Meeting notes - June 18th, 2024

[Chris-ECE]

Boundary between Specify Needs and Design and phases

• Regarding the query from France about the difference between 1.4 and 2.2, InKyung mentioned that in GSIM, a concept is related to conceptual variable via a unit type. In 2.2 when you design a variable, it needs a statistical unit, which may for example be contained in a metadata repository.
• InKyung also highlighted the feedback from the GSBPM “Task” task team, which pointed out that GSBPM talks about defining statistical units. So she suggested that 2.2 is where this should happen.
• Gabriel suggested it’s important to distinguish between the provider and the information unit to avoid confusion (as for example both could be households).
• Florian suggested describing statistical concepts in 1.4 if they differ from dictionary definitions for those words.
• He also pointed out that in defining units, its worth bearing in mind that, for example when using admin data, you may have to convert the admin unit to the statistical unit. (In the case of tax data there may be fiscal households which may not equate to real households.) Similar situations regarding units might also arise in the case of surveys, if questions are used to narrow down the group of interest.
• Changes were made to the description of 2.2, to make it clear that unit types (including observation units and statistical units) are designed there. The name of subprocess 2.2 was changed to indicate that it is variables that are being designed, rather than their descriptions.

• Regarding the query from France about the difference between agreements with the data providers in 1.5 and finalisation of agreements in 2.3, Florian said that in 1.5 he only envisioned new legal agreements, but the preparation of an agreement would be done in the design phase. – The description text for 2.3 was changed accordingly to clarify that it includes both the confirmation of existing agreements with data providers AND the creation of new agreements.
• There was also a slight change to the language used in 1.5 (provision agreements) to align with the terminology in GSIM, as well as clarification that such agreements could be legal or other agreements such as memoranda of understanding (the GSIM definition of a provision agreement is that it is a “legal or other” basis for parties to agree to data exchange).

Where are units defined?

• The already-discussed feedback from the GSBPM “Task” task team, has been addressed by the changes made to the text of 2.2 referred to above.
• Regarding the feedback from France that identification of target population is not exclusive to survey sources of data, a minor change was made to the text in 2.4 to accommodate this point.

2.1

• Regarding the feedback from France about revisiting 2.1 after the sub-process "apply disclosure control" is completed in 6.4, it was agreed that the more prominent emphasis on non-linearity (including an illustrative diagram in the introduction) is sufficient to address this point.