Sufficient Quality for Swedish Official Statistics

1
Sufficient Quality for Swedish Official
Statistics

• Mats Bergdahl, Statistics Sweden
• Eva Elvers, Statistics Sweden
• Ingegerd Jansson, Statistics Sweden
• Jens Olofsson, Örebro University
• Further participation three authorities
• responsible for official statistics - AROS

2
Outline

• Swedish System for Official Statistics
• Sufficient quality in what sense
• Criteria for sufficient quality
• The future

3
Swedish Statistical System

• 25 different AROS
• Statistics Sweden is one AROS, with a
  coordinating role
• The AROS decides which of its statistics that are
  official
• Council for Official Statistics for coordination
  laws, guidelines

4
Way of Working

• Task from the Council for Official Statistics
• 2 year period
• Thoroughly discussed
• Working Groups
• Conferences
• The Council for Official Statistics
• The Scientific Council
• Official Reference

5
Sufficient Quality in What Sense

• the quality is sufficient for how the
  statistics are used.
• the quality is sufficient for how the
  statistics are intended to be used.
• Manageable
• Under the authority's control

6
Types of Criteria

• Non-response below 20
• CV below 10
• Release monthly statistics within 6 weeks
• Reflections
• Significant variations between types of
  statistics
• Focus on factors of importance for quality
• Constraints for the production

7
Criteria for Sufficient Quality

• Support for the AROS and
• information to users and data suppliers
• Three groups
• Laws-Ordinances-Instructions
• User Contacts
• Design-Production-Evaluation

8
Laws-Ordinances-Instructions

• Obviously important
• Quality report/declaration mandatory

9
User Contacts

• The purpose of the statistics
• Documentation of important users, their uses and
  future needs
• The planned quality level is based on a dialogue
  with the main users
• Annual follow-ups of the statistical quality

10
Design-Production-Evaluation

• European Union regulations
• Classifications and standards
• Scientific principles

(to be ctd)
11
Design-Production-Evaluation (ctd)

• Existing sources of data
• Methods of data collection
• Distribution over respondents
• Estimation of time and costs
• Measurement instruments

(to be ctd)
12
Design-Production-Evaluation (ctd)

• Releases are announced in advance
• Regular quality studies
• Plan and implement for the sufficient level
• Review fulfilment of sufficient quality

13
The Future

• The Secretariat of the Council is the owner
• AROS can make a commitment
• Information
• Support
• Leading examples
• Review and revise

14
Sufficient Quality for Swedish Official
Statistics

• Contacts
• mats.bergdahl_at_scb.se
• eva.elvers_at_scb.se
• ingegerd.jansson_at_scb.se
• jens.olofsson_at_esi.oru.se