Environment for Development Dar es Salaam Energy Survey - Households

The sampling strategy was designed in such a way that the resulting data would be able to meet the following goals:
a) The sample should be as representative as possible of household energy use in Dar es Salaam.
b) The sample should be structured in such a way that enables an impact evaluation of the UNIDO bioethanol cookstove program in a future follow-up survey.
c) The household and enterprise surveys should take place in similar areas in order to inform an understanding of the energy enterprise landscape from a household perspective.
A multi-stage stratified random sampling design was followed in the selection of final wards, streets and households to include in the survey.
In the first stage, the intended sample size of 1000 was allocated between the three main districts of Dar es Salaam (Temeke, Ilala and Kinondoni). This was done in proportion to the population of each district, yielding the number of households to be interviewed in each district. The population numbers used were based on the 2012 Census data – the most recent census of households in Dar es Salaam
In the second stage, the survey team visited the offices of the District councillors of each district and asked them to rank all the wards within their district by socio-economic status (from richest to poorest). District councillors were asked to assign a number to each ward in their District, where 1 is richest and n is the poorest (n depends on the number of wards per-district).
This ranking was used to divide wards into three socio-economic status groups. These groups were "Relatively Poor" "Middle" and "Relatively Rich", corresponding to the bottom, middle and top thirds of the socio-economic status rankings assigned by District councillors . The primary reason for this exercise was to ensure the inclusion of households across the income distribution in the survey.
Following this, six wards were selected from each District. Two “Relatively Poor” wards, two “Middle” wards and two “Relatively rich” wards were selected in each district, yielding a total number of 18 wards in this survey.
The following process was used to select these 6 wards from each district: In order to ensure the sampling design would be compatible with a later impact evaluation of the UNIDO ethanol stove program, in each district, 3 of the 6 wards (1 poor, 1 middle, 1 rich) were randomly selected from the set of wards targeted by the Ethanol stove rollout. The other 3 wards (1 poor, 1 middle, 1 rich) were randomly selected from the set of wards not targeted by Ethanol stove rollout .
Within each selected ward, two streets were randomly selected . Within each street, the number of households to be skipped by enumerators was determined by the dividing the estimated number of households per street (derived from the estimated number of main streets per ward) by the number of interviews to be completed on that street. Enumerators were then instructed to skip this number of households before interviewing another household. For details, see the attached document "Basic Information Document", section "Sample Design". Visa mindre..