Modeling Simultaneity in Survey Data |
| |
Authors: | Timothy J Gilbride Sha Yang Greg M Allenby |
| |
Institution: | (1) Mendoza College of Business, University of Notre Dame, Notre Dame, IN, 46556;(2) Stern School of Business, New York University, 44 West Fourth Street, New York, NY, 10012;(3) Fisher College of Business, Ohio State University, 2100 Neil Ave., Columbus, OH, 43210 |
| |
Abstract: | Responses to questions in a survey can reflect a behavior process that influences multiple response items. Respondent ratings
of brand attributes, for example, can be affected by past purchases by making a brand more salient, or by respondents attributing
higher performance to justify their purchases. When multiple response items are influenced by a common underlying process,
there is simultaneity in the data. This paper proposes an approach to model the simultaneity in different survey responses
by using common parameters and structural relationships motivated by behavioral theories on how consumers respond to surveys.
Specifically, the proposed models show how brand usage and attribute perception responses are jointly determined by justification,
order, and brand halo effects in two brand positioning studies. We detect a significant tendency for respondents to inflate
their reported beliefs for particular brands as well as the selected brand across five countries in an international survey
as well as in a domestic study.
JEL Classification: C35, C53, D12, M31 |
| |
Keywords: | Bayesian methods endogeneity survey research data augmentation |
本文献已被 SpringerLink 等数据库收录! |
|