Formal Choice Models of Informal Choices: What Choice Modeling Research Can (and Can't) Learn from Behavioral Theory

Publisher:
M.E. Sharpe
Publication Type:
Chapter
Citation:
Review of Marketing Research vol 4, 2008, 1, 4 pp. 3 - 32
Issue Date:
2008-01
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail2008000295OKfixed.pdf13.31 MB
Adobe PDF
In this paper we illustrate the benefits of forging a better alliance among behavioral, economic, and statistical approaches to modeling consumer choice behavior. We focus on the problems that arise when building descriptive models of choice in evolving markets, where consumers are likely to have poorly developed preferences and be influenced by beliefs about future market changes. We illustrate how understanding the actual process that is driving preferences can provide analysts with both better a priori insights into the model structures that are likely to provide the best descriptive account of choices in such settings, as well as how stable these structures are likely to be over time. We show, for example, that analogical reasoning heuristicsa common strategy for making decisions under preference uncertaintycan produce choice patterns that resemble the output of complex nonlinear, nonadditive, multi-attribute utility rules. Likewise, because novice consumers are likely to display strong individual differences in the variance of unobserved components of utility, methods that fail to recognize such differences will tend to overstate the actual extent of taste heterogeneity that exists in a population. We also illustrate the benefits of a reverse dialogue, examining how economic theory can lead behavioral researchers to more parsimonious explanations for apparent anomalies in choice tasks where preferences are uncertain. We show, for example, that some ad hoc models that have been used to statistically describe the compromise effect in choice can be deduced from first principles of rational risky decision making.
Please use this identifier to cite or link to this item: