Understanding the 1996 census migration data

The Australian Census provides two approaches to measuring migration: indicators which distinguish movers from non-movers, and a geographic classification which identifies each person’s usual residence on census night, and their usual address one year and five years previously. Although these data represent a rich source of information, they contain several traps for the unwary. We show that differences in the variables and classifications used can result in marked variations in the apparent intensity and patterns of migration. The questionnaire and processing methodology used in the 1996 Census also resulted in a number of inconsistencies between the migration indicators and the usual address information. We examine the magnitude and source of these anomalies, assess their implications and propose a partial solution. The views expressed in this paper are personal and do not necessarily reflect those of the Australian Bureau of Statistics.