Since Yvonne brought this up, I thought I'd ask again. Does anybody know of any studies that used random sampling to recruit respondents but were not only survey-based, rather, included a more in-depth component such as interviewing or observation.
I ran a household panel study in the UK from 1998-2001 which did this. We recruited by random selection (well, post-code based random selection as the initial recruitment interviews were face-to-face CAPI). Once we had our sample of c 2500 individuals and the survey data we then selected a subset for in-depth qualitative interview. That subset were not randomly selected however - they were selected to give as broad a spread of IT ownership and social context as possible. We then followed the 2500 for 2 further years of data colelction (quant and qual) You can read about this research project here: http://www.bt.com/bttj/vol20no2/dlp/dlp.pdf in Lacohee H and Anderson B: 'Interacting with the telephone', in Kraut R and Monk A (Eds), Special Issue of the International Journal of Human-Computer Studies on: 'Home Use of Information and Communications Technology', 54, No 5, pp 665-699 (May 2001) and in American Behavioral Scientist November 2001, volume 45, issue 3 Ben -- Dr Ben Anderson www.essex.ac.uk/chimera +44 (0) 7710 187 806