Files in this item



application/pdf8026508.pdf (6MB)Restricted to U of Illinois
(no description provided)PDF


Title:The Situational Model: A Theory of Choice
Author(s):Hagerty, Michael Raymond
Department / Program:Business Administration
Discipline:Business Administration
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Business Administration, General
Abstract:This dissertation suggests a theory of how individual consumers choose among alternative brands, and presents experiments designed to explore the theory.
A number of methods employed in Marketing (Conjoint Analysis, LOGIT, Dollar-metric Analysis) use individual choice models. However, those models in use (Luce's Choice Axiom and the First Choice Rule) may predict market share incorrectly when, for example, a new product is introduced which is similar to an existing one (known as Independence from Irrelevant Alternatives (IIA) property in the economic and psychological literature). Testable properties of existing individual choice models are reviewed. An alternative, the Situational Model (SM), is suggested which avoids the restrictions in other probabilistic choice models, in particular avoiding IIA and the static Bernoulli assumption. The consequences of SM are derived and contrasted with Elimination By Aspects (EBA). Of particular interest is that EBA predicts Weak Stochastic Transitivity (WST), whereas SM does not. However, designing an experiment which vilates WST was found to be difficult and was not attempted.
Two experiments were performed to test whether the more restrictive EBA should be rejected in favor of SM. First, Tversky's experiment supporting EBA was replicated to test for the existence of (a) dynamic non-Bernoulli choices and (b) context effects in choice due to previous trials. Repeated choices among gambles were collected from 52 subjects, where the three trials previous to the repeated trial were manipulated to be either high or low in expected value, and either high or low in probability of winning the bet. Results showed no deviation from the Bernoulli model, and no effect of manipulated previous choices, consistent with EBA.
The second experiment investigates choice in a more "realistic" setting where repeated choice of the same product may result in variety seeking or satiation. Such a result would violate the static Bernoulli assumption and therefore violate EBA. Twenty-five women served as a panel who received the same fruit juice day after day over a period of two weeks. Subjects rated their preference for each juice on each day. Results showed strong violations of the Bernoulli assumption at the individual choice level. However, the Bernoulli model predicted aggregate market share almost as well as the Situational Model's dynamic mechanism.
In summary, a dynamic Situational Model seems unnecessarily complex to describe the experiments reported here. While the model seems to be a natural and simple explanation of choice under changing situations, simpler static models such as EBA could also predict well.
Issue Date:1980
Description:180 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1980.
Other Identifier(s):(UMI)AAI8026508
Date Available in IDEALS:2014-12-13
Date Deposited:1980

This item appears in the following Collection(s)

Item Statistics