Analysing decision behavior styles in contingent valuation: The latent class and the factor analysis

A better understanding of respondents' decision behaviors in contingent valuation (CV) is essential to reveal the true preferences of the public for environmental goods or services. Although the theoretical foundation of CV is based on the assumption of the full rationality of respondents, the literature provides various evidence of limited or partial rationality. In a CV survey of air quality improvement in China, we identified non-rational decision behavior style by adopting the latent class analysis and factor analysis methods, both of which are based on a series of questions related to decision behaviors initially proposed by Frör (2008). The application of the latent class model proposes the identification of two or three classes, with at least one more analytical reasoning group significantly differing from the other group(s). The factor analysis approach allowed us to identify two decision behavior factors, i.e., the analytical reasoning factor and the non-analytical reasoning factor. Our estimation results show that the analytical reasoning style is positively correlated with willingness-to-pay (WTP). Furthermore, the mediation tests conducted in the WTP determination models reveal that simply including respondents' socioeconomic, knowledge and perception characteristic questions in the survey to collect the information does not ensure that all the information conveyed by people's decision behavior style is captured.