Rural Youth Survey 2012

The Household Survey has a sample size of 1,400 households the entire rural area of Tunisia, as defined by the Tunisian Statistical Office, Institut National de la Statistique (INS). For the purpose of sampling, administrative governorates were grouped into 3 Survey Regions. The data is representative on the level of these Survey Regions, which largely correspond to socio-economically and geographically distinct rural zones. The first survey region covers the Coast and includes coastal governorates in the North and East of the country. The second survey region covers South and includes the southern governorates. The third survey region is covers the rural Interior of Tunisia and includes the remote areas of central and western Tunisia, incl. the Algerian border.

The sample was drawn from the latest available census, the 2004 General Census of Population and Housing, provided by the INS. This census also provided the sampling frame for the corresponding Urban and Peri-Urban Youth Survey. For determining the number of households in rural areas, proportionality of the possible locations was used to ensure representativeness. Because of the overall research focus on youth, the sampling design ensures representativeness of youth population, which is defined by ages 15-29. The proportionality to youth population size is based on the disaggregation of Tunisia into Enumeration Areas (EA). Each EA contains about 100-120 households. In total 70 EAs were randomly selected, with 29 EAs along the Coast, 10 EAs in the South, and 31 EAs in the Interior survey regions. The relative distribution between the survey regions corresponds to their respective shares of youth population. From each of these 70 EAs, 20 households were randomly selected, leading to a total sample size of 1,400 households.

The random sampling of PSUs was performed by experts from the INS, who were also responsible for the sample frame. The drawing of 20 households from each PSU is processed on a systematic and clearly defined approach. A random-walk procedure was conducted for each of the PSUs of the sample, which included 2 separate starting points at opposing ends of the east-west dimension of each PSU, and moving towards the population center of the PSU to allow a full coverage of both centrally and remotely located households.