Insights into the World Programme for the Census of Agriculture 2020

Regional Roundtable on World Programme for the Census of Agriculture 2020 Port of Spain, Trinidad and Tobago 22-26 May 2017 The Census of Agriculture 2016 in Canada: Key Methodological Considerations Technical Session 04 Etienne Saint-Pierre Statistics Canada etienne.saint-pierre@canada.ca 1

Background and Scope A separate Census of Agriculture began in 1896 with the first mid-decade Census of Agriculture in 1956 Classical approach as defined by the FAO s WCA 2020 Legal basis: Collected under the Authority of the Statistics Act Unique source for: small/custom geographic areas enumerating rare and emerging commodities Provides: frame information for the survey program comprehensive farm data on commodities, farm finances and farm practices and socio- economic data on operators and their families Meet users needs for agricultural and environmental programs, health programs, trade and crisis management (floods, disease outbreaks) Meet international commitments such as providing accurate measures for international comparability to the FAO and OECD 2

Integration with the Agriculture Statistics Program 3

Background and Scope (3) 4

Background and Scope (4) Databases created from the Census of Agriculture: Farm database Agricultural operation level Farm database Census Geographic Components level Operator database Agriculture-Population Linkage database 5

Integration with Census of Population Conducted simultaneously in May Timing not optimal for farmers (seeding season) but Reduction of costs Improve quality (coverage and response) Projects conducted with CEPOP: Public communications Census Help Line Questionnaire printing Mail-out, Data reception, Capture, Internet response AG-POP linkage Projects conducted independently: Content determination Census of Agriculture frame management Follow-up (Non-response & failed edit) Data processing 6

Methodological Considerations: Target Population All agricultural operations in Canada that produce any agricultural products with the intention to sell. Farms with very low farm revenues commonly called "hobby" farms are included as long as the agricultural products produced are intended for sale. 7

Methodological Considerations: Content determination Content consultation Consultation workshops across the country All requests for content require a written submission Content testing Qualifying content suggestions are tested with farmers using One on one interviews Focus Group Conducted by a group of experts Farm operators are selected Cabinet approval 8

Methodological Considerations: Frame From the Farm Register to Business Register (BR) New unknown population Frame coming from: Statistics Canada s Business Register Updated with signals from regular production and financial surveys Tax data Agricultural Frame Update Survey Multiple sources of administrative data Self identification on the Census of Population Sample from unknown population Population drawn from the BR in February for May collection 9

Methodological Considerations: Data Collection Use of Remote Collection Methods only No in-field enumeration Invitation letters with secure code to access electronic questionnaires (CASI) Mail-out / Mail-back Census Help line CATI for Follow-ups Collection done at the Headquarters and 5 Regional offices CASI and CATI applications are the same; only 1 questionnaire is used for all units in the Census Early collection for complex operations (3 months prior to the regular collection) Use of administrative data to prioritise follow-ups Regular Collection completed in 4 months 10

Methodological Considerations: Data Processing Data Reception and Capture Match / Unduplication Geo-edit Assigning a geo code to the farm HQ address provided Web mapping application for 2021 Error detection Micro-Edits Will be fully automated in 2021 Imputation Minimum content Donor imputation Use of integrated admin. and survey data in 2021 11

Methodological Considerations: Validation Records stratified based upon contribution to estimates size by geography, industry or by operation type Sort records based upon coherence with referential and administrative data sets using a score function Attempt is made to validate records by section of the questionnaire Place data validation calls to respondents for impacting results Clean the imputation donor pool

Methodological Considerations: Analysis Review of socio-economic context, outlook and development of anticipation of the results Consultation with industry experts Validation with numerous administrative data sources and Earth observation data Comparison with other production / commodities surveys Coherence with Supply-Use model Results for each variable and each province must be presented and formally approved by the Census of Agriculture Certification Committee 13

Methodological Considerations: Dissemination All data are disseminated exactly one year after Census day (available for free in StatCan s CANSIM database) Thematic maps and articles Provincial highlights Road shows across the country Custom Requests Additional products released later; Linkages with Census of Population Linkages with Tax data Agriculture at a glance-in depth analytical articles using Census data. 14

Thank you 1