Facing a Pervasive Bias in Warnings Research

Warning labels and signs are widely used and recommended despite a lack of evidence that they prevent accidents. Most studies of warnings use subjective judgments or responses or else measure compliance in laboratory conditions rather than real-world compliance, but many researchers, juries, and regulators continue to advocate development and deployment of warnings. Several lines of research suggest that people – including researchers and safety professionals – are prone to an attribution error, emphasizing individual rationality and character while underestimating the importance of situational influences. Recommendations are offered for a shift in warnings research and practice.