Qualitative studies are often described as small N studies because the number of respondents is small. I argue that this is the wrong perspective: What we really have in qualitative data, say interviews, is lots of data (points) clustered within individuals. Rather than focusing on the number of respondents, we should probably focus on the number of relevant statements (i.e. statements about our quantity of interest), and describe this number (along with the number of respondents). When computer aided qualitative data analysis (CAQDA) is used, I guess the number of tags relevant to our quantity of interest is that number. Seen this way, many qualitative studies are no longer small N studies, but we’re still faced with unstructured, messy data that may be difficult to analyse, and of course we don’t have independent observations — so generalization remains a challenge.