The disparity between self-reported and observed food safety behavior: A case involving consumers from rural China

Abstract

Self-report and observation are two common methods for learning about consumers' food handling behavior. Identifying the relationship between self-reported behavior and observed behavior, especially in statistics, contributes to decisions regarding the survey method, allowing a more precise estimation of causal effects. However, this relationship has rarely been investigated. Based on 900 consumers from rural China and taking cleaning cloth use behavior as an example, we explored the correlation between self-reported and observed behavior. The results showed that consumers' observed practices were significantly poorer than their self-reported practices. Moreover, the ordinary least squares (OLS) estimation results revealed that self-reported behavior was not a good proxy for observed behavior. The coefficients of the independent variable were obviously different if the dependent variable was simply replaced by self-reported behavior instead of observed behavior. We also found that the magnitude of the differences between self-reported and observed cloth use behavior was heterogeneous among consumers with different characteristics. Males and elderly individuals (aged 60 and older) demonstrated a significant disparity between self-reported and observed behavior due to their poorer behavior in daily life, whereas the disparity in female, young and middle-aged groups was not significant. Highly educated consumers (more than 9 years of education) also had a larger discrepancy between the two behaviors. To collect consumers' actual behavior, enriching new ways of observation, reducing memory bias and encouraging real answers warrant further study.