Since the mid-1990s, space-time sampling methods (or Time-Location Sampling) have been widely used in France in quantitative research of homeless populations (Marpsat and Firdion, 2000; Bush, 2002) and drug users (Emmanuelli et al. 2003; Jauffret-Roustide M. et al. 2006 1). Adapted indirect sampling methods initially developed in the United States (Burt and Cohen, 1989; Dennis and Iachan, 1993), they make it possible to construct representative samples of populations that are difficult to reach, for which there is no sampling basis, based on the aid schemes for the populations concerned: reception, catering, accommodation, care, travelling teams, etc. However, from the theoretical protocol to its practical implementation, various difficulties are emerging. In particular, the need to involve these aid services, the constraints associated with the issuing of questionnaires under sometimes very specific conditions (at night, in public spaces or in mobile teams, etc.), the unique organisation of each of the departments in which investigations are required, call for thought to be given to the issues of coverage of the field investigated but also to the representativeness of the sample, and to adapt the method to the various configurations encountered. With this in mind, we will show how the integration of qualitative survey methods, such as the observation of services and interviews with managers, has helped to improve the methodology and scope of quantitative surveys through capitalised collection and experience.