Workshop on Implementing Standards for Statistical Modernisation, 21-23 Sept 2016 in Geneva.

Submit your abstract by 6 May. More info here
Skip to end of metadata
Go to start of metadata

The definition of the Design Context according to the Annex C Glossary is "Methodological metadata that provide the basis for the specification of the information objects required as input to and output from the Process Step Design including Process Method and Rules." I think I can understand "Methodological metadata that provide the basis for the specification of the information objects required as input to and output from the Process Step Design." but the addition of Process Method and Rules in the defintion is slightly misleading. They are examples of information objects required as input to Process Step Design, but they are not Design Context so they should be removed. Instead we should include som examples of Design Context in the Explanatory text which is currently empty.

As a reminder the recommendations from ISO/IEC 11179-4: Formulating data definitions are

A data definition should: a) state the essential meaning of the concept, b) be precise and unambiguous, c) be concise, d) be able to stand alone, e) be expressed without embedding rationale, functional usage, or procedural information, f) avoid circular reasoning, g) use the same terminology and consistent logical structure for related definitions and h) be appropriate for the type of metadata item being defined

Can someone provide an example, or two, of Design Context?


  1. Notes from DDI group discussion on this object:

    GSIM definition needs to be revised when you create a statistical design, this is a rationale for why you decided to do that. It is not at the group level not the study unit. In GSIM, this is at the program level.

  2. Discussion in GSIM Implementation Group

    This is a placeholder for the statistical methodology.

    It has an association to the StatisticalProgramDesign.

    StatisticalProgram is hierarchical so the design and the DesignContext can be included at the appropriate level.

    A new name could be proposed.

    This i.o. is needed and an example will be provided.


  3. An example could be the SNA 2008.

    The SNA contains an identity for calculating GDP (para 16.47), which specifies inputs, the calculations to apply to these inputs, and the output. It seems clear to me that this identity would "provide the basis for the specification of the information objects required as input to and output from the Process Step Design[s]" involved in a number of national accounts processes. This is just one instance of many in the SNA.

    It suggests, however, that perhaps the definition should be broadened to include the various roles than methodological documents play. For example, the SNA could be an input to literally any one of the design sub-processes in the GSBPM (2.1 through 2.6).


  4. I agree that the definition should be updated.  Particularly considering Gareth's example, in addition to influencing its ProcessInputSpecification and ProcessOutputSpecification, DesignContext could influence the (high level) ProcessStepDesign associated with the StatisticalProgramDesign in terms of the ProcessMethod that is selected (very likely), the Rules (quite likely), the lower level Process Step Designs and Process Controls etc.  In other words, DesignContext would provide the methodological context for the design choices that have been made beyond just inputs and outputs as per the current definition.  

  5. This box is the methodology (the scientific basis)- not the method.

    (could be why you have done what you have done to align with SNA - to tie with Gareth's example).


    Option 1:

    Is it an object on its own? Maybe it is a reccommended paper to have? Is there really any reuse from this?


    Option 2:

    Redefine this object. Need some more examples of what one might be trying to document. What is an instance of it?

    Plus need a better name?

  6. No use cases to support this since last discussion. Does this suggest that we should remove it? (until a use case arise). There is definitely a requirement for this, but there is a lot of ambiguity. How to tie methodology to tie to actual method? Clear design context is not a useful name. Methodology is the top level method and people can capture the variation of the of the method that they use using the Method object. 

    Methodology discussion has to happen before we come back to this.



  7. New name: Methodology Design.

    Includes GSBPM 2.3 data collection methodology. 2.4 frame & sample methodology and 2.5  statistical processing methodology, but also any methodology you plan to use in Analysis ( e.g. seasonal adjustment) and Dissemination (e.g. statistical disclosure control).

  8. 2/10 Discussion:

    Action: We are not making changes to the methodology for this version. We should remove this item.