This article discusses how the Organisation for Economic Cooperation and Development's Centre for Educational Research and Innovation (CERI) addresses the task of conducting international policy research. The article begins with a descriptive account of CERI's work, including the way member countries shape the research agenda. Several issues which relate to how research evidence is compiled within an international context are addressed. First, why the supposed priority area of lifelong learning is only weakly supported by systematic research is considered. The author raises the question of how we are to judge the quality and impact of international research work, especially where it is policy-related. He suggests that an increasing focus on the outcomes of education raises questions about causality in a policy research context. This leads to some brief consideration of evaluation of research, and of the country as a unit of methodological analysis. Finally, he asks what might be meant by learning from international experience.Cross-national education research is carried out by a number of international organisations, as well as by individuals and units working within national boundaries. In this article I examine, from the inside, the role played by one such international unit, the Organisation for Economic Cooperation and Development's Centre for Educational Research and Innovation, where I worked for four years in the 1970s, and now work again since 2003. How does the operation of units such as CERI, and the OECD more widely, influence the agenda and the nature of international research, and what are the potential implications for European research?According to Rinne et al, the OECD 'has become established as a kind of "eminence grise" of the educational policy of industrialised countries', and 'has claimed for itself a central position in the collection, processing, classification, analysing, storing, supplying and marketing of education policy information -the extensive control of information on education ' (2004, p. 456). The first observation is a clear judgement on the part of the authors, i.e. it is their own view of where the OECD stands. The second is more ambiguous: does it mean that the OECD has actually achieved this position of control, or merely that it claims it? In either case these are fairly extensive statements, which are both flattering and challenging to those of us who work within the OECD's Directorate for Education.Academic interest in the OECD's role in educational policy analysis and formation is growing (see Papadopoulos [1994] for a historical account of the OECD's education work up to a decade ago). A similar study to the Finnish one cited above is being carried out at the University of Bremen, located within a wider study of international governance, though the results will not be available until 2006(Martens et al, 2004. Both these projects cover OECD work on education at a very general level and over a long time span.