Home Agricultural Data Concepts Table of concepts
 

Canada

Mexico

USA

Agricultural Units

Operation: A single business which produces some agricultural products for sale.
Operator: Person responsible for the day to day decisions made in/on the operation.

Ejido: Area of land, forest or water from which agricultural products are produced.
Small property: Area of land, forest or water from which agricultural products are produced.
Ejidatario: A farmer holding an "ejido" and recognized as such by the assembly of the ejido.
Proprietor: Owner of a small property

Farm: Any place which produces and sold (or potentially could sell) $1,000 worth of agricultural products during the year.
Farmer/Operator: The person responsible for the day to day management decisions of the farm.

Statistics Programs

Crop Surveys:

Crop Surveys:

Crop Surveys:

 

Crop Surveys: six times per year
Annual Fruit and Vegetable Survey:
Annual Greenhouse, Sod and Nursery Survey
Potato Area and Yield Survey: in June and November

Agriculture Information (SIACAP)
- includes planted area, harvested area
- observed and estimated yields
Regular probabilistic surveys on corn, beans and sorghum to obtain production area and prices
Monthly, annual and agriculture year information of the main products, areas and prices.

Crop/Stocks Surveys: quarterly
Agricultural Yield Surveys: monthly from May through November
Objective Yield Surveys: for major crops annually in major states
Floriculture Survey: conducted annually
Chemical Use Surveys: annually (suspended in 2013), fruit/vegetable alternating years

 

Livestock Surveys:

Livestock Surveys:

Livestock Surveys:

 

Livestock Surveys: in January and July
Annual Honey Survey
Production of Poultry and Eggs
Annual Aquaculture Survey

Livestock Information (SIPCAP)
- includes inventory, weight and prices
Administrative records and visits by experts in large areas
Monthly, annual and agriculture year information of production and prices, complemented with information from slaughter houses and producers associations

Cattle, Sheep and Goat Inventory
Cattle on Feed Survey
Milk Production and Dairy Products
Hog Surveys
Bee and Honey Inventory and Production
Broiler and Turkey Hatchery Surveys
Chicken and Egg Survey
Trout and Catfish Surveys (suspended in 2013)

 

Economic Surveys:

Economics Surveys:

Economic Surveys:

 

Farm Financial Survey: every two years
Farm Environment Management Survey: every five years
Agricultural Water Use Survey: every two years

Estimates are derived from the SIACAP and the SIPCAP
Administrative labor and production cost, technical coefficients and production cost of the main products by agriculture year.
Probabilistic surveys of prices, production area and technical coefficients of the main products by agricultural year

Agricultural Resource Management Survey (ARMS): conducted in three phases
Cash Rents: annual, July
Agricultural Labor Survey: semi-annual, starting in March
Dairy Product Prices: weekly
Prices Paid and Received Surveys: monthly and annually

 

List Survey:

List Survey:

List Survey:

 

Farm Update Survey (FUS)

Just the sampled units are updated

Criteria Survey
Agricultural Identification Survey: 2 years sequentially prior to Census year

 

Census:

Census:

Census:

 

5 year cycle

10 year cycle

5 year cycle

 

Administrative Data:

Administrative Data:

Administrative Data:

 

Used extensively to reduce cost and respondent burden when possible in place of agricultural surveys.

Used extensively for estimation of planted acreage and livestock inventory.

Used to supplement data collection and reduce respondent burden and survey cost.

Processing Cycle

 

Quinquennial Census and five major surveys conducted at least annually.

Census conducted every 10 years, administrative registers, seasonal surveys.

Quinquennial Census and multiple sample surveys conducted annually, quarterly, monthly, weekly, bi-annually and semi-annually.

Frame definition and exclusion:

Frame definition and exclusion:

Frame definition and exclusions:

Census targets every household in Canada
Surveys use the Business Register as a list frame
Sampling frames generally exclude farms with less than $10,000 CAN in sales.
Census offers the main source of Business Register updates

Sample surveys based on a list frame including active units from the Census, administrative registers and producer lists

The main sampling frame is the List Frame composed of known farms
The Area Frame is composed of all land in the US divided into segments
Multiple frame sampling design utilizes both frames for efficiency and completeness.

 

Sample Design and Selection:

Sample Design and Selection:

Sample Design and Selection:

 

Population generally stratified by geography, size and farm type
Smallest units placed in take-none stratum - not sampled
Univariate or multivariate sample allocation
Permanent random numbers used to control sample overlap

Population stratified by geography, production and crop performance
Multivariate allocation used to meet prescribed precision
Random numbers used to control sample overlap
Additional samples may be drawn from important populations

Population generally stratified by geography, size and farm type                       
Census performed for small target populations
Multivariate probability proportionate to size are used for surveys with multiple commodities.
Permanent random numbers used to control sample overlap

 

Collection Steps:

Collection Steps:

Collection Steps:

 

Census conducted via paper and internet
Surveys conducted primarily with CATI

Census conducted via paper
Surveys conducted via paper
Information collected on site by PDA

Census: paper, electronic (web-based), and CATI based enumeration
Major Surveys: paper (both mail and personal interview), electronic & CATI
Area Frame Surveys: personal enumeration only
ARMS: primarily personal enumeration, small mail version for select few

 

Data Editing:

Data Editing:

Data Editing:

 

Non-response edits
Historical edits
Consistency edits
Statistical Edits
Outlier detection methods
All records go through automated edits
Only larger units go through manual edits

Manual validation of data
Electronic edits
Critical fields
Internal consistency
Outlier detection methods

Non-response edits
Historical edits
Consistency edits
Statistical Edits
Outlier detection methods
All records go through automated edits
Larger surveys and some small surveys go through manual edits

 

Data Imputation:

Data Imputation:

Data Imputation:

 

Erroneous fields which fail edit are flagged for imputation
If the most important fields fail, record unusable
Useable records are imputed with deterministic, historical, or donor methods

Verify information with person that captured the information
Administrative data is used extensively
SIACAP and SIPCAP

Depends on survey design
A re-weighted approach is typically used to adjust for non-response
Imputation for missing data is typically done with the mean from positive reports
Census utilizes a "nearest neighbor" (donor) approach to imputation

 

Estimation:

Estimation:

Estimation:

 

Horvitz-Thompson estimator
Non-responses/unusable records are compensated through adjustments of estimation weights.
Variance and CV produced from statistical theory

Horvitz-Thompson estimator
Non-responses/unusable records are compensated through adjustments of estimation weights.
Major commodities: National estimates first, then State levels
Variance and CV produced from statistical theory

Estimator type varies depending on survey:                                               Major commodities: National estimates first, then State levels
Survey computations and estimation procedures vary depending on survey design (probability vs. non-probability, list frame only, area frame only, or both)
Variance and CV produced from statistical theory.

 

Disclosure Control:

Disclosure Control:

Disclosure Control:

 

Estimates will not be published if (a) the estimate applies to fewer then X operations or (b) the data for a single operation represent more than Y percent of the estimate.
Automated methods used to identify sensitive estimates, manual methods used for complementary suppression

Estimates will not be published if (a) the estimate applies to fewer then X operations or (b) the data for a single operation represent more than Y percent of the estimate.
Automated methods used to identify sensitive estimates, manual methods used for complementary suppression

Estimates will not be published if (a) the estimate applies to fewer than X operations or (b) the data for a single operation represent more than Y percent of the estimate.                                                                              Automated methods used to identify both sensitive estimates and perform complementary suppression.

 

Frame Updates:

Frame Updates:

Frame Updates:

 

Updates to the Business Register use survey feedback.  Some updates are survey specific, others are across all surveys.
The Farm Update Survey uses tax records and administrative data to identify new farms to be birthed on survey frames.

Updates to the List frame are done on a continuous basis as part of standard survey procedure.
The agriculture and livestock census outcome and income will be used to design the list surveys

Updates to the List and Area frames are done on a continuous basis as part of standard survey procedure.  Updates are typically completed following completion of a survey.

Adjustment Methods

Coverage:

Coverage:

Coverage:

All farm operations in Canada as defined by the Statistics Canada Business Register

Agriculture, Livestock, Aquaculture and Fishing Sectors

All farm operations in the U.S. that qualify as a farm based on the USDA farm definition.

 

Coverage Adjustment:

Coverage Adjustment:

Coverage Adjustment:

 

The Business Register is based on the quinquennial Census of Agriculture
Surveys benchmark to Census counts, updated with any birth records identified from Farm Update Survey or ongoing Business Register updates
An estimated Census undercoverage rate is calculated. The undercoverage is not accounted for in any benchmarks used for weighting

None, there is no exclusion.
Administrative information is used for the records in the non sampled areas for surveys

NASS maintains a list of farmers and ranches from which the census mail list (CML) is compiled. An area frame sample is used to account for farming operations not on the mail list (NML). This dual frame (CML + NML) accounts for all farm operations in the U.S.
Surveys for most major commodities as well as economic surveys utilize the dual frame approach.
The CML is adjusted for nonresponse (CML+NR) and combined with the summarized NML to produced state and national totals. An allocation process called "calibration" is used to properly distribute the dual frame estimates across states.

 

Nonresponse Adjustment

Nonresponse Adjustment

Nonresponse Adjustment

 

Generally non-response weight adjustments are used
Large non-responding units in some surveys may be manually imputed using historical and trend information
Partial non-respondents may be imputed to create a full record

Generally non-response weight adjustments are done using administrative information

Adjustments are made for both whole-farm nonresponse and item nonresponse.
Large nonresponding units may be manually imputed using historical information or reports from similar operations.

 

Other Adjustments for Concepts:

Other Adjustments for Concepts:

Other Adjustments for Concepts:

 

Farms excluded from survey collection (generally less than $10,000 CDN) are usually accounted for through benchmarking or modelling

None

To further improve coverage adjustment, a second set of variables known as commodity coverage targets are added to the calibration algorithm. These targets are commodity totals which can be validated and are strongly supported by administrative sources. The targets assure that commodity totals remain within reasonable bounds of established benchmarks.

Quality Assurance and Indicators

Quality Assurance Practices:

Quality Assurance Practices:

Quality Assurance Practices:

Questionnaires reviewed by design experts
Frames updated through survey feedback, the Farm Update Survey and Census of Agriculture
Interviewers specialized in agriculture surveys used for collection
Computer applications have built-in edits to identify inconsistencies
Processing system allows analysts to review and compare data with previous occurrences
Well-accepted methods for sampling and estimation used

Practices depend on the survey
Internal consistency and filters used to produce coherent data

Questionnaires designed and tested by Survey Methodologist.                 
Survey frames are updated continuously based on feedback from surveys, census and other approved list sources.                                           
Interviewers are trained for each survey.                                                  
BLAISE and CAPI data collection instruments include interactive edits to identify inconsistencies.                                                                 
Processing system using Statistical Analysis Software (SAS) to further edit and summarized data items.                                                                             
A SAS based Interactive Data Analysis tool is  utilized to facilitate data review and analysis.

 

Data Quality Indicators:

Data Quality Indicators:

Data Quality Indicators:

 

Response rates, both weighted and unweighted
Reasons for non-response
Census Coverage Evaluation Survey
Coefficients of variation calculated and compared to targets

Response rates, both weighted and unweighted
Reasons for non-response
Coefficients of variation calculated and compared to targets

Response rates, both weighted and unweighted                                              
Reasons for non-response                                                                                
Census coverage evaluation survey                                                                
Coefficients of variation calculated and compared to targets                                                      

Last Updated on Wednesday, 11 September 2013 10:20
 

Disclaimer

The information is presented as it is, and Statistics Canada, the National Agricultural Statistics Department of Agriculture of the United States, the Agricultural and Fisheries Information Service (SIAP) and the National Institute of Statics and Geography (INEGI), don’t offer any guarantees, express or implied, including but not limited to guarantees of marketing and adapt to a particular purpose. In any case the above agencies shall not be liable for any direct, special, indirect, consequential or other caused by the use of the information provided.