Page 97 - Special Topic Session (STS) - Volume 1
P. 97
STS407 Étienne Saint-Pierre
Framework. In the example, the concept used is ‘revenue’. In the IBSP, for all
statistical programs, the concept of revenue is identified by cell numbers
ranging from F43000 to F59999. The next component is the sub-concept. For
all IBSP programs, the sub-concept of revenue named ‘sales’ is identified by
cell numbers between F43000 and F45000. The third part call the modifiers
provide a description of the variables. In the example here, the modifier is fruit
market. The dimension component categorize the variables. The Census of
Agriculture 2021 has adopted the corporate naming framework used in the
IBSP. The 630 distinct variables of the Census of Agriculture cover 12 concepts,
21 sub-concepts, 86 modifiers and 15 type of dimensions and 331 unique
dimensions. The Census of Agriculture shared common cells with 13 different
programs.
This structure provides a fully coherent framework facilitating the
integration of other programs data in the ‘collect once-use multiple time’
strategy or in the use of other programs data for validation or horizontal
analysis. With this harmonized Naming Framework, the production of in-depth
and timely analysis required by users to make informed, evidence-based
decisions about increasingly complex issues that affect sectors bordering
agriculture (e.g., food processing transportation, environment, rural affairs,
international trade, and agricultural price indexes) is easier. There is no need
to build complex transformations or concordances between variables in
different programs or data sources for the integration of alternative data.
iii) Mandatory use of corporate tools and generalized systems
By adopting the corporate tools and systems used by the Economic
Statistics programs, the Census of Agriculture will benefit, from the most state-
of-the-art statistical leading-edge approaches introduced to improve
methods, data quality and efficiency without having to invest locally to
compensate for the obsolescence of systems. Whether it is for the design of
the collection vehicle or for the management of the frame, to link efficiently
datasets, to impute missing data, to tabulate estimates, to visualize and
analyse data, to manage disclosure avoidance activities, to generate quality
indicators and to publish results, the Census of Agriculture will now make use
of all corporate harmonized systems and tools available.
The next Census of Agriculture will benefit directly from these advances in
corporate systems:
a. The integration of the variance due to imputation methods into the
generalized estimation system has made it more accessible. Its
adoption by the IBSP as a standard output means that the Census
now has a way to quantify a non-sampling error notion that was
always known but never measured. Users will be better informed
about the overall quality of the estimates.
86 | I S I W S C 2 0 1 9