Back to contents index

During the pre-enumeration (including demarcation) and enumeration phases of Census ’96, it was not possible to map EAs and their boundaries onto a geographical information system (GIS) because of funding and time constraints. The definition of EA boundaries for Census ’96 was largely based on a combination of the following sources: hard-copy, conventional maps, aerial photographs, sketch maps and text descriptions.

However, after enumeration, it became possible to consolidate and formalise demarcation, by means of a GIS. The Independent Electoral Commission gave the financial support to Stats SA for such a system, since it needed both census and mapping information to be overlaid to prepare for the forthcoming elections in 1999. Stats SA in turn collaborated with the Department of Land Affairs, which was refining the computerised ‘backdrops’ of land parcels upon which the EA boundaries would be displayed.

Consequently, during the first quarter of 1997, Stats SA began to convert analogue enumeration area boundaries into digital format. The digitising of the 1996 census EA boundaries was done through ‘Project Eagle’.

The main aim of Project Eagle was the production of an integrated geographical database for the whole country. The forms and maps that were the original source of information for the demarcation phase of Census ’96 were also the starting point for Project Eagle. The forms contained the following information: magisterial district and code, EA number and type, name of the city, town or tribal authority, name of the suburb, tribal ward, administrative area, village or settlement and description of the EA boundary. Project Eagle used this information to start to digitise EA boundaries and to form polygons for the GIS. Stats SA is now, for the first time, in possession of a spatial EA database linked to a detailed attribute database that covers the whole of South Africa.

Project Eagle contributed significantly to the outputs of Census ’96. It was able not only to streamline the EA boundaries used in Census ’96, but also to provide a spatial backdrop for reporting census results. It also enabled Stats SA to investigate whether any entire EAs had been excluded from both the census and the PES.

For the purposes of this report, the main interest is the contribution Project Eagle made to solving problems concerning possibly-missed EAs in the census. Readers who require additional information on the management and conduct of Project Eagle can obtain documentation from Stats SA.

During the demarcation phase of Census ’96, the entire country should have been covered and divided into EAs. In some areas, however, this was not possible due to time constraints, and vacant land between settlements and villages in previously undemarcated areas was sometimes left uncovered. On the other hand, Project Eagle covered the entire country, including vacant land. As a result, there were more EAs in the GIS of project Eagle (about 94 000) compared to the census (about 86 000). The reason for this occurrence was investigated.

• The first step in the investigation involved identifying the EAs that were demarcated, but no questionnaires were processed, as well as the EAs that were not originally demarcated.

• The next step was to determine whether or not these EAs were populated at the time of the census. To do this, the EA was selected on the GIS database, and the background map was used to determine whether or not it was populated at the time of Census ’96. In cases where it was clear that an EA was not populated at the time of the census, the EA was classified as a zero EA. In cases where there was doubt or where it looked likely that there were people in a particular EA at the time of the census, the matter was referred to the provincial office, where the GIS information was checked against source documentation and maps and records from the data-processing centres.

• All EAs that might have been populated at the time of the census, but for which there was no record of a count, were identified in this manner.

• A sample of 167 of these EAs was drawn, and the households in these sampled EAs were visited. A short questionnaire was administered asking respondents whether or not the settlement existed at the time of the census, and whether or not the people in the EA were counted during Census ’96.

There were 94 268 EAs on the GIS database, constituting South Africa. But the census output database which is linked to the GIS contained only 86 200 EAs. This means that there were 8 068 EAs without census information linked to them. Of these 8 068 EAs, a total of 3 544 were classified as EAs with no population or zero EAs.

Stats SA therefore had to investigate the situation in the remaining 4 524 EAs. It did so by means of the sample of 167 of these EAs that was drawn. The fieldwork indicated the following:

• Thirty percent of these EAs were empty at the time of Census ’96, and were still empty at the time of checking.

• Another 5% were empty at the time of the census, but were populated at the time of checking. The people who had moved in after the census indicated that they had been counted elsewhere.

• In another 57%, the vast majority of the people who were living in them said that they were counted in Census ’96. They were therefore not reflected as living in the correct EA, even though they were counted.

• Furthermore, 6% of the EAs in the sample could not be identified on the ground, even though they existed on the GIS. They require further investigation.

• Two percent of EAs in the sample were commercial farms in North West, and the fieldworkers were unable to gain access to them.

From this exercise, it seems that those areas which were not demarcated in the census, according to the GIS, were either empty EAs, or else they were included in the census as part of other EAs. It is unlikely that any populated EAs were entirely missed during Census ’96. People or households that were missed were taken into account through the PES.

The GIS should, in future, make it easier to recognise EAs on the ground, and to identify their boundaries, particularly if ongoing maintenance work and updating is done.