Large-scale research projects, conducted in a cross-European context, are increasingly attractive to educational researchers and policy-makers. However, this form of comparative research across cultures brings problems concerning the standardization of data collection and analysis, particularly where ethnographic research is concerned, as it prioritizes a full range of qualitative research strategies. This paper outlines the use of a universal model and the approaches recently taken by two research teams and contrasts these with another recent nine-partner comparative European study that used ethnographic methods. We then describe the analytical procedures used in the project, which encouraged participant observation and individual researcher interpretation in order to generate grounded accounts and outline how they were culturally sensitive and meaningful to research teams who used varied analytical approaches. However, this raised difficult issues for the "final" analysis and the production of a loosely coupled research report. Our pragmatic solution was a process of "qualitative synthesis" whereby individual partner reports were collated by the Project Director and treated as data and a grounded theory approach was applied to generate tentative theory in respect of creative learning. The paper concludes by arguing that data generated by a loosely coupled approach to qualitative comparative research which uses a wide range of data collection methods can be effectively analysed with a qualitative synthesis. (Contains 1 table.)