between need-based or value-based criteria and predictive criteria.
Value (Interest) Based Criteria- Criteria representing the ultimate preferences of the decision maker or decision making group. In the Chiefland Case these might be quality care or cost effectiveness; State University, teaching effectiveness or research effectiveness; Auto Case, Safety, size, etc. These criteria represent actual end states desired.
Predictive Criteria- Predictive criteria are measures used in decision making to predict a desired preference or end state. For example, we might use GMAT score as a selection criterion, because we believe it predicts success in the program, or in the State University case we might use years of teaching experience as a selection criteria because we believe that it predicts teaching effectiveness.

Maximizing strategies work much better in cases where you can determine criteria weight empirically. This requires the use of historical data and a statistical process such as regression analysis. Therefore, when predictive criteria are used this strategy may work nicely.

When Value based criteria are used, no empirical technique can determine weights because weights in this case are a matter of decision maker values. Therefore, satisficing strategies tend to work better in this case.

 http://www.uri.edu/research/lrc/scholl/Notes/Decision_Making_Models.htm