Impact of satisficing behavior in online surveys on consumer preference and welfare estimates. (October 2016)

• Record Type: Journal Article
• Authors: Gao, Zhifeng; House, Lisa; Bi, Xiang
• Is Part Of: Food policy. Volume 64(2016)
• Page Start: 26
• Page End: 36
• Publication Date: 2016-10
• Subjects: Survey data quality -- Choice experiment -- Willingness to pay -- Seafood -- Validation question -- Satisficing behavior; Q13 -- D12; Food supply -- Periodicals; Food security -- Periodicals; Food -- Quality -- Periodicals; Food Supply -- Periodicals; Alimentation -- Périodiques; Electronic journals; 338.1905
• DOI: 10.1016/j.foodpol.2016.09.001
• Languages: English
• ISSNs: 0306-9192

Abstract

Highlights: We examined the impact of satisficing behavior on consumer preference and welfare estimates in online surveys. Respondents with higher incomes and education levels, and Caucasian are less likely to have satisficing behavior. Respondents with satisficing behavior are more likely to provide inconsistent choices. Models for respondents with satisficing behavior provide different estimates of willingness to pay (WTP) and consumer surplus (CS) from those without this behavior. Ignoring satisficing behavior may significantly affect the conclusions regarding consumer WTP and changes in CS resulting from policy changes. Abstract: Conducting online surveys through consumer panels has become increasingly popular for researchers to assess consumers' preferences and attitudes for the purpose of obtaining welfare estimates because they are inexpensive, flexible, and allow for fast feedback. Interestingly, few studies have examined the behaviors of online survey panelists, particularly their satisficing behaviors on welfare measures. This study demonstrates the use of validation questions (trap questions) to detect survey respondents' satisficing behavior and its impact on consumer choice, willingness to pay (WTP), and consumer surplus (CS) estimates. We find that respondents who fail a validation question (VQ) are more likely to violate the weak axiom of revealed preferences ( WARP ) in the choice experiment. The estimates for preference parameters, WTP, and CS are statistically different between those who pass and those who fail the VQ. In addition, the WTP and CS from respondents passing the VQ in general have smaller variances than those from respondents failing the VQ. These results indicate that without controlling for potential satisficing behaviors, online surveys may produce less efficient estimates (estimates with larger variance) of welfare measures.