Title : On some aspects of adequacy of the sample sizes at different stages of sampling for estimation of area, yield and production of food grain crops in India
India is a vast agrarian country having geographical area of 329 million hectares. It is divided into 28 States and 08 Union Territories (UTs are administered by the President through an Administrator appointed by him/her). These comprise of 750 Districts and a total of 628,221 Villages. Around 70 per cent of India’s population (around 1.360 billion) lives in Villages. There are good number of food grain crops and non-food grain crops grown during the entire Agricultural Year (01 July to 30 June of subsequent year). An Agriculture Year is divided broadly into three seasons, namely Kharif (July-October), Rabi (November-April), Summer (May-June). For effective planning concerning arranging feeding the vast population of human and livestock, the estimation of total production of food grains and non-food grains is of paramount importance. In India, the estimation of yield rates of food grain crops is done on the basis of crop cutting experiments (CCEs) conducted in majority of States/UTs under the National Programme of Crop Estimation Survey (CES). At district level, the sampling design adopted usually is stratified multi stage random sampling. Presently, around ninetyfive per cent of the total food grains production is estimated on the basis of yield rates obtained from these CCEs conducted on scientific basis spread over various States/UTs. Around more than one million CCEs (of different sizes and shapes) are conducted covering 52 food crops and 16 non-food crops. The conduct of such a large number of CCEs had been in question since long. Accordingly, a high powered committee constituted by Ministry of Agriculture & Farmers Welfare (MoAFW), Government of India (GoI), has recommended for reducing the sample sizes at different stages of sampling, which is highly being criticized and debated.
In India, the Directorate of Economics and Statistics (DES) under Department of Agriculture & Cooperation (DAC), MoAFW, GoI, releases estimates of area, yield and production of main food grain crops, oilseeds, sugarcane, fibers and important commercial and horticulture crops. These crops together account for nearly eightyseven per cent of the total output of agriculture. The estimates of crop production are obtained by multiplication of area estimates by corresponding yield estimates. Therefore, the estimates of area and yield rates assume immense importance in the entire gamut of agricultural statistics.
The need for timely, reliable and comprehensive statistics on area, yield and production of crops assumes special significance in view of the vital role played by the agriculture sector in the Indian Economy. The primary responsibility for collection of statistics of land use and area under crops following prescribed procedures rests with the various State Departments. The yield rates of principal crops are estimated through General Crop Estimation Survey (GCES) conducted by State agencies following scientific techniques of random sampling.
Field Operations Division (FOD) of the National Sample Survey Office (NSSO) under the Ministry of Statistics and Programme Implementation (MoSPI), GoI has the overall responsibility for providing technical guidance to States/UTs in developing suitable survey techniques for obtaining reliable estimates, assistance in training of staff and exercising supervision. Under the Improvement of Crop Statistics (ICS) Scheme which was taken up during 1973-74 with the objective of locating, through the joint efforts of NSSO and State Governments, the deficiencies in the system of crop statistics by exercising technical supervision over the primary field work and suggesting remedial measures for improving the system. To achieve this, sample checks on area enumeration, and area aggregation are carried out in a sample of about 10,000 villages in each season and on conduct of around 31,000 CCEs during an agricultural year.
Data pertaining to CCEs for different crops, having smaller sample sizes, pertaining to good number of States under ICS scheme obtained from NSSO was statistically analysed. Estimates of average yield pertaining to various food grain crops along with estimates of their percentage standard errors were worked out. It was observed that estimates of average yield for the two major crops viz. wheat and paddy have been obtained with suitable degree of precision, however, for minor crops like maize, barley, jowar, ragi etc., these were obtained with higher percentage standard errors. Sample sizes at the primary stage of sampling i.e. Villages have been worked out for estimation of average yield of different crops for different levels of margin of errors.
Data pertaining to area for different crops pertaining to different districts (20 survey numbers in a village) of some States under ICS scheme were obtained from NSSO and analysed. Estimates of total area under different crops were obtained with estimates of very high percentage standard errors. However, these have also been worked out, had the sample size would have been increased from 20 survey numbers to 100 survey numbers. In that case, the percentage standard errors decreased significantly for number of crops.
The main finding in this study was that on the basis of smaller sample sizes, the estimates of average yields of two major crops i.e. paddy and wheat may be estimated with suitable degree of precision in those States in which these crops are being grown as major crop. However, for other crops, these sample sizes may not be adequate for estimating the average yields of these crops with suitable degree of precision. For estimating these with permissible margin of errors, the sample sizes for these crops would have to be increased suitably.