While no metadata for the aggregated tables are currently available, we seek to explain potential misinterpretation of the results when navigating the data. The nature of the delusion might stem from various sources, such as ambiguous labelling of dimension, time breaks due to changes in legislation/thresholds, complex datasets (i.e high number of dimensions and categories with each dimension) or lacking awareness on the computation of the values.  Therefore, we present a number of examples highlighting the difficulties with the interpretation of the results and we provide further explanations to avoid drawing wrong conclusions. 

For the future, we aim on providing comprehensive metadata for aggregated tables to avoid any further confusions in the interpretation of the results. This should be put in place by latest for the next dissemination wave of the 2023 IFS survey. 

Variable dimension: The computation of values

It is crucial to note that every EF table has a variable dimension determining the values in the table. For instance, the variable dimension is CROPS in ef_lus, ANIMALS in in ef_lsk or WSTATUS in in ef_lf tables. The example below in TABLE XX further illustrates the generation of aggregated values for the table ef_lus_main. Depending on chosen category of UNIT, the values are computed based on the selected class of CROPS. The other dimensions serve solely to filter the data. 

TABLE XX. Computation of aggregated values for ef_lus_main

Ambiguous labelling of dimension

We have noticed that several labels of classes and dimensions are not very intuitive in EF tables. Therefore, we tend to improve this part when revising the tables. Until then, we would like to clarify this vague concepts by adding some information. 

• AGRAREA - Agricultural area:
• This is misleading as it can be any type of agricultural area (e.g farm area, utilised agricultural area, non-utilised agricultural area). It refers to utilised agricultural area and it will be replaced in the future with the dimension UAAREA - Utilised agricultural area
• HA_LT2 - Less than 2 hectares: From the label, it seems that 0 hectare is included, but actually it is not as the formula used is "UAA>0 and UAA<2". We will improve this label by naming it HA_GT0_LT2 in the future
• KE_LT2 - Less than 2000 Euro: From the label, it seems that 0 Euro is included, but actually it is not as the formula used is "SO>0 and SO<2000". We will improve this label by naming it KE_GT0_LT2 in the future
• LSU_LT5 - Less than 5 LSU: From the label, it seems that 0 LSU is included, but actually it is not as the formula used is "LSU>0 and LSU<5". We will improve this label by naming it LSU_GT0_LT5 in the future

Interpretation: Combination of dimensions

Given that the interpretation of the results can be misleading when selecting several dimensions, we will outline few examples to clarify on how to report the data. 

Example 1: Number of young farmers with full agricultural trainings (EF_MP_TRAINING)

Imagine that we are interested to know the number of young farmers as farm manager with full agricultural trainings by country.

To do so, we need to select the following specification: 

1) UNIT select "HLD"; TRAINING select "FULL"; AGE select "Y35-39"

2) Additionally we need to select: SEX="T" and SO_EUR="TOTAL" and UAAREA="TOTAL" 

Information will only be available for 2016 and 2020. In previous years, age was collected as categorical variable (e.g 25-34, 35-44) and therefore it is not possible to visualize data in this specific category Y35-39.

Example 2: Number of farms having no utilised agricultural area in pig farming in European Union between 2005 and 2020 (EF_LUS_MAIN)

Now we would like to know how many farms have no utilised agricultural area in pig farming in EU without UK between 2005 and 2020.

To do so, we need to select the following specification: 

1) UNIT select "HLD"; CROPS select "UAA"; AGRAREA select "HA0"; FARMTYPE select "FT51_SO"; GEO select "EU27_2020"; YEAR select ("2005","2007","2010","2013","2016","2020")

2) Additionally we need to select: SO_EUR="TOTAL"

Its important to note that UAA needs to be selected in CROPS as it is the variable dimension determining the values in the table. 

Example 3: Annual work unit of agricultural holdings across legal status of the holding in 2020 for FR

To see the distribution of the total sum of annual work units across legal status, we need to select the following specifications: 

1) UNIT select "AWU" and CROPS select "UAA" and LEG_FORM select ("PER_NTL","PER_LEG","GRP_HLD","UNIT_CML") and GEO select ("FR") and YEAR select 2020

2) Additionally, we need to hold constant SO_EUR="TOTAL" and UAAREA="TOTAL" and FARMTYPE="TOTAL"

Labour is a module in FSS and therefore it is not mandatory to collect labour for all farms in census years. This means that some countries only collect a sample and the data is extrapolated by the extrapolation factor. 

Main frame and main frame plus frame extension

According to Article 7(2) of Regulation 2018/1091, modules in 2020 shall be collected only on the main frame. However, where countries collect modules on the frame extension, Eurostat accepts such data in 2020. It's true that not all countries opt for main frame plus frame extension and this might lead to inconsistencies  for module specific variables (i.e labour, other gainful activities or manure management) in the comparability between countries in data dissemination. In addition, when a table includes both core variables and module specific variables (i.e labour, other gainful activities or manure management), for the country which collected core on main frame plus frame extension and module only on main frame, the totals of core variables are not computed over the same scope as the totals of the module variables. On the other hand, restricting the data dissemination, -in which core and module variables are used, to only main frame would result in loss of information and comparability over time, but it would increase the comparability across countries in a given year. So there is a clear trade-off. Yet, to increase the comparability of the data across time, Eurostat decided to disseminate main frame and main frame plus frame extension for specific countries in data publications. 

For the sake of clarity, we present a table comprising the overview of countries that apply for modules main frame plus frame extension and another table with the concerned variables/categories in the data publications. 

Table. Overview of the application of main frame plus frame extension in 2020 by country

Table. Concerned data publications of IFS 2020 data until 27 Apr 2023