India - Annual Survey of Industries: 2007-08

Sampling design adopted from ASI 2004-05 to ASI 2006-07.

i) Units with 100 or more workers will be categorized as census sector and the rest of the units will be treated as sample sector, without any change in the existing criteria;
ii) In the sample sector, the units will be stratified at 4 digit level of NIC-04 in each State separately and 1/5th of the units in each strata will be selected circular systematically for coverage in each ASI subject to a minimum sample size of 6 units in each stratum;
iii) This design will ensure that the whole universe of units is covered in five years;
iv) The classification of the units in the frame into census and sample sectors should be done in the beginning of the 5-year cycle and it should not be disturbed during the course of the cycle;
v) At the end of the cycle when the data on the all the units in the frame become available the frame should be updated and then the composition of census and sample sector should be re-drafted;
vi) In respect of the new units getting registered each year of the last 4 years in the 5-year cycle, a supplementary frame has to be prepared for each year and units for coverage from this supplementary frame of each year may be selected using the same criteria as was applied to the main frame.

A new sampling design was adopted for ASI 2007-08.

For ASI 2007-2008, the Census Sector has been defined as follows:
All industrial units belonging to the five less industrially developed states/ UT's viz. Manipur, Meghalaya, Nagaland, Tripura and Andaman & Nicobar Islands.
For the rest of the twenty-six states/ UT's., (i) units having 100 or more workers, and (ii) all factories covered under Joint Returns.
c) Strata (State by 4-digit of NIC-04) having less than or equal to six units after selecting the Census Sector units as defined above are also selected as census sector.
d) From the remaining merged frame (main frame 2004-05 plus supplementary frame 2005-06 plus supplementary frame 2006-07 plus supplementary frame 2007-08), samples were drawn considering higher allocation for states having relatively higher percentage contribution in sample sector in terms of GVA based on last 2 (two) years' available data. Accordingly, state wise census and sample sector contribution of GVA was calculated based on last two years' available data along with the RSE of GVA estimates in sample sector based on data of ASI 2005-06. In general, sampling fraction of 12% was considered for all the states within a StateXSectorX4-digit NIC with a minimum of 6 units evenly distributed in two sub-samples, except (i) Jammu & Kashmir, Himachal Pradesh, Chandigarh, Madhya Pradesh, Daman & Diu, Dadra & Nagar Haveli, Goa with 20% sampling fraction, (ii) Delhi & Rajasthan with 13% sampling fraction, (iii) Bihar & Pondicherry with 15% sampling fraction due to relatively higher contribution of sample sector to the total GVA and higher RSE.

The sampling design has undergone changes in the past on several occasions.