Distinguished Scholar Invited Essay Behavioral Decision Making: Implications for Leadership and Organizations - Aldag - 2012 - Article

What are the underlying assumptions of the traditional rational economic man model?

The rational economic man model of decision making views humans as capable of optimizing. The perspective holds the following underlying assumptions:

• The decision maker has full knowledge of relevant aspects of the environment (including alternatives, events, probabilities, and associated outcomes). 
• The decision maker possesses a well-organized and stable set of preferences. 
• The decision maker has computational abilities and is capable of optimization. 
• The decision maker is not influenced by emotions and stress. 
• The decision maker has immediate access to costless information. 

How did Herbert Simon intend to complement the theory of the rational economic man?

Herbert Simon viewed these assumptions as unrealistic. Humans experience cognitive, perceptual, and situational constraints. Searching and finding information is not free nor limitless. According to him, humans are generally not capable of optimizing their decisions, but instead, seek to satisfice and find the first acceptable alternative. 

What are the underlying assumptions of Simon's model of the administrative man?

Herbert Simon presented a so-called `new rationality` that kept in mind his above-mentioned concerns. The model of the administrative man is based on the following assumptions:

• The existence of a bounded rationality in which the decision maker tries to find satisfactory solutions within cognitive, perceptual, situational, and other limitations. 
• Aspiration levels are dynamic. Successes and failures may lead to changes in the levels of aspiration and changes in what is thought to be acceptable and unacceptable. 
• Gathering information is time consuming, effortful, and costly. The question of the best level of persistence in pursuit of a goal involves a trade-off between the potential costs and benefits of the search. 
• Preferences are fluid. They may change with time and maturation. Also, consequences can change one´s payoff function. Sometimes, the decision maker does not know his own preferences because of lack of experience and a reluctance to explore alternatives. 

What are policy capturing and process tracing approaches?

Policy capturing and process tracing approaches are used to carefully examine human decision processes and outcomes. With policy capturing, an individual is presented with a series of profiles giving the scores of alternatives on major dimensions. Then, regression analysis, analysis of variance, or related procedures are used to optimally relate inputs to outputs in order to `capture` the decision maker´s policies. Policy capturing in this way offers a model of the relationship of inputs to outputs, but without reflecting the decision maker´s actual decision processes. Process tracing on the other hand is used to examine the decision process itself. 

What are heuristics?

The many constraints on rationality result in the use of a variety of heuristics. Heuristics are simplifying rules of thumb. The following important heuristics can be identified:

• The availability heuristic refers to the tendency to estimate the probability of an event on the basis of how easy it is to recall examples of the event. 
• The representativeness heuristic is the tendency to place something in a class if it seems to represent the class. Whereas stereotyping views those within a class to share common characteristics, representativeness views those who share common characteristics as falling within a class. 
• Anchoring and adjustment is the tendency to use an early bit of information as an anchor and then use new information to adjust that initial anchor. It leads to giving too little weight to new information and insufficient adjustment. 
• The default heuristic refers to the tendency people have to simply accept the default presented to them. 

What are biases?

Constraints on human decision making also lead to a variety of systematic biases. The following important biases can be identified:

• Conservatism in information processing leads to inadequate response to changing situations. When presented with new information, people tend to under revise their past estimates. 
• Framing effects refers to the influence of the way information is framed on our choices. 
• Hindsight bias refers to the tendency for people who learn of the outcome of an event to believe falsely that they would have predicted the reported outcome. It leads to distorted views of the accuracy of past decisions. One explanation for hindsight bias is creeping determinism, a fast and unconscious process in which outcome information is immediately and automatically integrated in a person's knowledge of events preceding the outcome. 
• Confirmation bias is the tendency to seek, interpret, and recall information in ways that confirm our preconceptions. People prefer information sources that favor our preconceptions and avoid disconfirming sources. Ambiguous situations are more likely to be interpreted to favor our views, and people are more likely to remember confirming than disconfirming information. 
• Overconfidence bias occurs when people´s subjective confidence in their judgements is greater than their objective accuracy. 
• Illusory correlations is the tendency to see relationships between variables that do not actually exist (could be because of stereotypes or expectations). 
• Escalation of commitment, also known as the sunk cost fallacy, is the tendency to continue to invest in a course of action, not based on the future benefits and costs, but on the basis of sunk costs (like money and lives already invested). 

What is the clinical-actuarial controversy?

The clinical-actuarial controversy refers to the discussion about how the accuracy of human clinical decision making compares to that of actuarial statistical models. Clinical decisions are made every day that have direct consequences for the lives of many people. Research strongly indicates that actuarial models consistently outperform clinical judgements and that decisions should be based on actuarial methods. 

What are proper and improper models? 

In a proper model predictors are optimally related to outcomes. It is sometimes however not possible to create such a model, for example when there are not enough observations or a measurable criterion doesn´t exist. In that case an improper linear model can be used. An improper linear model is a model in which the weights of the predictor variables are obtained by some nonoptimal method (such as set to be equal, or determined at random). An example of an improper model is the model of man, which shows to have perfect reliability and enhancing validity. With this model policy capturing is used to develop a linear model of an individual's decision process. THat model is then used to make decisions in place of the individual. 

What is clinical synthesis?

Many people are opposed to replacing clinical decision making with actuarial models. This has led to the use of clinical synthesis, where output of actuarial models is provided to the individuals as input to their decisions. This improves their decision making, though it is still not as good as if they simply used the output without modification. Especially individuals with more experience and expertise are reluctant to use the output of actuarial models, and it often negatively influences their decision making. 

What is automatic information processing? 

A lot of decision making is not the consequence of a conscious deliberative process, but it occurs automatically at nonconscious levels. Individuals often behave without conscious consideration of elements of the rational economic model. When people repeatedly face a decision situation they may develop scripts. Scripts are models held in memory that specify behaviors or event sequences that are appropriate for specific situations. They can be effective when dealing with routine situations, but they tend to cause problems in new situations. 

How can statistical groups and prediction markets help overcome the limitations of human decision making?

Statistical models of decision makers perform at a level that is equal to that achieved by averaging many people´s inputs. Taking the average of many individuals´ judgements reduces unreliability. Using statistical groups can thus be useful. It then makes sense also to use prediction markets, which pool individual judgements to forecast the probabilities of events. In such markets, individuals bid on contracts that pay a certain amount if an event occurs. 

What is paternalistic intervention?

Paternalistic intervention refers to the process of using knowledge about human cognitive limitations and tendencies to nudge people towards desirable behaviors. This phenomenon causes a lot of controversy, as there are ethical questions regarding the use of unobtrusive nudges and the question of who determines what behaviors are desirable. Those who are pro-paternalistic intervention say that nudges are needed when decisions are difficult and rare, when there is no prompt feedback, in the case of investment goods for which costs occur now but aren't seen until later (e.g. exercise), and sinful goods for which pleasure is immediate but negative consequences are not seen until later (e.g. smoking). One example of a nudge is to automatically have all citizens be organ donors, unless they register their unwillingness. This generally leads to more organ donors as most people don't object to being organ donors, but do not normally make the effort to register themselves to be organ donors.