23. The organizer of this session was Mr. I. P. David, Assistant Chief Economist, Asian Development Bank; and the discussant was  Mr. álvaro González-Villalobos, Senior Officer, FAO Statistics Division.

 

24. Four papers were presented:

(i) “Data Quality Control in Different Phases of the Agricultural Census” (Zheng Zhihao);

(ii) “The Post Enumeration Survey of the Chinese National Agricultural Census” (L. Fabbris);

(iii) “Reconciling Official Statistics with the Agricultural Census” (F. Vogel);

(iv) “Reconciliation of the Current Agricultural Statistics with the Results from the Agricultural Census” (Zeng Yuping).

 

25. The first two papers paper described the quality control measures employed at key stages of the census, from planning, organization, questionnaire design, recruitment and training of census staff, pre-tests, publicity, field operations, to data processing. Among other things, the paper clarified the work done for the preparation of the list for the census using the list from the Population Census and from a pre--enumeration of the rural households and holdings. The conclusion was that the powerful leadership and scientific program ensured the success of data quality control in the census.

 

26. The characteristics and results of the Post Enumeration Survey (PES), conducted one month after the census enumeration, were presented. With 40,000 households, the PES would qualify as a large-scale sample survey in any country other than PRC. The measurement of interviewers′ errors and bias had been the main objectives. It was pointed out that a thorough analysis of the PES vis-à-vis the census had been undertaken, concluding that, broadly, the quality targets for the census data had been met. The results also had identified certain characteristics whose estimates should be released only at higher territorial levels, and had made suggestions for future activities.

 

27. The results of the census showed that there were discrepancies between the census and current statistics data. This problem is not unique to PRC, and the third paper described some of the discrepancies between US official statistics and census results and the ongoing efforts to reconcile them. Some discrepancies between Chinese current statistics and census results were described, foremost of which is on total cultivated area and livestock inventory. It was noted that information on cultivated area from the household questionnaires could be used as a basis for analyzing the reasons for the discrepancies. The paper concluded by suggesting some guideline principles for the future, especially to review the methodology used to prepare the current official statistics.

 

28. The fourth paper mainly described the issues on reconciliation of the current agricultural statistics with the results from the Agricultural Census in PRC. After a lot of testing, NBS was sure that the census data published were more reliable than that coming from the current statistics. On this basis, the indicators and methods to be used for the reconciliation of the current statistics to the census data were specified.

 

29. The discussions focused on the very strong component of the data quality control throughout all the census operation. Comments were also made on the design of the PES (selection of places, and not of households) and on the indicators used for the reconciliation of data, particularly for the sown area. There was also some discussion on the use of a modeling approach to adjust current official cropland statistics to the new sown area data as measured by the census.