Data quality of the 2002 population and housing census: a case study of Dar es Salaam Region, Tanzania

The study evaluate the data quality of 2002 Population Census results on timeliness and reliability in terms of coverage and content of Dar es Salaam region. The study premises focused on evaluating data production process and the significant difference in magnitude of sampling errors between national and sub-national level The findings were based on information gathered through documentary reviews, semi- structured interviews to stair involved in 2002 Census and Post Enumeration Survey (PES). The bulk of data were analysed using matching process and computation of coverage and content errors. The study established that 2002 census results were found to be of high quality on timeliness and reliability though cartographic work was poorly done. The improvement of timeliness was influenced by the utilisation of scanning machine and computers. While coverage and content of households, age and sex were found to be relatively better and acceptable though a bit lower than of national coverage. Therefore, it is recommended that future censuses should use well trained cartographers employed by the Government, as was in the 1978 and 1988 population censuses, to do cartographic work and the employment of field assistants should be done at district level. This will help to get appropriate field assistants who know the geographical area of a district. It is also recommended to consider monthly wages instead of piece work bases to pay field assistants as were in the 1978 and 1988 censuses to avoid unfaithful field assistants to delineate poor maps, each district/municipal of Dar es Salaam region should be treated as a region in order to reduce workload of the Regional Census Coordinator and utilise the GPS to record coordinates of delineated EA and document in the GIS software designed for drawing maps