«Tim Rosengart July 17, 2009 This is a Bucerius/WHU MLB thesis 12,832 words (excluding footnotes) Supervisor 1: Professor Dr. Bernhard Hirsch ...»
As said before, a basic principle of the normative decision making theory is to translate a concrete decision making problem which is faced in reality into a formal model with the purpose of reducing its complexity and being able to apply rational and logical [mathematical] means to find the optimal solution.14 The normative decision making model is influenced by two different categories of information on the problem structure. On the one hand, there is an objective assessment of the situation in which the decision is made. On the other hand, there are the influences of a subjective decision maker who has individual objectives and preferences which he wants to satisfy.
The objective assessment of the situation contains information about potential courses of action, which are called alternatives, and information about the conditions and potential constraints which influence any decision. Alternatives are controllable because they can be chosen and designed by the decision maker. Conditions, on the other hand, are given and uncontrollable which means that they have to be accepted by the decision maker.15 The combination or product of alternatives and conditions results in a set of potential consequences. This set is evaluated based on a hierarchy of the decision maker’s personal objectives and preferences.
After having evaluated all alternatives, the decision maker selects an alternative, which under the given conditions and evaluated on the basis of his personal preferences, promises the most optimal and hence rational consequence to him. In this context, Jost and Simon pronounce the particular importance of judging rationality in the context of the given conditions and the decision maker’s personal objectives.16 In his process of selection, the decision maker is supported by various decision making rules which will be discussed in part three of this thesis in the practical context of whether they are found to be suitable to explain observed decision making behavior of doctors’.
Conditions, alternatives and their respective consequences, as well as preferences and individual objectives have to be considered as the influences which affect decision makers in their behavior.
Influences which are of particular relevance in evaluating the decision making behavior of German doctors will be described in part two of this thesis.
Bartscher/Bomke (1993), p. 51.
Compare Bartscher/Bomke (1993), p. 53.
Grünig/Kühn (2009), p. 41ff..
Compare Jost (2000), p. 315; Simon (1986), p. 210; Simon (1978), p. 2f..
Here, it shall be sufficient to re-emphasize the importance of the degree of certainty, risk and emphasize uncertainty about possible alternatives as a substantial influence, which has already been described substantial in this text.
Although, the aforementioned normative model has as its purpose to reduce complexity to a minimum, the process of constructing the decision making model always has to ensure that the elements of any model correspond to the real problem structure. Oversimplification and neglect of the real problem structure have to be avoided.19 Figure 3: Summary of decision making phases – normative decision making theory Source: Own figure.
Kahnemann and Tversky won a Nobel Prize in 2002 for the development of their frequently cited Prospect Theory which has to be considered as highly relevant in the explanation of empirical observations of decision making under uncertainty. According to Hirsch (2001), p. 248 and Frey (1990), p. 172, the Prospect Theory can also be used to explain a limited amount of decision making errors observed in practice. The interested reader should refer to Kahnemann/Tversky (1979).
Bekmeier-Feuerhahn (2008b), p. 4.
Feuerhahn Compare Bartscher/Bomke (1993), p. 53; Bamberg/Coenenberg (1991), p. 12ff.; Schneeweiß (1992), p. 76f..
Figure 4: Normative Decision Making Model
As previously discussed, descriptive decision making theory describes decision making behavior which can be observed in reality and tries to explain why and how decisions are actually made. Insights derived from descriptive decision making theory are also used to predict future making decision making behavior.
Most errors can be attributed to the bounded rationality i.e. the limited information processing capacity of decision makers.20 Weber et al. take another approach and explain decision making errors another with deficits in the field of motivation wollen and ability können.21 The following table shall assist the reader as a summary of the concepts of normative and descriptive decision making theory.
Figure 5: Normative vs. Descriptive decision making theory Source: Similar to Bartscher/Bomke (1993), p. 51.
Compare McFadden (1999), p. 84ff..
Compare Weber et al. (2003), p. 10. The interested reader should refer to this piece of work to acqui a acquire comprehensive overview of an approach which recognizes wollen und können problems to be equally important to explain decision making errors. Also compare Weber (2005), p. 25ff..
When using the concepts of normative and descriptive decision making theory, it is often referred to rationally acting decision makers who pursue their personally optimal alternatives.
In economic theory the notion of rationality is inseparable from the definition of the homo oeconomicus22 which to a large extent is also accepted by the normative decision making theory. The homo oeconomicus as a decision maker, either group or individual, always tries to achieve a personally optimal solution, in which his personal input is minimal, whereas his personal output shall be maximum given the existing conditions and constraints. This kind of behavior is considered rational from the viewpoint of the homo oeconomicus.23 The assumptions underlying the homo oeconomicus are described by Eichenberger as follows.
First, the decision maker knows all possible alternatives and is capable of selecting one as the most optimal one. Second, the decision maker is in a position to attribute any alternative to the corresponding consequence [given the known conditions]. Third, the decision maker has noncontradictory, precise and stable preferences which are decision-relevant but do not only apply to particular decisions.24 Kirsch adds to this assumption that these preferences shall not be contrary to moral standards25. Kirsch also emphasizes the assumption that any decision maker possesses an unlimited information processing capacity.26 As previously stated, any “decision making theory cannot fully rely and use the definition of the homo oeconomicus”27 because in reality human decision makers do not meet all attributes of this “ideal type decision maker who is abstractly derived from economic decision making theory”.28 Simon, in particular, emphasizes the notion of bounded rationality which he bases on observed “incomplete and imprecise information of decision makers about their pursued [alternatives and their resulting] consequences” 29.
Bartscher/Bomke (1993), p. 83.
Compare Bartscher/Bomke (1993), p. 83.
Compare Eichenberger (1992), p.7; Suchanek (1994), p. 101 ff..
Compare Kirsch (1970), p. 27ff.; For a more elaborate and interesting discussion on moral standards in this context compare Kant (1983), p. 22f.; Kirchgässner (2000) p.16.; The conflicting views of Kant and Kirchgässner are also compared in a piece of work by Erni (2004).
Compare Kirsch (1970), p. 27ff..
Compare Bartscher/Bomke (1993), p. 83ff..
Becker (2003), p. 41.
Compare Simon (1997) p. 17 f.; Simon (1994), p. 266ff..
The concept of mental models as part of the descriptive decision making theory, accepts the assumption of “incomplete and imprecise information”30 which is proposed by Simon. The same model rejects the assumption of Kirsch on the “unlimited information processing capacity”31 of decision makers and expects them to be limited in their cognitive ability to process information.32 As a consequence of these two assumptions, the concept of mental models describes decision makers who create [mental] models which describe the environment in which decisions are made in a simplified way with the purpose of reducing complexity in decision making.33 Mental models are particularly relevant in the context of complex, constantly reoccuring decisions.34 Jaeger sees mental models as “a grid which provides a simplified structure”35 to decision relevant information. Hirsch interprets mental models as “a system of rules which is developed by a specific decision maker for his own use” 36 and which suports him in achieving a decision which is rational in the sense of the homo oeconomicus construct.37 Jost prefers the notion of “heuristic principles on which people rely in order to simplify the complex task of assessing probabilities of decision outcomes”.38 Because each decision maker uses his own individual mental model, it can be assumed that their heuristics are influenced by information which is particularly present to the respective decision maker. Nonexclusively, “information which the decision maker remembers easily”39, information which was studied by the decision maker inensively, information which threatened or benefited the decision maker’s well-being in the past and information which led to past positive and negative experiences can be assumed to be information which is particularly present to decision makers. This effect, which is used widely in decision making theory, is known as “framing”40.
In addition, the previously described threat of oversimplification has to be considered when decision makers tend to use heuristics in order to “minimize the effort of collecting and evaluating information”41, which might consequently lead to “wrongful decision making tendences”42.
Compare Simon (1997) p. 17 f.; Simon (1994), p. 266ff..
Compare Kirsch (1970), p. 27ff..
Compare Hirsch (2007), p. 239.
Friederich (2005), p. 90.
Compare Hirsch (2007), p. 241.
Jaeger (2003), p. 81.
Hirsch (2007), p. 240.
Compare Hirsch (2007), p. 240.
Compare Jost (2000), p. 191.
Compare Tversky/Kahnemann (1973), p. 453.
Compare Hirsch (2001), p. 130f; Tversky/Kahnemann (1981), p. 453.
Compare Hirsch (2007), p. 241f.; Jost (2000), p. 602.
Jost (2000), p. 193.
Furthermore, heuristics are only of limited use in environments which are subject to constant change because principles which were developed and good for one point in time and one particular environment are not necessarily suitable for another.43 For the discussion of the relationship between patients and doctors, which will take place at a later stage in this text, and also due to its importance as another specific concept of descriptive decision making theory, the principal-agent theory needs to be worked out.44 The principal-agent theory or agency theory is regarded by scholars such as Levinthal as a “response to questions raised many years earlier by March and Simon regarding the behavior of an organization of self-interested agents with conflicting goals in a world of incomplete information.”45 As a model, the agency theory is based on a contractual relationship between an agent and a principal. The principal, i.e. a patient or health insurance company, entrusts and pays an agent i.e. a doctor to fulfill a task to the principal’s best interest.46 The agent has the discretion to choose among a set of alternatives which then produce, together with the given conditions, the consequence which has to be coped with by the principal.47 Equivalent to the normative decision making theory, the agency theory assumes that the principal, as well as the agent, both act rationally with respect to the previously described underlying assumptions of the homo oeconomicus and therefore try to maximize their personal utility from any decision outcome. Both actors are expected to act opportunistically and are assumed to possess an unlimited information processing capacity.48 Although the principal and the agent both act according to the assumptions of the homo oeconomicus, none of the two actors can actually rely on complete and precise information [about the prevailing conditions, alternatives or the corresponding consequences and their respective certainty of coming into effect]. Consequently, the agency theory, as well as it was already derived for the normative decision making theory, is undermined by the observation of an incomplete and imprecise supply of information to the decision maker.
Compare Hirsch (2007), p. 242.