Research traditions exert a powerful influence over the thinking of academic researchers. In population geography an entrenched empiricism and a reluctance to engage in wider debates on theory and method within human geography have resulted in a degree of separate development. Such methodological complacency has, until recently, threatened to undermine population geography's role within the discipline as a whole. Thus the discussion of multi-method research designs is of particular significance because it offers population geographers an opportunity to break out from the confines of a dominant research tradition and participate with other human geographers in an important methodological debate. This paper seeks not only to provide a critical overview of the current debate within population geography (as represented in the preceding papers in this Focus section) but also to extend that debate by raising issues of much more general concern. I argue that there are dangers in drawing the terms of the debate too narrowly. If we are to understand the nature and potential of multi-method research, we must first pose fundamental questions about the interrelationships among methods, data, and research problems. Pragmatic views on the choice of research method are inadequate because they fail to recognise the theory-driven nature of research. Only once we have achieved a better understanding of the philosophical grounding of research strategies will the opportunities afforded by multi-method research be fully realised.