Survey ID Number
IND_2005_NSS62-SCH2.2_v01_M
Title
National Sample Survey 2005-2006 (62nd round) - Schedule 2.2 - Unorganized Manufacturing Sector in India
Sampling Procedure
One salient feature of the sample design adopted during the 62nd round was the use of list frame, in addition to the usual area frame, which was done to capture sufficient number of relatively 'bigger' enterprises with a view to improving the overall estimate of gross value added per worker, total number of workers, total input, total output, etc. A list of 8,000 big non-ASI manufacturing enterprises2 for the urban sector only was prepared as per the data of the census of manufacturing enterprises conducted by Development Commissioner of Small Scale Industries (DCSSI) in 2003. This list served as the list frame. All these units in the list frame were considered for survey without resorting to any sampling. For the coverage of all other enterprises in the universe, the usual area frame approach was followed for sampling of enterprises in stages. It is important to mention that this dual frame approach was experimented for the first time in the 62nd round. The effectiveness of using the list frame has been discussed under Chapter four.
In the area frame approach, the list of all the villages (panchayat wards in case of Kerala) / urban blocks of the country served as the sampling frame of first stage units (FSUs). Thus, the FSUs were villages (panchayat wards in case of Kerala) in the rural sector and urban blocks in the urban sector. The ultimate stage units were enterprises in both the sectors. However, in case of large FSUs requiring hamlet-group (hg) / sub-block (sb) formation, one intermediate stage in the sampling involved the selection of two hg's / sb's from each FSU out of a minimum of three hg's/sb's formed in the FSU. Of these two selected hg's/sb's, one was selected with probability '1' (termed as segment 1) and another one (termed as segment 2) was selected from among the remaining hg's/sb's of the FSU at random. The hg/sb selected with certainty (i.e. segment 1) was the hg/sb having maximum number of directory manufacturing establishments (DMEs) (or with maximum number of non-directory manufacturing establishments (NDMEs) if there was no DME, or with maximum number of own account manufacturing enterprises (OAMEs) if there was no DME/NDME, or with maximum population if there was no DME/NDME/OAME3 in the entire FSU). Smaller FSUs without any hg/sb formation were identified/categorized as segment 1 for the purpose of survey (segment 2 does not exist for such FSUs). As regards the first stage stratification, two basic strata were formed within each district of a State/UT: rural stratum comprising all rural areas of the district and urban stratum consisting of all urban areas of the district. However, each city with a population of one million or more as per Census 2001 was invariably treated as a separate stratum by itself. For details of stratification, sub-stratification and selection of sample FSUs, reference may be made to Appendix-B of of the final report no.526.
For each of segments 1 and 2 for the selected sample FSUs, a frame of eligible enterprises was prepared by the field investigators by visiting each and every house/household within the selected geographical area. While doing so, if any enterprise of the list frame was encountered, care was taken not to list it again within segment 1 or 2 as a part of the area sample / area frame to guard against duplication of enterprises between the two types of frames. Listing and sampling of enterprises in the area frame was independent for each of segments 1 and 2. In this context, it may be mentioned that for each selected FSU of rural sub-strata 1 and 2 only (see Appendix B for composition of these two sub-strata), segment 9 was also carved out within the FSU, which comprised top 10 big non-ASI registered SSI enterprises (identified by jointly considering the number of workers in the enterprise and gross value of output of the enterprise) located within the boundaries of the entire FSU. The list of such units for selected FSUs was made available to the field investigators in order to facilitate formation of segment 9.
Respective frames of segments 1 and 2 in these FSUs excluded the units listed under segment 9. The effectiveness of the formation of segment 9 has been discussed under Chapter four of the final report no.526.
All the eligible enterprises listed under each of segments 1, 2 and 9 were stratified into 2 broad second-stage strata by enterprise type i.e. OAME & NDME/DME. Each of these two broad second-stage strata was further divided into 3 broad manufacturing groups (BMG) i.e. BMG 1, BMG 2 and BMG 3. BMG 1 comprised eligible enterprises belonging to NIC codes 15 and 20. BMG 2 consisted of eligible enterprises belonging to NIC Codes 23, 27, 30-35 and 01405 while all other eligible enterprises belonging to the rest of the NIC codes under coverage formed BMG 3. Thus within a segment for any given sample FSU, six ultimate second-stage strata were formed by jointly considering the broad second-stage strata and BMG. A total of 12 manufacturing enterprises - two from each ultimate second-stage stratum - were selected for detailed enquiry. In case of shortfall of enterprises in the frame of any particular second-stage stratum, compensation was made by selecting additional samples from other second-stage strata so that a total of 12 enterprises were selected from each FSU for detailed enquiry. Data from enterprises were collected from books of accounts if those were available. Otherwise, it was through oral enquiry.
Design-based estimates of aggregates for any selected survey characteristic were obtained separately for list frame and area frame. Finally these two sets of estimates were added to get the pooled aggregate estimate for the combined frame. In the final report,
discussion will be focused on the pooled estimate based on two types of frames used in
the survey.
Sample size and its limitation
Out of 8,000 enterprises selected from the list frame, data could be collected from only 2,260 enterprises (for state/UT wise distribution of allotted and surveyed numbers of enterprises as per the list frame, reference may be made to Table 2 at the end of chapter 1 of the final report no.526. Thus a large percentage (72%) of the list frame enterprises became casualties (see Chapter four for the reasons for casualties). In the area frame, 80,637 enterprises (42,050 from rural India and 38,587 from urban India) spread over 4,798 villages and 5,125 urban blocks across the country were surveyed. It is important to mention here that even though a large percentage of list frame enterprises became casualty, theoretically the surveyed enterprises netted through the list frame and area frame represented the whole universe of the unorganized manufacturing sector. The estimation procedure developed for the purpose (see Appendix B) also ensured this. The sample sizes in terms of number of FSUs and enterprises are presented in Tables 1 and 2 respectively at the end of chapter 1 of the final report no.526.
Thus, altogether a total of 82,897 unorganized manufacturing enterprises considering both list frame and area frame were surveyed. In the appendix tables, although estimates are presented at disaggregated level i.e. sector (rural/urban/combined) x enterprise type (OAME/NDME/DME/All) by NIC code (for all-India) or by state/UT, sample size at the ultimate level of disaggregating may sometimes be very small. The users of the final report are advised to look at the related sample size while using the survey results particularly at the disaggregated level.
A detailed description of the sampling procedure can be found in section 2 Appendix B in the final report 526.