India - Survey of Disabled Persons , July-Dec 1990, NSS 47th Round

Outline of Sample Design:
A stratified multi-stage design was adopted for the conduct of survey of NSS 58th round. The first-stage units were census villages (panchayat wards for Kerala) in the rural sector and the NSSO Urban Frame Survey (UFS) blocks in the urban sector. The ultimate stage units were households in both the sectors.

Sampling Frame for First-Stage Units:
For the rural sector, the list of Census 1991 villages (panchayat wards for Kerala) and Census 1981 villages for J & K constituted the sampling frame. For the urban sector, the list of latest available Urban Frame Survey (UFS) blocks was considered as the sampling frame.

Stratification

Rural sector:
Two special strata were formed as given below at the State/ UT level on the basis of Population Census 1991 viz.

Stratum 1: all FSUs with population between 0 to 50, and
Stratum 2: FSUs with population more than 15,000

The special stratum 1 was formed if at least 50 such FSU's were found in a State/UT. Similarly, special stratum 2 was formed if at least 4 such FSUs were found in a State/UT. Otherwise, such FSUs were merged with the general strata.

From the remaining FSUs (not covered under stratum 1 &2) general strata (hereafter, stratum will refer to general stratum unless otherwise mentioned) was formed and numbered 3, 4, 5 …. etc. (even if no special strata have been formed). Each district of a State/UT was normally treated as a separate stratum. However, if the provisional population of the district was greater than or equal to 2.5 million as per Census 2001, the district was divided into two or more strata with more or less equal population as per population census 1991 by grouping contiguous tehsils. However, in Gujarat, some districts were not wholly included in an NSS region. In such cases, the part of the district falling in an NSS region constituted a separate stratum.

Urban sector:
In the urban sector, stratum was formed within each NSS region on the basis of size class of towns as per Census 1991 town population except for towns specified in Table 4. The stratum number and their composition (within each region) are given below:

stratum 1: all towns with population (P) < 0.1 million
stratum 2: all towns with 0.1= P < 0.5 million
stratum 3: all towns with 0.5= P < 1 million
stratum 4,5,6, … each town with P= 1 million

The stratum numbers was retained as above even if, in some regions, some of the stratum is not formed.

Total sample size (FSUs):
A total number of 8338 and 9076 first-stage units were selected for survey in the Central and State samples respectively. The sample size by State and Sector is given in the Annexure

Allocation of total sample to States and UTs:
The total sample FSUs was allocated to the States and UTs in proportion to provisional population as per Census 2001 subject to the availability of investigators ensuring more or less uniform work-load.

Allocation of State/ UT level sample to Rural and Urban sectors:
State/UT level sample was allocated between two sectors in proportion to provisional population as per Census 2001 with double weightage to urban sector.

Allocation of Rural /Urban sector level sample size to strata / sub-strata:
Both rural and urban sector samples allotted to a State/UT were allocated to different strata in proportion to population of the stratum. All the stratum-level allocations were adjusted to multiple of 2. Stratum-level sample size in the urban sector was further allocated to 2 sub-strata in proportion to the number of UFS blocks in them with double weightage to sub-stratum 1 subject to a minimum sample size of 2 or 4 to sub-stratum 1 according as stratum-level allocation is 4 or greater than 4. Sub-stratum level allocations in the urban sector were made even.

Selection of FSUs:
FSUs were selected in the form of two independent sub-samples in both the sectors. For special stratum 2 and all the general strata of rural sector, FSUs were selected by probability proportional to size with replacement (PPSWR) where size was the 1991 census population. For urban sector and special stratum 1 of rural sector, FSUs were selected by simple random sampling without replacement (SRSWOR).


There was no deviation from the original sampling design.