Affiliations: Bureau of Statistics, International Labour Office, 4 Rue des Morillons, CH-1211 Geneva 22, Switzerland
Note: [1] Based on a paper originally presented at the First Conference of the International Association of Official Statisticians, Rome, October 1988, and included in its Proceedings. The permission of the International Statistical Institute to make use of the paper in this form is gratefully acknowledged. Some of the modifications were inspired by the discussions at the ECE/CEE Work Session on Registers and Administrative Records for Social and Demographic Statistics, Geneva, 23–25 January 1995. The article can be seen as a first attempt to create a framework for the forthcoming ILO work to develop guidelines for the use of administrative registrations as basis for labour statistics. However, the views expressed are those of the author and do not necessarily reflect those of the ILO or its Bureau of Statistics.
Abstract: This paper argues that there are both inherent and operational issues of quality associated with the use of administrative registrations as basis for official statistics, with respect to coverage, timeliness, frequency, validity, reliability and consistency. It further argues that because increased indirect and direct use of administrative registrations seems inevitable, the official statisticians must apply and adapt the established practices and principles of their trade to get to know the data generating process in detail, to monitor the data collection process, to try to persuade the responsible agencies to make changes which lead to improvements in data quality, to calibrate the observations generated by the administrative registrations by the use of statistical surveys, to use the administrative data and the calibration results as a basis for estimating the statistical parametres to be published, and to explain to the user how to properly interpret the resulting statistics. Procedural and inherent weaknesses of administrative registrations can only be overcome by developing and using appropriate methodological counter-measures.
