Abstract: Starting from the perspective of evidence-based decision-making, this paper explores the relative value of different types of evidence in scientific decision-making by analyzing the population-wide salt reduction. This paper argues that a simple linear relationship cannot be established between "evidence" and policy outcomes. First, the nature of "evidence" from the perspective of evidence-based decision-making is broader than that of traditional scientific evidence, and should include non-scientific evidence including values and civic cognition. Second, when the scientific evidence involved in the policy is controversial, two conflicting claims are often generated from the relevant scientific research, and the relative value of scientific evidence is weakened. Third, policymakers need to carefully consider the various evidentiary value of policy issues, recognize the role that non-scientific factors should play in evidence-based policy making, and avoid being bound by methodological constraints.