Understanding GSIM
10. The documentation of GSIM v1.0 has a number of layers. Each layer provides more detail on the model and is targeted at different audiences. There is:
- A number of brochures which provide a high level introduction to GSIM
- A Communication document which gives an overview of the model. It also describes the scope and benefits, what the model means for staff in a statistical organization and relationships to other standards and models.
- A User Guide (this document) which explains how GSIM can be used and implemented.
- A Specification document which provides deeper detail about the model, an explanation of the GSIM Extension Methodology and descriptions of how GSIM relates to other standards and models at a detailed level.
- An Enterprise Architect file which contains the UML descriptions of the model.
11. To help target the efforts of staff in statistical organizations, Table 1 provides some guidance on which documents are most appropriate for different audiences.
Table 1. Reader Guides
Audience | Suggested documents |
Top & Senior managers | GSIM Brochures |
Middle Managers, | GSIM Brochure |
Architects, business analysts and metadata specialists | GSIM Communication document |
Solution Architects | GSIM Specification |
12. There are some terms and concepts related to statistical information which are important to a statistical organization which are not represented in GSIM v1.0 as specific information objects. In particular, these include Methodology, Quality and Reference Metadata. The following paragraphs explain why these are not included as separate information objects in GSIM v1.0. If an organization has specific unmet requirements for information relating to methodology, quality or reference metadata in the meantime, the extension mechanism that is provided within GSIM can be used during local implementation to fit those specific requirements.
Methodology
13.Methodology is a crucial consideration in production of official statistics, particularly in terms of determining methods to be used during the statistical business process.
14. Methodology designates the study, science and/or theory of "method"(s). Methods themselves are represented as an information object within GSIM but the science of determining methods is not. Methodology can be seen reflected in several GSIM information objects. For example, the rules and parameters that are defined from a methodological view point. The design of a statistical production process sets out the methods to be followed when the process is run. For example, when running a process such as editing, the Rules for the procedures in such a tool are defined by existing best practice methodology. The method used is therefore likely to be fairly specific either to an individual process or to a group of similar processes.
15. "Methodology" is not an information object in itself and therefore not modelled as such by GSIM. GSIM is intended to be a generic model, capable of supporting all current and future methods. An example on how the information in a common process conducted by methodologists (sample selection and estimation) is captured by GSIM can be found in Annex A.
Quality
16. Quality means different things in different settings. There is quality as an organizational aspect, the quality of the processes and the quality of the statistics. While methodology is embedded in the design of the statistical production process, quality is typically linked to the instance (i.e. to production runs) of the process.
17. Quality is relevant at a number of different levels of instances of information objects. For example, as an attribute to an information element (e.g. quality flag), as an attribute to a data set (e.g. status provisional data, final data, revised data). It also appears as process quality information. The product quality is laid down in a quality report, which is itself also a statistical product. It is not appropriate for quality to be represented within GSIM as a single information object, distinct from all other information objects. Various facets/aspects of quality are, instead, described, assessed and/or managed using various information objects within GSIM.
18. Quality reports traditionally mainly refer to the quality of the statistics. Quality information and quality reports can be tied to the production process as a whole and/or to parts of it. Quality is present in the inputs and outputs of process steps in the Generic Statistical Business Process Model (GSBPM), acting as process data to control and define rules for processes. The outputs from a GSBPM process step using GSIM information objects can be used as quality measures (in the form of Process Metrics).
19. Quality related to statistics can have many forms depending on its purpose. Depending on the scope, it will refer to different information objects in relation to relevant processes.
20. An example on how quality information is captured by GSIM can be found in Annex A.
Reference metadata
21. There is currently no globally agreed definition of the scope of reference metadata, and GSIM should remain generic rather than linked to one specific definition.
22. Conceptual metadata are represented in GSIM in the Concepts Group, methodological and procedural aspects of metadata are represented by GSIM Production Group. As described above, various information objects from GSIM can be referenced when describing the quality of statistics.
23. GSIM models connections between data and its associated metadata. For example, from a Data Set through its Data Structure Definition, relevant information from the Concepts Group (for example about Variables, Classifications, Populations, Concepts) can be discovered. If the data were being represented in SDMX then the conceptual information (and process information and quality information) relevant to that data would be presented as Reference Metadata.
24. When defining and managing conceptual content in its own right, such as negotiating the definition of a new standard classification, the content isn't typically referred to as "Reference Metadata". In that case the conceptual metadata isn't being considered in the context of describing the contents of a specific Data Set.
25. If the term "Reference Metadata" is considered more generally as the ability to refer from the definition of a particular information object (typically data) to other information which is relevant when considering that object then aspects of metadata may be modelled by means of object attributes. The extension mechanism within GSIM can be used to define additional attributes for information objects if these are required (including use of the extension mechanism).