Aim: The aim of this paper is to identify the broad epistemological debates which underpin conflicting statements on 'rigour' and 'good practice' in qualitative research; to relate divergences in statements of 'good practice' in focus group design made by the pre-eminent commentators on focus group methodology to these broader epistemological debates; and to stimulate further reflection on the range of possible uses for focus groups in health services research. Considerations of the analysis of focus group data are beyond the scope of this paper.
Discussion: Focus groups are a popular form of qualitative data collection, and may be defined as a particular form of group interview intended to exploit group dynamics. While qualitative research may be broadly characterized as concerned with exploring people's lived experiences and perspectives in context, it is a heterogeneous field incorporating many theoretical traditions. Consequently, qualitative researchers may be informed by a wide range of assumptions about the nature of knowledge (epistemology). These assumptions, whether implicit or explicit, have important consequences for claims about rigour and 'good practice' in data collection. Thus, while there is broad agreement over the general form of focus groups, statements of 'good practice' in terms of its application are varied. A close reading of texts by the two pre-eminent commentators on the practical application of focus groups identifies differences in 'best practice' focus group design related to their respective epistemological assumptions, and differences principally related to sampling techniques, composition of groups, the perceived role of group interaction and the nature of inference.
Conclusion: Explicit consideration of the epistemological basis of divergent statements of 'best practice' in focus group design forces health services researchers to balance the demands of theory with the practicalities of conducting focus group research within complex host organisations; and encourages readers to apply appraisal criteria appropriate to the stated intentions of researchers.