We find William Sutherland and Mark Burgman's advice on the complex social processes between evidence and policy decisions to be naive at best, and antidemocratic at worst (see Nature 526, 317–318; 2015).

Policymakers are influenced by a much greater range of factors than are considered by the authors — including public opinion, inheritance of policies and institutional rules, finance, unpredictable events, and trust in actors (see also W. Pearce et al. Evid. Policy 10, 161–165; 2014).

In our view, the authors perpetuate negative stereotypes of policymakers and academics, when in fact many examples of productive collaborations and hybrid roles exist. Their edicts seem to undermine colleagues who mobilize knowledge for policy, and to reduce the intricate relationship between evidence and policy to a linear, technocratic process. As they themselves attest, giving advice to policymakers or academics that is not evidence-based could hamper the formation of useful collaborations.