In order to capture representative samples of the different target populations outlined above, the survey used two different sampling approaches. A probabilistic sampling approach was used for IDPs from Somaliland, refugee returnees, economic migrants and host community living in settlements. The sample in the settlements was stratified by population group and was representative at the total population level. The distribution of each population group was proportionate to their size in each settlement and was considered when distributing the sample across the settlements. For the IDPs from South Central Somalia and refugees/asylum seekers, a non- probabilistic sampling approach was adopted due to the relatively small size of these population groups and scattered distribution in the urban area of Hargeisa. For the IDPs originating from south- central Somalia and residing in the urban areas of Hargeisa outside of the known settlements, the profiling deployed respondent- driven sampling, whereby the individuals being surveyed referred the data collectors to people they know. In turn, the next individuals surveyed referred the data collectors to people they know, and so forth (snowball sampling). Therefore, results cannot be generalised to the total population, but can give a good indication of the situation of the SC IDPs living outside of settlements. In addition, this method introduced a bias in the sample, as only the households linked to the initial network could be reached.
