Sampling
Sampling Procedure
The sampling frame for the Gender Protection Survey 2010 is based on the data and cartography from the 2008 Sudan Census of Population for the five state capital cities in Southern Sudan covered by the survey. A stratified three-stage sample design was used for the survey. The primary sampling units to be selected at the first sampling stage were the enumeration areas (EAs) in each of the five capital cities. The EAs were defined as small segments with about 100 to 150 households each, delineated for the census operations.
In general one census enumerator was responsible for completing the enumeration in one EA. These EAs were identified on census maps with well-defined boundaries, and conceptually the complete set of EAs was designed to cover the entire territory of Southern Sudan. The sampling frame for the survey was limited to the set of EAs that were identified as belonging to each of the five state capital cities included in the survey. The sampling frame for each city is in a spreadsheet format with information on all the geographic codes and the number of households enumerated in the 2008 Sudan Census for each EA.
The sampling frame of EAs for each city was used to select a sample of EAs systematically with probability proportional to size (PPS) at the first stage, where the measure of size is based on the number of households from the census frame. Following the selection of the sample EAs, a listing of households was conducted in each sample EA in order to provide the frame for the second stage of selection. The households were selected using random systematic sampling with equal probability. Within each sample EA 15 households were selected from the listing for the female interviews, and separate sample of 15 households for the male interviews. Within each sample household, the number of eligible female (or male) respondents 18 years or older were identified, and one eligible respondent was selected per household with equal probability at the third stage. The unit of analysis is the individual eligible female or male respondent selected for the interview within each sample household.
Weighting
In order for the sample estimates from the GPS Survey to be a representation of the population, it was necessary to multiply the data by a sampling weight, or expansion factor. The basic weight for each sample person was equal to the inverse of the overall probability of selection (calculated by multiplying the probabilities at each sampling stage).
The sample EAs for the GPS Survey was selected at the first stage within each stratum systematically with PPS from the 2008 Sudan Census frame. At the second stage 30 sample households were selected with equal probability from the listing for each sample EA. A systematic sub-sample of 15 households was selected for female respondents, and the remaining 15 sample households were assigned to male respondents. At the third stage one eligible female or male respondent 18 years or older was selected within each sample household.
