CENSUS QUALITY

In the field, the completed questionnaires were checked for completeness, accuracy and consistency before they were dispatched to the regional census offices. Supervisors together with interviewers did this. After the enumeration exercise, completed questionnaires were coded and again checked for errors by coders/editors together with regional and constituency supervisors at the regional offices. If any serious errors were detected, field revisits to the relevant households were made.

The demarcation of the whole country into enumeration areas was intended to ensure that everybody in the whole country would be enumerated. However, for one reason or another, it is possible either to over-count or under-count the population. The Post Enumeration Survey (PES), which was conducted, aimed at providing a measure of the degree of coverage as well as an evaluation of the overall quality of the information collected.

Post Enumeration Survey (PES) Objectives

The main objectives of the Post Enumeration Survey are:

Census Coverage Errors

There are two main types of coverage errors. These relate respectively to under-coverage and over-coverage.  Under-coverage errors occur when persons who should have been enumerated in the census are missed or the completed questionnaires relating to them are misplaced or lost.  On the other hand, over-coverage errors are caused by mistaken inclusions, such as multiple enumerations of the same persons and the enumeration of persons who were not in the country during the Census Reference Night.

Under-coverage errors may be an outcome of one of the following situations:

Over-coverage is likely to occur when:

The latter case may occur when the concept of the census reference night is not clearly understood by the respondents, or the interviewer fails to pose this question properly.

For this census, the PES was limited to the household population.  Under-coverage for the special population groups like institutions and the homeless was not included.  It is assumed that these are relatively small and with high mobility.  The cost of including them in the PES is not commensurate with their likely contribution to the coverage error.

Census Coverage

The Post Enumeration Survey covered a representative sample of enumeration areas, which was statistically selected in all the regions. Out of the 4042 enumeration areas, only 315, about 8 percent, were selected for this purpose. All households in the selected areas were enumerated.

Initial results, from manual compilations, indicated that the national coverage rate of the census enumeration was 97.8 percent of all the households. This implies that, on average, about 98 out of every 100 households were enumerated. The regional coverage ranged from 92.6 to 99.7 percent respectively for Karas and Caprivi regions. A breakdown by region is given below.

It should be noted that, although all possible measures are put in place to achieve 100 percent coverage, it is not possible for any census operation to achieve this goal. Usually households are missed due to the inability of the interviewers to find the household members at their homes.

 Enumeration Coverage Rate by Region, 2001 Census