Abstract
The big advancement in technology and the availability of 'big data', pose new demands for more detailed, more accurate and timelier official statistics. The target of the present article is to discuss some of the major new methodological challenges underlying the production of official statistics (POS) in coming years, and in some cases suggest ways of dealing with them. In particular, I consider the following challenges: collection and management of big data for POS with integration of computer science, increasing data accessibility but maintaining privacy and confidentiality, possible use of web-panels for POS, how to deal with mode effects, measurement of error in small area estimation in conjunction with modern censuses, and integration of statistics and geospatial information. In the last part of the article I confront the question of whether universities train students to work at National Statistical Offices.
Original language | English |
---|---|
Pages (from-to) | 425-467 |
Number of pages | 43 |
Journal | Journal of Survey Statistics and Methodology |
Volume | 3 |
Issue number | 4 |
DOIs | |
State | Published - Dec 2015 |
Bibliographical note
Publisher Copyright:© 2015, American Association for Public Opinion Research. All rights reserved.
Keywords
- Big data
- Coverage bias
- Disclosure control
- Integrated census
- Mode effects
- MSPE in SAE
- Propensity scores
- Sample distribution
- Web panel