The History of census taking in Botswana is traced back to the beginning of the twentieth century. The first census of the then Bechuanaland Protectorate was planned for 1901 but due to the perpetuation of the Anglo-Boer war it was held in 1904. It was conducted on a de-jure basis and the census night set to be the 17th of April. Being a first census, the 1904 census was characterized by many difficulties such as fear and suspicion on the part of the population, poor communication and the fact that the country was large and sparsely inhabited. Nonetheless, the exercise was considered a good beginning and that the insight gained would make the following censuses easier.

With no wars or disturbances in the Anglophone Africa, the Bechuanaland Protectorate joined the rest of the British Empire in conducting the 1911 and the 1921 decennial censuses. Although it was hoped that later censuses would be easier to undertake, the same administrative difficulties experienced in the 1911 census were still prevalent during the 1921 census. In addition to these difficulties, the administrators had to contend with inaccessibility to remote areas, fear and prejudice from the people as well as poor communication link with some areas.

In view of the worldwide economic depression of the late twenties and early thirties, there was no census in 1931. The census was held in 1936 instead. The reliability of the 1936 Census results was questioned, mainly because the district figures were too different from the 1921 Census figures. This could be because either the 1936 census had gross over-enumeration or the 1921 had gross under-enumeration.

The 1904, 1911 and 1921 censuses were limited in subject coverage since most information collected was on the total number, broad age group and sex. The questionnaire for the 1946 Population Census was long and complicated, a factor which dragged the tabulation preparation on towards the preparatory time for the 1951 census. As a result the 1951 census was delayed until 1956. Likewise the 1956 census questionnaire was complicated. It is believed that there was a gross under-enumeration and since it took seven years to process, the census results had become outdated by the time they were presented to Government.

With the urgent need for information upon which to base development plans and accurate population figures for constituency delimitation, the 1964 Population Census was conducted. Although the 1964 Population Census was conducted on de-jure basis like the first six, it was the first to be on a house-to-house canvassing method. The basis of enumeration for the first six censuses was group enumeration while in the 1964 census enumeration was on the basis of allegiance to major villages.

Two terms on the basis for enumeration i.e. de facto and de-jure need defining. The de facto method enumerates the people who spent a predetermined night (census night) in a particular house or location. In the de-jure method on the other hand, the population is enumerated at their 'usual' place of residence, irrespective of where they actually were on the census day. Although the 2001 Census taking is on de-facto basis, by asking questions about the ‘place of usual living', the de-jure figures are derived. No attempt will be made to discuss the pros and cons of these two enumeration methods in this instance.

Following the attainment of independence in 1966 it was found necessary to return to a year ending with 0 or 1 as recommended by the United Nations hence the next census was re-scheduled for 1971. The 1971 Population Census was the first to be conducted on a de facto basis with the idea of enumeration area being applied for the first time in Botswana's census history. In terms of the quality of data obtained and the wide subject coverage, it was an important census for future planning. It also provided a basis for sampling in later surveys.

The 1981 Population and Housing Census was the second census to be taken since Botswana became independent. It was the first one ever to include a housing component in its questionnaire. In view of better cartographic support, effective publicity campaigns and stronger administrative backing as well as improved level of education, the 1981 Census could be considered as having had better coverage than the 1971 Census.

The 1991 Population and Housing Census was yet another success story in conducting decennial censuses in Botswana with a more comprehensive and wide coverage of topics. Cartographic work had greatly improved resulting in very few incidents of omission of localities during enumeration.

The 2001 Population and Housing Census is the fourth census to be held since independence, the third to collect housing information and the eleventh in the series of censuses conducted in Botswana. Unlike the 1991 Population and Housing Census, the 2001 Population and Housing Census is part and parcel of the SADC 2000 Census Project. The aim of this SADC initiative is to harmonize all census-taking activities and collect information on similar topics in the region. The SADC 2000 round of Censuses objective is for all SADC countries to cover similar topics for better comparability of demographic characteristics in SADC. The SADC member states agreed that there was a need to pull together resources in working towards a common goal and this resulted in a resolution being taken in support of the harmonization initiative among SADC countries. Countries are still free to add-in their own country specific topics to the regional minimal set.

Population Census