Census of agriculture

Name: Census of agriculture
Topic: Agriculture, forestry, hunting and fishing

Division for Housing, Property, Spatial and Agricultural Statistics

Municipality. Other regions at lower level will be considered if it does not get into conflict with confidensiality.

Complete censuses have been conducted every tenth year since the beginning of the 20th century. All agriculture properties and agricultural holdings are included in this years census.

Information is forwarded to EUs statistical office, Eurostat in accordance with specifications given in the i regulation. Results will also be a part of data deliveries to FAO.

Deidentified data file with data from registers and questionnaires will be filed in Statistics Norway.

The aim of the census is to give a view of base of resources and production, and operational and environmental conditions in agricultural business. Additionally, the census shall give a historical view of the development and how farming is conducted in Norway.

The Census of agriculture is financed by own resources in Statistics Norway's budget as well as contributions from Norwegian Envrionment Agency and Ministry of local Government and Modernisation.

Important users of results are administrative agencies, the governments agricultural authorities and the farmers unions. Data will be used in planning, investigations, support for desicion making and research etcetera. Data is a very important contribution in the negotiations between the farmers unions and the Government

A census makes it possible to publish results on low geographical levels which makes the municipality bodies capable to study development at smaller regions than county.

Not relevant

The Census of Agricultural is close related to other statistics published annually. In the work with the census, one will make it possible to see the statistics in connection with each other.

Statistical act § 10-1

REGULATION (EU) 2018/1091 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 18 July 2018 on integrated farm statistics and repealing Regulations (EC) No 1166/2008 and (EU) No 1337/2011

Census of agriculture comprise all agricultural properties and all holdings with a minimum production (area, animal husbandry).

An agricultural property is a property with at least 25 decares (2,5 hectares) productive forestry area and/or at least 5 decares (0,5 hectares) agricultural area. An agricultural holding is defined as a enterprise with at least 5 decares of agricultural area in use and/or a minimum of animal husbandry.

Holdings with horticulture compromise holdings with at least 2.0 decares of field-grown vegetables, at least 1.0 decare of fruit trees (50 fruit trees in 1979 and 1989), at least 1.0 decare of field-grown berries, at least 300 m 2 of greenhouse area or at least 1.0 decare of nursery.

All information about farmers, owners and properties are collected from different registers. Some information from the holdings are collected on own questionnaires (unfortunately only available in Norwegian here: https://www.ssb.no/lt2010 ).

Data collection can be divided in two.

All holdings are obliged to fill in a questionnaire with questions about labour input, other gainful activities, drainage, irrigation and use of animal manure. In addition, all horticultural holdings were obliged to fill in a questionnaire regarding horticulture.

Many registers are merged with data from the questionnaire:

• Farm register
• Application for Governmental grants
• Business Register
• Education-register
• Tax register
• A-ordningen

Census of agriculture is a complete census. To be able to publish preliminary figures at an early stage, a sample is drawn from all received questionnaires. The sample comprise 7 850 holdings. The sample is stratified by size of agricultural area in use, what kind of production and county.

When final figures were produced, the figures for labour input were lower than expected. The difference was due to the sample that did not intercept changes and specialisation in the population. The difference is linked to "other prcductions" which includes to many different productions.

The questionnaires were sent to the holdings in the population. The counting date of the Census was 1st October 2020.

Information from registers are collected by different divisions in Statistics Norway.

Many different controls and checks have been conducted to make sure quality and consistence within one report and between different reports and registers. A new automatic data editing system was introduced to control, correct, edit and imputate missing or erroneous values. The system imputated values based on logical controls, average values, proportional frequency distribution and by "nearest neighbour" principles (impute values from a holding very similar to the holdings that is edited).

In a complete census the calculations are rather simple with addition, grouping and cross tabulating.

In the sample (preliminary figures) the population are stratified by county, size of agricultural area in use and production. A weight is calculted for each holding.

See also 3.3. Sample.

Conversion of energy contents to kWh in 2010 was based on factors described in About the statistics regarding the Energy balanse and energy account, see https://www.ssb.no/english/subjects/01/03/10/energiregn_en

Not relevant

Results from the census are restrained if there are to few observations behind the results, or if the results can be traced beck to single respondents.

Censuses in primary industries have been conducted since 1907.

The time series can not be compared back to 1907, but are in many ways comparable back to late 1940's. For some variables there are breaks in the time series. From 1999 all activity are registrered in the municipality where the agricultural area is situated. Before 1999 the activity was registered where the farm was situated.

The questions about foreign labour force in 2010 included more kind of labour force than 1999. In 1999 labour force were defined as "guestworkers staying in Norway for shorter periods to participate in farming, i.e. harvesting". The questions in 1999 were only asked to horticulture producers. Foreign labour force has increased since 1999, and there are reasons to beleive that foreign labour input in 1999 on the whole only consisted of persons that stayed in the country for shorter periods of time. Anyway a comparison of foreign labour force in 1999 and 2010 should be done with consciousness.

The questions about other gainful activities were introduced in 1999. In these 11 years, the more attention has been put on such activities and the questions have been adjusted. This might have influence on some of the results. For example the increase in number of holdings whice rent out hunting and fishing rights are not real. Probably the figures for 1999 are to low.

Primary data are collected from electronic questionnaires filled in by the respondents. The questionnaires can include errors done by the respondent filling in the form. Many of the errors are detected through the processing in Statistics Norway.

All surveys contains measuring errors, i.e. the respondent does not know the exact value or the questions are misunderstood.

An automatic editing system has been used in data editing, which can create new errors. The data editing is anyhow not done in a way that create new systematic errors. The data set is controlled and checked after data editing.

In data registers both measuring errors and processing errors can occur. Not all register data are input to administration routines, and are therefore not thoroughly checked. As far as possible, Statistics Norway corrects register data for statistical calculations. Statistics Norway has an overview for how register data are controlled by the register owners.

At last deadline for data reporting, 98 percent had responded. Values for no response will be imputated from other sources (registers, other surveys) and by other statistical ways of data imputation.

Partial no response will be corrected either manually or automatically. Several automatically controls are ran in connection with other ways of imputation (cf. 3.5.)

Sample errors is not relevant in a complete census, but will be handled when preliminary figures are published. Preliminary figures are based on an selection. In 2020 the selection size was 7 850 units.

Register errors: Quality of registers used to establish population might have influence on quality of the product. The quality of the registers is calculated as very good, and many resources were spend to establis a complete population.

All data on marital status are collected from registers. The register do not intercept all non marital cohabitations, so there might bed differences between information given on questionnaires and register. For example it might have been given information on labour input for both holder and spouse/cohabitant whilst the register "says" that the holder is single. When publishing results, we strictly use register data which makes it possible to link other information to the persons.

Not relevant