1.Survey process and timetable |
No |
Task |
Opening date |
Closing date |
1. |
Preparation work |
January 2015 |
May 2016 |
2. |
Survey design |
March 2015 |
May 2016 |
2.1 |
Set-up organization of the survey |
March 2015 |
December 2015 |
2.2 |
Design of data collection procedure |
December 2015 |
April 2016 |
2.3 |
Definition of survey variables |
March 2016 |
May 2016 |
2.4 |
Design of data processing procedures |
March 2016 |
May 2016 |
3. |
Survey promotion |
April 2016 |
June 2016 |
4. |
Data collection |
April 2015 |
July 2016 |
4.1 |
Sampling frame construction |
April 2015 |
August 2015 |
4.2 |
Framework design and sampling |
September 2015 |
September 2015 |
4.3 |
Recruitment of interviewers |
April 2016 |
May 2016 |
4.4 |
Training of interviewers |
May 2016 |
June 2016 |
4.5 |
Fieldwork |
01 June 2016 |
29 July 2016 |
Natural persons' farms |
01 June 2016 |
29 July 2016 |
CAII/CAWI |
01 June 2016 |
09 June 2016 |
CATI |
10 June 2016 |
22 July 2016 |
CAPI |
10 June 2016 |
29 July 2016 |
Legal persons' farms |
01 June 2016 |
29 July 2016 |
CAII/CAWI |
01 June 2016 |
30 June 2016 |
CATI |
01 July 2016 |
29 July 2016 |
5. |
Data processing and validation |
June 2016 |
May 2017 |
5.1 |
Data entry |
01 June 2016 |
29 July 2016 |
5.2 |
Data validation at record level |
June 2016 |
August 2016 |
5.3 |
Data correction and imputation |
August 2016 |
May 2017 |
6. |
Data compilation |
August 2016 |
June 2017 |
6.1 |
Weight calculation and estimation |
August 2016 |
May 2017 |
6.2 |
Calculation of quality indicators |
January 2017 |
May 2017 |
6.3 |
Aggregation and tabulation |
August 2016 |
June 2017 |
6.4 |
Data analysis |
September 2016 |
June 2017 |
7. |
Data dissemination |
April 2017 |
September 2017 |
|
2. The bodies involved and the share of responsibilities among bodies |
The bodies involved in the survey were:
- Central Statistical Office,
- Statistical Computing Centre,
- Regional Statistical Offices,
- Centre for Research and Statistical Education,
- Institute of Agricultural and Food Economics - National Research Institute.
The share of responsibilities among bodies involved was as follows: I. Within the CSO:
- Agriculture Department and Demography and Labour Department - all works related to methodology of the survey including the questionnaire content, trainings, elaboration of the algorithms for data control and corrections as well as data imputation, addition of administrative data, development of the assumptions for the Eurofarm, data analysis and publication.
- Methodology, Registers and Standards Department – sample design, sample selection, weight’s corrections, precision calculations.
- Surveys’ Programming and Coordination Department - was responsible for organizational issues, such as engagement of interviewers and arrangement of equipment for data collection, checking at all stages of the survey progress. In addition specialists from this Department were responsible for preparation of the geographical coordinates.
II. Statistical Computing Centre, branch in Radom – prepared – based on the assumptions from the Agriculture Department – the electronic questionnaire, developed the managing application (CORstat), IT personnel was also involved in the process of purchasing tablets for interviewers as well as supervising the functioning of the CATI centre. III. The RSOs The RSO was responsible for conducting the survey in individual voivodship from the organizational and content point of view.
- Organizational Coordinators - thanks to the CORstat application current observation of the course of survey was possible. Everyday pre-defined reports were prepared. In addition coordinators could prepare other necessary reports, using accessible filters. Coordinators could change channels for data collection.
- Content Coordinators – data control and analysis using the Content Module.
- External interviewers – data collection in CAPI (they were engaged from the resources of the Centre for Research and Statistical Education in Jachranka and were trained by the RSO staff both in the content of the survey and in the IT tools (tablet + electronic questionnaire)).
- Telephone interviewers – data collection in CATI.
- Statistical interviewers (permanent and supplemental) - in case of lack of the telephone number of a farm or preferences of farmers for the face to face interview (CAPI), employees of the RSO played the role of statistical interviewers.
The RSO in Olsztyn – apart from the tasks enumerated above, the IT employees elaborated the Content Module to be used in the survey, they implemented the assumptions for data validation, imputation, typology of farms and supplemented individual records with data from administrative sources. They were also responsible for data processing and preparation of the file to be transmitted to Eurostat (Eurofarm). IV. The Centre for Research and Statistical Education in Jachranka (the budget economy unit) was engaged as the subcontractor to organize recruitment of external interviewers. V. Institute of Agricultural and Food Economics - National Research Institute - in cooperation with the Agriculture Department was responsible for calculation of the Standard Output and preparation of the algorithms for typology calculation. |
3. Serious deviations from the established timetable (if any) |
Due to the delay of the transfer of administrative data to the CSO and the need for their corrections, the need for changes in the assumptions for data validation (domestic and Eurofarm) as well as problems related to implementations of assumptions, the scheduled date of data and NMR transmission to Eurostat has been postponed (from the end of May to June). |
|
1. Source of data |
The farm structure survey was designed as a sample survey on about 12% of farms included in the frame. In case of: organic farming, ecological focus area (EFA), support for rural development and common land the administrative data were used. |
2. (Sampling) frame |
The source of the frame was the statistical register. The register was built on the base of the AC 2010 results and administrative sources. The frame used to draw the sample for the survey is a list frame. The list frame included the following information: - identification number of an agricultural holding, - address of the holder, - address of the seat of the holding, - geographical coordinates of the address of a seat of the holder and of the seat of the holding, - telephone number, - e-mail address, - information on the holder: --- first name and surname/name, --- PESEL number for a natural person, --- identification number REGON for a legal person and an organisational unit, - information necessary for stratification: running of organic production, special branches of agricultural activity, area of agricultural land, livestock, economic size. The statistical register was updated every year with results of statistical surveys and administrative data: ARMA – Records of agricultural holdings, Records of producers; AFQI – Register of organic farms; Ministry of Finance - Special branches of agricultural activity, CSO – REGON, TERYT. Because of the organisation of statistical surveys and the schedule of the frame preparation as well as the dates of availability of statistical and administrative data necessary for updating, data included in the frame was delayed in comparison with date of survey from 0,5 to 2 years. |
3. Sampling design |
3.1 The sampling design |
The sampling design was based on a single-stage stratified random sampling of holdings with take-all strata. The broad subject matter of this survey caused some special approach to sample drawing and first of all it caused necessity of usage of stratification. Experience from previous farm structure surveys: 2005, 2007, 2010 and 2013 (experience of the European Union countries and Poland as well) was taken into account. The method of a simple random sampling without replacement was used in every stratum independently. |
3.2 The stratification variables |
Due to the fact that it was difficult to use the same variables for each type of holdings, therefore, before sampling, the population of holdings was divided into the following categories: A. Farms with agricultural activity, not classified as organic farms, of which: I. farms, which in the sampling frame had: 20 head of sheep or 20 goats or 2000 heads of hens or 500 head of turkey or 500 head of goose or 500 head of duck or 80 beehives and farms conducting activity connected with special branches of agricultural production and with minimum 250 thousand euro economic size, as well II. basic category of farms with agricultural activity but not mentioned above. B. Non-active farms. C. New created farms and farms without established type of farming and economic size. D. Organic farms. E. Agricultural holdings of legal persons and organizational units without legal personality. Each population category was divided by voivodships (NUTS 2) and strata were created using different criteria of stratification depending on the category of holdings. There are 24 strata in total across all population categories: Category A.I forms a single stratum - stratum h=01. Category A.II comprises 10 strata i.e. h= 02, 03, ... 11 Category B comprises 5 strata i.e. h= 12, 13, ... 16 Category C comprises 6 strata i.e. h= 17, 18, ... 22 Category D forms a single stratum - stratum h=23. Category E forms a single stratum - stratum h=24.
Samples were drawn only for: - category A.II, strata h= 02, 03, ... 10 - category B, strata h= 12, 13, ...15 - category C, strata h= 17, 18, ... 20 Category A.II - as criteria of stratification, we took two variables: the area of agricultural land (A_3_1+B_5_1) and the economic size. The economic size of holding was calculated in accordance with the Eurostat methodology. Category B - as a criterion of stratification, the total area of holding (A_3_1+B_5_1+B_5_2+B_5_3) was used. Category C - as a criterion of stratification, the area of agricultural land was used. Strata codes in database take into account intersection of original 24 sampling strata in each voivodship with NUTS2 codes, and additional strata with outliers (where weights are 1). |
3.3 The full coverage strata |
The sampling scheme took into account a complete survey of certain types of agricultural holdings. The farms included in category A.I and categories D and E were surveyed on a 100% basis. In addition, take-all strata were: - Category A.II: last stratum (i.e. h = 11) in each voivodship which consists of such sampling units for which the value of at least one of the variables adopted as the stratification basis is above the specified threshold. - Category B: last stratum (i.e. h=16). - Category C: last two strata (i.e. h=21 and h=22 (farms which in sampling frame had minimum 20 head of cattle or 50 head of pigs)). |
3.4 The method for the determination of the overall sample size |
The size of the sample was decided in accordance with financial and organisational possibilities and the precision requirements provided in the Regulation 1166/2008. The number of agricultural holdings in the population and in the sample by category is provided in the following table: Number of holdings in the population and in the sample by categories
Category of holding |
Population |
Sample |
Poland |
1547980 |
184129 |
A.I |
28152 |
28152 |
A.II |
1199303 |
112424 |
B |
44919 |
1839 |
C |
247679 |
13787 |
D |
23798 |
23798 |
E |
4129 |
4129 |
|
3.5 The method for the allocation of the overall sample size |
Sample allocation between voivodships and strata was differenly done depending on the category of holdings. Due to the importance of category A.II, in this case the rules of drawing the sample will be described more precisely (see below). The following assumptions were made while drawing the sample from this category of farms: (1) the size of n sample is established for the population of farms in Poland, and not for individual voivodships, where n consists of approx. 112 thousand farms, (2) the sample is drawn in individual voivodships according to the stratified and optimal sampling scheme of Neyman, (3) the population in each voivodship is divided into 10 strata (h = 02, 03, ... , 11), and the sample is simultaneously allocated between these strata, (4) last stratum (i.e. h = 11) includes the farms which are not drawn, but which are all included in the sample, (5) it has been assumed that the expected accuracy of the survey results with respect to this group of farms, measured with the coefficient of variation of the area of agricultural land and economic size, will be identical for each voivodship and will be equal approximately to 0.4%. For category B, the sample allocation considered the total area of holding and it was assumed that coefficient of variation for this variable should be for each voivodship equal 1.8%. For category C, the allocation of sample between voivodships and strata (17-21) considered the area of agricultural land. It was assumed that coefficient of variation for this variable should be for each voivodship equal 2%. |
3.6 Sampling across time |
For each occasion a new sample is drawn. |
3.7 The software tool used in the sample selection |
The samples in categories A.II, B and C were drawn with standard SAS system drawing procedures (procedure SURVEYSELECT). |
3.8 Other relevant information, if any |
Stratification and sampling allocation were carried out by means of the numerical optimisation method. The description of the solution to this problem was published in the article written by B. Lednicki and R. Wieczorkowski (2003) (see section 9.6 Documentation on methodology - item 2). |
4. Use of administrative data sources |
4.1 Name, time reference and updating |
Records of agricultural holdings benefitting from support in respect of the Rural Development Programme (used for replacing the values of characteristics) Legal base: EU - Regulation No 1305/2013 of the European Parliament and of the Council of 17 December 2013 on support for rural development by the European Agricultural Fund for Rural Development (EAFRD) and repealing Council Regulation (EC) No 1698/2005 PL - Law of 20 February 2015 on support for rural areas development by the resources of the European Agricultural Fund for Rural Development within the Rural Development Programme 2014-2020 Responsibility: ARMA Reference period: 2014-2016 Records of agricultural holdings (used for replacing the values of EFA characteristics, and validating and analysing the FSS data) Responsibility: ARMA Reference period: 2016 campaign Register of Organic Farms (used for replacing the values of characteristics) Legal base: EU - Council Regulation (EC) No 834/2007 of June 2007 on organic production and the labelling of organic products and repealing Regulation (EEC) No 2029/91, PL - Law of 25 June 2009 on organic farming. Responsibility: AFQI Reference day: 31 December 2016 Records of the Land Communities (used for replacing the values of characteristics) Responsibility: HOGC Reference day: 1 January 2016 |
4.2 Organisational setting on the use of administrative sources |
Law on official statistics: Article 5 1.The official statistics have a mandate to collect from all available sources, specified in detail in the programme of statistical surveys of official statistics or in other laws and to gather statistical data from and about business entities, other legal entities and entities without legal status and their activities, further referred to as 'individual data' and statistical data on natural persons concerning their life and status, further referred to as 'personal data'. Article 13 1.The data collected and stored by State administration bodies and units of local authorities, other governmental agencies, organs maintaining official registers and the National Bank of Poland, on the basis of Regulations other than this Law, are data of information systems of public administration, and are further referred to as the 'administrative records'. 3.State administration bodies and units of local authorities, other governmental agencies, organs maintaining official registers and the National Bank of Poland on a free-of-charge basis, shall submit to the official statistical services the administrative records stored by them in the scope, form and time specified each time in the programme of statistical surveys of official statistics, especially in the form of extracts from the registers, copies of data files, collected declarations, registration documents and other official forms, the results of measurements, and data on environment monitoring, and shall provide data from their computerised database systems. The CSO representatives participate in meetings concerning the content of administrative registers and emphasize statistics’ needs, also in the context of reducing the burden on the respondent. This cooperation and the arguments presented by the statisticians bring moderate results, since changes to the administrative registers are laborious and costly. Actions at EU level aimed at coherence of the needs (statistical and administrative data) reported to Member States would be helpful. |
4.3 The purpose of the use of administrative sources - link to the file |
Please access the information in the file at the link: (link available as soon as possible) |
4.4 Quality assessment of the administrative sources |
|
Method |
Shortcoming detected |
Measure taken |
- coherence of the reporting unit (holding) |
|
Differences in definitions of agricultural holdings (e.g. one holding could correspond to two or more holdings in the administrative sources). |
“Administrative” holdings were combined into the one that met statistical definition. |
- coherence of definitions of characteristics |
|
Differences in definitions of some characteristics, for example: Register of Organic Farms: Organic farming - other animals covers: equidae, bees, rabbits, deers, and are different from the categories in Other livestock (C_99). Records of agricultural holdings: Total agricultural land as well as Permanent crops are different from characteristics (A_3_1+B_5_1) and B_4. Different reference periods for statistical survey and administrative data. |
For the FSS purposes (especially when the data is replaced) variables which have the same definition were used. If possible, data was collected for individual characteristics so that allowed to aggregate data according to statistical requirements. Differences in reference periods were taken into consideration when comparing and analyzing the data. |
- coverage: |
|
|
|
|
over-coverage |
|
Not found. |
|
|
under-coverage |
|
Not found. |
|
|
misclassification |
|
Not found. |
|
|
multiple listings |
|
Register of Organic Farms: There were 5 duplicated units with different agricultural production. The units were under the control of two certification bodies for different production. The one concerns crops production, the other – beekeeping. |
The data were added for one unit (CSO). |
- missing data |
|
Register of Organic Farms: 68 records with missing data were found. |
The missing data were completed (AEQI). |
Register of agricultural holdings (EFA): 800 records with missing data for EFA (needed data was not provided by farmers). |
They were eliminated from the file (ARMA). |
- errors in data |
|
Register of Organic Farms: Some data on organic crops area were given in square meters. |
The values have been converted into ha (CSO). |
Register of agricultural holding benefitting from the Rural Development Programme: In the case of 161 records identifier number (producer number in the ARMA) was wrong. |
Wrong numbers were corrected (ARMA). |
Register of agricultural holdings (EFA): There were about 1 000 records, which did not fulfil EFA obligation. |
After checking, they were eliminated from the file (ARMA). |
- processing errors |
|
Not found. |
|
- comparability |
|
There were no other sources of data on organic production, EFA as well as support for rural development. |
|
- other (if any) |
|
The final validated administrative data are not in time to meet statistical deadlines. IT problems - the structure of the files does not always meet standards used in statistics, the lack of data formats, in the case of address data, the TERYT is not used. |
Different structure of the file, different cell’s formats needed additional transformation of files. |
4.5 Management of metadata |
Administrative metadata provided by ARMA and AFQI are stored and maintained in dedicated databases. |
4.6 Reporting units and matching procedures |
Records of agricultural holdings: agricultural producers benefitting from direct payments in a given year (including EFA); Register of Organic Farms: organic producers; Records of agricultural holdings, Register of Organic Farms and Records of agricultural holdings benefitting from support in respect of the Rural Development Programme: the common identifier was the producer number in the ARMA which is linked to the PESEL or REGON of the holder. |
4.7 Difficulties using additional administrative sources not currently used |
Not applicable. |
|
Data validation |
Data validation took place at the data collection level (interviewers), regional (voivodship) level (Content Coordinators) and central level (CSO experts). Questionnaire application contained the algorithms of accounting and logical control. At the same time, the applications did not allow the omission of the questions which were required to be completed on certain “paths” of the interview. The data was subject to control and approval by Content Coordinators in the Content Module. If any data was found to be potentially incorrect, a Content Coordinator could make the necessary adjustments (if on the basis of the available data such an adjustment was possible) or request that the interview be repeated. After loading the data obtained from the approved CAPI, CATI, CAII forms and administrative sources, as well as imputed data for units non-response, the central set was subject to automatic logical, accounting and range control. The algorithms of control were prepared by the CSO experts and also contained the rules for validation required by Eurostat (Manual for data suppliers, survey 2016, rev. 7). For each module (section) of the questionnaire, the rules for inter-section control were prepared. The control also covered the inspection of section interrelations and coherence. The modules were controlled in a particular order. After checking and making any potential adjustments to a given module, the next module was checked. The rules of control in individual modules also had a pre-determined order. If an error was identified, the application performed automatic adjustment. After the control was completed, a report was generated containing information on the used validation rules, occurring errors and applied adjustments. The records which contained errors and could not be automatically corrected were analysed by experts who decided on recognising the error or choosing the adjustment method on the basis of specialist knowledge and available data. Tools for data validation were: - at the stage of data collection - the questionnaire application, the Content Module (the Module received the data collected by interviewers/telephone interviewers, when a given farm was closed by the voivodship Organizational Coordinator. The Content Module, after supplementary validation, was used by Content Coordinators to analyse the collected data in detail. The analysis could use pre-defined inquiries (filters) and individually constructed queries (on the basis of SQL tools)). - at the stage of data processing (central dataset) - a dedicated application using SQL and SAS tools. |
|