Nordbotten and Denmark

1. Nordbotten and Denmark Lars Thygesen lth@dst.dk

2. 1960 1966 1964 Nordbotten in Nordic Statistical Meeting Helsingfors Nordbotten in Nordic Chief Statisticians Meeting Copenhagen Ohlsson in Nordic Statistical Meeting Copenhagen The impact of Nordbotten on statistical thinking in Denmark

3. Ingvar Ohlsson lectures, Copenhagen 1964

4. Poul Jensen Aage la Cour Kjeld Bjerke Henning Friis Erling Jørgensen Axel Holm

5. Milestones

6. Milestones #2

7. The basic ideas of Nordbotten – how they were implemented

8. Educa- tion Employ- ment CPR Person id: Person Number Inter- view Tax Question- naire Social Health etc Dwelling id: Address Enterprise id: CBR-No Cadastre BDR CBR The notion of a Statistical Information System

9. Reuse of data for multiple purposes • The Sample Archive from 1971 (originally the Mini Population Register) • a representative sample of the population from a variety of sources and all kinds of social and population statistics • longitudinal • aim: trying out different combinations of data, especially for analysis • a test bed for new combinations • The Law Model from 1980 onwards • aim: calculating the effects of contemplated legislation • contains micro data on a sample of the population • in cooperation with the Ministry of Economic Affairs • up and running 24 hours a day, 7 days a week

10. Law Model “...and when our expert has coded the proposed law changes into the terminal, we will in a few seconds witness the computer calculating the effects for a typical family - - based on a simulated model population”

11. Consequences of archive method • Data collection not linked to use • Key purposes are foreseen • Almost anything can be made on demand

12. SumDatabase Statistical registers - The filing or archive structure Administrative registers Web XML Paper Cleaning Estimation Aggregating Analysis Metadata Publishing StatBank Denmark dst.dk

13. Micro data and longitudinal analysis Elsebeth Lynge: Occupational mortality 1970-75 Researchers’ remote access: 300 institutes 3,000 researchers 400 active projects 2 M€

14. Confidentiality and data protection • Challenge: Keep public trust • data security regulations • strict organisational and technical measures • limiting internal and external access to data • need to know principle • 1978 Register Law

15. To link or not to link - that’s the question! The case of Denmark

16. Difficulties in practical implementation • Convincing ourselves • Champions of traditional statistics • University textbooks • Strong management commitment • Convincing the rest of the world • Suspicion • Bad experience history, especially WW2 • Survey is the truth • Jørgen Wedebye 1979: “On what grounds can anyone claim that there has ever been one single piece of correct information recorded on a Census form?” • International meetings, Eurostat book

17. Difficulties #2 • Confidentiality and access to data • The threat of a “Big Brother” society • Lawsuit against Statistics Denmark in 1978-81 by five municipalities • Data breaks • changes in the legislation • difficult to assess long-term trend • identify events

18. Difficulties #3 • Functional organisation • Nordbotten:“Today data for the survey of manufacturing and the wage statistics are collected separately. It will probably in many cases be more efficient to integrate the collection, i.e. collect as much as possible from a respondent when we have made a contact instead of repetitive visits or requests” • slow progress • Metadata • allow user to determine whether the data in the archive could be recombined to create new knowledge, e.g. to confirm or reject certain hypotheses • Internal as well as outside users claim this is the most important hindrance

19. Nordbotten 1960: “In 1970, there will probably not be any need to collect a census in the traditional way”

21. The End!