Calculating the undercount in Census ’96
Contents 2. The post-enumeration survey 4. Characteristics of undercount The data contained in this report were gathered in October 1996. Since then, there have been demographic changes in South Africa associated, inter alia, with internal and external migration, and population growth. This means that population profiles may have changed at differing geographic levels. Stats SA is not responsible for any damages or losses, arising directly or consequently, which might result from the application or use of the data gathered as part of the 1996 population census. The data have been evaluated and reviewed by a task team of the Interim Statistics Council. The Council’s statement and the task team’s report are available from the Statistics Council through Stats SA. Census ’96 was statistically adjusted for undercount on the basis of a nationwide post-enumeration survey, instead of being brought into conformity with a population-projection model and demographic assumptions reaching back 20 years, as was the case with the 1991 census. Consequently, Census ’96 results differ in some respects from those based on the projections of the model previously used, including:
Copyright notice: Information contained in this publication may be freely cited, provided Statistics South Africa (Stats SA) is acknowledged as the source of the original information; and that the information cited or otherwise utilised is not sold or offered for sale in any form whatsoever without the prior permission of Statistics South Africa. Introduction During October 1996, the first census of the new South Africa was conducted. Every person present in South Africa on census night (9-10 October) should, in principle, have been enumerated on a census questionnaire between 10 and 31 October. However, it is inevitable that some people will, in practice, not be reached by any census. In a regionally diverse, unevenly skilled, ethnically heterogeneous country like South Africa, and with a degree of political suspicion towards the new government in different areas or among different groups, the proportions of people not enumerated will vary in intricate ways. The challenges were heightened in Census ’96 by the short time (slightly more than a year) available for planning and implementation of the fieldwork. The post-enumeration survey (PES) was conducted immediately after the census on a countrywide basis to measure the proportion of the population not reached in Census ’96. This was the first time that a post-enumeration survey of this scope and nature was conducted in South Africa. This report details the methodology of the PES, notes some of the problems encountered and how they were addressed, and presents the estimates of undercount. |