statistical process description using gsbpm as a reference ... · quality documentation 1. the new...
TRANSCRIPT
1
Statistical Process description using
GSBPM as a reference – Challenges in
a process changing environment
Maria João Zilhão, Statistics Portugal, [email protected]
Magda Ribeiro, Statistics Portugal, [email protected]
Abstract
Statistics Portugal (SP) has a long tradition of statistical process description, starting with a generic model in 1997. Several editions were revised since then and the latest one was recently updated using the last edition of the Generic Statistics Business Process Model (GSBPM V.5) as a reference. Decisions had to be made to have a handbook in a generic format for all staff in statistical production environment, as well as an adaptation between the GSBPM model and all horizontal processes and internal standards in use by SP. The paper will describe i) the new version of the Statistical Process Procedures Handbook model produced by SP, ii) its challenges (strong and weak points) in conception while adapting the GSBPM with SP reality of statistical process implementation and iii) the subsequent levels of process description added to GSBPM (namely, geospatial statistical dimension), and much aligned with ISO 9001 Norms with respect to documented procedures.
Keywords: Statistical Process; Process description; Documented procedures; Quality Standards;
Quality Documentation
1. The New version of the Statistical Process Proce dures Handbook produced
by Statistics Portugal
Statistics Portugal has a long tradition in describing the Statistical process making
use of a process description model. The first edition is dated 1997, very much alike
the experiences that other NSI had in that respect, namely Statistics Canada. Up to
now other editions were implemented and evolved in various ways, namely in detail
(adding new features and dimensions) and flexibility. However, its background
approach always applied the logic of ISO documented procedures (namely,
procedures/tasks; responsibilities; documentation/registers to produce along the
process).
2
This type of documents have also been of extreme importance in moments of
changes in organizational structures within the office, namely when a centralized data
collection department was established.
As such, the motivation for describing and updating a Statistical Process Procedures
Handbook has various dimensions in what concerns Statistics Portugal experience:
− First it facilitates internal and external communication (national and
international) about statistical production amongst the staff, helping each unit
and statistician seeing him/herself as a player in the whole process;
− It promotes the standardization of specific terminology and the sharing of
knowledge, data and tools along the business process. As such, it fosters the
process of standardization as a whole.
− It is also extremely important in supporting the planning process of a specific
activity, making use of IT tools that might be available for that (which is the
case of Statistics Portugal).
− It provides a map of the entire business process, enhancing communication in
the organization.
− It allows for the identification, in a systematic way, of all necessary
documentation to be produced along the business process (e.g. related
procedures, reports and performance indicators).
− It allows for the identification of all possible synergies along the process and
between activities.
− It facilitates and supports the modernization of production systems.
− And finally, it is a relevant tool for quality management in what concerns
review activities, such as auditing processes.
2. The challenges in adapting the GSBPM 1 with Statistics Portugal reality of
statistical process implementation
After several years in improving the business process description a decision was
made along the last peer review process to completely align Statistics Portugal
handbook with the GSBPM. 1 Generic Statistics Business Process Model (GSBPM V.5), UNECE , 2013
3
Figure 1 – Adopted Business Process Model (adapted from the GSBPM V.5)
Similarly to the GSBPM, the overall model must not be followed in a strict order:
− It is flexible and interactive in what concerns phases and sub-process
sequence;
− There is interdependence among sub-processes;
− The applicability of certain phases and sub-processes depend on the type of
statistical activity, making the scope of the Handbook quite large.
1. Conceção/Identificação de
necessidadesSpecify Needs
2. Conceção/Especificações
Design
3. Conceção/Desenvolvimento
Build
4. RecolhaCollect
5. ProcessamentoProcess
6. AnáliseAnalyse
7. DivulgaçãoDisseminate
8. AvaliaçãoEvaluate
1.1 Identificar necessidadesIdentify Needs
2.1 Especificar resultados
Design outputs
3.1 Desenvolver suportes para
recolhaBuild collection
instrument
4.1 Criar base de amostragem e
selecionar amostraCreate frame and
select sample
5.1 Integrar dadosIntegrate data
6.1 Preparar resultados
Prepare draft outputs
7.1 Atualizar sistemas de difusão
Update output systems
8.1 Reunir elementos para
avaliaçãoGather evaluation
inputs
1.2 Confirmar necessidades
Consult and confirm needs
2.2 Especificar variáveis
Design variable descriptions
3.2 Desenvolver suportes para
tratamento e análiseBuild or enhance
process components
4.2 Preparar recolhaSet up collection
5.2 Classificar e codificar
Classify and code
6.2 Validar resultados
Validate outputs
7.2 Elaborar produtos de difusão
Produce dissemination
products
8.2 Executar avaliação
Conduct evaluation
1.3 Establecer objetivos
Establish output objectives
2.3 Especificar recolha
Design collection
3.3 Desenvolver suportes para
difusãoBuild or enhance
dissemination components
4.3 Executar recolha
Run collection
5.3 Validar microdados
Review and validate
6.3 Interpretar e explicar resultados
Interpret and explain outputs
7.3 Divulgar produtos de difusãoManage release of
dissemination products
8.3 Estabeler plano de ação de melhoria
Agree an action plan
1.4 Identificar conceitos
Identify concepts
2.4 Especificar universo/amostraDesign frame and
sample
3.4 Configuar fluxos de produção
Configure work flows
4.4 Finalizar recolhaFinalise collection
5.4 Editar e imputarEdit and impute
6.4 Garantir confidencialidadeApply disclosure
control
7.4 Promover produtos de difusão
Promote dissemination
products
1.5 Avaliar dados disponíveisCheck data availability
2.5 Espeficificar tratamento e análiseDesign processing
and analysis
3.5 Testar sistema de produção
Test production system
5.5 Calcular variáveis derivadas e
novas unidadesDerive new
variables and units
6.5 Finalizar resultados
Finalise outputs
7.5 Gerir apoio a utilizadores
Manage user support
1.6 Preparar processo produtivoPrepare business
case
2.6 Especificar sistemas e fluxos
de produçãoDesign production
systems and work flow
3.6 Testar processo de produção
Test statistical business process
5.6 Calcular ponderadores
Calculate weights
3.7 Finalizar sistema de produção
Finalise production systems
5.7 Calcular agregadosCalculate
aggregates
5.8 Finalizar processamento
Finalise data files
Figure 2 – Flexible framework of the adopted model (adapted from the GSBPM V.5)
One of the strong points about having the GSBPM as a model is a consequence of
the multiple European projects that use the model as a
and research. Making use of the same language and process terminology it simplifies
the conception of the handbook itself.
The weak point was the matching process between the sub
and the organisational characteristics of the process organization, that proofed to be
very unbalanced between sub
The fact that SP had already a previous version of
and hierarchy made the task a lot easier, but still the GSBPM structure proofed to be
somehow unbalanced in what concerns the detail of each sub
For that reason, Statistics Portugal applied a third layer in a s
with the tasks of a called “production process matrix” that goes into detail in what
concerns tasks, responsibilities, related documentation and other to produce, as well
as the respective linkage with the SIGINE IT system, an in
standardized tasks for the statistical activities planning process.
Flexible framework of the adopted model (adapted from the GSBPM V.5)
One of the strong points about having the GSBPM as a model is a consequence of
the multiple European projects that use the model as a reference for development
and research. Making use of the same language and process terminology it simplifies
the conception of the handbook itself.
The weak point was the matching process between the sub-processes of the GSBPM
and the organisational characteristics of the process organization, that proofed to be
very unbalanced between sub-processes.
The fact that SP had already a previous version of the handbook with different layers
and hierarchy made the task a lot easier, but still the GSBPM structure proofed to be
somehow unbalanced in what concerns the detail of each sub-processes.
For that reason, Statistics Portugal applied a third layer in a systematic way, aligned
with the tasks of a called “production process matrix” that goes into detail in what
concerns tasks, responsibilities, related documentation and other to produce, as well
as the respective linkage with the SIGINE IT system, an in-house It tool system with
standardized tasks for the statistical activities planning process.
4
Flexible framework of the adopted model (adapted from the GSBPM V.5)
One of the strong points about having the GSBPM as a model is a consequence of
reference for development
and research. Making use of the same language and process terminology it simplifies
processes of the GSBPM
and the organisational characteristics of the process organization, that proofed to be
the handbook with different layers
and hierarchy made the task a lot easier, but still the GSBPM structure proofed to be
processes.
ystematic way, aligned
with the tasks of a called “production process matrix” that goes into detail in what
concerns tasks, responsibilities, related documentation and other to produce, as well
se It tool system with
This was a major development since the update of the handbook was done in parallel
with the update of the IT system itself. Meaning that the tasks are similar, enha
standardization in the process at all levels.
Figure 3 – General organization of the adopted model and its alignment with the GSBPM V.5
3. The subsequent levels of process description added to the GSBPM
It was important to keep a more detailed level of process description (3rd level),
featured on previous versions of the handbook, for different reasons:
− Mapping important activities throughout the business process, resulting from
the process modernization;
− Defining responsibilities;
− Identifying relevant reference documentation and expected outputs;
− Demonstrate the applicability of the business process model regardless of the
type of statistical activity.
As the example shows for a single task (Fig. 4), we can see its relation
GSBPM phase and sub-process, the task description, the units that are responsible
for it, and the applicability of the task taking into account the type of a statistical
activity.
This third level also describes its relation with the IT planning
feature a major development, also relevant in a changing environment.
GSBPM
Phases
Sub-processes
This was a major development since the update of the handbook was done in parallel
system itself. Meaning that the tasks are similar, enha
in the process at all levels.
General organization of the adopted model and its alignment with the GSBPM V.5
The subsequent levels of process description added to the GSBPM
keep a more detailed level of process description (3rd level),
featured on previous versions of the handbook, for different reasons:
Mapping important activities throughout the business process, resulting from
the process modernization;
ilities;
Identifying relevant reference documentation and expected outputs;
Demonstrate the applicability of the business process model regardless of the
type of statistical activity.
As the example shows for a single task (Fig. 4), we can see its relation
process, the task description, the units that are responsible
for it, and the applicability of the task taking into account the type of a statistical
This third level also describes its relation with the IT planning system, making t
feature a major development, also relevant in a changing environment.
GSBPM
Phases
processes
Adopted model
Phases
Sub-processes
Businesss process matrixMain tasks of each subprocess
(150 tasks)
5
This was a major development since the update of the handbook was done in parallel
system itself. Meaning that the tasks are similar, enhancing
General organization of the adopted model and its alignment with the GSBPM V.5
The subsequent levels of process description added to the GSBPM
keep a more detailed level of process description (3rd level),
featured on previous versions of the handbook, for different reasons:
Mapping important activities throughout the business process, resulting from
Identifying relevant reference documentation and expected outputs;
Demonstrate the applicability of the business process model regardless of the
As the example shows for a single task (Fig. 4), we can see its relation with the
process, the task description, the units that are responsible
for it, and the applicability of the task taking into account the type of a statistical
system, making this
feature a major development, also relevant in a changing environment.
Adopted model
process matrixsubprocess
6
Figure 4 – The business process matrix (example for task 30)
The Geospatial information infrastructure dimension along the business process can
only be mapped on the business process matrix, that is within the third layer of the
model. Within this dimension, there are 13 tasks altogether, belonging to different
phases, stages and sub-processes, which are explicitly connected to the spatial data
infrastructure:
− Specify and characterise the dissemination formats and products, including
geographic information;
− Evaluate the ability of the spatial data infrastructure responding to
methodological needs;
− Specify the functional requirements for data collection;
− Specify the functional requirements for the disclosure of geo-referenced
statistical information;
− Test the IT system supporting data collection and reception;
− Test the data disclosure components, including the GIS components;
− Refresh or geo-reference the interviewer's residence;
− Calculate distances between the statistical units and the interviewer's
residence;
− Propose the creation or correction of statistical units in the statistical registers
and propose updates to the samples defined in the sample management
systems;
− Analyze the proposals for updating samples and creating or correcting
statistical units on the sample management systems and statistical registers
and make the respective adjustments;
Ce
nsu
s
Sam
plin
g
Collection instrument
30
Designing or reviewing the collection
instruments, specifying the attributes'
validation rules and the type of testing
to be applied
DMDRI
DMSI
New
OngoingYes Yes Yes
Primary
statistics
The output of this task is included in items V.5.1, VI of the methodological document
Collection instrument: survey form; electronic file register; administrative form
SIGINE tasks:
301 - Support for the designing, building and testing of the questionnaire
401 - Designing, building and testing of the questionnaire
Interrelated tasks from other sub-processes:
32 - “Formulating the instructions for the correct filling of the collection instrument” (3.1)
33 - “Finalizing the collection instrument” (3.1)
Design 2.3 Design collection
Span of
activity
New /
Ongoing
Survey
Ad
min
istr
ati
ve
sou
rce
Type of
activity
Primary /
Derivate
Comments GSBPM phase GSBPM sub-process
Tas
k n
um
ber
Task description
Res
po
nsi
ble
d
ep
artm
en
t
Par
tici
pan
t d
ep
artm
en
t
Applicability
7
− Prepare the indicators to be made available in the dissemination database and
associated metadata;
− Upload and update data on the dissemination database;
− Compose the publication or develop other dissemination products.
4. Next steps
− Regular updates to the description of the business process are expected,
following the streamlining of procedures;
− Contributions to model simplification are deemed useful, specifically in what
concerns the more detailed levels of the production matrix;
− Expanding the mapping of documentation to be produced in the context of
each task;
− Producing and standardizing reference documentation that is central to the
business process (work instructions, guidelines, templates);
− Describing and standardizing over-arching processes with a statistical
component (e.g.: quality management, metadata management).
5. References
UNECE (2013), Generic Statistical Business Process Model - Version 5.0. Available
at: https://statswiki.unece.org/display/GSBPM/GSBPM+v5.0
SP (2017), Manual do Processo de Produção Estatística (internal document).
SP (2010), Manual de Processos da Produção estatistica (internal document).
SP (1997), Manual de Procedimentos da Produção Estatística (internal document).