University of Botswana History Department

A report on Afrobarometer Botswana 2005

by B. S. Bennett

History Home Page  ||  Site Index  ||  B. S. Bennett home page


The recent Afrobarometer Survey of Botswana has produced some very interesting data. This survey, conducted in May-June 2005, surveyed a sample designed to be nationally representative. For example, it was geographically representative, being based on the Census Enumeration Districts.

In the following figures and tables, note that because numbers have been rounded to whole numbers, totals will not necessarily always add up to 100%, and similar small apparent anomalies may occur.

The tables appear by kind permission of the compilers of the Afrobarometer report:

Afrobarometer: Botswana Round 3: 2005

compiled by M.H Lekorwe, M.G. Molomo, W.B. Molefe, D. Sebudubudu, L.L. Mokgatlhe and K.K. Moseki.

The complete report can be downloaded free (see end of page.)

One should begin by noting that Batswana are cautious people. Asked "Generally speaking, would you say that most people can be trusted or that you must be very careful in dealing with people?", 94% said "You must be very careful" and only 6% "Most people can be trusted". Interestingly, urban people were more trusting than rural people, by 8% to 5%. (Table 83.) 38% trust their relatives "a lot" but 20% trust them "not at all". (Table 84a.). "People have to be careful of what they say about politics" (see below), and about half those surveyed thought that the survey was being done by the Government (table 100); so perhaps a little caution is in order about some of the answers. Nevertheless, this survey is a treasure trove of information, and deserves to be read with close attention. I am not going to attempt a general summary; I am simply going to summarize some of the findings that particularly interested me. I urge readers to download and read the full report. I may add more comments later.

Back to contents

Language and Ethnicity

The survey included questions on sensitive topics such as language and ethnicity. For readers unfamiliar with Botswana, note that most of the answers have been given in a Setswana-prefixed form; thus Kalanga, or Ikalanga, is given as SeKalanga (Se- being the prefix normally used for languages in Setswana). [See our page on place names and historical terms for a basic explanation of Setswana noun classes which will hopefully make things clearer.]

The results show that Setswana is dominant as a home language, possibly having displaced other languages significantly during the twentieth century:

  Urban Rural Male Female Total
English 0 0 0 0 0
Setswana 88 73 79 79 79
Sesarwa [1] 0 3 1 2 2
Sekgalagadi 2 5 4 4 4
Sesobea [2] 1 0 0 1 1
Sekalanga 7 11 9 9 9
Seherero 1 1 1 1 1
Sembukushu 0 3 2 2 2
Sebirwa 0 4 2 2 2
Sengologa 0 1 0 1 1
Seyei 0 1 1 1 1

Afrobarometer: Botswana Round 3 2005: table 3.

Note 1.  Note that "Sesarwa" is not actually the name of a single language, but indicates any of the Khoesan languages spoken by the peoples usually known in Botswana as "Basarwa".

Note 2.   Sesobea: more commonly spelt Sesubiya.

The distribution shows that Setswana is dominant in urban settings. The minority languages are all mainly rural. A total of 86% speak Sotho-Tswana languages (Setswana 79%, Sekgalagadi 4%, Sebirwa 2%, Sengologa 1%.)

The related Q. 79 asked "What is your tribe? You know, your ethnic or cultural group?" Again, readers unfamiliar with this part of the world should note that the answers all have the Setswana Mo- prefix.

  Urban Rural Male Female Total
Mokgatla 11 8 9 9 9
Mokwena 11 8 10 9 9
Mongwato 21 12 17 15 16
Mongwaketse 11 8 9 9 9
Motlokwa 0 1 1 1 1
Moherero 1 1 1 1 1
Morolong 3 3 3 3 3
Mosarwa 0 3 2 3 2
Mokalanga 14 16 16 15 15
Molete 3 5 4 4 4
Mmirwa 3 4 4 4 4
Motswapong 7 7 6 7 7
Mosobea 1 0 0 1 1
Motawana 2 1 2 2 2
Mokgalagadi 3 10 7 7 7
Moyei 2 2 2 2 2
Mohurutshe 2 2 2 2 2
Mombukushu 0 3 2 2 2
Monanjwa 0 0 0 0 0
Mokhurutshe 1 2 1 2 1
Motalaote 0 0 0 0 0
Motswana only, or doesn’t think of self in those terms 4 3 3 3 3
Refused to answer 0 0 0 0 0
Don't know 0 0 0 0 0

Afrobarometer Table 79.

For readers unfamiliar with Botswana history, perhaps I should say something here about the problematic nature of the term "tribe" in Botswana. When the British took over the territory, they took over a cluster of kingdoms, all of which were ruled by Tswana dikgosi but which varied in their composition. In he south-east the population was mainly Tswana, but to the north and west it was much more mixed. Under British rule these states became Reserves, in which the dikgosi were Paramount Chiefs. The point is, though, that the "Batawana Tribe" [Batawana, not Batswana] of the North-West was a political entity, in which the Batawana, as an ethnic group, were a minority. Most members of the "Tribe" spoke other languages. A "tribe" was multi-ethnic. Hence Isaac Schapera's famous study The Ethnic Composition of Tswana Tribes (1952), which sounds like a paradox to those who think of a tribe as meaning an ethnicity. Within a "tribe" such as the Bangwato (then more usually called Bamangwato) there were not only the core Ngwato group, whose totem was phuti, but other Tswana groups such as BaKaa, less closely related Sotho-Tswana groups such as BaBirwa, and non-Sotho-Tswana such as Kalanga.

Since independence and the decline in the power of dikgosi, however, "tribe" has tended to be understood in the sense of "ethnic group", and the question seems to make this sense explicit by referring to "your ethnic or cultural group".

Comparison with Schapera's analysis of the 1942 census is not entirely straightforward. I have some tentative ideas, but will write more on this later.

Taking the Afrobarometer results, we can group the numbers thus:

We thus have three large categories, and "other" :

Other Sotho-Tswana:18%
Other responses:10%

The proportions shown in the survey for the various groups among the "Other responses" may not be reliable as the numbers are too small in relation to the survey, so I will just leave them as "Other responses" in this grouping. (It is interesting to note that although the survey found 1% speaking Herero and 1% speaking Sesobea, neither "MoHerero" nor "Mosobea" show up as ethnic identities. Perhaps they are among the 4% who answered "Motswana only, or doesn't think of self in those terms".) Grouped as above, the proportions are shown in the following pie chart:

Botswana ethnic identities

(Chart produced using OpenOffice 1.1.4)

It is notable that Bangwato are the most urban ethnicity. Groups such as BaKgalagari and Basarwa are mainly rural. The survey indicates that this is also true of Hambukushu, but not of WaYei, but the sample is too small to be confident of these results.

I expect that there will be debate over the reliability and interpretation of these results. Comments are welcomed!

Back to contents

Politics and public affairs

73% of Batswana are interested in public affairs, indeed 44% are "very interested". The level of interest is roughly the same between town and country. (Table 16) But as a topic of conversation, 38% never discuss politics and 39% discuss it occasionally. Only 23% discuss it frequently. (Table 17.) This may reflect a lack of confidence: 65% agreed that "Politics and government sometimes seem so complicated that you can't really understand what's going on" while only 21% disagreed (Table 18a.).

Radio, television and newspapers are all significant sources of news. Radio remains dominant: 44% listen every day, and another 32% listen a few times a week, making 76% frequent listeners. Newspapers seem to be read on an intermittent basis: 23% read a newspaper every day, but another 26% read one "a few times a week" (thus 49% per week) and another 14% "a few times a month". Perhaps the most striking result is the degree to which television news has already become an important source.

Table 15b. How often do you get news from the following sources: television?

  Urban Rural Male Female Total
Never 25 60 42 48 45
Less than once a month 3 5 4 4 4
A few times a month 8 8 9 8 8
A few times a week 21 12 17 15 16
Every day 43 15 28 26 27
Don't know 0 0 0 0 0

Afrobarometer Table 15b.

36% of the population own TV sets, 51% in urban areas and 24% in rural areas (table 93c.) In many places, people will congregate to watch news reports, so television news reaches a wider group than those who actually own TV sets. BTV (state-owned public terrestrial television) now covers most of the country, and broadcasts news in Setswana and English. It should also be noted that after local programmes finish BTV transmits the BBC World news channel, so anyone with BTV will have access to this, albeit at a late hour. Satellite TV is also relatively widespread.

The report indicates that ordinary people experience low levels of corruption, a finding which was understandably given prominence in local media reports of the Afrobarometer report's release. Questions on party support show that the ruling BDP (Botswana Democratic Party) retained its clear lead at the time the survey was taken. Omitting the parties whose support did not show up at all, the results were:

  Urban Rural Male Female Total
Botswana Alliance Movement (BAM) 1 1 0 1 0
Botswana Congress Party (BCP) 9 9 9 9 9
Botswana Democratic Party (BDP) 49 56 49 57 53
Botswana National Front (BNF) 24 24 29 20 24
Botswana Independence Party (BIP) 0 0 1 1 1
Botswana People's Party (BPP) 1 0 0 0 0
Would not vote 5 4 4 5 4
Refused to answer 4 1 3 2 2
Don't know 7 5 6 6 6

Afrobarometer Table 99

This indicates that, as was already known, the BDP is relatively stronger in rural areas and among women. The extent of the gender gap is perhaps surprising: there are three male BNF voters for every two female.

Despite this high level of general support for the government, there is questioning of some policies. Some questions attempted to probe attitudes to the trend of government policy. Most (55%) agreed that it was necessary to suffer some hardships now for future benefits, 41% thought the price too high. However it was thought that the policies had not brought results yet: 60% agreed that "The government's economic policies have hurt most people and only benefited a few.", only 32% thought most had benefited. (Table 11).

There was strong opposition to the re-introduction of school fees. The survey posed the question in a form which gave maximum weight to the claims of school fees:

Which of the following statements is closest to your view? Choose Statement A or Statement B.

(Table 10).

Even confronted with the question in this form, which might be regarded as a false dichotomy, those surveyed refused to endorse school fees. (57% for free schooling, 33% for fees.) Opposition is stronger in rural than urban areas, but interestingly there seems no significant difference between male and female opinion.

Batswana endorse liberal democratic values. Equal rights for women, free speech, and the rule of law are all supported. 72% agreed that "The President must always obey the laws and the courts, even if he thinks they are wrong"; only 23% agreed that "Since the President was elected to lead the country, he should not be bound by laws or court decisions that he thinks are wrong." (Table 41). The consistency of these answers is interesting and suggests that Botswana's liberal-democratic culture has broad support. However Batswana do not seem entirely confident that they can rely on these values being implemented in practice at the day-to-day level. A significant minority (33%) believe that unequal treatment under the law is common (table 53d.) 53% say that "people have to be careful of what they say about politics". (table 53a.)

Nevertheless, Batswana regard their system as a successful democracy: only 4% thought the last elections were not free and fair (table 45). Botswana was classed as "a full democracy" or "a democracy with minor problems" by 73%, and as "a democracy with major problems" by 13%. Only 3% thought it was not a democracy. (Table 46). However, problems were evidently seen ahead: 43% thought that "If our present system cannot produce results soon we should try another form of government" (table 42). A sign of this willingness to innovate may be seen in the apparent preference for direct election of the President over the present system (63% to 33%; Table 89BOT.)

Back to contents  ||  See also Report on 2001 Census on religious statistics


The figures on religion are interesting because for a long time statements about Botswana's religious makeup have been very sketchily based. The survey results, reproduced below (table 91) show that Batswana overwhelmingly identify as Christian. The figure for Islam (below one percent) seems too low, and is perhaps a result of the concentration of small groups and the way the survey worked - I will leave this to the statisticians.

The first problem with the figures is how to interpret the 27% who describe themselves as "Christian (general)". Comparison with attendance figures suggests that this may represent partly the believers who are not very formally affiliated to particular churches. But they could also include people who are active church-goers but think "Christian" is the appropriate answer to the question"What is your religion, if any?" (Table 91). This interpretation is supported by a comparison with the Afrobarometer survey for South Africa released at the same time. There, the total for African Independent Churches was 4% and "Christian (general)" scored 57% (table 91, SA report). But 45% are active members of churches or mosques (Table 28a, SA report) and 24% inactive members (ibid.), a total of 69% being members, so clearly it is not a marker of inactivity. Probably it is to do with the fact that in South Africa Islam and Hinduism are much more prominent, and so "Christian" seems a relevant answer. In Botswana, where other religions are less prominent, a denominational answer seems more relevant to most, but probably an answer at the level of a religion seems relevant to a significant number of people.

In Botswana the survey shows more people identifying as adherents of the African Independent Churches than of the mainline churches. However, it is unclear which group is actually larger due to the large number (27%) who answered simply "Christian" (see above). If these are not evenly distributed between the two it could alter the result.

If this reservation is set aside, however, the African Independent Churches are clearly the largest group, with 21%. Even if "Protestant (Evangelical/Pentecostal)" and "Seventh Day Adventist" are added to "Catholic" and "Protestant (Mainstream)" to constitute an "Other Churches" category, they only add up to 17%. It would be interesting to see further research into the relative size of particular AICs: in view of their importance, they clearly need to be disaggregated, and analysed more particularly, like the "mainstream" churches.

Comparison with figures for church attendance suggests a picture of Botswana Christians something like the following: at one end of the spectrum an active membership with a clear church affiliation, at the other believers who do not attend services and lack affiliation, and in the middle people whose attendance is less frequent and whose level of identification with a specific church will vary (some will clearly be "inactive members" but evidently some do not feel they qualify).

It is interesting that while 28% say they have "No religion", very few accept terms such as "Atheist".

Traditional Religion, identified as such and practised as an alternative to Christianity, is claimed by only 5%. This presumably includes the Mwali Cult. However, it would be a mistake to think that traditional religious ideas are limited to 5% of the population.

  Urban Rural Male Female Total
None 25 30 35 20 28
Catholic 6 5 5 6 5
Protestant (Mainstream) 6 5 5 6 5
Protestant (Evangelical/ Pentecostal) 7 3 4 5 5
African Independent Church 19 23 18 24 21
Traditional religion 5 5 6 4 5
Hindu 0 0 0 0 0
Agnostic (do not know if there is a God) 0 0 0 0 0
Atheist (do not believe in a God) 0 0 0 1 0
Christian (general) 28 27 24 30 27
Muslim, Sunni 0 0 0 0 0
Muslim, Shiite 0 0 0 0 0
Jehovah's Witness 0 0 0 1 0
Seventh Day Adventist 3 2 3 2 2
Other 0 0 0 0 0
Don't know 1 0 0 0 0

Afrobarometer Table 91

38% are active members of churches or mosques, while if "inactive members" are included total membership goes up to 62%. Thus, Batswana can roughly be divided 40-20-40 as active church-goers, inactive members, and non-members. (See Table 28a. reproduced below).

"28a. Now I am going to read out a list of groups that people join or attend. For each one could you tell me whether you are an official leader, an active member, an inactive member, or not a member: a religious group (e.g. church, mosque)?"

  Urban Rural Male Female Total

Not a Member






Inactive Member






Active Member






Official Leader






Don't Know






Afrobarometer Table 28a

A later question on attendance at religious services gives a similar picture.

"Excluding weddings and funerals, how often do you attend religious services?" (Q. 92)

  Urban Rural Male Female Total







About once a year or less






About once every several months






About once a month






About once a week






More than once a week






Don't know






Afrobarometer Table 92

Attendance weekly or more is 43%, which probably equates to the "active membership". Attendance once a month or every few months is 19%, matching the "inactive membership" but probably overlapping a category of non-members who attend occasionally.

See also Report on 2001 Census on religious statistics

The tables and figures are reproduced from and by kind permission of the Afrobarometer report:
Afrobarometer: Botswana Round 3: 2005 compiled by M.H Lekorwe, M.G. Molomo, W.B. Molefe, D. Sebudubudu, L.L. Mokgatlhe and K.K. Moseki.

The complete report can be downloaded from the Afrobarometer site: - right-click the " Round 3 Summary of Results, Botswana 2006" link and save to disk. This is a 6.38 MB download, so over a slow connection it will take quite a long time.

For reports on other countries, see

Back to top

Copyright © 2006 B.S. Bennett
First uploaded 25 May 2006
Last updated 3 June 2008.