This preview shows page 1. Sign up to view the full content.
Unformatted text preview: AEM 414 BEHAVIORAL DECISION MAKING FOR MANAGERS
(AKA Behavioral Economics and Management Decisions) Why Behavioral DecisionMaking? Managers must understand human behavior to be successful—e.g., they must understand and be able to predict the behavior of competitors, workers, customers, suppliers, government regulators, etc. 8/24/06 Lecture 1 Theory
Two approaches to gain this understanding: Rational Choice Theory (Economics) – Rational choice theory asks: “Given a set of preferences, what should a totally rational being do?” Decision Theory (Psychology) – Decision theory asks: “How can we explain what people actually do?” 8/24/06 Lecture 1 Theory (cont.) Since you care about actual behavior, you should prefer decision theory (if you are rational). However, the first explanation that decision theory always tries to use to explain behavior, logically, is rational choice theory. Thus, decision theorists and cognitive psychologists are quite interested in economics. If rational choice theory fails, then we have a behavioral anomaly.
Lecture 1 8/24/06 Behavioral Economics A new field of inquiry was thus begun at Cornell in the Johnson School (by Richard Thaler and others) to study behavioral anomalies, called Behavioral Economics, that combines psychology and economics. An example of a behavioral anomaly is shown by the following experiment that asks consumers the question: – “Imagine that you are very pressed for time and need both a new car and a new TV, but do not have time to get both today. Both the TV you want and the car you want are on sale today only. The car is 10% off and the TV is 50% off. Which would you get?” 8/24/06 Lecture 1 Results of Experiment The results are absolutely astonishing: Most people choose the TV, foregoing the much larger cash savings on the car How do we learn about such anomalies, since humans are not always Homo economicus? ANSWER: WE CONDUCT EXPERIMENTS 8/24/06 Lecture 1 Why Experiments? The Scientific Method (Taken from Scientific Method in Practice by Hugh Gaugh) QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture.
QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture. Hypothesis Hi is the theory (e.g., consumers maximize expected utility) Deduction is used to work out consequences (predictions such as risk aversion) Note in deductive logic (theory) if premises are true, consequences must be true 8/24/06 Lecture 1 The Scientific Method (Taken from Scientific Method in Practice by Hugh Gaugh) QuickTime ™ and a TIFF (Uncompressed) decompressor are needed to see this picture. Hypothesis Hi which corresponds to a particular theory implies an experimental design, Di, to test the theory. Data from Nature contains noise. Now the scientist has both new and available data 8/24/06 Lecture 1 The Scientific Method (Taken from Scientific Method in Practice by Hugh Gaugh) QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture. Available and new data (both noisy) are used to compare to consequences derived by deductive logic from the theory using statistics. Since many theories can explain any observation, one can only reject theory if a consequence does not occur (induction). 8/24/06 Lecture 1 The Scientific Method (Taken from Scientific Method in Practice by Hugh Gaugh) QuickTime ™ and a TIFF (Uncompressed) decompressor are needed to see this picture. The process of induction is the derivation of theory from data which is inherently uncertain. If a theory fails, induction leads to a modified theory as a hypothesis Hi+1 and the process starts over again. Induction assumes two things (presuppositions) that require faith: – Uniformity of Nature, that results of an experiment are generally valid. – Parsimony, the most simple theory is best Lecture 1 8/24/06 Economics versus Psychology Economics emphasizes deduction Psychology emphasizes induction Both are needed 8/24/06 Lecture 1 Experiments Why experiments?
– Control – Causal ordering – Replication A theory that fails in the lab under ideal circumstances cannot succeed in explaining what happens in the field. Experiments complement surveys, field studies, and econometric approaches.
Lecture 1 8/24/06 Purposes of Experiments Falsify theory Stress test theory Search for empirical regularities Develop mechanisms for implementation Teaching and training 8/24/06 Lecture 1 Theory of Experiments Induced value theory Payoff dominance Saliency of rewardsSiegel at Stanford in 1960s Parallelism and faith 8/24/06 Lecture 1 Guidelines for experimental design Use clear and unbiased instructions Use salient rewards Start with a pilot experiment No deception Always have a baseline treatment Change one thing at a time 8/24/06 Lecture 1 Guidelines (cont.) Make the change in the independent variable as large as possible Control all controllable variable and randomize over the rest (shoe example) Make the experiment as simple as possible Account for learning and experience Consider using an unknown or random number of rounds to avoid end effects 8/24/06 Lecture 1 Statistical Testing Use within subject design whenever possible (AB, BA, or ABA, BAB) Specify statistical tests in advance and estimate sample n based on a pilot tstatistics are now viewed as preferable since non parametric tests require symmetrical distributions sign test
Lecture 1 8/24/06 Section sign up on Tuesday Section 1 meets Thursday 9:059:55 Section 2 meets Thursday 10:1011:00 Section 2 meets Thursday 11:1512:05 See me today if you want to be a TA for this class
Lecture 1 8/24/06 ...
View Full Document
This note was uploaded on 02/20/2009 for the course AEM 4140 taught by Professor Schulze,w. during the Fall '08 term at Cornell University (Engineering School).
- Fall '08