Systematic review (SR) is often promoted as a ‘best practice’ method to inform both policy‐making and policy‐evaluation in social policy in the light of the ever‐growing volume of research. This article considers an innovative use of the method to advance and refine academic knowledge and illustrates this through a small‐scale study of the literature on attitudes to welfare. SR relies on rapid, structured searches of large quantities of material. However, the method has encountered criticism. The article calls for a greater degree of reflection in terms of possible bias in SRs. A pilot using tools from SR methodology to survey attitudes towards social policy is used to demonstrate the problems. These include the US bias of major databases, and weaker reporting of book publications than of articles. SR may help to advance knowledge in social policy, but researchers need to be aware of its weaknesses and possible biases.