Survey ID Number
ZMB_1996_LCMS-I_v01_M
Title
Living Conditions Monitoring Survey-I 1996
Sampling Procedure
The LCMS1 covered all the nine (9) provinces of Zambia, both rural and urban areas on a sample basis. The domains of study and data disaggregation for this survey were:-
- Rural
- Urban
- Province
- District
Stratification
The country is made up of 9 provinces comprising 57 districts delineated by the local government administration. CSO has delineated the districts into Census Supervisory Areas (CSAs) and then these into Standard Enumeration Areas (SEAs). Each CSA is made up of about 3 SEAs. The districts are made up of 4,193 Census Supervisory Areas (CSAs) out of which 3,231 are rural and 962 are urban. Further stratification was done using urban/rural and centrality as stratifying variables. (See Appendix II for definition of Centrality). The urban SEAs have been classified into low cost, medium cost or high cost depending on the type of housing in the area. Within the selected rural SEAs households have been classified on the basis of the type of agricultural activity in the area into small scale, large scale, medium scale and nonagricultural households.
Sampling frame
The sampling frame consisted of 4,193 CSAs and 12,999 SEAs. It was obtained from the 1990 Census of Population and Housing. The computer package QUATRO PRO was used to sort the data by rural/urban, by centrality and by low cost, medium cost and high cost to form strata. All in all, the frame gives information on the population size of each SEA throughout the country, the number of households, information about rural/urban, centrality as described in Appendix II, stratum number, the number of SEAs to be selected from each stratum.
Sample size
A sample of 610 SEAs was selected. Only one SEA was selected from each CSA (262 SEAs were selected from the urban stratum and 348 from the rural stratum).
Allocation
Allocation of SEAs to provinces was done using the "modified equal allocation" method. This simply means that allocation is first done equally across all the provinces by dividing the sample size by the number of provinces. In this case it meant that each province was to get 67 SEAs. However, considering other factors such as the population size of the provinces, heterogeneity or homogeneity of the province and also using the probability proportional to size method, additions and subtractions were made accordingly for all stages of allocation. First, allocation was done on provinces considering the variables mentioned above. Then allocation was done at district level in the same way. Within the districts, allocation was done by rural/urban by the same method. Within the rural and urban strata, allocation was done considering centrality of the SEAs so as to increase the probability of including even the remote areas in the sample. The minimum size for each district sample was 7 SEAs (See Appendix I).
Listing
In each selected SEA, households were listed and each household given a unique sampling serial number. A circular systematic sample of households was then selected. Vacant residential housing units and non-contact households were not assigned sampling serial numbers.
The circular systematic sampling method assumes that the households are arranged in a circle and the following relationship applies (Kalton, 1983):-
Let N = nk
were, N = Total number of households listed in a stratum
n = Total sample required from a stratum
k = The sampling interval in a given stratum calculated as k = N/n
Therefore, for the urban strata, k = N/25, because 25 households were selected from each SEA in the sample.
And for the rural strata, k = N/15, as 15 households were selected from each rural SEA. In the rural areas, 7 households were selected from the stratum of small scale farmers, 5 from the medium scale, 3 from the non-agricultural and all the large scale households if any. Therefore, the number of selected households was more where there were large scale farmers. The N was different for each SEA depending on how many households were identified as large scale in the listing. At this stage, a random-start number was obtained using a table of random numbers. This number was between 1 and N. The household whose random number lay between 1 and the random start was the first to be selected. Then k, the sampling interval was added to the sampling serial number of each selected household in the respective strata until the required n was achieved. All in all 6,550 urban households were selected and 5,220 were selected from the rural stratum.
Panel design
The panel design of sampling was chosen to allow for monitoring of change. In this case, half of the households in each SEA were retained for re-interview in the next round of the LCMS. The N this time was the number of selected and canvassed households and the small n half this number. The households were selected using the circular systematic sampling method just as above. For example, if 25 households are canvassed in one urban SEA, then k = 25/13.