I though that there was some attempt to stratify the samples. Especially that of service users. The researchers gave a breakdown of the self-reported ethnicity of the user and professional group. They also went on to discuss how this might be a limitation to the transferability of results as other populations might not reflect this one in a south london borough (see my reply to Sadhana's post). I think that a stratified sample is a reasonable approach in qualitative research. As the BMJ article talks about, qualitative research depends on context. The age, gender, ethnicity and/or occupation of members will affect themes and issues discussed in the interviews through socio-cultural perspectives.
Its important also to highlight reflexivity and the lack of it in this study. In my opinion, the impact of the researchers age, gender, ethnicity, occupation and social status on the research process and subjects (ie during interviews) is a considerable one, particularly when discussing the topic of domestic violence. What do others think?
