Policy makers and researchers may have different agendas as to what constitutes "evidence". This paper reviews some of the issues currently debated in the UK and elsewhere about researchers' understandings of how social research can usefully inform the policy-making process, and their techniques for making it available. Using examples from two programmes of research synthesis in health promotion and education, the paper argues that methodological work advancing the systematic review of social research is more useful than continued debates about the relative advantages of "qualitative" and "quantitative" methods. The challenges in the path of this progress include ideological resistance among academics, the traditional culture of academia, which is opposed to collaborative and cumulative work, lack of methodological skills among social scientists, nonstandardisation of bibliographic databases, and a reluctance to accept the existence of problems in the way much social research is carried out and reported.