India - Debt and Investment Survey: NSS 48th Round, Schedule 18.2, January - December 1992

Sample Design
A stratified two-stage sampling design was adopted for the survey with the first stage units as census villages for rural areas and the Urban Frame Survey blocks for urban areas. Households formed the second statge units in both rural and urban areas.

Sampling frame for first stage units(FSU's):
In the rural sector, the sampling frame in most of the strata was the 1981 census list of villages. However, in Assam, where the 1981 census was not undertaken, and in a few districts of other states, where the available list as per 1981 census was incomplete, the 1971 census list of villages was used. In the urban sector, the sampling frames used in most cases were the lists of NSS Urban Frame Survey (UFS) blocks. However, the 1991 census house listing enumeration blocks were considered as the sampling units for some of the new towns declared as urban areas in the 1991 population census.

Stratification:
Each state/union territory (u.t.) was divided into one or more agro-economic regions by grouping contiguous districts which are similar with respect to population density and crop pattern. In Gujarat, however, some districts were subdivided for the purpose of region formation on the basis of location of dry areas and the distribution of tribal population in the state. The total number of regions formed in the India as whole was 78.

In the rural sector, within each region, each district with a rural population of less than 1.8 million according to the 1981 census formed a single basic stratum. Districts with larger population were divided into two or more strata, depending on population, by grouping contiguous tehsils, similar as far as possible in respect of rural population density and crop pattern. In Gujarat, however, in the case of districts extending over more than one region, the portion of a district falling in each region constituted a separate stratum even if the rural population of the district as a whole was less than 1.8 million. Further, in Assam, the strata formed for the earlier NSS rounds on the basis of 1971 census rural population exactly in the above manner, but with a cut-off of 1.5 million population, were retained as the strata for rural sampling.

In the urban sector, strata were formed, again within NSS regions, on the basis of 1981 (1991 in some of the new towns) census population of towns. Each city with a population 10 lakhs or more formed a separate stratum by itself. The remaining towns of each region were grouped to form three different strata on the basis of 1981 (1991 in a few cases) census population.

Allocation of sample:
A total all-India sample of 6,812 first stage units (4,328 villages and 2,484 urban blocks) - determined on the basis of investigator strength in different states/u.t.'s and the expected workload per investigator - was initially allocated to the
states/u.t's in proportion to central field staff available. The sample thus obtained for each state/u.t. was then allocated to its rural and urban sectors considering the relative sizes of the rural and urban population with almost double weightage being given for the urban sector. Within each sector of state/u.t., the allotted sample size was re-allocated to the different strata in proportion to the stratum population. All allocations were adjusted so that the sample size for a stratum was at least a multiple of 4 for the rural and urban sectors separately. This was done to accomplish equal sized samples in each sub-sample and sub-round. The only exception was Daman & Diu for which the first stage rural sample comprised 2 villages only.

Selection of first stage units:
The selection of sample villages was PPS (with replacement) with population as the size variable, in the form of two independent subsamples. The sample blocks were selected by simple random sampling without replacement, also in the form of two independent subsamples.

Selection of hamlet-groups/sub-blocks:
Large villages and blocks were divided into a suitable number of hamlet-groups and sub-blocks, respectively, having more or less equal population content. Two hamlet-groups were then selected from large villages, whereas only one sub-block was selected from the large blocks. The hamlet-groups were selected circular systematically and the sub-block with equal probability.

Selection of households:
Two different procedures of selection of households were used for the rural and urban sectors. Different procedures for the two sectors were necessary, since in the rural sector schedules of enquiry for LHS survey and Debt & Investment survey were required to be canvassed in two separate sets of sample households, while in the urban sector, both the schedules were to be canvassed in the same set of sample households.

In the rural sector, nine households were selected from each sample village/selected hamlet groups. For selecting a sample of nine households, each sample village/hamlet group was sub-divided into 7 AIDIS sub-strata on the joint consideration of "land possessed" and "indebtedness status" of the households; first, all the households of the sample village/selected hamlet groups were divided into four LHS sub-strata by area of land possessed by them. Households possessing either no land or land less than 0.005 acre were grouped in substratum 1. The rest of the households were
then arranged in ascending order by area of land possessed and classified into three substrata, 2, 3 and 4, such that the total area of land possessed by the households in each of the 3 sub-strata was nearly the same. Each of the LHS sub-strata 1 and 2 was further divided into "indebted" and "not indebted" groups to form AIDIS sub-stratum 1 to 4. AIDIS sub-strata 5 to 7 are formed by first merging LHS substrata numbers 3 and 4 and then sub-divided by the merged group into 3 classes, viz., (a) indebted to institutional agencies with or without being indebted to non-institutional agencies ( b) indebted to non-institutional agencies alone and (c) not indebted. Independent sample of size 1,1,1,2,1,1&2 were selected circular systematically from the AIDIS sub-strata 1,2,3,4,5,6 and 7 respectively

In the urban sector, a sample of 9 households was selected from each sample urban block/sub-block. The households of a sample block/sub -block were classified into 7 AIDIS sub-strata, considering the monthly per capita consumption expenditure (mpce) and indebtedness status of the households. For this, the households were first grouped in three mpce classes, viz., less than A, A to B and B & above. The cut-off points A and B were determined at the state-level on the basis of mpce obtained from the survey on consumer expenditure, NSS 43rd Round, such that the mpce classes, below A, A to B, and B and above, respectively constituted 30 p.c, 60 p.c, and 10 p.c. of the urban population of the state. These mpce classes were further sub-divided by indebtedness status of the households to form 7 AIDIS sub strata. Independent samples were selected circular systematically from each of the sub-stratum. The number of households was selected from sub-strata 1, 2, 3, 4, 5, 6 and 7 respectively 1, 1, 1, 1, 2, 2 & 1.

Sample size:
In all, the survey covered 57,031 households spread over 6,650 sample villages/blocks.

There have been no deviations from sample design.