Accession Number : ADA205749

Title :   When Choice Models Fail: Compensatory Models in Negatively-Correlated Environments.

Descriptive Note : Technical rept.,

Corporate Author : DUKE UNIV DURHAM NC CENTER FOR DECISION STUDIES

Personal Author(s) : Johnson, Eric J. ; Meyer, Robert J. ; Ghose, Sanjoy

Report Date : FEB 1989

Pagination or Media Count : 47

Abstract : Linear compensatory models, which make tradeoffs between product attributes, provide reasonably good predictions of choices made by non-compensatory heuristics which do not make tradeoffs. This robustness to miss-specification of functional form, however, may fail when there are negative correlations among attributes in a choice set. A Monte-Carlo simulation first demonstrates that certain non-compensatory rules are poorly fit by linear models, even in orthogonal environments. Two laboratory experiments then assess the extend to which such model failure might arise in natural contests. The first, a process-tracing analysis, examines the decision strategies consumers use in non-orthogonal contexts. We conclude with a discussion of the work's implications for current research in applied choice modeling. (KR)

Descriptors :   *COMPENSATION, *MATHEMATICAL MODELS, CONSUMERS, CORRELATION, DECISION MAKING, ENVIRONMENTS, FAILURE, LABORATORY TESTS, LINEARITY, MODELS, MONTE CARLO METHOD, ORTHOGONALITY, SIMULATION, STRATEGY, TRADE OFF ANALYSIS.

Subject Categories : Operations Research
      Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE