A NEW METHODOLOGY FOR PROCESSING AUSTRALIA 1991 POPULATION AND HOUSING CENSUS

Processing of Australia's 1991 Population and Housing Census was based on the use of optical mark recognition (OMR) for the capture of self-coded responses, and computer assisted coding (CAC) for the coding and capture of the remaining write-in responses, and for editing and amending the data. This combination was considered to be the most cost effective input processing methodology for the 1991 Census and reduced significantly the number of temporary staff required compared with that needed for processing previous censuses. Both OMR and CAC improved the quality of 1991 Census data by reducing processing errors. The strategy also provided for early capture of some of the census data during the OMR phase for use in timely revision to the population estimates and the early release of preliminary census results. This paper describes the methodology used in processing Australia's 1991 Census and discusses new methodologies for processing future Australian censuses.