Skip to end of metadata
Go to start of metadata

 

From Statistics Finland:

Define the boundaries of top level phases more clearly.
- For example, how far the data is processed after phase 5?
- Define key requirements to check for the completion of each top level phase, e.g. "response rate should be sufficient"
- This would be useful for defining internal requirements or service level agreements within NSIs



From Australian Bureau of Statistics:

Clarity around Process vs Analyse boundary

ABS staff found the GSBPM's split of output generation related activities (5.7, 5.8 and 6.1) across the Process and Analyse phases quite confusing and not very intuitive, particularly since they were previously using a process model (pre-cursor to the GSBPM) which had 'Transform inputs into statistics' as a separate phase.  If all of the output preparation or compilation related processes were contained within the Process phase, the main focus of the Analyse phase then becomes more related to its title, i.e. one of analysing, validating and explaining the statistical outputs, ensuring they are 'fit for purpose' for dissemination.  In fact, as data are analysed throughout many phases of the GSBPM, the ABS have recently renamed this phase 'Validate' to reflect more than just the examination or analysis of outputs, but also the treatments that need to be applied to make outputs valid (or 'fit for purpose').  In addition, given a change in scope and the future need to undertake more complex compilation activity (bringing together a wide range of input data sources to compile outputs, often in a dynamic way), ABS staff in many statistical business areas prefer the term 'Compile' to 'Process'.  This change would also help win over the areas delivering economic and environmental accounts (e.g. National Accounts), as they commonly refer to their work as compiling accounts (where processing and analysis processes are strongly intertwined).

 


Please indicate your support for this change using the stars and legend below

  • 5* (We should do this)
  • 4* (Good idea, but need to discuss)
  • 3* (I am not sure, we need to discuss)
  • 2* (Should not make the change, but need to discuss)
  • 1* (Should not make this change)

 

 

  • No labels

5 Comments

  1. ONS

    There are also dynamic elements to processing which imply iterative use of the model which again does not seem to fit with the current static model

  2. I recognise the problem of distinguishing between 5 Process and 6 Analyse. I usually explain in term of inputs: the process phase is restricted to the (stove pipe) data set and its series, whereas analysis combines with information from other sources. Sometimes I also use a process argument: the process phase is fully designed and build, with limited manual intervention, wheras the analysis phase uses interactive tools, driven not only by the input, but also by the current environment (economic, legal) and by information from other sources. (This explanation serves mainly to distinguish between the different validation levels).

  3. I agree with the ABS that it would be good to take a look at the "split of output generation related activities (5.7, 5.8 and 6.1) across the Process and Analyse phases". I think that in many cases the "production of additional measurements such as indices, trends or seasonally adjusted series, as well as the recording of quality characteristics", which is part of 6.1 Prepare draft outputs, in practice takes place alongside 5.7 Calculate aggregates activities. Maybe there is some way to recognize this in the model.

  4. Istat suggests to modify the sub-process 5.7 adding it sub-process 6.1 and rename it consequently.

    5.7. Calculate aggregates and other outputs- This sub process creates aggregate data and population totals from micro-data. It includes summing data for records sharing certain characteristics, determining measures of average and dispersion, and applying weights from sub-process 5.6 to sample survey data to derive population totals.

    +6.1

    Data collected are also transformed into other statistical outputs, through the production of additional measurements such as indices, trends or seasonally adjusted series, as well as the recording of quality characteristics.

  5. 15/10:

    The split is maybe not quite as simple as micro vs macro because it depends on the type of process survey data vs National accounts.

    A number of suggestions about changes to subprocesses so we should come back to this conversation.