The social construction of uncertainty plays a major role in environmental decision-making. Methods for assessing this aspect of scientific knowledge quality are lacking. Our analysis of the French debate on the risk that the insecticide Gaucho (R) (active substance: imidacloprid) forms for honeybees is particularly relevant to this theoretical and practical gap. Based on our analysis, we propose six knowledge quality criteria that can assist in assessing the information communicated in an argumentative public process: reliability of the information - it must be based on all available scientific knowledge; robustness of the information - it must take into account criticism; use of the information produced by other stakeholders; relevancy of the arguments for issue under debate; logical coherence of the discourse; and legitimacy of the information source. Further, our findings deepen the understanding of the relationships between the social, economic, and institutional stakes of the actors involved in the debate and their strategies of 'creating uncertainty'. Finally, we compare the findings of this case study with the twelve lessons drafted by the European Environmental Agency (EEA) in its report Late lessons from early warnings, and we draft two more lessons. These lessons can be applied to future policy in order to minimize the repetition of past mistakes.