Systematic quality work in official statistics Theory and practice

1
Systematic quality work in official statistics
Theory and practice
Hans Viggo Sæbø, Statistics Norway hvs_at_ssb.no
2
(No Transcript)
3
Issues

• Systematic approach to quality has been focused
• Principles and frameworks
• Measures
• European Statistical System (ESS)
• Statistics Norway
• Developing countries
• Experiences and the way forward

4
Model for Total Quality and Code of Practice
The point of departure for
Users
systematic quality work is
Perceived quality
the "user needs"
Needs and effects
The users demand "product quality"
Relevance
Accuracy and Reliability
Statistical Output
Timeliness and Punctuality
Coherence and Comparability
Product quality
Accessibility and Clarity
Study of processes is a precondition
for improvements
Statistical Processes
Sound Methodology
Appropriate Statistical Procedures
Process quality
Non-Excessive Burden on Respondents
Cost Effectiveness
Structural conditions provide a framework
Professional Independence
Institutional Environment
Mandate for Data Collection
Structural quality
Adequacy of Resources
Quality Commitment
Statistical Confidentiality
Impartiality and Objectivity
5
European Statistical System (ESS)

• Self assessments and practice Example
• 75 percent of European NSIs claim that they have
  systematic documentation schemes on web
• But in practice, only a few have such
  documentation linked to statistics!
• www.ssb.no/english/
• In European Code of Practice questionnaire (2005)
  it turned out that quality commitment indicators
  such as existence of processes to monitor quality
  (of processes) only exist to a small extent. But
  (on the top level) almost everyone is committed
  to quality!

6
Measures in Statistics Norway

• Information and training
• Management by performance indicators
• User surveys and service declaration
• Project and teamwork
• Current Best Methods (CBMs, e.g. handbooks)
• Staff surveys
• Documentation and metadata
• International cooperation

7
Challenges in Statistics Norway

• Systematic quality work should be more systematic
• Systematic quality work should be integrated in
  all planning and follow up routines
• Process variables are seldom measured
• CBMs (e.g. handbooks) are not known and followed

8
Challenges in developing countries

• Same challenges as in Norway, but more evident
  that quality efforts such as documentation tend
  to lose in the (normal) situation with scarce
  resources
• Organisational problems are more visible
• User dialogue is more limited
• Development projects are often highlighted and
  given priority, this is a challenge also for
  donors and advisers!

9
Conclusions - Experiences

• Systematic quality work is not always systematic
  enough too much emphasis on development
  projects?
• Identifying and measuring (key) process variables
  for monitoring, control and follow-up is
  difficult
• Even if CBMs exist, they are not well known and
  even less followed
• The quality organisation (coordinators and
  facilitators) often lack sufficient authority
• The needs of users that do not directly fund
  statistics (i.e. the general public) tend to be
  neglected
• Development of new statistics is given priority
  on the costs of quality assurance and
  documentation

10
The way forward

• Self-assessments should be followed by external
  reviews.
• Important activities and projects should be given
  priority by quality facilitators
• More pressure should be put on identifying and
  measuring key process variables for the most
  important activities and projects. These
  variables should be used by management in
  planning and follow-up
• Activities to promote and ensure the use of CBMs
  should be carried out
• The quality organisation should be linked to top
  management
• Attention should be given to all users, for
  example by user surveys and user groups with
  broad representation.
• The issue of different user needs and the need to
  balance quality versus new developments should be
  discussed with authorities funding statistics

11
Final remarks

• Concentrate on core principles, not on models!
• Management!
• High vision, but small concrete steps!

EFQM
TQM
Risk management
BSC
CAF
6sigma
ISO
12
Systematic quality work in official statistics
Theory and practice
Hans Viggo Sæbø, Statistics Norway hvs_at_ssb.no