Indicators for Managing and Improving the Data Collection Process

1
Indicators for Managing and Improving the Data
Collection Process

• Sindre Børke and Jonas Dahl,
• Division for Data Collection Methods,
• Statistics Norway

2
Looking around

• Eurostat focus on quality issues through several
  years
• Statistics Norway
• Quality Issues at Statistics Norway
• Systematic Quality Work in official Statistics
  Theory and Practice
• FOSS A Standardisation Programme
• Development of standardised working prosesses,
  methods and systems
• A system for systematic quality measurements and
  control
• Organisation and human resource development
  supporting this

3
Defining a Project

• Information for Different Levels of Management
• Continuous Quality Improvement
• Part of the Statistical Value Chain
• Questionnaire-based Data Collection, and Editing
  on Micro Level
• Limited Number of Indicators

4
Indicators

• Specific and measurable elements of statistical
  practice
• Defined by parameters
• Representative for the component it indicates
• Easy to interpret
• (Easy to collect data)

5
Indicators definitions
6
Groups of Indicators

• Illuminating stages in the value chain
• choice of instrument(s) and designing
  questionnaires
• respondent behaviour (incl use of
  support-telephone and e-mail)
• progress in data collection (effect from
  reminders etc)
• non-response
• the merging of data in multi-mode design surveys
• controls and corrections

7
Sick leave questionnaire respondents telephone
calls
8
Groups of Questionnaires

• Not all indicators are relevant on all
  questionnaires
• Use of indicators means comparing
• Clustering of questionnaires to create meaningful
  comparisons (benchmarking)
• Year, Quarter or Month
• Business or Huseholds/Persons
• Interview/selfadministration
• .

9
Response rates development over time
10
Recurrent surveys

• Comparisons/development over time
• Expected results of changes in questionnaire or
  process
• Stability in processes

11
Hotel statistics questionnaire Electronic Data
Delivery
12
Summing up

• No standards set, but observing level in each
  indicator
• Following recurrent surveys over time,
  identifying changes
• Benchmarking comparable surveys

13
Experiences so far

• Split up is necessary to establish relevance
• Small steps on the value chain
• Groups of comparable questionnaires
• Criterias for clustering are not established
• Some interesting indicators will be difficult
  (expensive) to establish
• Project challenge to keep focus on definitions,
  leaving the interpretation and action to others

14
The near future in the project

• Reformulating targets and ambitions
• Some more results in demo-version
• Defining indicators
• Clusering questionnaires
• Data collection design for defined parameters
  (implementing project results)