Main steps in developing and implementing the census of agriculture Regional Workshop for Monitoring the SDGs related to Food and Agriculture Sector and on the WCA 2020 Nadi, Fiji 6-10 November 2017 Main steps in developing and implementing the census of agriculture Technical Session 8 Jairo Castano Senior Statistician Leader, Agricultural Census and Survey Team FAO Statistics Division

Contents Planning and preparation of the agri-census Field work Role of the census in the integrated system, strategy and objectives Methodology Work plan & budget Census legislation, census office & staff Publicity campaign Data quality framework Frames & mapping Tabulation plan Questionnaires (design, pre-test, pilot, fine-tuning) & manuals Data processing system Field work Organization of field work, staff and training Census enumeration PES Processing, analysis, publication Data processing, tabulation, archiving Analysis, reports, dissemination Dara reconciliation

Steps in developing and conducting the AC 1. Identify the role of the AC in the system of integrated agricultural censuses and surveys 2. Determine overall strategy for and objectives of the AC 3. Define the methodology 4. Develop work plan and budget for developing and carrying out the census 5. Prepare census legislation, if required 6. Create the Agricultural Census Office and recruit staff 7. Develop and implement the census publicity campaign 8. Design data quality assurance framework 9. Prepare frames and mapping 10. Develop the tabulation plan 11. Design, pre-test, pilot, fine-tune questionnaires; Prepare field instruction manuals 12. Design/test the computer processing system, incl.data entry, editing and tabulation 13. Organize field work (including monitoring system); recruit and train field staff 14. Census enumeration 15. Post enumeration survey (PES) 16. Data processing, tabulation and archiving 17. Analysis and prepare census reports, disseminate and publicize results 18. Reconcile current statistics with census data Design & planning Field work Processing, analysis, publication

1. Identify the role of the AC as a component of the system of integrated agricultural censuses and surveys The census of agriculture must not be carried out in isolation but as a component of an integrated system of agricultural censuses and surveys Development of the system of agricultural censuses and surveys within the National Statistical System (NSS), in line with the National Strategy for the Development of Statistics (NSDS) A comprehensive multi-year census-survey programme ensures efficient use of available resources, standard definitions/classifications and avoids duplication of data collection or releasing conflicting statistics.

2A. Define the specific objectives of the AC Specific objectives - to provide key data (mainly on the structure of agriculture): At the national level and for small administrative units/geographic areas; and at community level if needed. as benchmarks for and reconciliation of current agricultural statistics To provide frames for agricultural sampling surveys and data needed for registers of agricultural holders To collect some additional data on non-agricultural production households (in a widened AC).

2B. Define the overall strategy to implement the AC Relationship to other censuses: linked to or jointly with pop or economic censuses Census scope: Crop and livestock; and Aquaculture, and/or Forestry, and/or capture fisheries Census coverage: holdings in the household sector holdings in the non-household sector Non-agricultural production households in rural areas National or regional geographic coverage Minimum thresholds . Way to implement the census: Methodology (see next slide)

3. Define the methodology Define the statistical unit (e.g. agric. holding) Reference period (census reference year) Approach: classical, modular, integrated census- surveys, use of registers Enumeration period(s) and number of phases Design collection: Establishment of census frame (including listing operation) Use of maps Data collection method (e.g. PAPI, CAPI, etc.) Sample design (if relevant) Fine-tuning of methodology based on pilot census results Design of Post-enumeration Survey (PES)

4. Develop work plan and budget Elaboration of detailed work plan with the different stages of implementation clearly established. Development of detailed budget where different components of spending are estimated and time-table of expenditures and funding. All census operations including planning, cartography, enumeration, processing, PES, analysis and dissemination carefully budgeted from the beginning and efforts made to mobilize the required funds. Establishment of a monitoring system and periodic review of work plan, budget and expenditure control. Outsourcing of some activities under Census Office’s (CO) supervision: Layout and printing of questionnaires, other census material; reports and other printed census dissemination products Packing and delivery of census materials and equipment Census mapping Publicity campaign Data collection, processing and dissemination systems

5. Preparation of census legislation One of the first aspects to be considered when starting to plan the AC. One of the most important instruments for facilitating the census work. If a country lacks an appropriate legal framework for taking periodic censuses, it is important to act early to establish it. For the census legislation, subject matter specialists from the census agency should work closely with admin officers and legal experts in order to ensure coherence of census acts with other relevant legislation. In general, legal provisions are required for: establishing and regulating primary admin responsibility, e.g. agency responsible for the census organization; obtaining the necessary funds; determining the general scope and timing of the census; placing a legal obligation upon the respondents to cooperate and provide truthful information; placing a legal obligation upon the census agency/personnel to ensure confidentiality and protection of holders’ information, etc.

6. Census Office & Institutional framework The census agency should be specified in the census legislation. Ensure strong political support and efficient coordination with other agencies. Create the AC Office (central & provincial) and recruit necessary staff. Establish a National Census Committee and subcommittees to oversee and coordinate AC activities.

7. Develop and implement the census publicity campaign A well-planned census communication and publicity is essential to sensitize the public about the purpose of the census of agriculture and, ultimately, to ensure the cooperation of holders to provide complete and accurate data. A communication and publicity strategy involves: Situation analysis to identify opportunities or issues that need to be taken into account Definition of the target audiences (who) Messages to be communicated, logos & slogans (what) Communication channels and promotional tools (how) Implementing the strategy (when)

8. Design data quality assurance framework The main goal of quality assurance for the CA is to prevent and minimize potential errors at design stage and detect errors as soon as possible so that timely remedial actions can be taken even as the census operations continue. A quality assurance framework aims at achieving an appropriate balance between the needs of users, costs, respondent burden, and the various dimensions of quality. The six quality dimensions of statistical data (or outputs): Relevance Accuracy and reliability Timeliness and punctuality in disseminating results Accessibility Coherence and comparability of statistics Accessibility and Clarity/Interpretability. Quality management of a census of agriculture should be comprehensive and should cover all activities including planning, development, data collection, processing, evaluation and dissemination of census results. The main techniques applied during data collection and data processing phase: Complete verification Sample verification: acceptance sampling or statistical process control

9A. Prepare the AC frames Sources of information for the construction of frames The statistical data and maps of the EAs of the most recent PHC and a list of EAs along with number of households/housing units. The statistical data and maps of the EAs of the latest AC including a list of EAs with their number of holdings. Statistical farm registers, cadastral and other cartographic materials, other statistical and admin data sources, which include listings of holdings and/or holders' addresses or their approximate location. When lacking reliable information, prepare census frame by conducting a listing operation (prior to starting the census enumeration).

9B. Prepare maps for census field operations Cartographic work should be conducted during early stages of preparatory census activities. It should start with the inventory of available geospatial information and maps and the evaluation of their suitability for AC purposes. B Types of maps: topographic sheets, other government maps, satellite images, aerial photography, communication maps, land-use maps, economic maps, city and tourist maps, EA maps from PHC. A large proportion of the cartographic preparations for an AC consists of delineating and identifying the enumerators' areas (EAs) of work. GIS provides computer-based design of EAs and significant automation of map production tasks. In many countries EA maps are prepared as part of the cartographic work of the PHC. In some countries the cartographic work is conducted for both PHC and AC (e.g. Brazil). Adapting, revising and updating available maps to census requirements.

10. Develop the tabulation plan The tabulation plan is a set of prototypes of statistical tables (dummy tables) prepared to present the main census results, based on users’ primary needs. Should be based on user consultation. Needs to be undertaken at early stages of census preparation to define census content and developing the census questionnaires. The preparation of the tabulation plan is an iterative process: census questionnaire and methodology are conditioned by the data to be tabulated, and vice-versa.

11A. Design and test questionnaires Design supported by a working group and users-producers consultations (to ensure ownership) Design and characteristics of the census questionnaires depends on: the census approach/modality the type of holding (household sector, non-household sector, community level) method of data collection: PAPI (face-to-face interviews or mailed) or electronic questionnaires (CAPI or CASI) Pre-test of questionnaire(s), revise and pilot test

11A. Design and test questionnaires (cont’d) Aspects to be taken into account Concepts and definitions in questionnaire harmonized with other agricultural statistics programmes. Questions, concepts and definitions used in the questionnaire should be easily understood by the holder and census field staff. Reasonable size (length). Different languages in multi-lingual countries.

11B. Prepare instruction manuals Manuals establish criteria and procedures for supervisors and enumerators, roles and the work expected to be carried out during the census. They pursue quality of work and standard procedures at all levels Purposes: - training - field reference: enumerators’ manual supervisors’ manual provincial coordinators’ manual. Other manuals are for training, listing operation, key-punchers, data editing, data cleaning, PES, etc. Formats Printed manuals On-line manuals Manuals included in mobile devices or Any combination of them.

11C. Census testing Questionnaire pre-tests Pilot Census Small scale Table 14.1 Census testing 11C. Census testing Questionnaire pre-tests Pilot Census Small scale Test the suitability of the intended census questions, including their formulation and the instructions provided questionnaire design Test with holdings, including special holdings Estimate time requirements in enumeration Several rounds may be done Test the data collection methods (e.g. paper, tablets) Larger scale Test the entire census infrastructure Cover one or more sizeable administrative divisions Test all stages of a census: preparatory enumeration, processing and dissemination Best if conditions in the pilot census are close to the conditions present during the actual enumeration Ideally conducted exactly one year before the planned census, if not at least 6 months before Pilot census data no usable substantive data. But analysis of data errors may be informative for identifying problems

12. Design and test the computer processing system Considerable time is required to design computer programmes (CPs) for error identification, automatic error correction, tabulation, calculation of sampling errors (when relevant), etc. Initial test of CPs using questionnaires with artificial data. CPs should be tested with data from pre-tests and/or pilot census. CP should be tested, normally by verifying results of both error detection and tabulations for a group of 100‑500 questionnaires. Data used for such tests should be tabulated manually to check each item or its classification in the tabulations. Useful to enter erroneous data to test the full range of error detection Tabulation process to be simulated during the test. Data transfer to be tested during the pilot census (for CAPI, CATI, CASI).

13A. Organization of field work and field staff Staff play a critical role in AC and duties should be clearly defined. Advisable that supervisors and enumerators live in places of work, are familiar with local conditions and able to communicate easily with holders. Extension assistants or field officers of the Ministry of Agriculture could be a good source for enumerators or supervisors. Usually, not more than 100-200 holdings should be assigned to an enumerator, and only 20-50 if objective measurements are made. Commonly, one supervisor effectively supervises 5 to 10 enumerators Different ways of organizing enumerators’ field work: the supervisor gives certain number of EAs to each enumerator - used mainly when enumerators with local knowledge are recruited and can work alone in their EAs a small team of supervisor and enumerators work together in all assigned EAs to the supervisors; reasonable in remote areas with poor transport and communication facilities.

13B. Training The organization and delivery of training courses should be included in the planning and budgeting stages of census preparation. Diversity of training programmes/courses: training of trainers, training of supervisors, enumerators, office staff (e.g. programmers and system analysts, editors, coders, computer operators). Establishment of training venues all around the country. Content and timing of training programmes should be appropriate for the level of personnel to be trained. Developing proper training material for the training classes: training manuals, aids (audio-visual aids) and new multi-media technologies. Selection of appropriate training techniques. Uniform delivery of training in all venues. Assessment of trainees. Census personnel should be screened and a final selection made on the basis of a written examination plus an interview.

13B. Training (cont’d) The training can from up-to-down in a “cascade” manner: in first place, high-level census organizers responsible for census organization and administration; main technical staff at the central census office is charged with the training of trainers for different type of tasks (training for census enumeration and supervision, training of office staff (programmers, coders, data-entry operators or operators of data recording equipment, cartographers); training for using advanced technologies; the staff so trained will deliver the training courses down to provincial coordinators and office staff; provincial coordinators so trained have to deliver the training to field supervisors; field supervisors must train the field enumerators.

13A. Organization of field work (cont’d) All questionnaires should be recorded irrespective of the outcome of the interview (completed, no respondent, non-existing, etc.) using a specific code, to ensure that all holdings within each EA are accounted for. Set up monitoring mechanism, including periodic reports with key data regarding census coverage to the Census Office to enable timely actions for critical problems. For PAPI, special control measures on questionnaires flows are needed : Completed questionnaires returned by enumerators through supervisors to processing center, grouped by geographical areas and properly filed. During processing: questionnaires are removed from storage for manual coding and editing, data entry and verification. Very rigid control over this flow is needed; periodic reviews in order to detect misplaced questionnaires; An adequate physical storage space is also required. When CAPI or CASI is used, the activities related to monitoring questionnaires, data entry, editing and coding are done during the enumeration.

14. Census enumeration Rigorous procedure for census enumeration should be established with clear responsibilities of enumerators, supervisors and other field staff. Enumerators are responsible for accurately recording all required information on the agricultural holdings in an assigned area and reporting the progress to their supervisors. Supervisors responsible for performing quality checks according to the census’ quality assurance plan on enumerators’ work. Monitoring enumeration on a daily basis: information regarding AC coverage (numbers, %), average time taken, problems faced. Reports on the outcome of field data collection to the Census Office (CO) to solve critical problems. Enumerators’ kits according to data collection method (PAPI, CAPI).

15. Post-enumeration survey (PES) Timing: it should be carried out soon after the census enumeration is completed. In the cases of the modular and integrated sample/survey modalities, soon after the core modules. New listing: agricultural holdings must be listed again in sampled areas (e.g. EAs). Data collection: on key selected census variables. It should attempt the use of physical measurement of area and actual count of livestock and trees. Reference period: must be the same as for the census enumeration. Objective: to assess the magnitude of non-sampling errors in terms of i) coverage errors and ii) content errors (quality of census data collected). Plan: should be carefully planned and synchronized with the overall planning. Staff: the best supervisors and enumerators assigned to other EAs to ensure the best quality of data. Design: a sample survey to be conducted independently from the AC enumeration. Further info: Reader is referred to the UN PES Operational Guidelines (2010).

16. Data processing The ICT strategy for the census should be part of the overall AC strategy. Hardware requirements Main characteristics of data processing: Large amount of data to be entered in a short time with multi-users processing in parallel Large amount of data storage required Relatively large numbers of tables to be prepared Extensive use of raw data files which need to be used simultaneously. Method of data capture chosen by the census office Basic hardware equipment: Many data entry devices (PCs, hand-held devices, depending on data collection mode) Central processor/server and networks Fast, high-resolution graphics printers Number of PCs /handheld devices to be carefully considered Software Preferable to use standard software maintained by the manufacturer with available documentation (to allow for data portability).

16. Data processing (cont’d) Data coding Data entry: manual, optical scanning, hand- held devices, CASI and CATI. Data editing and imputation: manual, automated Validation Tabulation Calculation of sampling error.

16. Data processing (cont’d) Examples of data scanning and computer assisted systems use for WCA 2010 round (FAO, Metadata reports: http://www.fao.org/economic/ess/ess-wca/en/ ) TECHNOLOGY Country Optical character recognition (OCR) Albania, Czech Republic, Greece, Ireland, Malawi, Norway, Philippines, Sweden Intelligent character recognition (ICR) Tanzania, Canada, Cook Islands CAPI Argentina, Colombia, Cote d’Ivoire, Finland, France, French Guyana, Iran, Jordan, Lithuania, Malta, Namibia, Martinique, Mozambique, Slovenia, Thailand, Venezuela PDA Mexico, Brazil CAWI, CATI, CAPI combined Australia, Austria, Canada, Estonia, Finland, Iceland, Italy, Latvia, The Netherlands, Poland, Spain, Sweden, USA

16. Archiving It is important to physically secure census data. Important for wider use or reuse of census data, time series, historical analysis and justification of the high cost of the census. Backup copies of data. Guarding against hardware and software obsolescence, such as outmoded floppy disks and unreadable file formats, as well as physical threats, such as natural disasters, theft or sabotage. Data types include: census microdata, final published aggregate data, transitory data file, unstructured data (documentation, records of decision, work plan, budget, manuals, questionnaires, etc.). These data types can be archived and may have different archiving, retention periods and metadata requirements.

17. Analysis, census reports and dissemination Design the dissemination plan during census preparation Census reports: On preliminary results On final results On quality evaluation of census results (including results of PES) Analytical/Thematic reports Technical report (census operation, methodology, concepts and definitions) Other census products: Brochures and flyers Atlases and other geographic products Methods and tools for dissemination (printed, digital, online, social nets) Providing access to census databases Providing public access to microdata Dissemination workshops (national, regional)

18. Reconcile data from the system of current statistics with census data The existence of discrepancies between new census results and the previous published estimates from agricultural surveys performed during the inter-censal period is a common issue. Important gaps may be observed when comparing indicators like crops area and production, agricultural population or quantity of fertilisers. Reconciliation of census and surveys data consists in comparing estimates from previous surveys and the new census results regarding a number of important agricultural indicators and correcting the discrepancies between them.

18. Reconcile current statistics with census data (cont’d)

Thank you