Abstract:
|
Ken Brewer joined the Australian Bureau of Statistics (ABS, then the Commonwealth Bureau of Census and Statistics) in 1954 and left it in 1974. Over those twenty years he was either directly or indirectly responsible for many of the new methodological developments at the ABS. In this talk I will focus on aspects of Ken’s research at the ABS that, between 1970 and 1975, underpinned what was, to the best of my knowledge, the very first attempt to build a complete system for specification, design, implementation, analysis, and maintenance of a program of official sample surveys based on population modelling ideas. Given the limited computational power available at the time, the Generalised Survey System (GSS) represented a massively optimistic undertaking, and it was no surprise that it never achieved full operational status. In this talk I will provide some personal insights into the GSS, its links to Ken’s ideas and its continuing influence on ABS survey methodology.
|