The sampling design of the Annual Survey of Industries (ASI) has undergone considerable changes from time to time keeping in view the technical and other requirements. However, the sampling design was broadly standardised from ASI, 1973-74. The salient features of the sampling designs of ASI, 1973-74 to ASI, 1998-99 are discussed in the following paragraphs.
For ASI, 1973-74 to ASI, 1986-87, factories employing 50 or more workers and using power and those employing 100 or more workers but not using power constituted the census sector. A systematic sample of one half of the remaining factories (i.e. those employing 10 to 49 workers and using power or 20 to 99 workers but not using power that constituted the sample sector) was covered in the first year of the cycle and the remaining in the second year.
For ASI, 1987-88 to ASI 1996-97, all factories in the frame were first grouped into the following four broad classes:
Factories employing 100 or more workers; All electricity undertakings irrespective of the number of workers; Factories located in twelve less industrialised States and Union Territories; and The rest.
The rest of the factories under (iv) were further classified within a State or UT into different strata by treating each of the 3-digit industry groups as a separate stratum. All strata having 20 or less number of units formed Category I. While strata having 21 to 60 units formed Category II, those with 61 or more units formed Category III. All the factories under (i), (ii), (iii) and Category I under (iv) were surveyed on a complete enumeration basis. A fixed sample of 20 factories from Category II and one third of those from Category III under (iv) were selected for the survey. In each of the strata of Categories II and III, sampling was done circular systematically with a random start after arranging the factories in a certain order. The constituent factories were first listed in the ascending order of their 4-digit industry codes. Within each 4-digit code, the factories were arranged in ascending order of district codes. Within the district, the factories were finally arranged in descending order of employment.
In ASI, 1997-98, the following categories of units were surveyed on a complete enumeration basis:
All units having 200 or more workers;
All electricity and public sector undertakings not covered under (i);
Additional units having significant input or output or gross value added (GVA) or net value added (NVA) which jointly contributed with (i) and (ii) not less than 90% of NVA within a State or Union Territory; and
All units located in each of 12 less industrialised States and Union Territories.
The remaining units within each State or Union Territory were grouped into different strata by treating each 4-digit industry code (National Industrial Classification, i.e. NIC 1987) as a separate stratum. From each stratum, a sample of factories was selected circular systematically with a random start after arranging the units or factories according to district x worker categories.
ASI, 1998-99 broadly followed the design of ASI, 1997-98 with the following changes:
Electricity, gas and water supply as well as cold storage sub-sectors were kept out of the coverage of ASI, 1998-99 onwards; Public sector undertakings were not given any special treatment for the purpose of complete enumeration; All units in the frame of 5 only States and Union Territories namely, Manipur, Meghalaya, Nagaland, Tripura, and Andaman and Nicobar Islands were surveyed on a complete enumeration basis; and While forming strata of sample sector units within a State or Union Territory, 4-digit industry code as per NIC 1998 was considered in place of NIC 1987.