17. The organiser of this session was Mr. F. Vogel, Deputy Administrator, National Agricultural Statistics Service, United States Department of Agriculture (USDA); and the discussant was Mr. A. D. Marshall, Senior Officer, FAO Statistics Division.

 

18. Three papers were presented:

 

(i) “China Agricultural Census Design and Data Collection: Issues and Lessons” (L. Biggeri);

(ii) “Data Processing and Dissemination” (A. Giusti and Li Wenhai);

(iii) “A Concise Analysis of the Main Results of the First National Agricultural Census in China” (Zhu Xiangdong).

 

19. The most interesting points made in the first paper were about the statistical aspects of the census. One issue was that the preparation of a census programme was a continuing scientific experiment. The paper provided an excellent illustration of the careful planning and preparation required to conduct a census and pointed out the main reasons for its success and, above all, the problems met and the solutions adopted. It also presented many recommendations for future agricultural census and surveys, stressing the importance of updating the knowledge of the new methodologies and technologies for the collection of data in the strategic planning of all the agricultural surveys. To this end, further research and analysis on agricultural census data were suggested. It was recommended that a more detailed version of this paper be developed as a case study that other countries could use as a guide.

 

20. The second paper illustrated the most important tasks performed in the field of data processing in the census, taking into account that the amount of raw data from the census was up to 210 GB. The main strategies for electronic data dissemination were also presented. Some important issues about the use of Optical Character Readers (OCR) were raised. One was that careful consideration has to be given to the questionnaire design. In addition, considerable training of interviewers is required to ensure the data are recorded accurately for imaging. It was noted that PRC was probably the only country to have used OCR on such a scale, and so there was probably much to be learned from this experience. An interesting feature of the dissemination policy raised was that a one percent sample of the household form had been prepared and made available for analysis through the Internet.

 

21. The third paper provided a good overview of the highlights of the results of the census and illustrated the wide scope of the census. The paper also discussed the findings, noting that the area of cultivated land had been under-estimated and livestock inventories over-estimated in the official statistics. A significant point was that some crop yields would be revised downwards by as much as 20 percent when crop areas were increased. The area per capita had not been affected by the census results, because the quality of land would affect the results in terms of output. The paper also discussed the significance of the labor force information that was available for the first time. This will allow good policy decisions to be made, for example, about what to do with the surplus labor force, estimated to be about 150 million persons.

 

22. There were several questions from the participants about the implications of the large increase in cultivated area shown by the census and whether historic data series would be revised. A related question was whether the new information would be integrated into the Ministry of Agriculture’s data series. It was noted that this had already been done. A comment was made that the one percent sample was over 2 million records, and it was questioned whether the sample needed to be that large, particularly for analysis at the national level. It was noted that this sample size was necessary to provide estimates at provincial and prefecture levels. Another comment was made about other countries having problems with OCR, and training interviewers to accurately record the data for scanning. Finally, in response to a question about the kind of users who would be involved in the census analysis, it was noted that both policy makers and researchers had been involved in determining the content of the census and were working on the results, at national and international levels.