Sampling Procedure
In 51st round the total number of villages and blocks surveyed at the national level was 13690 for the central sample. These samples were classified into two categories as "Sample-1" and "Sample-2" in both rural and urban areas. The samples of sample-1 were distributed over the four sub-rounds in equal numbers, whereas for sample-2, there was no sub-round restriction. The survey period of 51st round was July 1994 - June 1995.
A stratified two-stage sampling design was adopted. The first stage units (FSUs) were the villages (panchayat wards in Kerala) in rural areas and urban blocks in urban areas. In particular, for the urban areas, enumeration blocks (EBs) were taken as FSUs where 1990 economic census (EC -90) frame was used and NSSO urban frame survey (UFS) blocks otherwise. The second stage units (SSUs) were the manufacturing and repairing enterprises in the selected FSUs. 2 Allocation of sample FSUs between rural and urban areas State/UT level total sample size (i.e.,total number of villages and blocks to be surveyed) was allocated between rural and urban areas in proportion to population as per 1991 population census with double weightage to the urban areas. The total sample size for both rural and urban areas, as mentioned above, was equally divided into two sample types, namely Sample-1 and Sample-2.
Sampling frame: For rural areas, list of villages showing number of OAMEs, NDMEs and DMEs as per 1990 economic census was used as the sampling frame for selection of FSUs, for both the sample types, in different States/UTs except Jammu & Kashmir(J & K), Arunachal Pradesh and Nicobar district of Andaman & Nicobar Islands. For the state of Kerala, however, list of Panchayat wards, giving count of OAMEs, NDMEs and DMEs at panchayat level as per 1990 economic census, was used. For Jammu & Kashmir, list of 1981 population census villages, while for Nicobar district of Andaman & Nicobar Islands, list of 1991 population census villages was used. In Arunachal Pradesh, list of villages as per 1991 population census was used as the sampling frame for selection of sample "nucleus" villages around which clusters were formed. In urban areas, list of EBs as per 1990 economic census constituted the sampling frame for Class 1 towns (except in Jammu & Kashmir) of sample type 1 and for all towns (except J & K) of sample type 2. The list of UFS blocks was used as the sampling frame for all towns of Jammu & Kashmir and other than class 1 towns of the other states for sample type 1, and for all towns of Jammu & Kashmir only in the case of sample type 2.
Stratification: In rural areas, for samples of type 1 each district generally formed a separate stratum. However for the state of Gujarat, where NSS regions cut across some district boundaries, parts (viz., groups of taluks) of each such district belonging to different NSS regions formed separate strata. If any district (or part thereof falling in an NSS region in case of Gujarat) had a smaller number of manufacturing enterprises, it was clubbed with the neighbouring district within the same NSS region to form a separate stratum to ensure a minimum allocation of 8 villages at the stratum level. Each district as a whole was always taken as a separate stratum for samples of type 2. In the case of urban areas for samples of type 1, strata were formed within each NSS region by grouping cities/towns according to the population of the cities/towns viz., p<=0.5, 0.5<=p<1, 1<=p<5, 5<=p<10 and p>=10 (where p stands for 1991 census population of the town in lakhs). Each city with population 10 lakhs or more formed a separate stratum. For samples of type 2, strata were formed within each district by grouping cities/towns according to the population, viz., p<1, 1<=p<5, 5<=p<10 and each city with p>=10. However, no grouping of towns was made for the states of Assam, Haryana and Pondicherry.
Sub-stratification: For rural samples of type 1, where 1990 EC frame was used, the FSUs in a stratum were grouped into 3 sub-strata as detailed below.
(a) sub-stratum 1 consisting of FSUs having at least one DME;
(b) sub-stratum 2 consisting of those of the remaining FSUs which had at least one NDME; and
(c) sub-stratum 3 consisting of all the residual FSUs in the stratum.
There was no sub-stratification at the stratum level for the States/UTs where population census frame was used for selection of FSUs. For these States/UTs, all FSUs in a stratum were identified with substratum 3 for operational convenience. In urban areas for samples of type 1 three sub-strata were formed in class 1 towns (except for J & K) as follows :-
(i) sub-stratum 1 consisting of EBs having at least one DME;
(ii) sub-stratum 2 consisting of those of the remaining EBs in the stratum having at least one NDME; and
(iii)sub-stratum 3 consisting of all the remaining FSUs in the stratum.
For other towns, includiing all towns of Jammu & Kashmir, two sub-strata were formed, where sub-stratum 1 consisted of all UFS blocks identified as Industrial area (IA) and the remaining UFS blocks in the stratum constituted sub-stratum 2. For samples of type 2 stratum level substratification was not done in both rural and urban areas.
Allocation of sample FSUs among strata and sub-strata: The total rural samples of type 1 for a State/UT were allocated to the constituent strata/sub-strata in proportion to weighted sum of the number of manufacturing enterprises of different types, as available from EC 1990, the weights being 16, 4 & 1 for DME, NDME and OAME respectively. A minimum allocation of one sample was ensured at sub-stratum level and efforts were made to make the stratum level allocations as multiples of 4 to allocate equal number of samples in each of the four sub-rounds. However, the total rural samples of type 2 for a State/UT were allocated to the constituent strata in proportion to their total number of manufacturing enterprises (weights being equal for DME, NDME & OAME) as available from EC 1990. A minimum allocation of 2 FSUs was done for each stratum. In urban areas, for samples of type 1, allocated sample size of FSUs for a State/UT was further allocated to the constituent strata in proportion to weighted sum of the manufacturing enterprises (in proportion to population for J & K) as was done in the rural areas. For class 1 towns (except for J & K), the stratum level FSUs were further allocated to three sub-strata again in proportion to weighted sum of the number of manufacturing enterprises in the ratio 16 : 4 : 1. The allocation at the substratum level had been kept at a minimum of 2 FSUs. For other towns (including all towns of J & K) the UFS blocks of substratum 1 were completely selected for survey, subject to a maximum of half of the stratum allocation of samples of a particular type; and the rest to sub-stratum 2. The total urban samples of type 2 for a State/UT (except for J & K) was allocated to the constituent strata in proportion to their total number of manufacturing enterprises. For the state of Jammu & Kashmir stratum level allocation was done in proportion to 1981 census population.
Selection of first stage units (FSUs): The sample FSUs for rural samples of type 1, at stratum x sub-stratum level, were selected with probability proportioned to size (PPS), size being the number of enterprises/population as detailed below :-
(A) For States/UT where 1990 EC frame was used, size was number of DMEs, NDMEs and OAMEs in the FSU belonging to sub-stratum 1, 2 and 3 respectively.
(B) For other States/UTs viz., (i) Arunachal Pradesh, size = 1 for each FSU (ii) Jammu & Kashmir, size = population in the FSU as per 1981 census; (iii) Nicobar district of Andaman & Nicobar, size = population in the FSU as per 1991 census.
The stratum x sub-stratum level rural samples of type 2 were selected by PPS with replacement, size being the total number of manufacturing enterprises (i.e. total of DMEs, NDMEs and OAMEs) in the FSU for the States/UTs where 1990 EC frame was used and for other States/UTs where 1990 EC frame was not used size was same as in rural samples of type 1. In urban areas sample blocks of type 1 were selected from each sub-stratum of class 1 towns (except for J & K) with PPS with replacement taking number of DMEs, NDMEs and OAMEs in the FSU as size for sub-stratum 1, 2 & 3 respectively. For other towns (including all towns of J & K) the sample blocks from each sub-stratum were selected circular systematically with equal probability. Sample blocks of type 2 were selected with PPS with replacement, size being the total number of OAMEs, NDMEs and DMEs for all the Sates/UTs except J & K for which simple random sampling without replacement technique was followed for selection of sample blocks. Both in rural and urban areas, for samples of type 1, size of an FSU of substratum 3 was taken as 1 if there was no OAME or no information about number of enterprises due to incompleteness of the frame. The FSUs of sample type 2 having no manufacturing enterprises or no information about them were also considered as FSUs of size 1.
Hamlet-group/sub-block formation : Large villages/blocks having approximate present (at the time of survey) population 1200 or more (600 or more for rural areas of Himachal Pradesh, Sikkim and Poonch, Rajouri, Udhampur & Doda districts of J & K) and/or approximate present number of non-agricultural enterprises more than 200 were divided into a suitable number (say,D) of hamlet-groups (h.g.)/sub-blocks (s.b.). The survey was conducted in two h.g.s/s.b.s selected from D hamlet-groups/sub-blocks in the following manner. The h.g./s.b. having maximum number of manufacturing and repairing enterprises (OAMEs, NDMEs and DMEs taken together) was always selected and allotted a serial number '0'. From the remaining h.g.s/s.b.s, one more h.g./s.b. was selected at random and allotted a serial number '1'. In case there was not even a single manufacturing or repairing enterprise in a large village/block, the h.g./s.b. having maximum population was selected and allotted serial number '0' and from the remaining h.g.s/s.b.s another was selected at random and allotted serial number '1'. In the case of a small village/block not requiring h.g./s.b. formation, the whole village/block was treated as a h.g./s.b. bearing serial number '0'. The minimum number of h.g./s.b. formed in large villages/blocks was 4.
Sampling of enterprises: All OAMEs, NDMEs and DMEs which operated for at least 30 days (15 days in case of seasonal enterprises) during the last 365 days were considered for sampling. The enterprises registered under Factories Act, 1948 and Bidi and Cigar workers (condition of Employment) Act, 1966 were not considered for survey purpose. For FSUs of sample type 2, all the eligible manufacturing and repairing enterpriseslisted in the village / block /h.g. /s.b. were taken up for detailed survey. However, for samples of type-1, a maximum of 16 enterprises were selected for detailed enquiry.
All eligible manufacturing and repairing enterprises of different types were separately arranged by broad industry division (2-digit) codes as per NIC, 1987. This grouping of enterprises by NIC codes was done independently for each of the two selected hamlet-groups/sub-blocks for FSUs requiring hamlet-group/sub-block formation. The required number of sample enterprises were then selected circular systematically from the rearranged frame, separately for OAMEs, NDMEs and DMEs.