2001 POPULATION AND HOUSING CENSUS
ADMISTRATIVE AND TECHNICAL REPORT
Price P30
ADMISTRATIVE AND TECHNICAL REPORT
Published by CENTRAL STATISTICS OFFICE Private Bag 0024, GABORONE Phone: (267) 3952-200, Fax : (267) 3952-201 Email: [email protected] Website: www.cso.gov.bw/cso
Contact Statistician: Dabilani Buthali Census Office Private Bag BO 339, GABORONE Phone: (267) 3188-500, Fax : (267) 3188-610 Email: [email protected]
Printed and obtainable from Department of Printing and Publishing Services Private Bag 0081, GABORONE Phone: (267) 3953-202, Fax: (267) 3959-392
May 2003
COPYWRIGHT RESERVED
Extracts may be published if source is duly acknowledged
CONTENTS Preface ...... i Acknowledgement ...... v Population Density Map ...... viii 1 Background ...... 1 2 Introduction ...... 8 3 Funding of the Census ...... 11 4 Census Programme of Activities ...... 17 5 The Institutional Framework of the Census ...... 21 6 Publicity and Census Education ...... 33 7 Census Cartography ...... 44 8 Pilot Census ...... 64 9 Development and Design of Census Questionnaire ...... 77 10 Preparatory Activities and Census Enumeration ...... 92 11 Quality and Security Measures...... 103 12 Post Census Enumeration Activities ...... 109 13 Census Data Processing ...... 115 14 Conclusions and Recommendation ...... 139 Appendix 1: Timetable of Census Activities ...... 145 Appendix 2: Census Order and Regulations ...... 149 Appendix 3: Index of Maps ...... 153 Appendix 4. Members of Various Committees ...... 166 Appendix 5: Training Schedules (Enumerators and Supervisors) ...... 171 Appendix 6: Guidelines to Training Officers ...... 175 Appendix 7: Training Centre Packaging Schedule...... 180 Appendix 8: Profile ...... 181 Appendix 9: Vehicles Deployment List ...... 183 Appendix 10: Control List ...... 184 Appendix 11: Prelisting Form ...... 185 Appendix 12: Enumeration Check Control ...... 187 Appendix 13: Household Questionnaire ...... 188 Appendix 14: Revisit Record ...... 189 Appendix 15: Institutional Questionnaire ...... 190 Appendix 16: Document Control Form (Main) ...... 191 Appendix 17: School Essays Competition...... 193 Appendix 18: Census Budget 2001/2002 ...... 196 Appendix 19: Cost of 2001 Census Printed Jobs ...... 197
ii Preface
1. The Administrative and Technical Report is essentially the methodological report emanating from the 2001 Population and Housing Census. This report is the fourth in the series of outputs from the census.
2. This report provides a historical account of the census programme, tracing activities from project launching to the release of outputs. With regard to administrative issues, the report gives details on the organisational issues, institutional arrangement as well as the funding. In addition, a very lucid account on technical issues is provided, which are: development of questionnaire, cartography, publicity, census enumeration, data processing as well as details on overall quality assurance measures.. The instruments and materials used in the census are given as appendices, for ease of reference.
3. Some of the reports in the series of publications based on the 2001 census results so far released are: • 2001 Population Preliminary Results (September 2001) • Population of Towns, Villages and Associated Localities (April 2002) • District Census Officers and Technical Officers Reports (May 2002) • Parental Survival/orphan hood (Monograph) – January 2003
4. Reports still to be released, based on the 2001 census, include the following: • Summary Statistics on Small Areas • Census Analytical Report • Population Projections • Monographs on various topics o Census Data Evaluation o Fertility and Mortality Levels o Migration o Economic Activity o Household Agriculture Activities o Population and Social Characteristics o Housing o Profile of Men and Women in Botswana o Summary of Findings
2001 Population Census – Preliminary Results
5. The publication entitled “2001 Population Census Preliminary Results” was issued as a Stats Brief on 19 September 2001. It highlighted some preliminary results of the Census. It discussed, broadly, the 2001 provisional figures in
i comparison with those of 1991 and the projections. It discussed the rate of population growth over the past ten years, and compared population distribution and growth in 1991 and 2001, at both national and district levels.
Population of Towns, Villages and Associated Localities
6. This is the second time the report titled “Population of Towns, Villages and Associated Localities” is released. This edition is based on the 2001 census while the first was published in August 1992, based on the 1991 Census. It contains the names of all localities identified through the census exercise. The localities are associated with villages they owe allegiance to, in terms of economic, social and cultural aspects
District Census Officers and Technical Officers Report
7. This report gives a narrative account of the overall census enumeration as experienced by the census at the district level. This report has two parts – the first part is on district reports by District Census Officers while the second part contains the views of Census Technical Officers.
It is the first time that such a compendium on census administrative and technical issues has been published. This might offer users a balanced picture that could enable a self-evaluation of the census results.
National Statistical Tables Report
8. National Statistical Tables Report provides statistics on all the various topics covered in the census exercise. The cross-classification of information is also done in respect of all variables, according to administrative districts as well as sub-district, where necessary.
Census Analytical Report
9. The plan for the publication of the census data analytical report is to prepare chapters on subject basis. Analysts from CSO and various other institutions will contribute chapters to the analytical report. Such analysts will be experts in the chosen area of analysis.
Population Projections 2001 - 2031
10. This report will contain population statistics estimated on the basis of future demographic trends in the country. These population estimates are presented at national, district and village levels.
ii Summary Statistics On Small Areas
11. “The Summary Statistics on Small Areas” is a compendium of statistics on settlements of 500 or more persons. This is the third edition; the first and the second were part of the series of 1981 and 1991 Census publications, respectively. The statistics covered are in two parts – one part deals with population characteristics and the other with household characteristics of the settlement concerned. A sizeable number of settlements, urban and rural, are covered by the publication. Among others, the topics covered under each part are:
a) Population Characteristics:
- Population size by sex, age and citizenship. - Population 12 years and over by sex, age and marital status. - Women aged 12 years and over by age, number of children borne alive and number dead. - Economically active population aged 12 years and over by age and sex. - Economically inactive population aged 12 years and over by age and sex. - Botswana citizens abroad by sex and reason for absence. - Population aged 2 years and over by age, sex and school attendance. - Population aged 2 years and over that ever-attended school by sex and highest educational attainment. - Population aged 12 years and over that ever-attended training by sex and highest level attained. - Number of orphans. - Disabled population by sex and type of disability. - Population aged 12 years and over by sex, major occupation and industry.
b) Household Characteristics:
- Number of households - number of households by type of housing unit and duration of occupation. - Number of households by method of acquisition of housing unit. - Number of households by household size, sex of head of household and several characteristics such as: • having one or more cash earning member • having no cash earning member • receiving remittance from within Botswana or abroad • receiving pensions / rents / interests / Relief / Aid, etc. - Number of households by size and number of rooms.
iii - Number of households by principal source of: • water • fuel for cooking, lighting and for heating - Toilet facilities. - Number of households by material of construction of dwelling units: walls, floor and roof. - Number of households owning one or more types of livestock. - Number of households that planted one or more type of crops last agricultural season. - Households by ownership of durables. - Number of households owning livestock and or planted crop by type of acquisition of grazing / planted land.
Guide to the Villages of Botswana
12. This will be the fourth edition of the Guide to the Villages of Botswana. The first was published in 1973 as part of the 1971 Census series of publications, the second was published in January 1983 and the third based on the 1991 census.
13. The data presented in this publication are collected in two phases. The first phase was during the mapping fieldwork, between October 1999 and April 2001. At this time, data collected included a list of villages and associated localities such as cattleposts; lands areas; freehold farms, total number of dwellings and facilities in each of the localities. The second phase was the main census enumeration, from which data on enumerated population and number of housing units has been compiled.
Anna N. Majelantle GOVERNMENT STATISTICIAN
May 2003
iv
Acknowledgement
1. The conduct of a national census is a mammoth exercise because it calls for extensive collaboration between many parties. Indeed every individual as well as a myriad of organisation play a role in order for a national census to succeed. It is the extent of varying levels of involvement of the various players that renders making an acknowledgement a difficult task.
2. With that background, I wish to acknowledge the participation of various players towards the success of the 2001 Population and Housing Census. In the interest of brevity of the report, I wish to mention but committees and few organisations for their contribution.
3. Committee are:
• Central Census Committee • Technical Advisory Committee • National Communications Committee • District Communications Committees • Census Standing Committee (who did the planning of census activities)
My appreciation is extended to members of these committees for overall direction regarding the census. Membership to these committees comprised officers from government ministries and non-governmental organisation. In addition, I wish to commend government ministries as follows:
4. The Ministry of Local Government for being the perennial pool from which the Central Statistics Office drew District Census Officers. Local authorities (District Administration, District Councils, Tribal Authority and Land Board) provided the needed officers for the day-to-day running of the census.
o The District Administration, through the District Commissioner played a very important role in providing direction.
o Tribal Authorities, through the willing cooperation and support by Chiefs in addressing Kgotla meetings to conscientise and educate their subjects on the importance of participation in the census.
5. Ministry of Lands and Housing for support rendered through
o Providing a pool to draw District Census Officers.
o Surveys and Mapping which provided all the base maps which were used in the census cartographic exercise.
v
o Land boards which assisted in mapping especially on verification to district maps.
6. The Ministry of Presidential Affairs and Public Administration, for the support rendered through the:
o Department of Information and Broadcasting, which assisted in the planning and implementation of the publicity and education campaign of the census.
o Government Printer, which had to cope with the printing of numerous items for the census, particularly for giving printing for the census a priority status, despite several competing demand on its limited service.
o Department of Police, who in spite of their limited resources, did their level best to provide necessary security for the field staff and availing themselves whenever their assistance was required in cases of lack of cooperation.
o Botswana Defence Force, who provided the helicopters and aircrafts that were used in enumerating the Okavango Delta. In spite of the last minute1 arrangement, resources availed within the limited time were well appreciated.
7. The Ministry of Education, for providing the bulk of enumerator and supervisor field force, through both primary and secondary schools teachers and the availing accommodation and training facilities. In spite of the fact that boarding facilities had been phased out in many schools, the few facilities availed made a big difference.
8. Ministry of Works, Transport and Communications, through the Central Transport Organisation (CTO), as the census enumeration could not have been a success without provision of vehicles. The CTO also made excellent arrangements for emergency repairs, vehicle coordination and providing trucks used for fuel dumps in order to expedite re-supply. The speed and accuracy with which this task was accomplished is appreciated.
9. Government Computer Bureau, which tirelessly provided the necessary professional advice on computing matters, through which the CSO was able to effectively manage the outsourced resources.
1 Wildlife Safaris (Pty) Ltd was earmarked for the task but were engaged by Ministry of Agriculture for Tsetse-fly control.
vi 10. Department of Supplies, which supplied most of the furniture, stationery and field equipments on time. Without such an effort the census could not have been accomplished on schedule.
11. The Attorney General’s Chambers which worked most of the time under pressure for putting in place Census Regulations, without which the legal framework for the taking of the census would not have been completed.
12. The Ministry of Agriculture, who let the Census Office, use some of their facilities, such as the District Agricultural Shows, to publicise the census. It was also through this Ministry that communication was established with the District Agricultural Committees thus facilitating rescheduling of agricultural fairs to accommodate census needs.
13. Various other government and non-governmental agencies contributed to the census in one-way or the other. The cooperation of these organisations is gratefully appreciated.
14. Kgalagadi Breweries Pty Ltd and Botswana Breweries contributed towards the publicity campaign by allowing their products to carry a census message and the logo.
15. The private media did a commendable job in providing good coverage of the census
16. All the enumerators and supervisors who worked tirelessly collecting the information required; for without their honest efforts we could not boast of a successful census.
17. Members of the public and the civil society for their assistance and patience in providing the information required without which the undertaking would have been inadequate.
18. All District Census and Technical Officers, who coordinated and executed the exercise at the various districts.
19. And last but not least, all staff of the CSO; the typists, the personal secretaries, the administrators, supply officers and the professionals for the dedication and thoroughness in their application to the entire exercise; from its planning stages to its conclusion.
Dabilani Buthali CENSU OFFICER
vii Population Density Map
viii Population Distribution Map
ix 1 Background
1.1 The Country
1. The Republic of Botswana occupies a land area of 582,000 square kilometres and lies at the centre of the Southern African Plateau at a mean altitude of 1,000 metres above sea level. Much of the country is relatively flat with gentle undulations. Along the eastern margin, adjacent to the Limpopo drainage system, the land rises above 1,200 metres and the Limpopo valley itself gradually descends from 900 metres in the south to 500 metres at the confluence of the Limpopo and Shashe Rivers.
2. Botswana is landlocked and is bounded by four countries - Namibia in the West and North-West, Zambia in the North, Zimbabwe in the North East and South Africa in the East and South. Botswana lies approximately between Latitudes 18 and 27 degrees South and Longitudes 20 and 29 degrees East. Most of the country lies to the North of the Tropic of Capricorn, which passes just south of Dinokwe (Palla Road). In the North West the Okavango River drains inland from Angola to form an extensive swamp. In the West lies the Kalahari Desert, which supports a vegetation of scrub and grasses with almost complete absence of surface water, although pans hold water for sometime after the rains.
3. The climate is described as continental, semi-arid, sub-tropical but generally considered healthy. Rainfall is low, erratic and unevenly distributed. It ranges from 650 mm in the North East to less than 250 mm in Kgalagadi. More than 90 percent of the rain falls in the summer months between November and April.
4. Mean temperatures vary according to region, with maximum temperatures reaching 40 degrees Celsius. During winter overnight temperature may fall to 0 degrees Celsius in the extreme south.
1.2 The Economy
5. At independence in 1966, Botswana was classified among the least developed countries in the world. The mainstay of the economy was the beef industry. The discovery and exploitation of minerals in the early seventies, as well as the re- negotiation of the Southern African Customs Union Agreement in the late sixties, led to the transformation of the economy, which currently is described as one of the fastest growing economies in the world.
6. The revenue from diamonds contributes significantly to Gross Domestics Product (GDP). There is concern regarding the dangers of over-dependence on a single commodity. Government has put in place ways to diversify the economy. Tourism management services are now considered as possible new engines of economic growth.
1
7. During the inter-census period the GDP is estimated to have registered a 5.4 percent average growth per annum (1991/92 to 2000/2001); and GDP per capita in (current prices) was estimated at P17, 396 in 2000/2001. This remarkable trend has helped to transform the socio-economic set-up of the country. It has influenced the settlement patterns and has led to highly urbanised population.
8. The transformation of the economy has brought considerable investments in terms of Health and Education. Transport and communication networks have also improved tremendously.
1.3 Settlement Patterns
9. The population density map on page 8 shows the distribution of the population in August 2001. A major feature depicted by the map is that the population is concentrated around four major centres of the country. There is a heavy concentration of the population around Gaborone and its surrounding areas, Serowe- Palapye, Francistown and Selebi-Phikwe, as well as the Okavango Delta area (Maun). Close to 40 percent of the total population is now living within a 100 kilometres of Gaborone, which is the capital. The heavy concentration of the population has also led to a high urbanization of the population, with about 54 percent of the population being classified as urban.
1.4 Constitution Obligation
10. The census is an important project through which data on various disciplines are collected. Therefore the conduct of censuses in Botswana is a constitutional requirement.
11. Census information is central to planning through which policies are initiated, monitored and reviewed based on information. In Botswana censuses are conducted under provisions of the census statutes hence it is mandatory. The statute guarantees privacy in the handling of data obtained through the census. In essence, divulging of data by census officials is a punishable offence.
1.5 Census Legislation
12. In terms of Section 64 of the Constitution, the Judicial Service Commission shall appoint a Delimitation Commission at any time after a comprehensive national population census has been held in Botswana. The interval between appointing one Delimitation Commission and another shall not be less than five years nor more than ten years. Implicitly, therefore, a comprehensive national census shall be held at regular intervals to satisfy the requirements of the Constitution.
2 13. In terms of the census, the need for a legal provision serves two purposes: -to make it obligatory for respondents to provide the required data. -to balance the obligation with the individuals’ right to privacy.
14. Just like the history of census taking, Census Legislation dates back to 1904. Proclamation Number 10 of 1904, issued on the 6 April 1904, provides for the taking of a Census in the Bechuanaland Protectorate. Each respective census had a Proclamation or Statutory Instrument issued to carry out a census.
15. The content of the first Proclamation of 1904 does not differ much from the Census Act of today. The major difference is that today's Act provides for penalties for refusal to provide information to a Census Officer. There were no such penalties in 1904. As noted in the section on the History of Census taking, it was thought that providing penalties for the general population would only complicate the process of Census taking. On the other hand, severe penalties were imposed on Census Officers who neglected their duties. Section 10 of the Proclamation of 1910 reads: "Any Census Officer who-
(a) without sufficient cause refuses or neglects to act as such or to use reasonable diligence and care in performing any duty imposed upon him;
(b) wilfully puts an offensive or improper question or knowingly makes any false return;
(c) asks, receives or takes from any person otherwise than an authorised officer of the Administration any payment or reward;
(d) divulges any information obtained during the performance of his duty as an officer of the census;
shall be guilty of an offence and shall be liable to a penalty not exceeding one hundred pounds and in default of payment to imprisonment with or without hard labour for a period not exceeding one month or to both such fine and such imprisonment"
16. The current Census Act, CAP 17:02 provides for a fine of P400 or imprisonment for a term not exceeding one year, or both in respect of the same offence.
17. The Census Act of 1967 provides for a penalty to be imposed upon those who, being obliged to provide information to a duly authorised Census Officer, fail, without reasonable cause, to do so.
18. All census officers were made aware of the census Act and its regulations by officials from the Census Office during the training sessions of District Census Officers, supervisors and enumerators. The section dealing with penalties for census officers is part of the training manual. The Census order 2001, made as a provision for the
3 carrying out a census between 17 and 26 August 2001, was published (as statutory Instrument No.26 of 2001) on 22 June 2001. The census regulations, 2001 itemising the topics to be covered, the penalties for refusing to cooperate with Census Officers and the duties of Census Officers, were published as Statutory Instrument No.29 of 6 July 2001.
1.6 Constraints, Limitation of Census Legislation and Recommendations
19. As noted during the preceding censuses, the provisions of the Act and the regulations thereto are adequate for the actual enumeration in the census exercise. However, bearing in mind that the Census exercise has many phases, legislation has to encompass all the phases, especially pre-enumeration
20. Whereas there were no serious incidents during the mapping and house listing, the 2000 Pilot Census, which is a very important preparatory activity, was not covered by the Act. In an attempt to enact regulations for the Pilot Census, the Attorney General’s Chambers advised the office that the Act was silent on the matter. As a result, the 2000 Pilot Census was conducted without the requisite statutory instrument. It was conducted with reliance on the Statistics Act and goodwill of sampled respondents.
21. It was due to such limitations that prosecution of the racist white farmer who insulted and detained pilot census officials on her farm took so long to resolve.
22. There is the incident of a certain couple who Police Officers detained for refusing to be enumerated twice because they had been enumerated in a different locality. Although the couple eventually agreed to be enumerated they refused to pay admission of guilt fine but opted to take the matter to Court. The Court ruled that having been enumerated was a “sufficient cause” as dictated by the Census Act. Notwithstanding the Court ruling, in terms of a population and housing census exercise, the Act dictates that the couple should have been enumerated again alongside their housing unit, which had not been enumerated.
23. Considering all the loopholes, it is important that the Census Act is amended to cover deficiencies and loopholes that were revealed during the handling of the incidents under question..
1.7 History of Census Taking
24. Historically, census taking began in ancient times for tax and military purposes. It is recorded in history that a population count was taken in Babylonia before 3800 B.C., followed by that in China in 3000 B.C.
4 25. In Botswana, the first census was conducted in 1904, with April 17th as the census night. This Census was held as part of the Census of British South Africa with the three High Commission Territories (Bechuanaland, Basotholand and Swaziland) and the South African colony participating as well.
(a) The 1904 Census
26. The Bechuanaland Protectorate was a very large area, which was sparsely inhabited. In 1904, the country was full of endemic (as well as epidemic) diseases. The literacy level of the people was very low. In view of the poor communication between the people and the administrators, the people were full of fear and suspicion of these administrators. The administrators knew little or nothing about the people's way of life. The authorities therefore decided that legislation (with attendant fines in default of regulations) would only complicate matters and arouse more fear and suspicion. Therefore the Census of 1904 was voluntary and not mandatory.
27. Bechuanaland Protectorate had Tribal Reserves, Crown lands, etc. The Census of 1904 covered the Reserves only and estimates were made for other areas. The enumeration was carried out by Chiefs and Headmen (free of charge). The only payments made were grants to the officials of Government who took part and these amounted to £155.9.1.
28. The data collected during this Census were on total numbers, sex, adults and children. Other data also collected was on wagons, cattle, horses, donkeys, sheep goats owned by individuals or households.
(b) The 1911 Census
29. The second census was carried out in 1911 as part of decennial census of the British Empire. The information sought in the Bechuanaland Census was restricted to number of people, sex and broad age group (adult/child) . The reasons for insisting on a simple return were compellingly put to the authorities in Pretoria by Mr Panzera thus:
30. "I need hardly point out the difficulties of carrying this work in a Territory nearly thirty times the area Basotholand with a tenth of officials and an almost entirely uneducated population."
31. The cost of the 1911 Census was £500 plus £138 in gratuities to officials. In Basotholand, where more detailed information was collected, the cost was £1,775 plus £1,000 for clerical assistance in handling the data the South African Bureau of Census and Statistics.
5 (c) The 1921 Census
32. The Bechuanaland Protectorate took part in the 1921 Census of the British Empire, with plans coordinated from Pretoria. The Bechuanaland authorities again insisted on a simpler questionnaire for the African Population; administrative difficulties were as great as ever. The forms for Ngamiland were sent by trader's wagon on a month-long journey from Serowe to Maun. In some parts of the country, the forms did not arrive on time. The census was heavily criticised for providing minimum information. It was reported by Sir Alan Pim, writing 1933, that in some districts the census papers consisted pieces of string in which the headman tied knots to indicate the number of people in his charge. The costs were some £800 against £3,500 for Basotholand and £400 for Swaziland.
(d) The 1936 Census
33. There was no census in 1931 because of the worldwide recession. The next census was in May 1936, at a total cost of £684.10.11. There were no illusions about its accuracy. The Resident Commissioner, Mr. C.F. Re reported to Pretoria that he was far from satisfied at the reliability of the figures. For example, the Basarwa population of the Ghanzi District entered on a single line was estimated at 7,000 males, and 3,000 females, which he remarked as palpably absurd. He continued;
34. "Moreover if these figures were accurate the population of the Ghanzi District would have shown an increase of nearly 500 percent - a result which needs no comment. It may be noted in this connection that only 44 Bushmen have been numbered in Kgalagadi District."
35. A Comparison of the 1921 figures with the 1936 count suggested that the very large increases recorded were a result of under-enumeration in the 1921 Census.
(e) The 1946 Census
36. One other census was conducted in 1946, and was the most ambitious. No less than 49 questions were asked and a large number of tables were published. The census work was coordinated from Pretoria and the Union Bureau of Census and Statistics carried out the analysis. The work on the tabulations was not finished in time for preparations for the 1951 census to begin. Thus the 1951 census was postponed until to 1956.
(f) The 1956 Census
37. Professor E. Batson of the University of Cape Town conducted the 1956 Census on behalf of the government. It was conducted on a sample basis using tax registers as a frame. For the purposes of this Census, a household was defined as comprising any person whose name appeared in a tax register together with his dependants or household associates of that person, whose names did not appear in a tax register. In
6 effect, this meant that every male aged 18 years or more together with his family if any, and a number of widows together with their dependants if any, comprised the African households of the country. These registers were, however, reported to have been deficient.
38. The processing of this Census took a very long time and the results were never published. The Administration Secretary writing to Professor Batson on 12 October 1964, regarding government's reluctance to finance the publication of the census results, had this to say:
39. "Viewing the matter broadly, the main consideration from this Government's point of view must be whether the results of the 1956 Census would now, if printed, be of any practical use. Seven years have now elapsed since the Census, and so severe a time lag has deprived government most of the original usefulness of the data, 7 years in planning time have now been lost and the data, even if otherwise accurate, is now largely obsolete".
40. The Government took this strong view for the very fact that another Census had already been carried out during 1964. Thus there was little point in financing the publication of Professor Batson's work.
(g) The 1964 Census
41. The 1964 Census, the seventh in the series, was the first census conducted on a house-to-house campaign. It was mainly intended for the delimitation exercise in preparation for Independence. Successive censuses have been improved based on the 1964 census. A major improvement to note is that the 1981, 1991 and 2001 Censuses, which attempted to cover the nomadic Population. In the previous census only estimates were provided.
(h) The 1971, 1981, 1991 and 2001 Censuses
42. The methodologies of the post-independence censuses, as well as that of 1964, are well documented in the respective reports.
7 2 Introduction
2.1 The Census
1. According to the United Nations definition, a population census is the total process of collecting, compiling, evaluating, analysing and publishing or otherwise disseminating demographic, economic and social characteristics pertaining, at a specified time, to all persons in a country or in a well-delimited part of a country. Population characteristics, which are usually collected in most countries are: geographic (place of usual residence and previous residence), migration, household characteristics (relationship to head of household, household structure), demographic, social characteristics (age, sex, marital status, citizenship), fertility, mortality, educational characteristics, and economic characteristics (whether in the labour force, occupation, industry, employment status).
2. In Botswana, as in most developing countries, a census involves enumerators visiting each dwelling in the country for purposes of enumeration. In the developed country questionnaires are usually left for self-enumeration by the households and are then posted back to the authorities.
3. Conducting a census in Botswana is a constitutional obligation, the constitution requires that a Parliamentary Seats Delimitation Commission meet after a comprehensive national population census has been conducted. Thus the primary purpose of a census is to provide the commission with the information it needs. Other planners, researchers, administrators and professionals usually take advantage of the census and add on questions of interest to the census questionnaire.
4. Censuses are very complex exercises, which directly or indirectly involve every individual in the Country. A census is time consuming and expensive. Unfortunately, the expenditures that are easy to quantify are those that are incurred by the census organisation and it’s other affiliated agencies. It is not easy to cost the time that the households and individuals avail to the census takers.
5. The process of taking a census may be classified into three broad stages:
(a) Pre-enumeration Phase
6. The pre-enumeration phase is the preparatory stage during which a number of issues are considered including the scope of the census, the budget and the design of the census schedule. This is followed by mapping or the cartographic exercise, the aim of which is to produce detailed up-to-date maps for the whole country.
8 (b) Enumeration Phase
7. The enumeration phase is the data collection stage, when census enumerators complete census schedules for all households and institutions in the country.
(c) Post - enumeration Phase
8. This phase entails coding, editing and processing of the schedules. Other very important activities that are an integral part of this phase are; evaluation, analysis of the census data as well as the preparation of census reports
2.2 Need and Usage of Census Data
9. Population censuses provide information on the size, distribution and characteristics of country's population. This information is essential to describe and assess its economic, social and demographic circumstances and to develop sound programmes aimed at improving the welfare of a country and its people. The census provides comparable basic statistics for a country as a whole and for each administrative unit and locality therein, and makes an important contribution to the overall planning process and the management of regional and national affairs.
10. A population census provides indispensable data for scientific analysis and appraisal of the composition, distribution and prospective growth of the population. The trend in the indicators to socio-economic development such as the occupational or industrial structure can also be examined using data from two or more population censuses.
11. Population censuses have many important uses for individuals and institutions in business, industry and labour. Reliable estimates of consumer demand for an ever expanding variety of goods and services can be made based on information on the size of the population at district level, urban or rural setting, age and sex composition, etc. These characteristics heavily influence the demand for housing, furnishings, food, clothing, recreational facilities, medical supplies and so on.
12. Some specific uses of the age-sex structure of the population are:
o preparing population estimates and projections. o examining the supply of the labour force. o examining the dependency ratios. o calculating morbidity and mortality rates that can be used as a guide for authorities in charge of public health as a measure of their success. o calculating survival rates that are used in the area of life insurance and actuarial work.
9 13. With regard to the project initiation, planning for the 2001 Population and Housing Census started in 1999 with the preparation of the Project Document, which was subsequently submitted to Government. The project was scheduled to last five years. Unlike in previous censuses, the 2001 Census was funded solely by the Government of Botswana.
10
3 Funding of the Census
3.1 Preamble
1. When preparations for the 2001 Population and Housing Census started, it was succinctly clear that the traditional support and funding extended to Botswana by development partners (UNFPA in particular) in previous census would not be available. Such support and funding was based on the level of economic development as measured by indicators such as GDP per capita. It was on such basis that Botswana was no longer eligible for support – financial or otherwise.
2. The conduct of censuses in Botswana being a constitutional obligation meant that the Government of Botswana had to fund the 2001 Population and Housing Census.
3.2 Direct Costs
3. Although the initial Total Estimated Cost (TEC) was P38, 750,000 it rose to P47, 887,000 largely due to unforeseen changes resulting from an increase in the number of enumeration areas. This cost had to be 2 borne by the Government. Suffice it to note that the amount in reference is, in essence, just the direct costs. Without doubt indirect cost could well be much more than that amount.
4. Weak capacity in the Administration Unit made documentation of census costs very difficult. Census ledgers contained not only incorrect entries but also very misleading figures in some instances. This was further compounded by the fact that the ministry planning officers who took over during the course of the project were not very competent. At least two project addendums requesting for adjustments to the initial project cost were never processed thereby creating serious shortages that eventually led to the gross over expenditure.
2 This is the direct cost, which excludes costs of resources that are not specifically aimed at the census (DCOs, TOs, equipment and materials, printing, etc).
11
Summary of Direct Costs
Item Description Amount (P) Proportion (%) Wages and Salaries 5,178,750 10.81 Allowances 19,729,314 41.20 Equipment and Supplies 7,154,772 14.94 Protective Clothing 1,076,485 2.25 Transport Cost 6,890,062 14.39 Advertising and Publicity 2,270,477 4.74 Data Processing 2,600,000 5.43 Training 2,564,659 5.36 Other 422,481 0.88 Total 47,887,000 100.00
a) Wages and Salaries
5. Close to 11 percent of the census budget was spent on wages and salaries for temporary staff – field mapping assistants, data processing staff (data entry operators, manual editors, coders and online editors) and other ancillary staff.
b) Allowances
6. Allowances accounted for the largest share of the census budget, at over 41 percent. The largest proportion of the cost on this item was spent on census allowances for enumerators and supervisors3 as well as subsistence allowances and overtime for drivers4.
7. District Census Officers census allowances and subsistence allowances were covered under this item. Also included in the cost are the subsistence allowances for CSO Technical Officers and support staff deployed countrywide at the time of enumeration.
8. Lack of training facilities within some districts resulted in huge subsistence allowances during the census-training period. With training being conducted outside the district, all the trainees attracted subsistence allowance. Furthermore, lack of proper management of vehicle logbooks resulted in huge overtime costs. Drivers, as with previous exercises, were very cunning and thus ensured that
3 There were over 4500 enumerators and over 1000 supervisors 4 There were close to 1000 drivers
12 knock off times were tied to the stipulated census working hours even though it was very unlikely that enumeration could take place as late as 2200hrs.
c) Equipment and Supplies
9. Among the major equipment in this group were vehicles, two-way radios, digital cartographic equipment, t-shirts, census satchels, computers and printers, data processing facilities (cabling, trunking, hubs, file servers, etc), access control system, new office furniture and training equipment. A total of P816, 210 was spent on satchels and t-shirts, P1,652, 758 for the purchase of 16 vehicles required for mapping, and P1,680, 000 on digital cartographic equipment. Costs relating to important supplies for census enumeration such as stationery and adhesive labels are also covered in the group. Overall this item accounted for close to 15 percent of the total cost. d) Transport Costs
10. Of the major changes that occurred following 1991 census was the introduction of the Central Transport Organisation (CTO) charging system. Hitherto to this, the budget for the running of transport rested with CTO. The introduction of the new CTO charging system meant that any cost relating to transport had to be budgeted for. Consequential to this development, the largest proportion of the cost under this item was attributed to the CTO vehicle and driver hire, fuel and mileage cost.
11. Much as the census was a major project, it had to compete with other government projects for the limited CTO vehicles. The shortage of vehicles resulted with a decision being taken to request for vehicle assistance from Land Boards and District Councils. Even with assistance from local authorities, the demand for vehicles was still more than the supply. This therefore resulted with CSO Management making a decision to allow census officials in selected areas using personal vehicles for a fixed standard rate.
12. The cost under this item cover CTO charges, Land Board and Council vehicles, and the usage of private vehicles by selected census personnel is specific areas.
e) T-Shirts
13. For ease of identification and furthermore to look more official, all census field personnel were provided with T-shirts. Although the aim was to provide two T- shirts, “one to wear and one to wash”, some shortages were experienced and this resulted with a few field staff (especially drivers) being issued just one. Overall the cost of T-shirts was 2.3 percent of the overall census budget.
13 f) Printing Cost
14. The Department of Printing and Publishing Services printed most of the census materials, at no direct cost to the CSO. However, private printing companies were contracted to print a few odd jobs. The use of incompatible desktop publishing software by the Department of Printing and Publishing Services resulted with private companies being contracted to design (and sometimes printing) and convert supposedly camera ready output. Overall cost for this item was close to P70, 000.00.
g) Advertising and Publicity
15. It is generally thought that the 2001 Population and Housing Census was one of the most publicised censuses. This was based on public awareness of the census as well as feedback on census education. Among the mediums used to publicise the census were a census film (with a cost of over P600, 000.00), a radio drama with theme centred on the census, radio jingles, census flyers, cartoons, census song, print media. In addition there were occasional radio and television interviews aimed at educating the general public on the census. For a full detail on the mediums employed refer to Chapter 6. With all these activities, over P2m was spent on census advertising and publicity.
h) Training Cost
16. Relative to previous censuses, training costs during the 2001 census were very high. Whereas in previous censuses, the Central Statistics Office took advantage of free boarding schools, almost all such facilities were abolished during the 1991 – 2001 inter-census period. This therefore gave rise to serious shortage of training venues necessitating conduct of training outside the respective districts. Where venues were secured within the districts, such facilities were very costly relative to boarding schools. Further compounding the problem was the fact that trainees could no longer be taken for granted in the name of patriotism5, like before. It was against such background that training accounted for over 5 percent of the total census budget.
i) Data Processing
17. The cost for data processing of the pilot and the main census results was about P2.6 million. CSO Management took a decision to outsource given the office’s inability to attract and retain suitably qualified IT personnel. Although a second consultant was engaged when the first absconded following processing of the pilot census data, the cost for the two consultants was within the budgeted figure.
5 Acceptance of basic facilities such as dormitory accommodation
14 j) Other Costs
18. A host of activities was covered under this item, these included aerial photographs, consultants engaged to offer professional advise on census operations, office maintenance, telephone costs, etc.
3.3 Indirect Costs by Government
19. Botswana censuses have succeeded largely due to effective mobilisation of government resources. Similarly, the establishment of the Census Office requires deployment of resources at the disposal of the department. With this organisational set up, there are many hidden costs. If the salaries of District Census Officers, Technical Officers, as well as all permanent staff of the Census Office were quantified, as part of the census cost, there is no doubt that the overall cost would almost double the figure quoted above.
20. In the discharge of their duties DCO often used office stationery, their departmental vehicles, telephone, faxes and other resources for the benefit of the census. Again these have implicit costs even though they are not included as part of the direct cost. Suffice it to mention that some DCO often raised concern regarding the fact that piggybacking on their department’s resources affected their budget quite significantly.
a) Printing and Publicity
21. Almost all costs for the print jobs were borne by the Department of Printing and Publishing Services. Being aware that plans are underway for each government department to bear the printing costs, the Department of printing and publishing Services was requested to cost all the print jobs related to the census to enable budgeting of this item in future censuses. Indeed the total cost of P904, 236 is somewhat an underestimate given the fact that most of the materials used for pre- census training were not included. Accounting for such materials6 could raise the printing cost to well over P1 million, which, without doubt is a huge amount.
22. A few jobs were printed through private printing agencies and most of these involved publicity materials. This option was necessitated by the fact that some of the jobs required more specialised designing skills. Suffice to note that Government paid for the design of some of the materials even though such jobs were eventually sent to the Department of Printing and Publishing Services for printing. This arose because the software used at the Government Printer was not compatible with standard word processing software such as Microsoft Word, Word Perfect etc.
23. A detailed account of all the major printing jobs is given as Appendix 19.
6 These are detailed in Chapter 10
15
b) SADC Collaboration Effort
24. Although the government fully funded the census, the SADC 2000 Collaboration project also contributed towards the 2001 Population and Housing Census. This was in the form of capacity building in overall census management and specific areas such as cartography, data processing and analysis. A number of CSO officers attended workshops and seminars and therefore benefited immensely.
25. Contributions from the project came in the form of purchase of air tickets, officers’ per-diem, accommodation and related costs.
c) UNFPA Missions
26. Despite the fact that Botswana was not eligible for funding of the 2001 Population and Housing Census, the UNFPA provided support to the project in the form of observation and monitoring missions. In the case of cartography, three such missions were undertaken by UNFPA – CST (Harare) at various stages of the project. Costs of services provided by the experts were borne by UNFPA while Government paid the daily subsistence allowance, airfares and hotel expenses.
16 4 Census Programme of Activities
4.1 Introduction
1. The programme of activities for the 2001 Population and Housing Census followed the tradition established in the three preceding post-independence censuses. With a lot being documented on issues affecting the timing of the census enumeration indeed there was no justification for change.
2. The Botswana government decided to follow the UNFPA recommendation that a national population census be carried out every ten years. The reasons for a ten- year interval include the following:
o updating the data every ten years is reasonable,
o a ten-year period is long enough for work on the previous census to be completed and preparations for the next one to start,
o to meet the requirements of the Delimitation Commission who must meet at intervals of not more than 10 years.
3. In preparing the timetable of operation of activities for the census programme, the timing of the period of census enumeration was the most central consideration. All the other activities were built around the period of census enumeration. The timetable was thus accordingly divided into into three phases:
o all census pre-enumeration activities o the census enumeration activities; and o the post-census enumeration activities.
4.2 Timing Period of Census Enumeration
4. The period of census enumeration must satisfy certain requirements. Some of these conditions are:
o It must be possible to move easily throughout the country. In order to contact the persons to be enumerated, enumerators move either by foot, motor vehicles, river transport or any combination of the three.
o The movement of the population to be enumerated should be as minimal as possible during the period of enumeration. Thus it is important to time enumeration to avoid times of major social or cultural activities in the country that would entail larger movement of people. Such type of
17 activities may include religious festivals of national importance, national holidays, agricultural shows, international trade fairs or sporting activities.
o The census draws heavily from educational facilities available in the country. Teachers are used extensively as enumerators and supervisors. Schools are used as training centres for enumerators and supervisors and also for storing census enumeration material. It is therefore important that the enumeration period coincide with the school holidays in order to minimize interference with schools’ normal activities.
4.3 Climate Condition.
5. The fulfilment of the above conditions depends on appropriate administrative measures being taking but largely on climatic and cultural situation in the country.
a) The Summer Period
6. The summer, November to January, is also the period of maximum rainfall. Consequently, there is a lot of undergrowth and vegetation. Travelling by vehicle in many parts of the country becomes hazardous especially in the North-West and the North-East because of the flooding of the flood plains and fast flowing rivers respectively. In the Kgalagadi, for example, traversing the pans from the South to the North is particularly treacherous because of the wetness and flooding of the pans. In the Ghanzi District, vehicles may have to pass through floodwater for several kilometres in low-lying areas.
7. Furthermore, the thick vegetation cover during the period could also inhibit the movement of enumerators from dwelling to dwelling in the rural areas and may even lead to some dwellings being hidden from view and thus not enumerated. On account of the foregoing, the summer period is not ideal for census enumeration.
b) The Autumn Period
8. In autumn, February to April, it is still generally hot throughout the country although the evenings are cool to sometimes cold. Thus this period is unstable for any tasks that require considerable physical efforts over long periods of the day. Enumerators have to walk long distances during the course of the day. Further, in the North-West and to a limited extend in the North-East the ground is till wet and vehicular activities on a major scale could be hazardous.
c) The Winter Period
9. In winter, May-July, the weather is generally too cold for door-to-door canvassing by enumerators. Since the daylights hours are shorter in winter, a considerably longer period would be required for enumeration compared to the period required
18 if the census were to take place at any other time of the year. The longer the period of enumeration the more likely the conflict with school’s calendar. Furthermore, during this period, a considerable proportion of the population lives at the lands, but for the most part in the fields, most of the day. Contacting them for enumeration purposes would be tenuous.
d) The Spring Period
10. In Spring, August - October, the weather is generally mild. The slight cold conditions in the early morning change to warmer ones by mid-morning. There is always gentle breeze that reduces the effect of the upcoming heat of the day. Generally harvesting is over and most of the farmers have returned back to the village. The vegetation cover would have dried up and travelling conditions improved. In the extreme Northwest, the floods are receding and no longer hazardous to vehicular activities.
11. Based on the above summary of the country’s climatic conditions, the best time for the census is the spring school holiday period. Usually this covers most of august, the beginning of the spring season. Thus the censuses of 1971, 1981, 1991, and 2001 have been timed to coincide with the spring school holidays.
12. Notwithstanding the problems enumerated above, with the continued improvement of roads through out the country, another possible alternative period could be autumn. Thus end of autumn coincides with school second term holidays.
4.4 The 2001 Census Timetable
13. The above discussion demonstrates how limited the options are for the choice of enumeration period. Thus any significant delay in carrying out any of the census pre-enumeration activity may lead to a postponement of the enumeration for a considerable time. A delayed census enumeration implies additional cost. In order therefore to ensure that all census pre-enumeration activities were carried out as scheduled, the Government made the census a priority. I A presidential Directive was issued ensuring that additional staff, vehicles, equipment and services required for the census from other government agencies could be obtained on a priority basis.
14. Appendix I show the details of the timetable of activities with respect to the 2001 Census. As mentioned earlier, the activities were divided into three phases: the census pre-enumeration activities; the census enumeration activities and the post- enumeration activities. The table also shows the time each activity was scheduled for and the actual period when the activity was carried out.
19 15. From the timetable it can be seen that all the pre-enumeration and the enumeration activities were carried out as scheduled. However, the slippage of time started to mount with the post-enumeration activities. The data processing took longer than it had been anticipated.
20 5 The Institutional Framework of the Census
5.1 Introduction
1. Within the legal framework of the Statistics ACT CAP 17:01 and CAP 17:02 of the Laws of Botswana, the Central Statistics Office (CSO) is responsible for providing all organs of government, non-governmental organisations and members of the general public with statistical services in areas of data collection, compilation, processing, analysis and dissemination. Encapsulated within that mandate is the conduct of census and therefore the 2001 Population and Housing Census was conducted by the CSO.
2. The CSO is a department in the Ministry of Finance and Development Planning (MFDP). It is currently composed of seventeen Statistical Units namely, Agricultural Statistics, National Accounts, Prices, Household Surveys, Labour Statistics, Demography, Environment Statistics, Health Statistics, Transport Statistics, Industrial Statistics, Education Statistics, External Trade Statistics, Computing Services, Cartography, Research, and Census Office. The Government Statistician, who reports to the Permanent Secretary of the Ministry of Finance and Development Planning, heads the department.
3. In spite of the long history of decennial population censuses carried out during the post-independence period, the Central Statistics Office has never had a permanent Census Office. Notwithstanding that, efforts were made following the 1991 census to maintain a functioning Census Office between censuses but such were thwarted by staff turnover. Thus whenever the need for a census arose, an ad-hoc Census Office was established, and then disbanded as soon as the major post- enumeration activities were completed.
5.2 Census Office During Preceding Censuses
4. When the 1971 Population Census was carried out, a Census Office was established and the Census Officer recruited externally under the provisions of the Special Commonwealth Assistance to Africa Plan. A cartographer-draughtsman was recruited on a part-time basis. Due to unavailability of data processing facilities in the country, the census data was processed in Zambia.
5. In 1981, again an ad hoc Census office was instituted and a national Senior Statistician (Mr. Maitlhoko Mooka) was temporarily designated the Census Officer. A UN Technical Advisor was recruited to advise on the technical aspects of census taking. A UN Cartographer was recruited to take charge of the geographic preparations for the census. Two draughtsmen were seconded to the office from the Department of Town and Regional Planning and the Department of Geological Surveys, to assist the UN Cartographer. For data processing, a UN Census Data
21 Processing Expert was recruited. This time, the data was processed through the Government Computer Bureau mainframe computer, with the UN providing 15 ICL Key-Edit Machines for data capture. For the temporary post of Publicity Officer/Field Organiser, an ex-senior secondary school head-teacher, Mr David Maine, was recruited on a contract basis. To augment the staffing requirements of the Census office, some junior statisticians from the various units of the CSO were seconded to the office. The recruitment of the UN expert was part of the technical contribution by the UNFPA.
6. Upon the Government reaffirming her commitment towards maintenance of a continued programme of decennial population censuses in 1988, a temporary Census Office was established. It was charged with the day-to-day professional and administrative responsibility of designing, coordinating and executing the Census. Consequently in 1989, temporary posts of Census Officer, Publicity/Education Officer, Census Advisor, Census Cartographic Advisor, Statistical Officer (Cartography), Senior Administration Officer, Senior Supplies Officer, Assistant Administration Officer and Personal Secretary were created. The post of Census Data Processing Expert was already created as a permanent post from 1981. The Post of Census Officer was, upon review of it responsibilities, upgraded to the same level as that of Principal Statistician II. It was in July 1989 that the Senior Statistician then heading the Surveys Unit (Mr Chris Molomo) was appointed to this post.
7. The Census Officer, in accordance with the organisational structure of the census is the de facto head of the Census Office, reporting directly to the Government Statistician, who by right of position as the head of the Department is the de jure head.
8. The Census Office was set up by selecting a number of the existing professional staff of the CSO from the Household Surveys Unit, to form the nucleus of the office. Inevitably, this caused a major disruption to the work of the Surveys Unit and consequently all household survey activities were suspended for an indefinite period.
9. Before the establishment of the Census Office, the Survey and the Demography units of the CSO had to fill the vacuum created by the non-existence of such an office. This involved performing preliminary preparatory activities such as circulation to users of proposed topics for the census, holding consultative discussions with major users, preparation of inventory of available materials, identifying core staff for the office, etc.
10. As part of the contribution by the UNFPA, three UN Experts were contracted to the Government of Botswana to provide technical assistance on this census. At the head of the UN Team was the UN Chief Technical Advisor, who joined the Census Office on November 14, 1989. Other team members were the UN Cartographer and the UN Census Data Processing Advisor, who joined the office on 6 October1989 and 29 October 1990 respectively.
22
5.3 Establishment of Census Office (2001 Census)
11. As noted in the foregoing section, efforts to maintain the Census Office functional following the 1991 census were marred by staff turnover. Around the mid nineteen nineties the then Census Officer took sabbatical leave and joined the University of Botswana. Apart from cartographic activities, there was no definite programme of activities and therefore the office made no commitment towards addressing the staffing needs of the Census Office.
12. The establishment of the Census Office for the 2001 Population and Housing Census followed the same pattern as for the 1991 census. With the exception of the Census Officer, all the posts that were created for the preceding census were no longer available at the time of launching the 2001 Census Programme. Although cartographic work continued during the inter-census period, the critical posts were vacant. Botswana was no longer eligible for UN technical support at the time of preparation for the 2001 census and thus all the UNFPA posts were non-existent.
13. Temporary posts of Publicity/Field Organiser, Administration Officer, Supplies Officer and Personal Secretary were created. In the place of a Census Data Processing Expert this time there was a Data Processing Manager whose services were provided through a private consultancy. DCDM7 Botswana was contracted to manage the data processing.
14. In April 1999 a Senior Statistician who was heading the Prices Unit (Dabilani Buthali) was appointed to the post of Census Officer. Initially the post of Census Officer was graded at Principal Statistician II but upon review of it responsibilities, it was upgraded to the same level as that of Principal Statistician I. Other officers were deployed to the Census Office in a similar fashion. Thus the nucleus of the Census Office was established by deploying professionals from other units of the CSO.
15. For ease of operation and space consideration, a one-storey building was secured through the Department of Lands. This office accommodated the Census Office, Household Surveys and Cartography Unit. The three units occupied the new building in October 1999.
5.4 Organisational Structure
16. The Census Office was functionally subdivided into the following sections:
17. The Statistical Planning/Methods and Standard Section was responsible for the overall design, co-ordination and implementation of the exercise. This included
7 De Chazal Du Mee
23 preparation of quality control procedures, coverage and content error evaluation programme, supervising the work of all the other sections, and coordinating the funding and the finances of the whole exercise. The Census Officer coordinated the work for this section. Originally it was intended that Methods and Standards be a separate section from the Statistical Planning Section, but due to shortage of manpower they were combined.
18. The Cartography Section was in charge of all the cartographic requirements of the exercise, including recruitment, training and daily supervision of the cartographic personnel. Cartographic work being an integral part of any successful census, the Cartographic section was upgraded to a fully-fledged unit. The senior statistical officer who until then was in charge of the section was appointed head of the unit (Census Cartographer), at the level of Principal Statistician II. The Census Cartographer coordinated responsibilities of this section, being assisted by six cartographic assistants who had a certificate in draughtsmanship.
19. The Data Processing Section took responsibility for all the data processing aspects, including training and daily supervision of its staff. A decision was taken to outsource data processing management. Consequently DCDM Botswana8 was awarded the tender for the data processing consultancy. A Data Processing Manager from the consulting firm headed the section, being assisted by a graduate programmer from the same firm. CSO attached a statistician to the section for capacity building and also provided all the auxiliary staff.
20. The Publicity/Field Organisation Section was responsible for all the publicity and educational needs of the census, at both national and sub-national levels. This section was head by the Publicity Officer, a retired (re-employed on contract) District Officer (Mr. Israel T. Matenge) whose post was equivalent to a senior statistician. A statistical assistant was deployed to the section to assist him. On suspension of the household surveys early 2001 and realising the burden of that dual role, the Head of Surveys Unit was deployed to play the role of Field Organiser.
21. The Administration Section took charge of all the daily coordination of personnel administration, payments, bookkeeping requirements and general financial administration of the project. This section was headed by an Administration officer in close liaison with the Principal Executive Officer based at the head quarters. The Administration Officer was fairly inexperienced while the Principal Executive Officer, though experienced, was recruited during the course of the project.
22. The Supplies Section was responsible for the procurement and administration of equipment, materials and stationery and control of stores. At the head of the section was a Supplies officer who worked in liaison with the Senior Supplies Officer based at the CSO. Although the Department of Supplies was approached very early regarding the provision of a supplies officer, the fact that the post was temporary
8 A subsidiary of a Mauritius company
24 resulted in protracted arguments regarding the fate of the post on completion of the project. The Supplies Officer was deployed to the Census Office within a month of the start of the actual enumeration. By this time a lot of materials had been procured. This again was a very weak area and therefore Census Office professionals were forced to shoulder the additional responsibility of purchasing materials to ensure the success of the project. It is against such background that is it is doubtful whether the Supplies Unit was able to account for all materials and equipment procured for the project.
5.5 Census Committees
23. In view of the importance attached to this exercise and furthermore its multi- disciplinary nature, a number of national committees were established to give it the desired direction9, in both technical and policy-related matters. Committees formed were as follows:
a) The Census Central Committee
24. Being the most important of the committees, the Census Central Committee was one of the first committees to be formed. It dealt with all major policy issues related to the census taking. Its composition is given in Appendix 4. This is one committee that neither met too often nor had to. Any matter however, which required cabinet approval had to be vetted by this committee.
b) The Technical Advisory Committee
25. This committee was constituted mainly to advise on technical matters of the census. Among other things, the role of this committee included, reviewing content of the questionnaire, phrasing of questions, concepts and definitions. The broad objectives of this committee were the setting of broad parameters for field application, tabulation and classification variables, evaluation, analyses and publication programmes. Membership of the committee is given in Appendix 4. In spite of its wide representation, this committee did not function as effectively as expected, due to continued absenteeism of some members at meetings. In fact, some members sent substitutes to meetings or never attended any at all. This inevitably resulted in the Census Standing Committee taking over the responsibilities of this Committee, a situation that the CSO had made a conscious effort to avoid. Such a situation has the potential to marginalise the objective to have census data being demand-driven. Suffice to note that the same sentiments were expressed in the 1991 census.
9 At advisory and decision-making levels
25 c) The National communications Committee
26. The National communications committee was formed in July 2000 with its first meeting taking place later the same year. Details regarding its composition are given in Appendix 4. The main function of this Committee was to design, co-ordinate, supervise and implement a publicity and education programme operational at the national level. It was charged with the responsibility of reviewing the effectiveness of the 1991 census education/publicity programme, and on that basis generate new ideas for an effective programme of education and publicity for the populace, on all relevant matters regarding the census. Effective publicity was seen as a major factor in ensuring a successful census count. The National Communications Committee, just like the Technical Advisory Committee, was not as effective as it should. Although the chairperson of the committee was decided upon in view of the strategic10 role of the department from which he belonged, the level of commitment was not that desired for a project of that magnitude. Notwithstanding that, the few members who attended meetings regularly made valuable input in terms of the overall direction of the publicity strategy.
d) District Communications Committees
27. These committees were responsible for designing, coordinating and implementing the education and publicity programmes at district and sub-district level. In practice, however, due to their composition, these committees performed more than publicity and education, since they played an advisory and coordinative role on all matters concerning the administration of the exercise at district level. In view of these roles, membership to these committees was mostly drawn from the most senior cadres of the districts’ personnel (see Appendix 4).
28. The composition of these committees however, varied from district to district, depending upon the availability of staff. In some districts politicians, especially councillors, and chiefs were drafted into the committee depending on the strategic role expected of them. There was a formal coordinative and consultative link between these committees and the national committee through the Publicity Officer/Field Organiser from the Census Office.
5.6 Role of District Administrations
29. Although District Commissioners11 had an executive and legal responsibility to coordinate all census activities at district level, it was however recognised that they could not effectively apportion their already busy schedule in favour of the demands of the census. Therefore some senior persons in the district administration were
10 Strategic in terms of resources for publicity 11 The functional role of the District Administration is presented in the District Census Officer Manual.
26 designated District Census Officers (DCOs). A DCO was delegated the coordinative authority and responsibility for all census activities in the district.
30. In their capacities as secretaries to the communications committees, the DCOs worked very closely with the Publicity Officer/Field Organiser from the Census Office. Administratively, the DCO reported through the District Commissioner/Officer to the (national) Census Officer, while the District Commissioner reported to the Government Statistician through the Census Officer. In practice however, lines of communication were less formal and more direct, thus increasing the efficiency with which responsibilities were achieved. However for protocol reasons a few district executives preferred to deal with the Government Statistician thereby retarding efficiency.
31. The DCOs were drawn from both Central and Local government in the districts through the District Commissioners' offices. A Cabinet Directive from the Office of the President was issued to all Central and Local Government offices through Permanent Secretaries, to the effect that the census exercise was to be given priority in the allocation of existing resources. This referred mainly to transport, time and human resources. With this backing, the Census Office communicated their needs through all government ministries to all government departments, district and town councils. Notwithstanding a few isolated cases, the response from all concerned was that of enthusiasm. The quality of the census cooperation and overall results bear testimony to the success of this institutional co-operation. In spite of the authoritative support of the said directive, much of the success could be attributed to the dedication and quality of staff from both central and local government in the districts.
5.7 Role of CSO Technical Officers
32. The Central Statistics Office provided a pool of Technical Officers’ (TOs), drawn from its professional cadre of Statisticians, to back up the DCOs. These ranged in rank from Statistical Officers to the Government Statistician level. This group of professionals were responsible for assisting the DCOs on all technical matters relating to the census. In addition, they were however expected to take charge or assist in any other matters concerning the census wherever there was need. Due to shortage of staff it was not always possible to allocate a Technical Officer to each of the DCOs. In most cases one TO was assigned to more than one DCO. This was seen, as it became evident later, to be one of the weakest links in the organization of the census.
33. As a cadre of professionals, the TO started participating in the field alongside the DCOs, at the beginning of the final preparatory activities of the main enumeration period, when they assisted DCOs in the recruitment of enumerators and supervisors, did the training of trainers, supervisors and enumerators and assisted with the employment of all field staff. Most of the TOs had taken part at all planning stages of the census, and were therefore well versed in all census matters.
27
5.8 Administrative Districts and Census District
34. Taking into account the district size and furthermore to ease the workload on the DCO, some districts were subdivided into census districts. These subdivisions were ideal for census operation purposes. The census districts, however, were in consonance with the administrative districts boundaries. They were either whole districts or sub-districts, where sub-districts were already institutionalised administratively. The towns, regardless of size, were regarded as separate census districts.
35. In total, therefore, there were 28 census districts served by 53 District Census Officers. Administrative districts and the towns were organised into census district as follows:
o Kgalagadi District was sub-divided into Kgalagadi South and Kgalagadi North census districts, conveniently coinciding with the already existing administrative sub-divisions. The South was served from the district headquarters of Tsabong and the North was serviced from Hukuntsi as the district sub-headquarters. Each sub-district had at least two DCOs.
o Central district was already administratively sub-divided into sub-districts of Serowe-Palapye, Central Mahalapye, Central Bobonong, Central Boteti and Central Tutume, hence the same sub-districts were used as census district, each with two DCOs.
o North-West District was organised into Ngamiland South (including the Okavango Delta), Ngamiland North and Chobe census districts. Ngamiland North was composed of all those areas, which fell under the administrative jurisdiction of the Okavango sub-district with headquarters situated at Gumare. Chobe census district maintained the same boundaries as the Chobe sub-district with its headquarters at Kasane. Each census district had two DCOs except Chobe, which had one.
o Southern District was sub-divided into three census districts, namely Kanye/Moshupa, Barolong and Mabutsane. Considering the size of Barolong and Kanye/Moshupa, each of these was assigned two while Mabutsane had one DCO.
o Ghanzi district was sub-divided into two sub-districts of Ghanzi main (including the Central Kgalagadi Game Reserve) and Charleshill. This arrangement was just for administrative convenience since district authorities could not identify the actual boundary between the two. Considering the distances between certain areas in the two-census district, each with peculiar
28 problems, particularly the Central Kgalagadi Game Reserve, two DCOs were assigned for Ghanzi main district.
o South East District was subdivided into two census sub-districts, corresponding to Ramotswa and Tlokweng each with a DCO.
o Kweneng District was sub-divided into two census districts corresponding to the formal subdivisions of Kweneng East with headquarters in Molepolole and Kweneng West whose headquarters is in Letlhakeng. Each of these was assigned two DCOs.
o Kgatleng District was also not formally sub-divided into census districts. However, it was assigned two DCOs whose areas of responsibility were demarcated into east and west.
o North East District was designated a census district with two DCOs.
o Each of the towns; Lobatse, Selebi-Phikwe, Orapa, Jwaneng, Sowa Town and the Cities of Gaborone and Francistone each formed a census district.
o Gaborone was subdivided into four census sub-districts namely Gaborone Central, Gaborone West, Gaborone North and Gaborone South/East, each with a DCO.
o Francistown although not subdivided either formally or informally into census districts, was assigned three DCOs. It was not the area size that was the factor here but the workload.
o Lobatse was a census district served by two DCOs and similarly Selebi- Phikwe was a census district with two DCOs. Orapa was assigned one DCO who also helped the DCO for Central Boteti.
o Sowa Town was considered a census district and was served by one DCO.
o Jwaneng, though relatively small, was assigned two DCOs
5.9 Field Organisation
36. As a usual practice, each of the 28 census districts were in turn divided into statistical areas referred to as Enumeration Areas (EAs). Each EA was composed of approximately 100 to 150 dwellings or malwapa, and was the responsibility of one census enumerator.
37. Over 4500 enumerators (including Botswana Defence Force enumerators) were deployed nationwide. Each enumerator was responsible for enumerating all persons
29 found in a single enumeration area. However for supervisory purposes enumerators were organised into teams of four or five. A supervisor, who reported directly to the DCO for that census district, looked after each team. Only in the absence of the DCO would the supervisor report to the TO.
38. The enumerator to supervisor ratio was on average 5: 1 in towns while it was 4: 1 in the rural areas. The ratios varied because EAs in towns are more compact as compared to the ones in the rural areas where the enumerator had to travel relatively longer distances between dwellings and sometimes between localities. Over 1000 supervisors (including BDF supervisors) were deployed.
39. Special Botswana Defence Force (BDF) personnel were trained (together with the civilian enumerator force) specifically for enumerating the BDF camps. BDF supervisors reported directly to the DCO in the district in which the camps were located. The institutional co-operation with the BDF was very good.
5.10 Census Standing Committee
40. In addition to the national committees referred to in the foregoing section, the Census Office also established a technical working group referred to as the Census Standing Committee. This Committee was an internal arrangement comprising of all CSO subject matter specialists. It was an internal advisory of the Census Office whose responsibility was to scrutinise and advise the Census Office on all technical matters pertaining to various subject matter areas, before presenting to the Census Central, National Communications and Technical Advisory Committees.
41. Membership of this committee was drawn mainly from heads and assistant heads of unit within the CSO. The Government Statistician chaired the Committee while the Census Officer was the Secretary. Recording responsibilities were rotated amongst other junior members of the committee, as a means of on-the job training.
42. The Committee constituted one of the best and most effective links between the Census Office and the rest of the CSO. It was a very effective forum for uninhibited professional discussions. It also provided very good on-the-job training for the less experienced members of the committee.
5.11 Constraints
43. It is indeed appreciated that efforts were made to maintain the Census Office functional after the 1991 census. Whilst staff turnover was a major impediment to the realisation of the intention, CSO’s commitment towards that objective should have been demonstrated by recruitment or re-deployment of staff to the Census Office on time rather than wait until the time of the 2001 census. The absence of a Census Office between the inter-census period impacted negatively on the operations of the
30 office, especially on development of expertise to build on the experience acquired during the censuses.
44. Ad hoc census offices have resulted in:
o inability to retain and sustain the experience gained by temporary and seconded personnel.
o lack of satisfactory accountability once the office has been disbanded.
o lack of continuity and follow-up actions in a field which is inherently dynamic.
o lack of adequate research in areas of sampling for censuses, development of frames for future studies, questionnaire design and other methodological aspects.
o inadequate preparations due to time constraint, lack of experienced personnel and researched information bases.
o enormous overhead costs each time the office has to be established.
45. While these sentiments were also raised following the 1991 census but the importance of an adequately staffed and operational Census Office cannot be overemphasised. To the extent that these reports are aimed at users, it would seem the observations and recommendations contained therein might have missed the attention of authorities12 charged with the responsibilities of the institutional machinery of the Census Office.
5.12 Recommendations
46. A yardstick by which the importance of an institution could best be measured is the level of empowerment given to the staff in charge of such an institution. Literally put, being in charge should be understood as such rather than supposedly so. Although the recommendations from the 1991 census might not have been succinctly clear on the matter, the issue regarding the empowerment of the Census Officer was clearly implied. It was recommended following the census that the Census Officer's post be graded to a level higher than the level of a Principal Statistician. Experience from the 2001 census further confirmed the justification for this upgrading
47. A census is a very enormous undertaking in terms of operation - mobilisation of resources as well as the general public, financial management, material procurement and accounting, spatial coverage. The fact that the Government Statistician is the de- jure head of the census has resulted in the overall responsibilities of the Census Officer being grossly underrated. A comparison of the volume of work between Census Officer and some government agencies13 would reveal that the former carries
12 CSO Management 13 For example election, executive role in small administrative districts, etc.
31 more burden than the later. Perhaps these issues need to be well articulated to enable an effective evaluation of the responsibilities.
48. In mobilising resources that are part of the census institutional framework, there is a lot of interaction with officers of various levels and responsibilities. The level of people in such structures may not14 allow direct communications with the Census Officer therefore creating serious administrative bottlenecks that hamper the effectiveness of the census machinery. At departmental level there are also some serious impediments regarding issuing instructions to fellow officers if the office carries relatively little or no authority. It is therefore recommend that the post of Census Officer be raised to the level of Deputy Government Statistician.
49. There are many potential research areas to keep the Census Office fully engaged in between censuses. In view of that, consideration must also be given to the needs of training and retention of relevant manpower beyond production of the main tabulation. These areas include research on suitability of specific questions, HIV/Aids projections, updating of sampling frames, use of emerging technologies (scanning and related issues) in censuses, evaluation of the census results (using survey results, queries from users), evaluation of methodology, following up issues emanating from DCO and TO reports, exploring new report format, etc. In addition there are many areas that are not usually covered in the analytical reports, in-depth analysis of such could be assigned to the Census Office. There is also a lot that could be learnt from other well-established census institutions, e.g. US Bureau of Census and related entities.
14 Because of protocol and the rather junior level of Census Officer
32
6 Publicity and Census Education
6.1 Introduction
1. The Publicity Officer/Field Officer for Census was appointed in March 2000 on a 24 months contract. Preparations for the 2001 Census were already under way such as Cartographic work, which indeed was at an advanced stage. Not very much publicity was accorded this very important aspect of 2001 Census. However, a Maokaneng radio programme was launched for this endeavour and many people across the country called to appreciate the role of cartographic work, which was on-going around the country.
6.2 Objectives of Census Publicity
2. Publicity was to be launched slowly on all various social structures in Botswana as evidence from the previous censuses proved that some members from the low, middle and the upper classes misinterpreted the role and the intentions of census and its questions as they appear in the Census Questionnaire. It was therefore planned to go slow from the beginning and gather more momentum as the Enumeration period got nearer. The intention was to educate everybody in Botswana to know and understand the followings: when the Census Enumeration was coming; the importance of census and the uses of data collected; why some seemingly queer questions are asked contrary to our norms and also to emphasize the importance of confidentiality by all those engaged in the Census operations. All persons, young and old had to be educated on the importance of the Census so that everyone could appreciate the census and thus cooperate and supply correct information.
3. The target was to educate the youth and adults, Batswana and the non-Batswana about census enumeration through the two official languages used in Botswana to communicate with members of the population as varied as it is, through English and Setswana languages.
6.3 Formation of Census Communication Committees
4. Additional expertise was needed from outside Central Statistics Office, therefore committees were formed at national level and others at district level. These played an advisory role at the national and district levels. This was felt necessary to avoid any misunderstanding that could jeopardise the implementation of the 2001 Census programme.
6.4 National Communication Committee
5. This Committee was to formulate ways on how to go about census and also to:
33 o Advise on census Publicity o Address transport problems. o Guide districts on how to operate on census publicity.
6. The Committee comprised the Director of Information and Broadcasting as chairperson, Census Publicity Officer, as the secretary and other members were; the Government Statistician and the Census Officer, Schools Broadcasting Unit, Curriculum Development Unit, Officer from the Daily News, Department of Tourism, Department of Wildlife, Central Transport Organisation, Department of Secondary Education, Department of Primary Education. The Botswana Teachers Union, University of Botswana Statistics Staff and some NGOs were invited but did not nominate representatives.
7. This committee met just twice and was chaired by the Census Officer each time. The substantive chairperson was reluctant to come to the meetings and never appeared in spite of the numerous invitations made through letters and telephone calls.
8. In brief, the committee was a disaster as it met only twice and never lived for the duration of the census period. It never made any impact on census planning and operations.
6.5 District Communication Committees
9. These committees were chaired by either the District Commissioner/Officer or Council Secretary/Town Clerk or by the Assistant Council Secretary in a particular district or sub-district. The District Census Officer of the area was the secretary of this committee. Membership of these committees varied, as it could comprise representatives from all the Government Departments found in the area and even from NGOs or VDCs. The Districts knew better about the people they could use in these committees, so selections were done by the district. The composition of the committee included personnel from strategic Departments such as Police, District Administration, Tribal Administration, VDC, Land Board etc.
6.6 Achievements of District Communication Committee
10. These committees were a success as they tailored programs to the district and helped the DCO in performing their roles. Kgotla meetings were addressed and publicity days were also organised. Publicity days were in the form of a marches or motorcades from one point of the town/village to the other. This would be followed by a public address about the census and serving of refreshments. The public rally would be made with the help of census posters and placards. These activities helped a lot in as far as sensitising certain members of the society about census. In some localities drama groups were utilised to sell the Census. These committees held meetings at which plans on the publicity programs were discussed.
34
11. The District Communication Committees were a success and lived up to their expectation. It is therefore a recommendation for future censuses these should not be phased out. a) Address of Political Bodies
12. Political forums such as Full Council, Parliamentary Population Committee and indeed all the Parliamentarians were addressed so that they could understand the roles of census and why certain seemingly queer questions are asked. This was very important since politicians were expected to sell the census to the people at Kgotla and at “freedom squares”. This was a good campaign for the census operations.
13. Parliamentarians and Councillors were expected to know and fully understand the Census Operations because this knowledge would equip them to explain further to their electorate the role of the census. This in turn enabled maximum participation and full understanding of the census by members of the community. b) Address of Full Council Meetings
14. The Census Publicity Officer addressed full Council meetings throught the country, where the objectives of the Census were explained to the Councillors. The questionnaire was explained, as well as the use of data that was to be collected. c) Parliamentary Population Committee
15. This committee was addressed and the job was easier because the chairperson of this Committee was once a Census Officer. Being well versed on technical matters, indeed helped to articulate issues to fellow parliamentarians. d) Address of Parliamentarians.
16. The Minister of Finance and Development Planning played a positive role to the publicity and education of Census by addressing Parliamentarians on the 2001 Population and Housing Census. The Minister was accompanied by the Permanent Secretary, Government Statistician and Census Officer. e) Results of Public Addresses
17. These addresses made a lot of impact in the census publicity as Parliamentarians and Councillors started going around addressing Kgotla meetings and “ freedom squares” within their constituencies and wards respectively, explaining why everyone should participate and cooperate.
35 f) Production of Census Pamphlets and Booklets
18. The following booklets and pamphlets were produced in both Setswana and English for use in publicising and educating people about the Census.
o Guide to the 2001 Population and Housing Census o Advocacy for Census o Population and Housing Census - August 2001. o Secondary School Lesson Notes o Primary School Lesson Notes
19. The first three were each available in both Setswana and English in one copy while the last two were only in English. Reference copies have been deposited with the CSO Library and some are with Archives for safekeeping.
20. School Lesson Notes were produced as a collaborative effort between the Census Office and the Department of Curriculum Development. The Departments of Primary and Secondary Schools helped to distribute the notes to all schools around the country through the Education Officers. However, some Education Officers did not pass on these materials in good time. This delay was one reason why some districts did not participate in the Census Essay competitions.
6.7 Impact of Publicity Documents on Society
21. These booklets and pamphlets were designed with eye-catching pictures. It provoked the curiosity of the mind thereby arousing the urge to go through the pamphlets. This most certainly encouraged one to read on all the information therein contained. These documents were not seen scattered all over the shore as part of the litter and this suggested that they were treasured and kept nicely as part of the family library.
6.8 Assistance from Government Departments and Other Institutions
22. Assistance was sought from various bodies to help in the publicity of the Census. It was the intentions of the Census Office that no stone should be left un-turned in publicising and educating people about the Census and its value to society. Help was therefore solicited from a number of organisations to assist in educating people about the census.
a) Department of the Accountant General
23. The Salaries Section of the Accountant General Department was approached to print a census message on all Government salary pay slips in order to advertise the coming of the Census. This was done from April to the end of July 2001 covering public
36 officers’ and pensioners’ salary pay slips. The message was: “Be there to be counted” 17 – 26 August 2001 “Nna teng gore o balwe”.
b) Botswana Post
24. The Botswana Post agreed to make a canceller that was stamped on all locally posted letters. The period covered was from May to July 2001. This was also one way of selling the census to the people. The Botswana Post charges were only for artistic work done to produce the canceller.
c) Radio Botswana
25. The Radio was always at our disposal for participating in programs such as Maokaneng and Masa-a-Sele. Census jingles, testimonials and the radio drama were all broadcast on Radio Botswana on fixed times and days of the week. News items on the Census were widely covered including all preparations that were being made throughout Botswana on the 2001 Census.. The radio did commendable publicity for 2001 Census.
26. This coverage greatly augmented the publicity and education campaign that was going on. Through Radio Botswana, a wide audience of Batswana in towns, villages, cattle posts and lands were informed about the census.
d) Botswana Television (Btv)
27. Btv was highly instrumental in educating and publicising the coming of the Census. Strips from the census video were shown on daily basis just before the 1900 hours news. Setswana and the English versions of the census video were also played alternately on Btv screens. The staff of Btv adequately covered news items on the Census. Members of the public with television sets watched this video from their homes while those without watched from public places such as bars. Many commented that the video helped to open their scope and understanding about the Census. The Documentary-drama film was both entertaining and educative.
e) Botswana Daily News
28. This paper did a remarkable job through the production of Mabijo Cartoons and features. These cartoons were produced and appeared twice in a week in the paper. The paper also covered several news items on Census operations around the country such as the addresses of Kgotla meetings by Members of Parliament and/or District Census Officers.
37 6.9 Private Sector Assistance
29. It was not only the Government Institutions that were solicited to cover publicity and education about census. Private companies were also approached and some responded positively.
a) Kgalagadi Breweries (Pty) Ltd
30. The managers of this company agreed to sponsor a public awareness on census by inscribing a census message on the Saint Louis Beer Cans and the Coco-Cola Beverage Cans. They printed the Census Logo and a message on the cans as follows; “ Be there to be counted” “August 17 to August 26 2001”, “Nna teng gore o balwe”. These products were distributed all over Botswana to all the liquor and grocery outlets. In this way a large proportion of residents of Botswana got the message regarding the census enumeration.
b) Botswana Breweries (Pty) Ltd
31. Likewise this company was also approached and it agreed to inscribe the census message on its Chibuku cartoons. All the Chibuku brewing depots around the country produced and distribute cartoons all over the country. The Brewery produced 300,000 cartoons per day to be distributed all over Botswana. It is therefore assumed that the message displayed on the cartoons reached many people.
32. The Brewery decided to write the message in Setswana only as most consumers are Setswana speaking and it read, “Nna teng gore o balwe” “17 – 26 Phatwe 2001”.
c) Launch of BBL and KBL Products for Census Publicity
33. Launching of all the products that were used to publicise census by the above companies was staged at the Grand Palm Hotel in May 2001, the two companies sponsored the launch. These products were Chibuku, Coco-Cola and Saint Louis beer. The ceremony started with speeches by the Government Statistician and the Managing Directors of the two companies. Snacks and refreshments followed this launch.
34. The Press, i.e. Radio, Botswana Television, Botswana Press Agency (BOPA) and Mmegi Newspaper were present at the launching and gave coverage about the occasion. The Mmegi Newspaper produced a centre page report on the proceedings of this launching.
35. All the costs were shared by the two sister companies as these were organised to assist the Census Office with the publicity.
38
d) Lack of Support by Some Companies
36. Some companies objected to participate in the publicising of census through the use of their products. They cited that some individuals might be against the census and this would impact negatively on their sales.
6.10 Census Education And Publicity Activities
a) 2001 Census Song
37. A Census song was produced and sung by Dinare Choir of Ramotswa. This song was recorded at the Radio Botswana Studios. Soul Power Records produced a compact disc of the song and this has been deposited with the Central Statistics Office Library for future reference. The song had a lot of impact in publicising the census.
b) 2001 Census Video
38. A company by the name of African-American Satellite Communications won the tender to shoot this film. A story line was produced to guide the company as to what kind of video film was required. The plan was to avoid a documentary film like that produced for the 1991 census. It was felt that a docu-drama film would not be boring but would entice watchers to follow-up with the film instead of slumbering or dozing off. The company was assisted to come up with an appropriate script written for the film. It was also emphasized to the company that testimonials should be sought from across a wider spectrum i.e. the film should be well balanced, politically.
39. The film was produced and copies were distributed to Btv and all the District Census Officers. Btv received a master copy in Setswana and English and the film was shown during the President Day Holidays and again during the first week of August 2001. The District Census Officers played the video at Kgotla meetings around the country to educate and sensitize people of Botswana about the coming census. The census song was played in between episodes in this film. Schools and other organisation wishing to view the video were loaned the video for general showings.
c) 2001Census Radio Drama, Testimonials and Jingles
40. A company known as Soul Power Records produced these and they were played in RB1, RB2 and GAB-Z FM radio stations. The drama was in Setswana only while testimonials and jingles were in Setswana and English. Only GAB-Z FM broadcasted the jingles for a price while RB1 and RB2 broadcast for free. The census song also featured in these products, as an introduction or between episodes of the drama.
39
d) Billboards Publicity
41. Inter-Africa, a local private company was engaged to publicise the census through its billboards found around Botswana. A picture of a young lady standing at an open door of a house with census messages, which read: ”Be there to be counted” 17th – 16th “ August 2001”, “Nna teng gore o balwe” were displayed in billboards around the country along highways and at shopping centres. Motorists and shoppers could definitely catch up with this message. Small villages, cattle posts and lands were not covered by the billboard advertisement. It was assumed that residents of such places travel to larger villages and used the highways on which the advertisements were posted.
42. Billboards were displayed rather late and they covered a period of one-and-half months before the actual census enumeration. It could have been better if they were displayed for a period not less than 3 months before the enumeration. This would have allowed many people to see them. In spite of the short period they made an effect on the publicising of the Census.
e) Other Census Publicity Materials
43. The Census Office decided that some publicity materials should be produced in the form of human attire and accessories. The following items were considered as suitable; car door stickers, bumper stickers, annual date calendars, T-Shirts, caps, brief cases and pens. They all had a census logo and some census message printed on them.
44. Some white T-Shirts, and caps were given as presents to winners of the schools’ essay competitions together with pens and portfolio bags at the district levels while radios were awarded as prizes at national level.
45. Blue T-shirts and caps were supplied to everyone involved in the census enumeration. Technical Officers, DCOs and Supervisors were in addition supplied with portfolio bags. All these items were not to be returned. Enumerators and Supervisors were again supplied with satchels. The satchels had bright colours15 for ease of identification even from a distance. Satchels therefore served a useful purpose in publicising the census.
46. These items were to be used during the enumeration period as last minute publicity and also to identify those involved with the census enumeration. It was important that people knew who was a census officer and who was not. So the blue colour was reserved for those directly involved with the census enumeration and the white colour was for general publicity and competition prizes
15 Yellow for enumerators and green for Supervisors
40 47. There were other materials such as car door stickers. These were used mostly in all the census vehicle fleet to distinguish them from any other Government Vehicle. Car bumper stickers were supplied to everyone and thus some batches were sent to District Census Officers for the general distribution to members of the public.
48. The telephone on hold message at the Census Offices was one good way to sell the Census to the people. Everyone who called this Office would after speaking to the telephone operator listen to a brief message narrating that the census month was August 17 – 26, 2001. Callers commented that this was one good way of advertising the census.
f) Mabijo Cartoon Strip
49. Mr. Tebogo Motswetla, a cartoonist, was contracted to produce Mabijo Cartoons Strips and these appeared twice a week on the Daily News over the period March to August 2001. The aim was to educate and publicise the census in a manner that would amuse people as well.
50. The first issue of Mabijo portrayed him being employed by the Census Officer as an Enumerator, taking the Oath of Secrecy and changing his trademark shirt for the Census T-Shirt.
51. Mabijo, the Enumerator, went through different ordeals through his work as an enumerator. This was to educate and advertise the census by portraying the ordeals he underwent as an enumerator - depicting all kinds of characters that are expected to be found across segments of the society.
g) Schools’ Essay Competitions
52. Census education and publicity efforts were not only centred on adults but the youths as well. This was the reason why schools’ essay competitions were organised. School Lesson Notes for primary schools and secondary schools were produced and distributed to schools for coverage in the school curriculum. Later on pupils and students entered district level competitions on the census.
53. The number ones from each district were later made to compete nationally for positions one, two and three. This effort was considered another way of making the youths ambassadors who would persuade their parents to cooperate during the census enumeration. Not all districts participated in these competitions- some DCOs were too busy to find time to organise this event and furthermore most schools in their districts did not receive the Lesson Notes.
54. The Primary Schools essay winners were as follows:
41
Position Name School District
1 Oitsile Madu Kanzungula Primary Chobe
2 Winnie Babusi Masedi Primary Tutume Sub-District
3 Mogomotsi Kamogelo Lady Mitchison Primary Kgatleng
55. The Secondary Schools essay winners were as follows:
Position Name School District
1 Donald Molosiwa Molalatau CJSS Bobirwa Sub District
2 Susan J. Dingalo Bokamoso CJSS Gaborone South 3 Thabo Mokowe Ipelegeng CJSS Lobatse
56. District Census essay competition winners were awarded presents as follows, primary and secondary school alike:
o No. 1 A portfolio bag and a Census ballpoint pen o No. 2 A T-shirt, Cap and a ballpoint pen o No. 3 A T-shirt and a ballpoint pen
57. These items had a Census Logo and a Coat of Arms engraved for perfect visibility.
58. The national winners for both primary and secondary schools were selected after the census enumeration period. This was because of delays in forwarding scripts of districts winners to the Census Office. The presents for national winners were as follows:
o No. 1 Radio with CD and Cassette player o No. 2 Radio of less quality with the CD and Cassette player o No. 3 Just a radio without any extras.
h) Census Slogans
59. Census slogans were coined deliberately to incite the population into supporting the census and cooperate during enumeration. Such slogans were; “Count me too” “Le nna mpala”, “Be there to be counted” “Nna teng gore o balwe””, “Palo batho ke e o”,
42 as derived from the census song. These slogans were very popular around Botswana and within all the socio-economic strata of the Botswana society.
60. A slogan that is well accepted by members of the society serves as a good tool to sell any product and these slogans did exactly that for the 2001 Census publicity. People around all corners of Botswana would be heard reciting or humming the slogans in appreciation when they saw census vehicles. These slogans went a long way in advertising the 2001 Census.
6.11 Negative Census Publicity
61. Some church pastors found it necessary to jeopardize the census. They claimed to be prophesizing to their congregation that the census enumeration is coming. They urged their congregants that they should not accept to be counted, as they are not cattle. They told people at funerals that those who allowed members of their families to be counted would die together with members of their families as this was against the will of ‘God’.
62. These statements were made before the census publicity and education were intensified to cover all residents of Botswana. However, such damaging statements were overcome by the intensity of Census education and publicity that followed.
6.12 Recommendations
63. Census publicity and education can not be handled solely by the Census Publicity Officer alone and therefore effort should be extended to include District Offices, the press and large Corporations.
64. Road shows should also be tried as they are both entertaining and educating. They could do well at small villages where there are not enough entertainment facilities. They were not engaged for the 2001 Census because authorities considered that enough was achieved in terms of publicity, which had that far been launched.
65. It is also advisable that a Census Publicity Officer should have the necessary qualifications and must be a sociable person. That is a person who can mingle with all members of the Botswana society.
66. The Publicity Officer should be prepared to travel extensively around Botswana and should also understand Government structure and procedures in the public service. The two official languages of Botswana are a must including knowledge of Botswana cultures. This will enable such an official to know the taboos of Botswana societies and thus would not offend nor antagonise any group within the society.
43 7 Census Cartography
7.1 Cartographic Field Operations and Delineation of Enumeration Areas
7.1.1 Introduction
1. A permanent cartographic unit within the CSO was established immediately after the 1991 Population and Housing Census to service the cartographic needs relating to the Household Survey Programme, the annual Agricultural Surveys and the future Censuses. A chief cartographic officer heads the unit with an assistant statistician and six cartographic assistants, with certificate in cartography and draughting.
2. The 2001 Population and Housing Census preparatory work started on the 7 October 1999, in Gaborone. The reason for starting in Gaborone was mainly to monitor the mappers’ work progress. Each team was assigned two vehicles fitted with multi-functional high frequency (HF) radio telecommunication system with a build in Global Positioning System Receiver (GPS).
3. Cartographic preparations constitute a vital part for the implementation of a modern Population Censuses. It is a basis for ensuring that every person in the Country is counted with minimal possibilities for over-enumeration or under- enumeration. It is one of the links between population censuses and other related activities as it is an essential input into the development of several types of statistical sampling frames.
7.1.2 Objectives
4. The cartographic exercise was conducted to satisfy the following objectives.
o To up-date and correct the1991 Population and Housing Census maps by listing all localities by ecological type, all dwellings (malwapa) and recording the number of persons residing in those dwellings.
o To delineate new enumeration areas on the basis of reliable census maps that would be produced to constitute the framework within which an orderly enumeration could be conducted with minimum possibility for over or under-counting.
o To collect information for estimating the number of enumerators and supervisors, in order to determine workloads at enumeration time.
44 o To collect information at community level, on certain selected socio- economic characteristics, such as religion, home based business and language to be used in determining the type of enumerators to be recruited for the area.
o To collect information that would form part of the input for one of the most useful publication “ Guide to the Villages of Botswana” and the construction of the sampling frames for subsequent statistical exercises.
7.1.3 Composition of Mapping Teams
5. In order to accomplish the above stated objectives, seven mapping teams were formed, each team comprised five mappers and one supervisor. Supervisors were chosen amongst the newly recruited personnel, after a five-day training. Each team was assigned two vehicles fitted with high frequency (HF) radio telecommunication system, mostly four-wheel drive vehicles. At the beginning all the seven teams were assigned to work in Gaborone, mainly for close observation of their work. On 10 December 1999 all the teams were deployed to rural Districts. It took eighteen months to complete the cartographic work countrywide.
6. For operational convenience, each supervisor split his/her team to work on different localities in the assigned enumeration areas. The supervisor checked the work of each mapper at the end of each working day. All the work for each team was in turn checked regularly in the field as well as in the office by the census cartographic control teams comprising cartographic assistants.
7.1.4 Equipment for the Fieldwork
7. Equipment supplied to each team was as follows:
o Introductory letters to the District Commissioner o Copies of the 1991 District Maps o 1: 10,000 Village maps o Pre-listing forms o Sheets of numbering metal plates o Sets of figure punchers o Staple guns o Pencils and ball- point pens o Clip-boards o Mappers manual o Metal box for stationary o Daily workloads forms o Pair of scissors o Tents
45 o Camping beds o Spare tyre, jack, tyre lever, air pump, repair kits for mending tyre punchers and other related tools per vehicle. o Axe and spade per vehicle o Torch per vehicle
8. Prior to the commencement of the cartographic fieldwork programme, an inventory of all existing cartographic materials was prepared. Available materials were evaluated for their suitability as cartographic base material for the 2001 cartographic preparatory work. These were mainly the 1991 District maps with scale of 1: 100, 000 and 1:350, 000 as well as village sketch maps. Sketch maps were to be avoided as much as possible, opting instead to blow aerial photographs to approximately 1:10,000, resulting in 1:10,000 scale in most villages.
7.1.5 Inventory of the 1991 census maps
9. At the conclusion of the 1991 Population and Housing Census copies of maps were deposited with the department of Surveys and Mapping. These were District maps at scale 1:100,000 and 1: 350, 000 respectively and village maps with a scale of 1:10,000 as well as plot numbers and street names for urban centres with scale of 1: 5000. This arrangement was made because the demand for these maps is very high. Moreover the Central Statistics Office is not mandated to sell maps. For the 2001 Population and Housing Census similar arrangements were made.
10. The District Maps mentioned above were evaluated and satisfied to be suitable for use in the 2001 cartographic preparatory fieldwork. The scales chosen were fairly suitable scales while the reduction method was retained, as it proved convenient.
11. For those areas not covered by 1: 50, 000 topographical sheets series but by the 1991 census smaller scale maps, ranging from 1:500,000 to 125, 000 such scales were maintained.
12. The 1: 50,000 map series had been compiled for other purposes (e.g. geological and underground water surveying) other than the census exercise. Therefore to adapt them to census requirements, all the information from the 1991 census base maps were manually transferred to the 1:50,000 and reduced to a 1:100,000 scale before dispatching them to the field for updating. Mappers were also requested to take the 1:50,000 copies with them for ease of plotting geographical co-ordinates. The information transferred included enumeration area boundaries, directional positioning of localities, their names and codes, physical landmarks and number of dwellings.
46
7.1.6 Mapping work.
13. The mapping work was done in two stages. The first stage was the updating of the rural area base maps in the office and in the field. The fieldwork involved preparation of listing of dwellings, these was done by filling the site condition form for each EA, correcting and updating the district and village maps containing information from the 1991 census. The second stage consisted of EA delineation and creation of supervision areas on the basis of the updated information from the field. This also included reproduction of the required number of copies of updated census maps for the main enumeration.
14. Updating of the 1991 census maps involved teams visiting each and every locality in the district. A locality being defined for census purposes as “ any human settlement with a name and identifiable boundaries.” The teams would check whether the locality already exist on the district map, its position and the name spelling. If the locality was not on the map, it was entered on the map at the correct position using the Global Positioning System Receiver relative to those adjacent localities already on the map. Since some maps do not have gridlines, this system was not possible in all the districts except only those districts covered by the 1:100,000 scale. All localities were assigned geographic codes, which were also entered on the map. Localities that no longer existed were not deleted from the map but the number of dwellings was zeroed and locality names were left for further verification during enumeration. As the mappers were updating the list of localities, they indicated new roads and other important features not included on the map. By all possible means the teams were advised to use vehicles odometers to estimate distances between localities.
15. According to the definition of a locality, a locality could be a town, a village, a lands area, cattle post, a freehold farm, a camp, a border post, etc. In the listing and mapping of localities, each locality was assigned a geographical identification code in such a way that it uniquely identified the district to which it belonged. Built into the code was also the ecological category for the locality. As part of the locality identification, the district to which the locality belonged had its own unique code.
7.1.7 Geographic Coding System of Rural Districts
16. Census districts fall within the administrative districts boundaries. Those considered sub-districts administratively are uniquely coded and treated as census districts. Each administrative district is given a unique two-digit identification code. The first digit (1-9) identifies the administrative district.
47 o Southern District was assigned code 1, and was subdivided into census districts of Kanye/Moshupa coded 10, Barolong coded 11 and Ngwaketse West coded 12 o South-East District was assigned code 2, and was itself a census district coded 20 o Kweneng District was assigned code 3, and was subdivided into census districts of Kweneng East 30 and Kweneng West 31. o Kgatleng District was assigned code 4, and was itself a census district coded 40 o Central District was assigned code 5, and was sub-divided into census districts of : Serowe/ Palapye coded 50, Central Mahalapye coded 51, Central Bobonong coded 52, Central Boteti coded 53 and Central Tutume coded 54. o North-East District was assigned code 6, and was itself a census district coded 60. o North-West District was assigned code 7, and was sub-divided into census districts of Ngamiland South 70, Ngamiland west 71, Chobe 72 and Delta 73. o Ghanzi District was supposed to have been divided three times but because of boundary problem, it was not possible to do so. It was assigned code 8, and was sub-divided into census district of: the rest of the district was coded 80 and the Central Kgalagadi Game Reserve was coded 81. o Kgalagadi District was assigned code 9, and was subdivided into census districts of Kgalagadi South coded 90 and Kgalagadi North coded 91.
a) Geographical Coding System of Rural Localities
17. The coding for localities was as follows:
1 = Villages 2 and 7 = Lands Areas 3 and 8 = Cattle Post Areas 4 and 9 = Freehold Farms 5 = Mixture of lands areas and cattle posts 6 = Other (Camps, safaris lodges etc )
18. Though a village is defined as a locality, it is superior in the geographical identification hierarchy. This is because it is a locality, which apart from the district, is associated to itself, and in turn other localities are administratively associated to it. This relationship is clearly accounted for through the coding system.
19. There is, however no official definition of a village. Population size alone is not sufficient to classify a settlement as a village. Therefore a village is usually classified by the presence of a headman and availability of facilities such as
48 schools, clinic and water reticulation etc. In general, three local government authorities are involved in designating a settlement as a village and these are the Tribal Administration, District Council and the Central Government District Administration. In preparing the list of villages the three bodies were consulted for their comments.
20. All settlements, which were designated villages, were given a code 100. To distinguish them from other localities in the district, they were serially coded from 01-98. The serial code 99 was reserved for any settlement not associated with any village (mostly freehold farms, camps and safari lodges). The localities were serially numbered by locality type from 1 to 9.
21. The system as described above can be illustrated by the following examples
o Bobonong is a village in Bobonong District, coded 52-01-100, where , 52 is the district code, 01 is the serial village code and 100 is a code to signify that Bobonong is a village
o Marutlwane is a lands area in Central Mahalapye, coded 51-28- 225, where, 51 is the district code, 28 is the code for Shoshong village to which Marutlwane is associated and 225 signifies that it a lands area (first digit 2) assigned the serial code 25.
o Xhugana is a cattle-post in Central Tutume, coded 54-30-305, where, 54 is the district code, 30 is the code for Gweta village to which Xhugana is associated and 305 signifies that Xhugana is a cattle-post (first digit 3) assigned serial code 05.
22. The combination of these codes made the identification codes for localities unique within and between the districts. It also made it possible to uniquely associate every lands area, cattle post-areas and other settlement with the correct village.
b) Assignment of New Geographical Codes.
23. For ease of comparison with the previous census, the 1991 district, village and locality code numbers were maintained whenever possible. A few localities had to be assigned new codes in response to changes in the sub-district boundaries; changes which occurred either by administrative decisions, corrections or alignment of boundaries or settlements being upgraded to villages. For example the boundary between Kanye/Moshupa census district and Barolong in Southern District was shifted so that Mmathethe and associated localities, which were in the Kanye/Moshupa census district in 1991, were moved to Barolong in the 2001 Population and Housing Census.
49 c) Enumeration Areas and Localities
24. The 1991 maps were used to update the locality lists, names and codes. The enumeration areas (EAs) had numbers between 0001 and 9999. The relationship between EAs and localities largely depends on the type of the locality. If the locality is a village, the EA could be part of the village, so that the village is composed of more than one EA or could be a whole village in the case of small villages. If the locality is a lands area, cattle-post, camp, freehold farm or a mixture, then the EA will comprise more than one locality.
d) Numbering of Dwelling/ Malwapa
25. When the mappers got to a locality, they numbered all the dwellings/ malwapa they found in that locality, regardless of their occupation status (empty, vacant, under construction or occupied) and usage either as a private resident, commercial premises etc. Each dwelling was assigned a five-digit serial number. The district, the village and the locality codes formed part of the identification of dwellings. For example, the first dwelling listed in Xhugana in Central Tutume, which has code 54-30-305, was coded 54-30-305-00001. All the subsequently listed dwellings were numbered serially from 00002 in the same manner up to last dwelling in the locality.
26. The assigned number for each dwelling was punched on a blank aluminium plates obtained from the Government Printer. These were used aluminium printing sheets, which were otherwise to be disposed. They were cleaned in the Census Office and some of them were cut in the office although most were cut in the field by mappers. Those that were cut in the office were approximately size of 4.4cm x 6 cm. Such metal plates were later sent to the field before the numbers were punched on them. These were affixed on the doors of each and every dwelling using a heavy-duty staple-gun.
27. In villages, the district, the locality and the village codes, and the dwelling number were punched only for the first dwelling in the locality. For subsequent dwellings in the same locality only the dwelling number was punched on the metal-plate. For other localities e.g. lands areas, cattle-post and mixture all the codes were to appear from the first dwelling to the last dwelling in the locality.
7.1.8 The pre-listing form
28. This form (See Appendix 11) was designed to fulfil two purposes: i) firstly to record the code number assigned to each dwelling for control purposes, ii) to record additional information that would help enumerators in locating dwellings during enumeration period (e.g. name of the head of the household at the time of pre-listing), iii) information about number of persons living in that dwelling (this information is very useful in determining EA size as measure of enumerators’
50 workloads) as well as iv) collecting information on informal businesses, operating within the dwellings or compounds.
29. The entire geographical identification codes for the dwellings were recorded on this form. At the top of the form there is provision for district name and code, village name and code, locality name and code and the 1991 EA number. It is further divided into eight (8) columns, the first is for the pre-printed serial numbers from 1 to 50, the second column is for recording the dwelling/ malwapa numbers, the third is for describing the use of the dwelling or building (shops, churches, etc) and occupancy status (i.e. occupied, vacant and under construction). The fourth column is for recording the name of the head of the household, the fifth is for recording the estimated number of persons residing in the dwelling, the sixth is for recording home business ownership, the seventh for recording economic activity codes and the eighth for mappers comments that would assist enumerators during enumeration period.
7.1.9 The Site Condition Form
30. When carrying out the pre-listing of dwellings, mapping teams also collected additional information at locality level, besides the dwelling level information. This was done using the site condition form.
31. Information collected was as follows
o Summary from the pre-listing form: it provides information on number of persons from the pre-listing form, total residential places occupied, total residential places vacant and total residential places.
o Type of locality- whether a village, lands areas, cattle-post areas, freehold farms, mixture of lands and cattle-post areas, camps, border post and Remote Area Dwellers Settlement or other.
o Main language spoken in the locality: Setswana, Kalanga, Seherero, Seyei, Sembukushu, Sekgalagadi, Seburu, Sesarwa, Sesubia, Sekgothu or other.
o Main source of water supply (throughout the year), whether located within the locality or outside. If inside state whether, is Government/ Council owned standpipe, Private owned standpipe, Government/ Council owned Borehole, private owned Borehole, Dam/ Pan, River, Riverbed and other. If it is outside state as above and estimate the distance.
o Main source of public transport used, state whether regular (almost daily), irregular and not existing.
51 o Road Network whether tarred road, gravel road only used by two-wheel drive and tracks negotiable by four wheel drive.
o Type of facilities used – health, educational, religious, agriculture, commercial or other.
o The information was collected for use in the preparation of the census publication “Guide to the Villages of Botswana” and for the general planning of the census operations.
7.1.10 Sketch maps
32. In addition to updating the information on census district maps, the mappers were given 1:10,000 maps for villages, which replaced the sketch maps that were prepared in 1991. While the whole idea this time round was to avoid sketch maps, this has not been possible as some villages did not have recent aerial photographs and thus sketch maps were resorted to. Though mappers were instructed to use GPS co-ordinates to produce sketch where necessary, the instruction was not followed properly.
33. Sketch maps are normally prepared for those villages where more than one enumerator will be required. Such situations arise where the number of dwellings in those villages is 220 dwelling or more. These types of maps show rivers, streams, roads as well as location of dwellings by their range and important buildings such as schools, post offices, hospitals, clinics and other features serving as landmarks.
7.1.11 District Maps
34. District maps were updated in the field using the 1:100,000 scale. Where the area was covered by the 1:50,000 scale, the 1:50,000 topographical sheets were also used for ease of plotting geographical co-ordinates. These maps show the physical features of villages (roads, rivers, dams and hills). These maps were photographed on a plastic film, which could be reproduced at reasonable cost on a Xerox printing machine.
35. Latest information collected at locality level was transferred to the 1:50,000 sheet upon completion of the district. Supervisors obtained some of the information from the district authorities. The supervisor transferred the information onto the field copies before the mappers visited the field. This information was used to guide the teams around the locality, particularly information on new settlements. The verification of this information was done in the field by mappers. In addition to the supervisor copy and those used by each individual mapper, the supervisor
52 kept a master copy where all changes would be recorded. This is the copy that was returned to the head office, together with other documents.
7.1.12 Mapping work in Towns
36. For the 2001 Population and Housing Census, there were seven urban centres, Gaborone, Francistown, Lobatse, Selebi-Phikwe, Orapa, Jwaneng and Sowa, which are administratively referred to as urban areas. All these urban areas are mapped at scale of 1:5,000 by the Department of Surveys and Mapping, except for Orapa, which is mapped by mining authorities at scale 1:25,000 The problem with this map is that plot numbers on the ground do not match those on map, so street names were used when transcribing onto control forms.
37. In Gaborone there was a small problem with regard to undeveloped plots, especially in Gaborone west block 8, 9 and 3. This was resolved towards enumeration time as developments were now in progress. Where there were no plot numbers pre-listing and mapping procedures for the rural areas were applied.
38. Where plot numbers were available, no new numbering was done, but only pre- listing of dwellings in order to get an estimate of the number of persons residing in the plot. A listing of blocks of apartments and town houses was compiled and recorded on the map.
39. There were houses that occupied more than one plot and in those situations only one plot number was recorded. There were cases where plots had changed their status from private residences to commercial and these indeed necessitated some cartographic updating of the maps.
40. Towns were designated census districts like the rest of the rural districts. The coding system was slightly different from the rest of the districts and their localities
7.1.13 Geographical Coding System in Towns.
41. The coding system was arranged as follows:
o Gaborone: 01 district code and 01 village code. o Francistown: 02 district code and 01 village code o Lobatse: 03 district code and 01 village code o Selebi-Phikwe: 04 district code 01 village code o Orapa: 05 district code and 01 village code o Jwaneng: 06 district code and 01 village code o Sowa Town: 07 district code and 01 village code
53 42. Extension areas were designated as localities and were assigned a three-digit code between 001and 999.
7.1.14 Delineation of Enumeration Areas
43. Delineation of enumeration areas for both urban and rural areas started mid- December 2000 on a smaller scale because the fieldwork was still on going. It went into full scale by January 2001 and was completed mid-May 2001. Reproduction of enumeration area maps started immediately. Delineation started on time for distribution to the districts and the maps were then sent to the district authorities for verification of name spellings, missed localities, positioning of localities, etc.
44. An enumeration Area (EA) is defined as a smallest geographical unit that represents an average workload for an enumerator during a specified period of time (which is the enumeration time). In delineating an EA, a number of parameters were considered:
o non-overlapping of EA boundary into other EAs within and between localities and districts. o Geo-physical conditions of the area containing that EA and generally, anticipated climatic conditions at the time of enumeration. o The 10 days available time expected of an enumerator to finish enumeration of their assigned areas. o The need, as expressed by planners in the districts that enumeration area boundary should where possible follow the ward boundaries. This was not possible because the district authorities could not identify those boundaries on the ground. o Relating the 2001 census EAs to the 1991 EAs and by making efforts to maintain the same boundaries where possible. o The need to minimize instances where villages and other localities are contained in the same EA, particularly with lands areas. a) Geographical Coding System of Enumeration Areas.
45. Once information from the field reached the office, it was remodelled to what was desired for the 2001 Population and Housing Census and new enumeration area boundaries were drawn on the maps. The 1991 and the present EA codes were a four digit code. The 1991 EAs were maintained for comparison, subdivision being effected where necessary.
46. In rural areas, EA sizes ranged from 75 to 150 dwellings, or a population size of between 300 and 600 persons, depending on the Population density of the area and the number of vacant dwellings. In most densely populated areas like Mogoditshane, the average size of an EA could be 60 dwellings, in Towns and
54 major villages the EA size ranged from 110 to 220 dwellings or population size of 440 to 880 persons. In total there were 4100 EAs in 2001 as compared to the 1991’s 2627 EAs and that represent an increase of 57 percent. This increase was not only a consequence of increased population size but also of other factors such as the need to ensure comparability between the results of the 1991 and 2001 censuses. Furthermore there was need to arrange the EAs in such a way that the ecological character of an EA was not a mixture of different ecological type. b) Enumeration Area Map
47. After all the EAs had been drawn and reproduced in their final stage or format, separate copies of EA maps were reproduced for use by enumerators and supervisors during the enumeration period. A special file called the enumeration area file was compiled for each enumeration area. Amongst other documents such as the control form, this file contained a cutout map of the EA to be enumerated and was highlighted in red. This cutout map was glued to the inside of the front cover of the enumeration area file.
48. The map showed the EA, the EA boundaries denoted by + + + +, and all localities contained in the EA. It also showed parts of adjacent EAs, particularly those that had common boundary with the assigned EA. The assigned EA was also distinctly marked in red to distinguish it from any other EA, which appeared on the same cut-out map. The main purpose of these cutout maps was to enable the enumerator locate the EA assigned to him/her. Secondly to know the boundaries of the EA, to ascertain whether or not a locality belongs to his/her EA. Lastly to enable him/her to find way around the Ea.
49. In addition to the EA file, which was prepared for the enumerator, there was another file prepared for the supervisor, which was referred to as supervision area file. This file contained, where possible, a district map showing all the EAs under his/her supervision. The EAs were arranged in such a way that they were adjacent to each other. This map was called the supervision area map and its boundaries were highlighted in red. The supervisor was required to know boundaries of the supervision areas of all the enumerators under his/her control as well as EA and the district boundaries. The supervisor was also given a copy of the district map for him/her to be able to identify district boundaries, where the situation dictated so, as well as assigning new localities.
7.1.15 Constraints
50. Locality names: In a number of cases, localities had two different names depending on the political harmony or differences within the community. Mappers and enumerators were either threatened or being turned away or accused
55 of being partisan in local politics if they did not use the name preferred by either group. One of the serious issues was the boundary dispute.
51. Spelling of locality names: Spelling of locality names was a problem, particularly where the mappers and enumerators were not literate in the vernacular of that community. In certain cases, the pronunciation of such names by the locals differed from one member of the community to the other thus confusing the enumerators. Certain localities may appear to have been missed when in fact they appear under a different name.
52. Association of localities to village: Lack of documentation made the association of localities very difficult during the cartographic work, to the extent that mappers relied entirely on the district administration, local authorities and local people. It was very difficult to determine which localities were associated to which village. This led to conflicting assertions, in most cases all these problems had tribal and political connotations for which the Central Statistics Office or Census Officer should not associate.
53. A lot of localities had changed drastically in terms of status and size since the 1991 census; some lands areas had become villages, and some cattle-post areas had become mixture of lands and cattle-post. Redefining these localities in line with their level of development depended on mutual agreement amongst the District Council, Tribal Administration and the District Administration. How the authorities arrived at the conclusions is unknown because there were no official guidelines for such decisions.
54. Suitability of census vehicles: The four-wheel drive Nissan and the Isuzu vehicles proved unsuitable for the terrains that had to be traversed and mechanically they were overheating to the extent that a lot of time was wasted in waiting for the vehicle to cool down, resulting in a considerable delay in the work progress.
55. Repairs of census vehicles: Prior arrangement was made with Central Transport Organization to give census vehicles first priority in servicing and repairs. Surprisingly this did not work according to plan. The disappointing part was that some CTO workshops did not have the money to send the vehicles to private garages and even where the money was available the services at private garages left much to be desired.
56. Camping equipment: Some of the equipment used was repulsive, which resulted in some members of the team having to do without such equipment for a very long period. This indeed delayed our progress in a way because this lowered the moral of the mappers. Honestly the supplies section did not do a good quality job in this area
56 57. Mapping skills: Though mappers were given sufficient training on map reading, as well as sketch map drawing and plotting of Global Positioning System co- ordinates, provided with GPS receivers for collecting co-ordinates to plot them and produce well-scaled maps, the work in totality was not up to the mark. With GPS co-ordinates, it is possible to prepare a well co-coordinated map, something that mappers were instructed to do but was never done.
58. Office delineation of enumeration areas: Delineation of enumeration areas was done in the office instead of the field. It would have been ideal to delineate EAs in the field. This has its weakness in that consistency is also likely not be maintained, particularly in the size of EAs and other technical issues that needs to be taken into consideration in the field.
59. Geographical coding system: The three-digit coding system used for localities, where the first digit identified a locality by ecological type (i.e. whether a locality is a lands area, cattle-post, freehold farm, etc) and the last two digits sequentially numbering the localities from 01 to 99 could not cope with increasing number of localities (especially in lands areas and freehold farms). This occurred in districts, where localities associated to a particular village were more than 99. To maintain the three- digit code, a different code number for the first digit had to be used, so that two different numbers were identifying the same locality type.
7.1.16 Recommendations.
60. Official definitions of different settlements: There is need for the development of official definitions of various settlements to ensure consistency and uniformity in classification and application between any two operations.
61. Publicity: Though cartography forms an integral part of any successful census, it has never been adequately publicised. Mappers themselves had to publicise the cartographic work through the district authorities. Cartography needs to be well publicised for the success of the census exercise.
62. Suitability of Census Vehicles: Vehicles used for the census cartographic work must be very strong. Therefore in future Toyota 4x4s and Land rovers are recommended. It must be understood that cartography forms the backbone of any successful census. Using unreliable vehicles could paralyse the whole operations of the census.
63. Repairs of Census Vehicles: This is one area that needs serious attention. Though prior arrangement was made with CTO to give census vehicles first priority in repairs and servicing, the services were totally not pleasing. Lack of proper and good running vehicles could compromise the whole exercise..
57 64. Training of cartographic staff at all professional level: There is an urgent need for training of cartographic staff at professional and semi-professional level given the sophisticated GIS equipment and software used.
65. Geographical information system is a new field in this country and furthermore there are no trained people in this area in the department. As a result, the GIS positive results will not be as good as anticipated without the requisite training
66. Locality Association: This is one area that needs immediate attention because it is not clear as to whether association of localities should be based on proximity or origin, by origin we refer to the village where people originate.
67. Cartographic Committee: There is need for a cartographic committee during the census period. The committee will advise on cartographic matters relating to the census. Although for the 2001 Population and Housing Census Technical officers somewhat played that role, such efforts were not enough.
68. Revision of geographical codes: The present codes were prepared a long time ago and therefore do not represent a real life situation. For instance Central District has been divided into five census districts, as such there is need for a district code e.g. (50) for central district and Serowe/ Palapye (51) etc.
7.2 C artographic Overview.
69. This report gives an overview of the some problems encountered during census preparatory and the actual enumeration period.
70. Gaborone (01): Gaborone being the capital city was assigned a geographical identification code 01, mapping in Gaborone was not as difficult as in rural areas, as street names and plot numbers maps prepared by the Department of Surveys and Mapping were used, at scale of 1:5,000. The main problem with these maps was coverage, e.g. Gaborone north, around Ledumang area was not covered on the map. The area around the Dam, though surveyed and plots allocated and numbered, there is no map covering it, as a result a sketch map was prepared.
71. The developing areas of Gaborone west, Phase 4, Block 3 and Block 8 were pre- listed towards enumeration period, but still at the time of pre-listing there were so many empty plots. The other problem was inconsistency of plot numbers, duplication of extension numbers and maps with insufficient information. All these problems have since been communicated to the department of Surveys and Mapping through the map users committee. Most of these maps are not in digital form.
72. Francistown (02): The pre-listing started on 12 November 1999 and was completed on 10 December1999. The situation in Francistown was much better
58 than 1991 as squatters of Somerset were being relocated to a new area called Gerald Estate. The relocation indeed delayed our work progress as it was carried out few weeks before enumeration. In Francistown 1: 5,000 street names and plot numbers maps prepared by the Department of Surveys and Mapping were used. There were problems as in some areas numbers only existed on the ground but not appearing on the map. This indeed posed lot of problems during enumeration delineation. But otherwise all the mentioned problems were addressed.
73. Lobatse ( 03 ): The pre-listing in this town started on 12 November 1999, and was completed on 7 December 1999. One old location of Peleng did not have street names and plot numbers as a result the rural system of numbering was used. One location named Thema is well developed but had no maps showing plot numbers except a street plan, which was not suitable for our work. Plot numbers were appearing twice on the map and this indeed caused lot of confusion to enumerators.
74. Selebi-Phikwe ( 04 ): The town is also mapped at scale of 1:5,000, thus street names and plot numbers prepared by the Department of Surveys and Mapping. Like in other towns the problem of out-dated maps existed. There were squatter areas of Mekoro, which were pre-listed during the cartographic exercise and demolished before enumeration. The other major problem in this area was the BDF camp, Prisons and Water Utilities, which are outside the Selebi-Phikwe planning area. Surprisingly they get their services from Selebi-Phikwe but during election time, the polling station for a member of parliament is in the Mmadinare constituency and for the council in Selebi-Phikwe constituency. There was another problem particularly in BCL houses and hostels where plot numbers on the ground were different from those on the map.
75. Orapa (05). The maps for this township prepared by township authorities, were difficult to use as they do not have plot numbers. Instead the numbers are designed such that they are numbered from one to the last number for each street.
76. Jwaneng (06) The area is mapped at scale of 1:5,000, thus street names and plot numbers prepared by the Department of Surveys and Mapping. The maps for this area were fairly okay - the only problem in this area was the squatters. These squatters were built in developed commercial plots and thus were treated like dwellings in rural areas.
77. Sowa (07): This soda ash mine is mapped at a scale 1: 5,000, thus street names and plot numbers prepared by the Department of Surveys and Mapping. The area is fairly straightforward.
78. Southern District (10): This is one of the biggest districts in the country, it has been sub-divided three times. The district is partly covered by the 1:50,000 topographical sheets and the other part is covered by 1:250,000, the 1:50,000 have been reduced and mosaic to 1:100,000. Boundaries within the district were a
59 problem, especially between Kanye/Moshupa and Barolong. Association of localities also proved difficult, as it was not clear on how it should be done. During the cartographic inspection by CSO senior staff it was reported that some dwellings were missed within the EA, in some cases the whole EA was missed. Apart from missed dwellings and localities, the cartographic work went well because the numbering and coverage was within the accepted limit.
79. Barolong (11): Barolong sub-district was fairly small in 1991 but for 2001 Population and Housing Census the boundary was extended. The major problem in this district was the extended boundary. Missed dwellings within the EAs and localities were reported, some of the reported problems were taken on board, as the team was still in the area but some were very difficult to deal with as they concern boundaries, and CSO has no mandate to deal with boundary disputes. This indeed was left on the hands of district authorities and it proved unsuccessful. Otherwise the cartographic work went okay, the district is mapped at scale of 1:100,000.
80. Ngwaketse West (12): This is a new census district within the Southern Administrative District that was part of Southern District in 1991 census. Though the district is small and fairly straightforward, missed dwellings within the EAs and localities were reported. The district maps are at scale of 1:100,000.
81. South-East (20). This is one of the smallest districts in the country. In this district only missed dwellings within the EAs were reported, the only problems was in farms as most of them were found locked and therefore access was not easy. Whites who own most of these farms who were totally not co-operative. The district was supposed to have been divided into two census districts, as there is a DO’s office in Tlokweng, this was discovered after the pre-listing. The district is mapped at scale of 1:100,000.
82. Kweneng South (30): This district has so many problems such as boundary problems etc - the district is subdivided into two census districts but it was not clear whether the other part should be called Kweneng south or East. After lengthy discussions with the district administration, it was decided that it should be called Kweneng South. A number of omitted localities in this district were reported during the cartographic field checks. One of the major problem in this area was the size of lands areas, most of the lands areas in this area had more than one EA and this indeed had a bearing in EA delineation as maps prepared for lands areas are in district maps. This makes it very difficult to show or draw a boundary between the two EAs. Mogoditshane was one of the problematic areas in this district because of its multiple households - in one instance one dwelling mbcontained 65 households. These problems resulted with a very serious underestimation of the number of enumerators
83. Otherwise the cartographic work went okay. The district is mapped at scale of 1:100,000
60
84. Kweneng West (31): This census district was there in 19991 but surprisingly it was not assigned a geographical code. The boundary between it and Kweneng South was not clearly identified. Cattle-post areas dominate the district, as such lot of omission was reported during the cartographic inspection, the reason being that cattle-post are not easy to locate by outsiders. Inside the district there is a very big army camp, access and listing in the camp was not difficult as there was full co-operation of the army authorities. Otherwise the cartographic work went well in the district except the abovementioned problems. The district is mapped at scale of 1:100,000
85. Kgatleng District (40): There were no reports on the missed localities except missed dwellings within an EA in the western part of the district. The other major problems were developments at the edge of the village as there were so many undeveloped plots during the pre-listing exercise. These indeed caused lot of problems during EA delineation. Otherwise the cartographic work went okay. The district is mapped at scale of 1:100,000
86. Central Serowe/Palapye (50): Many problems were reported in this district during the cartographic inspection. Along the Mmashoro area it was reported that some cattle-post were missed out and duplication of geographical codes was a concern, and a number of houses within EAs were found missed. On the eastern side of the Railway line only missed dwellings within an EA were reported, this indeed was very serious as in some cases half the EA was missed. Boundaries between the villages were another major problem. Otherwise the cartographic work went okay. The district is mapped at scale of 1:100,000
87. Central Mahalapye (51): Missed houses in this census district were reported particularly in the western and eastern part of the district. A few cases of missed localities were reported in the western part of the district. Another issue was the duplication of geographical codes in this case locality was assigned code similar to another locality. Apart from the abovementioned problems the cartographic work went well. The district is mapped at scale of 1:100,000.
88. Central Bobonong (52): The district is covered by the 1:100, 000 map series. Duplications of geographical codes was observed during the cartographic inspection and missed dwellings within EAs were discovered. There were no reports of missed localities except change of geographical codes particularly within the Mmadinare area was observed. Otherwise the cartographic work went well in the district. The district is mapped at scale of 1:100,000
89. Central Boteti (53): This census district is covered by 1:250,000 map series. Not many missed localities were reported, only missed dwellings within the EAs were reported in the district. Boundaries in some of the areas were indeed a problem, particularly where the village and lands areas are close to each other. But otherwise the cartographic work went well.
61
90. Central Tutume (54): This is one of the most difficult and problematic districts in the country in terms of boundary disputes. A number of localities were observed missed during the cartographic inspection and these were mostly cattle- post areas. The villages in this district are close to each other to the extent that determining village boundaries is extremely difficult if not impossible and this indeed resulted in geographical coding extending to other villages or dwelling numbers. Ward boundaries in major villages were also a big problem. Missed dwelling within the EAs were reported. The district is mapped at scale 1:250,000.
91. The issue of locality association in this district was a problem. Boundary disputes are very sensitive, as such, they are better handled by the villagers or communities themselves. The boundary between Tutume sub-district and Chobe district need to be re-visited or clearly defined and that between Tutume district and Ngamiland East district is also problematic e.g. Phuduhudu administratively is in Central Tutume but for operational purposes it falls under Ngamiland South as it is near Maun. For operational convenience, people in that area were enumerated under Ngamiland East in 1991 instead of Central Tutume. Boundary problems in this district need immediate attention as they proved very serious in the 2001 Population and Housing Census.
92. North-East (60): The district is mainly dominated by village areas and freehold farms. There were few reports of missed dwellings. Other problems were in the farms as most of them were found locked. It must be emphasised that though the district is small with less problems the mapping team assigned to this area did not do a quality job. The district is covered by 1:100,000 map series and almost all villages are covered by 1:10,000. In spite of such problems the cartographic work went well.
93. Ngamiland South (70): There were some boundary problems between Ngamiland South and West. Even the naming of Ngamiland South, at first they wanted it to be called Ngamiland East. There were reports of missed localities and dwellings. Duplication of locality codes was another problem reported. Misplacement of localities on the map was another matter and this indeed was prompted by the fact that the district map for the area does not have a grid, only spherical co-ordinates values. As a result of that plotting of GPS co-ordinates was not easy. The district is mapped at scale of 1:350,000. Otherwise the cartographic work went well.
94. Ngamiland West (71): Like other districts, there were reports of missed localities and missed dwellings. Missed localities reported were mostly in cattle-post and lands areas, otherwise the cartographic work went well in the district. The district is mapped at scale of 1:350,000.
95. Chobe (72): The district is mostly dominated by villages and safari camps, as there are no cattle-post and lands areas. Only missed dwellings and camps were
62 reported. There are commercial farms in the district, some of those farms were reported to have been missed during the cartographic assessment because they were locked at the time of pre-listing. Apart from the abovementioned, the cartographic work went well. The district is mapped at scale of 1:350,000
96. Delta (73): This area was not pre-listed because the water level in the delta was high and thus boats could not be used because it was risky. The only means of transport in this area were boats or air. The pre- listing was done during enumeration period using a hired Botswana Defence Force helicopter. Five temporary cartographic staff were used as enumerators so that they could do the listing at the same time. The district is mapped at scale of 1:350,000
97. Ghanzi (80): This census district is dominated by cattle-post and freehold farms. It was supposed to have been divided into three census districts Ghanzi, Charles Hill and the CKGR but because of boundary problems that was not possible. It was therefore divided into two. This is one of the smallest district in terms of population but vast in area. There were a number of problems in this district, such as boreholes constituting localities and worst still these are not easy to locate. Also, Land board and the Council do not have proper records for these boreholes. Most of these boreholes were missed during the cartographic work. The other major problem was spelling of locality names. Of serious concern to be noted was the mis-positioning of localities on the map. The cartographic work in this district was not up to the mark. The district is mapped at scale of 1:500,000
98. Central Kgalagadi Game Reserve (81): There was no cartographic inspection carried out in this census district. Reports from the pilot census and main enumeration indicated that the work was well done except that metal plates were affixed on trees because of the type of structure in the area. There are only five localities in the game reserve, mapped at scale of 1:500,000.
99. Kgalagadi South (90): Like other districts there were reports of missed localities, the missed localities were mainly cattle-post areas. The other major concern was insufficient information on the district map, like major roads not shown on the map or shown on incorrect position. The EA size for the CDC or Banyana farms was said to be too big, taking into consideration the distance that had to be travelled. The district is mapped at scale of 1: 500,000.
100. Kgalagadi North (91): There were reports of missed localities in this district, like in other districts these were mainly lands and cattle-post areas. Indeed this was prompted by the fact that cattle-post and lands areas are difficult to locate without a guide.
63
8 Pilot Census
8.1 Background and Objectives
1. A Pilot Census was conducted in selected areas of Botswana from 23 August to 31 August 2000. Overall, 61 Enumeration Areas were purposefully16 selected, covering all the districts, with the exception of a few sub-districts: - Barolong, Central-Boteti, Central-Tutume and Central-Bobonong. The criteria in the selection process were based on the fact that areas, which are likely to pose problems in the main census, could provide a learning experience. Although it was initially planned that the sample should cover all the districts and sub- districts, this was not possible because the mapping and house listing exercise had not been started in the four sub-districts at the time of the Pilot Census.
2. As with all pilot censuses, the main objective of the 2000 Pilot Census was to assess the adequacy of preparation for the 2001 Population and Housing Census. To the extent that not all the operations of the pilot census were smooth, the exercise was very useful in directing the course for the main census.
3. Over 60 percent of the close to half a million Pula budget for the Pilot Census was spent on personal emoluments – as either subsistence or census allowance. Maintenance of census accounts ledgers was not very satisfactory hence the actual expenditures for the exercise are rather difficult to disentangle
8.2 Census Questionnaire
4. As part of the 2000 Census Project initiative, a minimum set of topics (core questions), which were mandatory to all member states, was prepared. Countries were given the latitude to phrase the questions the way they deemed appropriate and furthermore to include additional topics as dictated by their needs.
5. This collaboration effort brought in many topics some of which had not been tested before. These include Place of birth, Parental Survival (Orphan hood), Place of usual living five years ago, Usual Economic Activity, Space heating energy source, Refuse disposal and Ownership of household durables. Through the usual process of consultation and dialogue with users regarding their data needs, a few more new topics were included. Such topics were Language, Religion, Training (post secondary), Ownership of land (for cropping), and Home based Care. Suffice it note that, apart from the SADC requirement, there was demand for data on the extent of orphan hood at the national and district level.
16 In essence this seriously limits the extent to which the Pilot Census data could be analysed
64 6. In the light of the AIDS scourge, data to enable the assessment of home-based care programs was required. Closely linked to that is the need to measure accurately the number of orphans, the majority of which are due to AIDS. In spite of pre-pilot apprehensions that the public would find the questions on those topics objectionable, there were no reports of refusal.
7. With regard to parental survival, many members of the public questioned the logic of asking for information about the survivorship of biological parents. This seemed quite valid since many children whose biological parents (whilst still alive) have absconded could not qualify to be orphans on the basis of the technical definition. Indeed such children could be worse off than those who met the definition.
8. There were no reports to suggest that questions on language and religion were not well received. Notwithstanding that, there were however grey-areas regarding the boundary between a language and a dialect. To that extent, it is possible that some enumerators might have misclassified some languages, especially with a pre- coded list.
8.3 Personnel
a) Preamble
9. The socio-economic development that was witnessed during the inter-census period had an effect on the structures and institutions that have traditionally been relied upon in the conduct of censuses. For instance, with the then levels of unemployment, school leavers were a pool from which to draw field personnel. Furthermore, because of relatively high per capita GDP, Botswana was not eligible for the UNFPA17 technical support enjoyed in past censuses. Against such background, the Pilot Census was very important in providing the needed training whilst offering an opportunity for an assessment of structures put in place for the census.
b) CSO Staff
10. The Central Statistics Office has been plagued by mass exodus of trained personnel in recent years and this was further compounded by the inability to attract staff of the required calibre. As a result of the exodus, efforts to maintain the Census Office functional during the inter-census period proved futile – the office was set up late 1999 just in time for the 2001 Population and Housing Census. Notwithstanding that, the Census Office core staff benefited immensely from the numerous training workshops and seminars held within the region as part of the SADC 2000 Census Project.
17 Through the provision of technical advisors and payment of their salaries and per-diem
65 11. Most of the staff that comprises the pool from which to draw Technical Officers18 was relatively young and consequently inexperienced – 25 of the 54 CSO technical officers were officers at Grade C3 (Assistant Statistician or Statistical Officer). The poor quality of supervision that was manifested through the quality of the data could be partly explained by the composition of the census Technical Officers.
12. Without adjudging the success of the pilot exercise, optimising on the available personnel to conduct the Pilot Census was perhaps a good beginning. It was felt that with close nurturing and provision of supplementary training on weak areas, demonstrated effort and commitment shown would guarantee the success of the 2001 Census.
c) District Census Officers
13. Against the background of success stories achieved by mobilising local authorities, it was thought prudent to follows the same practice. Pursuant to that, a countrywide tour was conducted in March 2000 to brief the various stakeholders on the census programme and also to solicit for support. This was followed by countrywide recruitment and training of 45 officers who played the role of District Census Officers. Such officers were drawn from officers of the rank C3 to D319. In order to encourage participation in census activities, authority was granted to pay such officers “Census Allowances” whenever they were involved in census work for a reasonable period of time. Needless to say, what constitutes a reasonable time has always been a very dicey matter.
14. The level of participation demonstrated during the Pilot Census was generally good and therefore very encouraging. Suffice it to mention that District Commissioners and Council Secretaries, whilst not directly involved in the day- to-day activities, played an important role in overseeing census operations in their districts. Notwithstanding that, in some districts, some discernable indifference to the census was witnessed among a few district chief executives.
15. Commitment by District Census Officers was generally very good. However, participation by District Census Officers in a few districts was reduced to mere formality – in essence discharging an obligation without any display of commitment. For instance, in some districts enumerators were recruited with very little or no screening. Consequently the quality of enumerators was below expected standard. This occurred in spite of the fact that instructions and guidelines regarding recruitment were very clear on the matter.
18 Grades C3 and above 19 The pool comprised Assistant Planners, District Officer (Development, Lands), District Agriculture Officers, Education Officers, etc
66 d) Enumerators
16. Of the about 100 enumerators recruited for the pilot census training, the majority were teachers. Although past censuses have used teachers as enumerators, the high levels of unemployment provided yet another pool from which to draw enumerators. A decision was taken to ensure that half the recruits were teachers with the other half being constituted by eligible unemployed.
17. Whilst considerations were being made regarding possible use of the unemployed, there was, sadly, a political dimension to the matter. Politicians and traditional leaders alike were very vocal in public forums regarding the use of the unemployed in national exercises such as a census. The argument was that teachers already had jobs and therefore they felt that efforts must be made to cushion the hardships experienced by the unemployed. However, the general feeling in the SADC region was that there is less commitment on their part (unemployed) – a factor with a bearing on the quality of the results. It was thought, generally, that civil servants would do a better job owing to the fact that they have a job attachment and therefore could be traced and disciplined in case of poor quality work. The problem therefore reduced to that of optimising on quality of census results whilst at the same time minimising on unemployment levels.
18. In view of the foregoing, the pilot census therefore presented an opportunity to test the effectiveness of the two sources. The few reports, which made comments on the effectiveness of the two groups, preferred teachers to the unemployed. However, to the extent that most of the reports were silent on the evaluation of the two groups, it was initially thought a 50/50 split might be a solution. During a nationwide tour of the districts conducted in November 2000, local authorities in most districts were apprehensive about the use of the unemployed. In a few districts, it was felt that with good screening and supervision the unemployed could perform very well.
e) Support Services
19. Services by the Administration and Supplies sections of the Census Office were far below expectations and consequently affected the Pilot Census Operations in a number of ways. Needless to say, services by these sections are very essential to the success of both the Pilot Census and the main census.
20. The laizef-faire approach by CSO’s Supplies Unit resulted in untold frustration because procurement and delivery of materials was only possible after numerous follow-ups and constant reminders. In some instances the core staff of the Census Office had to play a leading role in the procurement of materials and supplies. Because of red tape and bureaucratic procedures, efforts to recruit a Census Supplies Officer before the pilot census could not bear fruit. Belabouring this
67 point is perhaps not necessary since the quality of service delivery could best be assessed against this background.
21. Yet another setback was in the delivery of administration services. This was due in part to protracted process to terminate employment of an inefficient officer (whose post was earmarked for the Census Office) and the lengthy process to replace the Principal Executive Officer who transferred just a few months before the pilot census. Although the two posts were filled eventually, this was too close to the pilot census to allow for a good service delivery. Inefficiency in payment of field staff was a major concern at the end of enumerator training and after the enumeration. Worse still, the delayed payment of CTO drivers overtime upon completion of the exercise proved quite costly to the office. This was due to the fact that, whilst awaiting20 the processing of overtime payment drivers attracted subsistence allowances. Furthermore the office was paying for both the driver and vehicle hire rates. Apart from the resulting unnecessary payments, the delay in the return of vehicles caused a lot of inconvenience to other government departments.
8.4 Training a) Preamble
22. A lot of attention must be paid to the training of field personnel if the enumeration must succeed. A five-day training period was planned for the Pilot Census but this was changed on the realisation that schools were closing later than anticipated. Censuses rely on school facilities for training and furthermore, enumerators are drawn from among schoolteachers. Late closure of schools impacted on both the duration of training and the timing of enumeration. The training period was reduced effectively to three days, which was just insufficient. Consequently materials that were covered during the pilot training were more than the time allocated. This resulted in making some sessions to continue beyond 2000hours. Even with the most attentive and conscientious participants, beyond a certain level, attendance is reduced to mere formality. Consequently, the quality of training was far below expectation.
b) Strategy
23. Although the reading of manuals enables the coverage of all aspects of the enumeration process, this approach may not allow the participants to grasp and synthesise all the important points. More emphasis must be placed on balancing theory with field practice.
20 Drivers loaned from outside Gaborone were not willing to return back to their departments before payment of overtime.
68 24. To the extent that training was conducted over a weekend, pre-occupation with weekend related activities reduced the concentration of participants. For optimal concentration training must be conducted during the week.
25. Class demonstration of interviews is very useful but has serious limitations arising from the fact that they present a simulation in a controlled environment, which is not necessarily in congruence with the practical situation. By asking for volunteers to play the role of an interviewer and a respondent the process is reduced to a self-selection sample of only those participants who feel confident, albeit not necessarily fully conversant with concepts and definitions. Even if the rest of the participants complete the questionnaire individually and furthermore the results are discussed, experience has shown that the majority of participants are just sailing along without understanding important concepts and definitions. An alternative approach, which was tried at one of the training centres, is that of participants asking the questions in turns. In this case the respondent is drawn from either drivers or someone other than a participant.
26. Interview profiles have serious limitation in that information about household members is so structured that it literally gives the response to be recorded by the interviewer on the questionnaire. The other limitation is that since there is no probing, any answer can always be justified. Given such background, it was felt that in the main census training, enumerators use profiles only for homework exercises.
27. Experience from the Pilot Census has showed that the first two days of enumeration were reduced virtually to training sessions. Consequently, there has been a change of strategy in favour of more emphasis on practical interviews. In the main census training, two days were set aside for interviewing of households in localities near the training centres. Unfortunately this strategy meant over- burdening such households at the time of enumeration. However, efforts were be made to explain to affected respondents regarding the need for such trials. c) Duration of Training
28. The duration of training was too short – a factor that could have compromised on the quality of the pilot census results. Reports from the field indicated that the first two days of enumeration were reduced to training session. At the end of the three-day training some enumerators were reported to be blank literally.
29. On the basis of field reports, it was decided to have a five-day training for the main census. Provision has also been made for an extra day for administrative logistics and the deployment of field staff to areas of assignment.
69 d) Quality of trainers
30. Reports on the Pilot Census training, especially from the District Census Officers, indicate clearly that preparation for lessons by some Technical officers21 at some of the centres was below expectation. This actually points to the fact that the composition of trainers for some centres was not looked at closely. In view of the fact that the team for Gaborone comprised a residual22 of Technical Officers, the quality of training was reportedly below par.
31. Two factors may have undermined the quality of trainers; i) the duration of the Technical Officer’s Training was too short and ii) lack of commitment23 on the part of participants. Furthermore, the assumption that many members of the Census Standing Committee understood the questions, concepts and definitions was probably a gross overstatement. To the extent that for the main census, trainers were to be drawn from non-census officials, the importance of selecting of trainers of the required calibre was to be given serious attention. This therefore raised the need for re-training, not only of the technical officers but the District Census Officers as well. There were reports of conflicts between District Census Officers and Technical Officers regarding the interpretation of concepts and definitions. Worse still, in some areas it was reported that the DCOs understood the concepts and definitions better than the Technical Officers in spite of the fact that the realm of their work was confined to administrative matters.
8.5 Publicity and Census Education
32. Information, education and communication are important vehicles to reach the various segments of society. Various communication mediums were used to publicise the census. The Census Publicity Officer addressed almost all district full council meetings, as well as District Development Committees and Planning and Management Committees. In turn the councillors and parliamentarians made their constituents aware of the census and the preparatory activities such as mapping and house listing.
33. An incident of negative publicity was observed in Botalaote in the North East district where a local priest who, it was reported, not only prophesied the coming of census officials but also raised false alarm regarding the fact that all those enumerated will perish. While the enumeration of the village took place with full co-operation from households, this was an important lesson to be guarded against in the main census.
21 Technical Officers played the role of Pilot Census Enumeration Trainers. 22 These were Technical Officers who could not be posted elsewhere for special reasons. 23 Many participants missed some training sessions
70 34. A note advocating for census support24 was prepared and published just before the pilot census. This note served a very useful purpose in addressing pertinent issues likely to be raised by respondents. In retrospect, a wider coverage could have been achieved if the note had been translated from English to Setswana. With hindsight advantage, the note was eventually translated into Setswana.
8.6 Enumeration
a) Preamble
35. Ideally a census must be conducted in the shortest possible time to avoid double or multiple counting. With a moving census night such as the case in Botswana, the situation could be worse than in the alternative approach of a fixed25 reference night. With the idea to reduce the enumeration period, it was planned to complete enumeration in seven days rather than the usual duration of 10 days. The fact that the size of the EA was not reduced accordingly, shows clearly that the plan was not only ambitious but was probably not well conceived. Consequently, by 29 August 200026 reports reaching the Census Office indicated that enumeration in more than half the selected areas was not complete. A decision was then taken to continue enumeration for two more days, redeploying personnel from the completed EAs to assist in areas where enumeration was behind. Even with the extra days and redeployed personnel, enumeration could not be completed in some areas.
b) Pilot Census Timing
36. Apart from the disproportionate size of the EA to the enumeration period, there was the issue of timing of the Pilot Census. Actual censuses are usually conducted in a more-or-less controlled environment with regard to the requirement of the census – schools calendar made to be in consonance with the enumeration period, international trade fair and agricultural shows, large scale congregations, etc scheduled outside the census period. The Pilot Census could not enjoy the privileges of the actual census hence the scheduling of activities listed in the foregoing affected the enumeration process quite a lot. Schools closed late and the fact that censuses rely on schoolteachers meant that the enumeration period was pushed towards month end. Month end has its own problems, people travel a lot, and some respondents have a faire share in alcohol intake thereby making enumeration difficult if not impossible. In order to conduct a successful interview many call-backs were necessary. This again had the effect of lowering the daily output.
24 Census Advocacy Note 25 However this approach also has a limitation since respondents will not recall correctly when the interview day gets further away from the reference night. 26 The seventh and last day according to the initial plan.
71 37. Staging27 the International Trade Fair during the Pilot Census period presented a lot of problems since the two activities compete for the same people. Even if people are aware of the census and are willing to be enumerated, the trade fair will always win. Publicity efforts in the rural areas met similar fates because people attended agricultural shows during the week reserved for census education and familiarisation tour.
c) Enumerating the Working Population
38. The working population proved to be very difficult to enumerate since they spend the whole day at work. Although the Census Act provides that enumeration could go on until 2200 hours, without provision for lights and the fact that some areas (e.g. Tsolamosese) are dangerous at night, enumerating the working population was very difficult. Even though provision of torches was accommodated in the census budget, the main census still had to contend with access to dangerous places at night.
d) Large EA Size
39. Many Enumeration Areas comprised well above the stipulated size. For instance in Lehututu the piloted EA contained 249 dwellings, much higher than the maximum size range of 110 – 220 dwellings in urban areas. Indeed there were many other areas were the size of the EA was about one-and-half times as much as the normal size. With oversized EA it was impossible to complete enumeration within the seven-day enumeration period that was being piloted. Steps were taken to ensure that EAs are of an acceptable size in the main census. Some of the piloted EAs were re-delineated and downsized.
40. Acute shortage of accommodation in towns and cities has generally resulted in many households per dwelling. Given that the size of the Enumeration Area is, to a large extent28, based on the number of dwellings rather than households, the enumerator’s workload in such areas was too high. In delineating EAs this factor was finally taken into consideration.
e) Effect of Morbidity on Enumeration
41. Yet another complication arose because of the level of morbidity in the country, which has reached unprecedented proportions. A conscious decision was taken to cover the Princes Marina in the selected areas since it was hoped that the hospital would provide its own learning experience. With the bed occupancy beyond the holding capacity of the hospital, interviewing inpatients in such a situation was
27 Or should we say making Pilot Census enumeration coincide with the fair since the two activities were scheduled independently.
28 Distance between localities, type of terrain, etc are also considered.
72 difficult moreover those patients demanded privacy during interviews29. Although the usual plan is to obtain basic data from hospital registers for patients who are seriously ill, there was a sad incident of a patient who died during the course of what seemed to be a normal interview. With the current state of our health in mind, a decision was made to hold discussion with health officials with the view to make arrangements safeguarding privacy during interview of inpatients.
42. Whilst enumerating inpatients has its own share of the problems in the enumeration operations, deaths affected the enumeration quite a lot. In small villages, almost all households attend death related formalities thereby slowing down the enumeration process because of revisits to households. Being recruited for the census does not stop close relatives of such persons from dying. With a culture such as Botswana’s, it is inevitable that some census personnel will be excused during the course of enumeration to attend funerals of close relatives. In order to accommodate problems associated with deaths, it was resolved to have extra personnel on standby and this had serious cost implications.
f) Enumeration in Lodges
43. Whilst lack of co-operation was not reported in the enumeration of lodges and hotels, the fact that all the questions in the institutional schedule were to be answered by hotel guests resulted in irrelevant data being collected. For instance the economic activities, occupation, industry, etc of tourists relate to their countries hence was irrelevant for the purpose.
44. Another problem experienced was with regard to the fact that some visitors could not communicate in English or Setswana. This resulted in many guests failing to complete the forms delivered to hotel managers. Moreover, by the time the enumerator revisited to collect the completed forms, some guests had already checked out of the hotel or lodge. It does seem the enumerators did not account for all the forms delivered to the hotel since some forms arrived by mail two weeks after the end of enumeration. This also pointed to lack of support from hotel/lodge management in ensuring completion of the forms.
g) Enumeration in the Delta
45. Following the heavy rains early 2000; access to some parts of the delta by boats was impossible. With the water level in the delta higher than normal, some islands were submerged under water – a factor that made travel by boat a dangerous undertaking. On realising the extent of the problems, efforts were made to hire helicopters or fixed-wing aircraft for use during the Pilot Census enumeration. The BDF helicopters were not available until late during the enumeration period.
29 This Census Act also dictates so
73 A fixed-wing aircraft was hired for a day but then it landed in a different place – the selected area was Jao Flats but the airstrip is located at Jao Camp. With no means of transport from the camp to the selected area, enumeration failed as a result. This was a useful lesson because provision was eventually made for use of helicopters and fixed-wing aircrafts for enumeration of the delta.
h) Enumerating Central Kgalagadi Game Reserve
46. Because of heavy sands, traversing the CKGR is indeed an arduous task. Firstly the area is sparsely inhabited thereby necessitating extensive travels through very heavy sands. This type of terrain therefore requires very robust vehicles that could withstand driving at low gears for a very long period. Secondly, there are no fuel points and furthermore food, water and other survival essentials are not available hence must be arranged in advance. Even with enough supplies of those essentials, rough vehicle turns and twists spoil all the food commodities. On account of the problems outlined above, neither the familiarisation tour nor the enumeration could be completed.
47. Yet another problem was that relating to the effect of the relocation of the inhabitants of the CKGR outside the game reserve. Resistance to the relocation complicated matters because; people viewed any government officers as perpetuating or being part and parcel of the relocation process. The census team was made aware of the fact that they were likely to encounter some resistance to enumeration if the Chief of New Xade did not accompany them. Last minute arrangements were therefore made for a vehicle to be provided to transport the chief during both the familiarisation tour and the enumeration process. This had cost implications since the Chief was paid both the census and subsistence allowance.
8.7 Cartographic Work
48. In conducting the Pilot Census, the intention was to select a sample to cover all administrative districts and sub-districts. However, while all the main districts were covered, a few sub-districts were not covered during the exercise because mapping and house listing had not been done in those areas. These were Central – Tutume, Central Boteti, Central Bobonong and Barolong. With the aim to cover Chobe in the pilot at all costs, only targeted areas were listed and mapped. Reports from some Technical Officers and District Census Officers indicated that the quality of mapping was poor in some areas. Indeed some of the cases might be true but it may also have been due to map reading skills on the part of both the Technical Officers and the District Census Officers.
49. Mapping in the Chobe District was done on the eve of the Pilot Census, specifically to enable the exercise. Reports from the pilot indicated that the quality of house listing was very poor and therefore recommended a re-listing of
74 the district. It was reported that for given localities only the first dwelling was numbered, thus creating a problem if that dwelling was not identified during enumeration. Faced with such problems, the enumeration team tried to improvise even though the resulting numbering proved quite problematic as manifested during machine editing (Consistency and Corrections - ConCor runs). Even though the mapping work was done rather in a rush, there was no justification for the poor work.
50. Some cattlepost around the Mashoro areas were reportedly missed during listing. In order to address such problems, the Census Office had planned to use borehole location maps since settlement patterns in cattlepost areas is closely linked to the location of boreholes. However, a serious limitation of the maps is that they do not indicate which borehole is operating and which one is not.
51. A few reports indicated that some localities contained dwelling numbers supposedly for different localities. Quality assurance measures were taken to check, verify and correct such mistakes in readiness for the census.
8.8 Census Data Processing
52. Computer hardware was not delivered until one and half months after completion of data collection. By this time data entry was complete, with the exception of machine editing. This was made possible through makeshift arrangements using almost obsolete computers awaiting disposal at the Supplies Department. Other equipment such as heavy-duty printer and the file server were not available two months after data collection. Although data processing could still have awaited the arrival of the equipment, processing under the conditions of the actual census could not be simulated.
53. There were other logistical problems, for instance the room designated for data processing was changed during the course of processing of data. With the required infrastructure not available in the new room, the issue of tendering for cabling and trunking arose. Having contracted a private consultant for the data processing, government bureaucratic procedures brought in a tale of frustrations on the part of the consultant. All this logistical bottlenecks were a very important lesson and were useful in directing the course for the main census.
8.9 Pilot Census Results
54. Being a purposive sample, the inherent weakness in the design seriously limited the analysis and consequently the usage of the Pilot Census results. Indeed the purpose of the exercise was more to do with testing the operational aspects of the census preparations rather than produce useful results. Results produced were merely a bi-products of efforts to check the effectiveness of data processing
75 mechanism – data entry and tabulation systems as well as data analysis. In that respect, not much could be read into the results beyond the foregoing. Notwithstanding that, when measured against the fact that there were lot of lessons learnt and the experienced gained, the pilot census met the desired objective.
76 9 Development and Design of Census Questionnaire
9.1 Introduction
1. The development of the census questionnaire is without doubt a very elaborate process. Censuses serve diverse and myriad users whose needs as well as those linked to important contemporary issues must be addressed. In developing the household questionnaire, the following factors were taken into consideration:
o regional requirement (SADC 2000 Census Project) o data needs for contemporary issues o basis of census enumeration, whether “de facto” or “de jure” o selection of the topics to be covered in order not to over-load the questionnaire o the need to reconcile demographic terms and concepts to the prevailing cultural and social situations o the necessity to structure the questions into a logical pattern that would be easy to administer to respondents by enumerators in order to minimise operational errors. 2. An in-depth discussion of these steps is presented in the sections that follow. Suffice it to note that, in practice, the steps are not independent of one another, as each decision taken has a strong bearing on others that follow.
9.2 SADC 2000 Census Project
3. An important consideration in the development of the 2001 Population and Housing Census questionnaire was the SADC 2000 Census Project. This initiative came about against the background of waning donor funding and technical support. Consequently, SADC member states resolved (through the SADC Statistics Committee) to pool resources together in the 2000 Round of censuses. Member states also felt that, as a prelude for further regional integration, there was need for collaboration and harmonisation of concepts as well as synchronised data provision. Towards that end, a minimal set of topics, mandatory for all SADC member states, was agreed upon.
4. Most of the SADC minimal set included questions, which were carried during previous, except the following:
o Place of usual residence five years ago. o Place of birth within the country/Country of birth o Usual Economic Activity (12 months before census) o Parental survival/Orphan hood o Source of energy for Heating
77 o Refuse Disposal
9.3 Stakeholder/User Needs
5. As usual, topics for inclusion in a census originate from the users since initiation, evaluation and monitoring of various policies rest on reliably data. The census being the vehicle through which data on a wide variety of topics could be gathered, inviting for topics was the natural way to start. In order to give users the desired direction, the list of topics, which were covered during the 1991 census were sent to users together with the invitation for new topics.
6. The response from users was quite phenomenal in view of the suggested topics. However, many users suggested a host of topics, which were not suitable for coverage in a census. These topics were screened by the Census Standing Committee for suitability for coverage in a census. This process involved a lot of consultations with users in a effort to get detailed background on the topic and also making users aware of what was suitable for censuses and surveys. The following is a list of topics suggested by stakeholders but excluded from the census because of the level of depth required.
o Maternal mortality o Duration of residence in Botswana (for non-citizens) o Previous residential status o Type of agricultural activity undertaken o Number of housing units owned countrywide o Date of taking residence in Botswana o Demand and supply of housing o Dwelling construction costs o Housing stock by agency o Contribution of wild life and veld products to household income o Causes of morbidity and mortality o Fertiliser and pesticide usage o Use of wind energy
9.4 Basis of Census Enumeration
7. In conducting a census, it is important to ensure complete coverage of the country and thus ascertain that everybody in the country during the period of census enumeration was counted. A particular concern is that movement of people within the country can be very fast and frequent. People may, for instance, spend the day in an urban area, village, cattle post, or lands area and travel overnight to other settlements. The speed of this movement, especially between villages and cattle posts and lands is indeed very fast. Thus the usual census anchor question - who spent last night with the household - is of social and technical importance. In this
78 regard, special consideration had to be given to the basis of enumeration in order to ensure the achievement of the census objective - of complete coverage of the country and complete enumeration and also of comparability of ensuing data with the data from the 1991 census.
8. Although in recent censuses the basis of the count was “de facto”, inevitably the basis of some other questions, such as fertility - children ever born living in the household or elsewhere, absent members of household and number of deaths in the household was “de jure”. It is probably necessary to define the two alternative basis of a census enumeration before proceeding further:
o a “de facto” count is a census conducted on the basis of where the persons to be enumerated are found. In order to have a tidy operational base, this is modified to where the persons to be enumerated slept or spent the night previous to the census.
o a “de jure” count is a census conducted on the basis of a person’s place of usual residence. This basis for a census count has the disadvantage of persons who are not at their usual place of residence being counted where they are during the enumeration, thus leading to possible over enumeration.
9. Against the background of fast and extensive free movement of people between the urban and rural areas, more so within the rural areas (largely between the villages and cattle posts, villages and lands), a de jure count would lead to ambiguities with respect to place of usual residence and double enumeration for some people. Consequently, a de facto count has always been perceived to be more appropriate.
9.5 Operational Problems Associated with “de facto” Enumeration
10. In order to enumerate all persons who slept or spent the night prior to the enumeration with a particular household, the “de facto” enumeration basis was modified. In Setswana the concept “ who slept or spent last night in this household” translated into “who woke up in this household this morning”. From a cultural and social viewpoint this is more acceptable to the respondents than the literal translation of “who slept or spent last night in this household”. Three sets of operational problems arose with this modified approach to “de facto” enumeration:
o the need to define the term “night” became crucial to the accuracy of enumeration. There is always the possibility that some people may die during the course of an ordinary night. A decision would therefore have to be made, for enumeration purposes, whether or not such persons should be recorded as living or dead on the census night. On the other hand, some children might
79 also be born during the course of the night thus requiring a decision whether or not their existence should be recorded for census purposes.
o there is a considerable number of persons who during the course of a night never slept/spent the night with any household nor for that matter woke up in any, for example persons travelling through the night, vagrants, etc. In a fewer cases, there are persons who may sleep or spend the census night with more than one household.
11. These operational problems are considered in turns and the solutions adopted for them in recent censuses. Notwithstanding the slight differences in interpretation and usage, the three concepts of - slept the night, spent the night or woke-up in a household were used interchangeably.
a) Definition of a “night”
12. With regard to the problem mentioned in the foregoing, a night was defined as the period between 600.p.m. and 6.00 a.m. Thus a person in the household during that interval was regarded as having spent or slept the night or woken-up with the household (See Enumerator’s Manual paragraphs 31-34)
b) Persons who did not spend a Census Night with a Household
13. Inevitably, some persons who would normally have slept the night with a household but were known not to have spent the night in any other household. Such persons were enumerated with the household with whom they would normally have spent the night. Examples of such persons include night shift workers; hunting or fishing parties who slept in the bush or by the riverside; persons away cutting thatch; persons at overnight prayer meetings and persons at funeral wake-keeping, etc.
14. Persons in transit were the problem group among this category. If a person spent the night or part of the night with a household before commencing a journey, the person would be enumerated with the household. There is, however, the possibility that the person could be enumerated at either both of the households, where he or she started the journey or where the journey ended; or be missed out. Although this was noted in 1991, it also remained an unresolved source of error to the 2001 Census. It is hoped that incidents of double enumeration and complete omission would cancel each other out. Moreover, it has been observed elsewhere that the net result is always an undercount.
9.6 Selection of Topics for the 2001 Census
15. Censuses in Botswana are conducted as a fulfilment of Section 64 of Botswana’s constitution. To that extent, therefore, the primary purpose of a population census
80 is to provide statistical information regarding the number of persons in the country, their distribution and their characteristics in order to enable the equitable distribution of parliamentary seats. All other considerations are subsidiary to this legal reason for the census. Hence in the midst of competitive requirements to include various topics in the census, the main purpose of the census - to determine the number of people in the country, their sex, age and geographical distribution over the country - must take priority.
16. With the legal requirement fulfilled, the other topics to be carried by the census were decided on technical considerations and mutual agreement between the census organisers and the users of the census data. Some of the guidelines, which are used in determining which topics are to be carried by the census, are:
o need for continuity and hence production of serial data and some topics
o need to accommodate changes in the socio-economic situation in the country and hence admission of new topics.
o the sensitivity of respondents (and the Government) to some topics, (for example topics on tribe, personal income, number of cattle owned, etc) has to be taken into consideration. The inclusion of such topics could affect the outcome of the census to the extent of compromising the quality of the returns.
o avoidance of topics that could cause or aggravate social, religious or political tension in the country.
o ease with which respondents can answer questions, if necessary after prompting or assistance from the enumerator.
o the topics should not be so numerous as to require undue demand of the respondents’ time.
17. Much as some of the guidelines are straightforward, others are subjective and in order to ensure universal and more importantly political acceptance of the results of the census, all the topics selected for the census must be politically acceptable to the Government. While some topics might be technically compelling to carry, the Government’s view must be final, even then.
18. The guidelines above were adopted in deciding on the topics carried in the 2001 Census. The topics covered by the 2001 Census and previous censuses are shown in the table below.
81 Table 1: List of Topics Carried by Botswana Censuses, 1904, 1956 -1991
Census Years
1904 1956 1964 1971 1981 1991 2001 CENSUS TOPICS
Name - x x x x x x Household Relationship - x x x x x x Sex x x x x x x x Age (Broad Age) x x x x x x x (Single Age) - x x x x x x Race x x x - - - - Citizenship - - - x x x x Place of birth ------x Residential Status - - x - - - - Place of Usual Living Now/A Year Ago/5 years ago - - - - x x x Village of Allegiance - - x x - - x Literacy - x x - - - - Highest School Grade Passed - - - - x - x - Highest School Grade Attained - - x x - x x Training ------x Language Spoken - x - - - - x Religion / Religion of Head - x - - - - x Marital Status - - x x x x x Working on Family Lands - - - x x x x Working Status - - x - x x x Occupation - - x x x x x Industry - - - x x x x No. of Children Ever Born - - - x x x x Particulars of Most Recent Birth - - - x - - - Particulars of Children born a Year to the Census - x - - x x x Deaths a Year to the Census - - - - x x x Survival of Mother / Father - - - x - - x Absent Batswana during Census - x - x x x x Blindness Disability (including blindness) - - - - - x x Livestock Ownership x - - x x x x Poultry Ownership x - - - - x x Household Farming Status - - - x x x x Arable/Grazing Land Tenure ------x x Household Cash Activities - - - - - x x Household Cash Receipts
82 (remittances, rents, etc) - - - - - x x Ownership of Radio Household Durables - - - x - - x Type of Tenure of Dwellings - - - - - x x x Number of Rooms in Dwellings - - - - - x x Material of Construction of Dwelling - - - - x x x Principal Water Supply Source - - - - x x x Principal Source of Fuel for Cooking / Lighting/Heating - - - - x x x Toilet Facilities - - - - x x x Refuse disposal - - - - - x Home-based care ------x x = topic was carried in the census - = topic was not carried in the census
9.7 2001 Census Questions
19. The household questionnaire for the 2001 Census is shown as Appendix 13
a) Name
20. In writing names, enumerators were instructed to write the name of the head of household (whose name should appear first) in full. The first names of the other members of the household were judged sufficient. The reason was that previous census experience has shown that writing the names of members of a household in full could slow down the pace of interviewing.
b) Relationship of Members of Head of Household
21. In framing this question it was very clear that the concept associated to relationship by respondents is much looser than the technical concept. Thus for example, a brother or sister of a respondent may not have any blood or family relationship to the head of the household. He/she may just be a close friend of the family or a kinsman from the same village or town or even country as the head of household. This is the more important for determination of level of fertility. Without proper probing, a niece or nephew or even grandchild could be wrongly recorded as a child.
22. Further compounding the problem is lack of uniqueness in the type of relationship within the household due to the fact that enumeration is on a “de facto” basis. The relationship type recorded is dependent on the vagaries of which member of the household (subject to that person being at least 12 years of age) is answering
83 the enumerators question as the relationship is recorded around the then head of household. Consequently, the statistics derived from this question has not always been fully utilized. c) Age
23. In spite of the importance of age as a demographic variable, data on age is extremely difficult to collect. Therefore a considerable amount of time was also devoted to the question on age during both publicity and training. An Events Calendar was prepared for each administrative district to aid the enumerators in administering this question. This document was not as useful as before mainly because of improved literacy level and the introduction of national identity cards “O mang”.
24. Processing of the census data has revealed that probing or requesting for supporting evidence with regards to the age of person absent at the time of interview was either not done or was inadequate in many areas of the country. This also points to the amount of emphasises made in this area during training. d) Citizenship.
25. Considerable economic growth in Botswana over the years, coupled with economic slowdown in neighbouring countries has resulted in an increase in the number of non-citizens coming into the country. Correspondingly the number of Batswana going abroad to work has been decreasing significantly. Since this situation has a lot a influence on population-age distribution and have other social implications, the importance of such data can not be overemphasised.. e) Place of Birth
26. The question on place of birth was introduced as part of the minimal set for the SADC 2000 Census. It was felt that responses to the questions would provide insight into lifetime migration. The place of birth was deemed to be the usual place of living of the mother at the time of the birth rather than the actual place. This was necessary since most births occur in hospitals.
f) Place of Usual Living
27. Given the level of mobility throughout the country, it is essential to obtain, as accurately as possible, information about the location of the population in the country. In addition to information provided by the de facto basis of enumeration with regards to the location of the population, questions on the place of usual living at the time of enumeration and the place of usual living a year previous to the census enumeration were asked.
84 28. It is possible, from the way the questions were framed, to obtain information on the number of persons who have remained in the same dwellings since a year previous to the census and those who have moved only within the same locality. It was decided to track movement across districts rather than intra-district movement. g) Education
29. The level of education attained, access to education facilities are important yardsticks for a nations’ socio-economic development. The socio-economic development witnessed during the inter-census period shaped the way the question was asked. For the first time there was demand for data on pre-primary schooling and non-formal education. These were new introductions in the 2001 census. For these reasons, these questions were asked of all persons aged two years and over.
30. Unlike before, interest was on education up to secondary school level rather than tertiary institutions. Furthermore emphasis was on the highest level completed, regardless of whether such level was passed or not.
h) Training
31. This topic was separated from education this time round to enable in-depth analysis on training. Emphasis was on training of any nature that lasted for at least three months. In essence this covered tertiary institutions, the university as well other less formal institutions offering mostly computing and business courses. Such information was deemed important for understanding skills training vis-à- vis labour demand and supply.
i) Language
32. Language has been a topical issue is recent years. Information on language will assist the Ministry of Education in refining the Revised National Policy on Education’s recommendations for the establishment of the Botswana Languages Council. The Language Council will have the responsibility of developing a comprehensive language policy. Also, it was felt that there is need to know all the languages spoken in Botswana, the numbers of people speaking different languages and the geographical distribution of these languages.
j) Religion
33. Religion has, for time immemorial, played a significant role in society’s moral development and evolution. The extent to which various segments of society pay allegiance to religious affiliations countrywide is of major importance in
85 community development. In addition, information regarding the dominant religious affiliations could be important to some users. k) Marital Status
34. Once again, the changing social status and habits were taken into consideration in drawing the various categories of marital status. It was against such background that the status of couples living together without undergoing formal marital status was included. In spite of such a habit not being legally recognised, the number of couples doing so is increasing. It was therefore necessary to admit this group into the marital status classification. l) Questions on Economic Activity Status
35. A better picture on economic activity is often derived from data relating to activities performed over a longer period. With this in mind, one question was included to gather data on usual economic activity. Usual economic activity collects data on main activities done a year prior to the census.
36. With regard to current economic activity, an attempt was made to move as close as possible to International Labour Organisation (ILO) recommendation. Of importance in this instance is that “work” was considered to be any activity for pay, profit or home-use over the past seven days that lasted for at least one hour. Thus any work a person does, no matter how short the duration, is or how little the pay is, takes precedence over all other activities.
37. In determining the activity status, it is very important to understand that people consider many activities, especially agricultural related, to be of no economic value or as non-work. A probe was included to ensure that such cases were reclassified accordingly after establishing that indeed a person participated in an economic activity. For those persons who reported that they had worked, there were follow-up questions to obtain information on the work done during the reference period by persons who were reported as neither employee nor as self - employed. Information was sought as to whether they were working in a family business or whether they were doing any manner of work at the family lands or cattle-posts.
m) Occupation and Industry
38. Data on occupation and industry is susceptible to many errors. These could arise because; the questions were not understood by the respondent, wrongly recorded by the enumerators or wrongly coded. Furthermore, respondents who might not have adequate information on absentee members provide proxy responses. In order to give a clear picture of the economic activity situation, it was important to
86 collect data that are as detailed as possible. In this regard, it was decided earlier on in the preparatory phase of the census that the coding of responses to questions on occupation should be at the four-digit level and responses to industry questions at three-digit level (ISIC Rev 3).
39. In order to justify these detailed levels of coding of occupation and industry, the quality of data had to be good. To ensure this, the probing of respondents had to be thorough and the recorded answers must be explicit enough to justify the coding effort. Sufficient space was therefore provided on the questionnaire for the enumerators to record full details of occupation and industry. n) Fertility
40. An improvement was made on the 1991 census in as far as obtaining information by sex in respect of the number of children ever born as well as births a year prior to the census was obtained. The information on the sex of children ever born was required, among other things, to improve the estimates - by sex, of the level of infant mortality and life expectation at birth.
41. Another improvement made was the inclusion of a follow up question to the births a year prior to the census to establish the number of such births still alive. This could enable estimation of infant deaths, as a residual. o) Batswana Citizens Absentees
42. In spite of improved employment opportunities in recent times, the topic of absent Batswana continues to have social and economic relevance. Apart from people going to South Africa to work in the mines and on the farms, many other people are going to South Africa and other countries for various other reasons. Hence, critical attention was paid to the framing of the question on why they were absent from the country.
43. Although a lot of attention has been paid to absentees citizens outside the country at the time of the census, there is need to consider obtaining limited details about household members away from home but within the country during census enumeration. This would have allowed better information on household composition. More accurate information on the headship of household would be obtainable. As of now, it is not possible to associate such persons to their actual households. p) Disability
44. Disability is largely a medical condition, thus the question on disability was limited to the type of disabilities that are observable. Extensive consultation with user departments enabled the office to obtain very useful information on some
87 terms and concept used in framing the question. Areas to mention are mental retardation (moderate or severe) and mental illness. q) Births and Deaths Reference Point
45. The reference point for the question on births and deaths was Botswana Day 2000 (30 September 2000). As the period was less than one year, a raising factor would be needed so that the births and deaths rates are brought to a yearly basis. For this purpose the mid-point of the census enumeration period, 22 August 2001 could be treated as the theoretical Census date. r) Household Agricultural Activities
46. Censuses provide baseline data for further surveys or studies. In this regard questions on the household agriculture activities in the 2001 census were to provide a basis for the Agriculture Census, initially planned for 2003 but re- scheduled for 2004.
47. Notwithstanding the foregoing, the importance of this topic to a wider base of users was recognized, hence the need to frame the questions on the ownership of livestock, crops planted and tenure on farming land was common and thus binary digits had to be used. With grazing being largely on communal land, it was felt that the tenure of grazing land was probably not appropriate and therefore the question was not carried in the 2001 Census. s) Household Cash Activities and Cash Receipts
48. Data from a de-facto census has serious limitation in that it is not always possible to obtain data pertaining to all household members. These topics are important in that they enable collection of an array of activities households participate in to subsist. Thus in spite of the fact that absentee members are not enumerated along with their household, their contribution will be captured in the section. Furthermore, there are economic activities or functions that are performed by the household as a collective unit. Information on such type of activities should therefore be obtained separately, at the household level, hence this section of the questionnaire.
49. The principal aim of this section was to obtain information on sources of household income, other than formal or wage employment. Such sources of income may be classified under two broad headings:
o production activities o transfers
50. The first would include economic activities such as;
88 o production of cash crops - maize, sorghum, beans, etc o raising of livestock - cattle, sheep, goats, etc. o hunting, fishing and gathering ; o running store/bakery, truck/taxi, boats/canoes, etc o making handicraft , brewing traditional beer, bee-keeping etc.
51. The second would include non-factor incomes such as interests and royalty receipts, pensions, remittances aid/relief etc.
52. In view of the need to avoid seasonality and other variable factors, the reference period for the questions on household economic activities was made longer than the reference period on individual economic activities, one year for household activities compared with one month for individual activities. Further, because of the possibility of multiple responses, binary digits were again used in recording the response.
53. The reference period for this question was similar to that of births and deaths i.e. 30 September 2000. This decision was made with the intention to adjust the figures accordingly. In hindsight, adjustment was not possible because the responses could not be subjected to mathematical manipulation since they are in essence just categorical. t) Housing Unit Type, Tenure, and Facilities Available
54. A housing unit was defined as the unit for accommodation for a household. Thus, in general, the number of housing units should be the same as the number of households. The responses to each question was expected to be unique for each household, hence binary digits were not used with regard to the questions on housing.
55. Information on the type of housing unit the household occupies, the mode of acquisition as well as the tenure of housing unit, is required in the formulation and evaluation of housing policies. Also of particular importance are data on the number of rooms, source of water supply and toilet facilities in urban and rural areas. The material of construction of the housing unit is an indicator of the quality of housing units in a locality. u) Refuse Disposal
56. Refuse disposal is a very important environmental issue. With the recent concern over pollution of the environment and depletion of the ozone layer through improper disposal of litter, this question collected information to assist the planners, sanitation officers, etc to evaluate or initiate appropriate environment policies.
57. This topic was covered for the first time in a census in 2001.
89
v) Ownership of Durables
58. This topic was a new introduction to census topics. Household ownership of durable goods is a useful indicator of household income. Affordability of luxury goods indicates better income or credit worthiness of the household. Items like Radios, TVs and Computers could be used for educational purposes and therefore their availability or unavailability indirectly gives an indication of studying environment. Furthermore, Radios, TVs, telephones and computers are used as mediums of communication.
w) Home based Care
59. This again was one of the new introductions to the list of topics carried in censuses. Initially the intention was to collect data relating to HIV/AIDS related illnesses but this was later watered-down to home based care.
60. The AIDS pandemic has overstretched available health facilities and thus resulting with some patients being cared for by their families and regular visits from community based health workers. The extent of the occurrence of home- based care had not yet been fully documented and therefore there was need to have baseline data to plan for improvement of services at community level.
9.8 The Design of Enumeration Questionnaires
61. The state of technology at the time of development of the 2001 census questionnaire made physical cutting and pasting a thing of the past. A form- designing software package, tele-Form was used for developing the 2001 census questionnaire. While tele-Form was very useful, the fact that it was not compatible with other software was a major limitation. For instance in designing census manuals there is need to have illustrations relating to the section in reference and therefore copying and pasting would have been useful. Furthermore, the 2001 census questionnaire was huge (A2) and therefore was printed as two A3, which had to be pasted on an A2 background paper to produce the questionnaire.
a) Physical Consideration: Fonts, Ink and Paper Size
62. In designing the 2001 census questionnaire, the fonts, ink and paper size were dictated by the number of questions. The number of questions carried were very many resulting with a paper size of A2 being considered appropriate to accommodate the questions, for a given font size. The questionnaire was therefore huge and the resulting font, though equal to that used in 1991 census, was considered small and not very user-friendly.
90
63. The type of ink used in printing is also of utmost importance. Documented historical accounts note that under dusty and poor lighting conditions green ink is much preferable to either black or blue ink. Green ink was therefore used in the printing of the 2001 Census Questionnaires. b) Printing On Both Sides of Questionnaire Form
64. Printing of questionnaires back-to-back is also a subject that requires considerable attention when designing the questionnaire. Indeed possible advantages of printing on both sides of the questionnaire form are many. There could be more space for the wording of the questions; more questions could be carried in the census; considerable saving in the quantity of the paper required to print the questionnaire could be achieved. However, it is possible that enumerators could occasionally omit or forget to ask questions at back of the questionnaire form. Similarly, coders could omit or forget to edit the questions on the back of the form. Errors could also be made at the data recording stage.
91 10 Preparatory Activities and Census Enumeration
10.1 Introduction
1. Preparation for census enumeration involved, mainly, the following leading to the census enumeration.
1) Procurement and Printing
2) Determining the number of District Census Officers (DCOs) per census district on the basis of district size in terms of population projected, and also their appointment.
3) Training of DCOs and Technical Officers (TOs) for pilot census and also for the actual census.
4) Determining the number of trainers and supervisors by census district, and also their recruitment and training.
5) Determining the number of enumerators by census district on the basis of actual number of enumeration areas delineated, and coordination of their (enumerators) recruitment.
6) Coordinating the securing of training centres, including food and accommodation of participants.
7) Census enumeration and associated issues
8) Transport arrangements in all districts and also determining terrain suitability of specific vehicles.
9) Fuel arrangements in each census district during enumeration.
10) Strategies to ensure complete coverage during enumeration.
10.2 Procurement and Printing
2. With the exception of a few jobs, printing of all census materials and documents was done at the Government Printer. Although this was done through their own budget, costing of such materials is presented in Appendix 20. The Census Office only provided the Government Printer with camera-ready copies and specifications of what needed to be printed.
92
3. Like in previous censuses, the period immediately prior to training and enumeration was marked by some hectic logistical work. The enumeration questionnaires reached the Government Printer early January 2001 and thus questionnaires were printed by end of February. However, the page numbering was not correctly done i.e. pages were serially numbered just like the practice with receipt books. Efforts to save cost by renumbering delayed the final release of the questionnaire. By the time a decision was made to print afresh, it was already late. As regards costs, all was not lost since most of the wrongly numbered sheets were used for training.
4. Some materials were needed for the training of enumerators while some were required for the actual enumeration itself. A schedule was developed such that materials needed for training were printed first, to give the Census Office enough time to do quality checks, packing and dispatching for the various centres in time for training. a) Printing and Packing of Training Materials
5. The Household Questionnaire which was designed as A2 size booklet of 25 pages with a front and a back cover was printed on loose forms of A1 and A2 sizes for the training of enumerators. For the inside pages a total of 100,000 A2 size, and 3,000 A1 size prints were requested. For both the front and the back cover, a total of 1000 A1 size and 10,000 A2 size prints each were printed. The A1 size copies were used as classroom-boards, while the A3 size copies were used for individual classroom exercises during training as well as for replacing spoiled enumeration check controls.
6. An Institutional Questionnaire was designed for the enumeration of persons who were found residing in institutions. This questionnaire was designed as an A2 size booklet of 10 pages. For training purposes, it was also printed in loose pages; 1000 in A1 size and 10,000 in A2 size.
7. Other materials printed for training were as follows:
¾ 6000 A4 size copies of training schedules ¾ 7000 A4 size copies of enumerator’s manual ¾ 7000 A4 size copies of comprehension questions ¾ 500 A4 size copies of comprehension answers ¾ 7000 A4 size copies of household profiles ¾ 10000 A4 size copies of control lists.
8. The materials described above were packed in cardboard boxes, which were organised according to designated Training Centres. The amount of material for each centre depended on the number of trainers, supervisors and enumerator
93 trainees at the centre. Each centre was given an extra ten percent of the total requirement in order to allow for defective documents.
b) Printing and Packing of Enumeration Materials
9. As regards materials for enumeration, the following were prepared by the Census Office and printed by the Government Printer:
¾ 30,000 A2 size 25 page Household Questionnaire booklets ¾ 6000 A2 size 10 page Institutional Questionnaire booklets ¾ 6, 800 A4 size booklets of Locality Coding list ¾ 6,800 A4 size booklets of Events Calendar ¾ 1,500 A4 size Supervisor’s Manual ¾ 7000 A4 size Enumerator’s Manual ¾ 300 A4 size District Census Officer’s Manual ¾ 10,000 A4 size pre-addressed census envelopes
10. In an effort to ease pressure at the Government Printer, the following were prepared and duplicated by the Census Office:
¾ 6,800 A4 size Individual Questionnaire for hotels, motels and safari camps ¾ 10,000 A4 size letters to hotel managers ¾ 6,000 A4 size letters of appointment for supervisors and enumerators.
11. The table below summarises the information on all the printing jobs done by the Government Printer, for both the training and the main enumeration.
c) List of Documents Printed for the Census
DOCUMETS Size of No of Sent to Documents Copies Govt. Printer
Locality Coding List A4 6, 800 May 2001 Events Calendar A4 6, 800 May 2001 Enumerator’s Manual A4 7,000 May 2001 Supervisor’s Manual A4 1,500 May 2001 DCO Manual A4 100 April 2001
Household Questionnaires: - Loose sheets A2 100,000 Jan 2001 - EA Books A2 30,000 Jan 2001 - loose sheets A1 3,000 Jan 2001 - front cover A1 1, 000 Jan 2001 - back cover A1 1, 000 Jan 2001
Control List A1 1, 000 May 2001
94 Returns Envelopes A4 10, 000 July 2001
Institutional Questionnaires - loose sheets A1 1, 000 Jan 2001 - loose sheets A2 10, 000 Jan 2001 - booklets A2 6,000 Jan 2001
Guide to the Census A4 25,000 March 2001 Advocacy for Census A4 27, 500 March 2001
12. The following materials were prepared and duplicated within the Census Office.
DOCUMETS Size Copies Prepared
Hotel Guests Forms A4 10, 000 July 2001 Letters to Hotel Managers And Guests A4 10, 000 July 2001 Household Profiles 1&2 A4 7, 000 July 2001 Training Schedule A4 7, 000 July 2001 Comprehension Questions A4 7, 000 July 2001 Comprehension Answers A4 500 July 2001
13. A local company, Feat Holdings was contracted to print and supply adhesive rolls, supervisor bags, T-shits and Satchels. It turned out these were printed in India. Suffice it to say the quality of all materials was sub-standard. It would seem there was an intention to cheat on the part of the company since sample presented for tendering was not the same as that finally supplied. Not much could be done to penalise the company because such materials were delivered just at the time of enumeration.
¾ 13, 500 rolls of printed white and 8, 000orange adhesive labels used for indicating the status of enumeration in any particular dwelling – printed by Associated Enterprises.
¾ 6,000 census T-shirts used both as a visual aid in locating enumerators from a distance
¾ 6,000 enumeration satchels
14. Another local company, The Trio Investment, was contracted to supply 1300 brief cases for Technical Officers, District Census Officers and supervisors.
15. Estimation of the required number of materials was made difficult by the fact that the exact number of Enumeration Areas and Supervision Areas was known rather late. Notwithstanding that, available information was used to apportion the
95 enumeration materials according to Supervision and Enumeration Areas. In reality the exact number of enumeration areas was only known after completion of enumeration. 16. Each enumerator and each supervisor was assigned a satchel containing all the materials necessary for enumeration in the assigned areas of work. The unit of packing was therefore a satchel. Each enumerator satchel contained the following items of equipment:
¾ 1 Enumerator’s Manual ¾ 6 Enumeration Area Books ¾ a file containing the EA map and control lists ¾ 1 or more Institutional Questionnaire Books (if the EA was known to have an institution) ¾ 1 copy of the Locality Coding List ¾ 1 copy Botswana Events and Age Calendar ¾ 3 Roll of printed white adhesive labels ¾ 1 Roll of printed orange adhesive labels ¾ 1 Census Badge ¾ 2 blue or black ball-point pens
17. Each supervisor was given a green satchel containing:
¾ 1 Supervisor’s Manual ¾ spare Enumeration Area Books (Household and Institutional Questionnaires) ¾ spare rolls of adhesive labels ¾ 2 Green ball-point pens ¾ spare blue or black ball-point pens ¾ a file containing maps identifying the areas of responsibility (and in some cases a district map) and spare blank control list form.
18. Each supervisor was assigned four to five enumerators. This, in a way, dictated the way the packing was eventually organized. All enumerator satchels were packed together with the supervisor satchels in the same cardboard box according to supervision area. The boxes were in turn arranged according to training centres within a census district. All spare enumeration materials for census district were packed in one box and addressed to the DCO responsible for the district.
10.3 Retraining of DCOs and TOs
19. In early March 2001 retraining of DCOs and TOs was conducted at Oasis Motel in Tlokweng. The participants were in two groups. The first group comprised DCOs from the south and half of TOs, while the other group comprised DCOs from the north and the other half of TOs. Each group took three days during which enumerator, supervisor, and DCO’s manuals were reviewed in detail.
96 Changes that were made in the manuals after the pilot census were also highlighted.
20. The training alluded to above was a follow up of trainings conducted in 2000 in preparation for the pilot census. Following inter-ministerial consultation regarding the role of District Census Officer, names of prospective DCO were forwarded to the Census Office. Five training session were conducted in May 2000 - Serowe, Francistown, Gaborone, Kanye and Maun. This training was aimed at equipping trainees with the background information on the census and the institutional arrangement as well as other important details that could enable them to effectively discharge the responsibility as DCOs. Bearing in mind transfers, resignations and other career prospects, over 100 trainees participated in the training. On completion of training and in consultation with district executives about half the number were appointed District Census Officers.
21. Training of Technical Officers was conducted at the end of July 2000, just a few weeks before departure to the field. Although the TOs were responsible for technical aspects of the census, it was generally thought that the separate training of DCO and TO resulted in different interpretation of the concepts and definitions hence the arrangement for a re-training in March 2001.
10.4 Recruitment & Training of Trainers and Supervisors
22. Trainers were recruited from among secondary school deputy headmasters and senior teachers.. The number of trainers was determined on the basis of the estimated number of enumeration areas (EAs) in the census districts. Training of trainers took place at various centres around the country in late March 2001 from 26 – 30 March 2001. At this training, TOs were presenters/trainers in class sessions while DCOs assisted in general class discussions and profiles, which profiles, were given as homework and discussed in class the next day. The supervisor and enumerator’s manual formed the core of materials covered in this training. The training lasted for five days, which included practical interviews, assessment on the practical, and a written test.
23. The selection criterion was based on best-performing candidates, in terms of participation in class, practical interview assessment and written test. Those whose behaviour indicated some possibility of being irresponsible were dropped regardless of their overall performance.
24. Supervisors were recruited from among secondary school teachers, head teachers of primary school, and government employees. All supervisors were required to have a minimum of diploma in any field of study. The number of supervisors per census district was determined on the basis of estimated number of EAs in that district.
97 25. Training of supervisors was conducted in various centres around the country. It took place over five days from 23 – 27 April 2001. The training followed the same schedule as the trainers’. Trainers conducted training while TOs and DCOs observed and offered assistance when necessary. Team leaders TOs were overseeing all the classes. Assessment and selection criteria were the same as those used in the training of trainers.
26. The number of supervisors that would be required in each district was not yet known at the time of training. As a result, no selection was made until EA delineation exercise was completed in July 2001.
27. Toward the census, it was realised that in some districts, e.g. Kweneng West and Barolong, there was shortage of supervisors. This was due to insufficient number of recruited trainers in those districts, which number, was as a result of the fact that the actual number of EAs were not yet known. To meet the shortfall, the district had to draw from extra supervisors in adjacent ones.
10.5 Recruitment of Enumerators
28. Recruitment of enumerators started towards the end of June 2001 and was completed by the end of July. Each census district recruited its own enumerators with Census Office only coordinating. The required qualification for enumerators was a minimum of form five. Most enumerators were drawn from teaching professionals. Those primary school teachers who did not have form five were required to have minimum of teacher training college diploma. The Census Office recommended enumerators recruitment quotas of 75 percent teachers and 25 percent school leavers. Most census districts conformed to the recommendation, except in a few cases, where there was shortage of teachers and had to slightly increase the number of form five school leavers.
29. It was also recommended that recruitment of census personnel should be in such a way that as much as possible enumerators and supervisors are assigned areas where they usually reside. This could be cost effective in that subsistence allowance payable to those outside their normal duty station would be minimised.
30. Botswana Defence Force (BDF) personnel residing in military barracks were required, for security reasons, to be enumerated by BDF soldier personnel. This condition necessitated recruitment of BDF soldier enumerators and supervisors. These enumerators also followed BDF anti-poaching camps around the country and enumerated soldiers living in them.
98 10.6 Training Centres
31. Each census district secured its own training centre for training of enumerators. Secondary schools with boarding facilities and colleges of education were used. There were problems however, in some districts finding training centres. This seems to have been caused by late request made to the secondary schools, whose administrators had already made arrangements for other activities to take place during the school holidays. An example of such activities is renovation. Other schools had been booked for other activities by the Ministry of Education. Alternative centres were found and where it was not possible trainees were allowed to find their own accommodation and be given 75 percent of subsistence allowance, especially those who came from outside the host locality. Food was provided in all centres.
32. The problem of boarding accommodation did not feature in urban census districts as all census personnel were recruited from within the city or town.
10.7 Re-orientation of Trainers
33. After training of trainers and supervisors a few changes in the enumerators manual were made. The possibility of trainers having forgotten some concepts and procedures of census could not be ruled out, given that census taking was not the day-to-day job of the trainers. Thus, it was necessary to give orientation to remind them of the concepts and highlight changes that had taken place after training of supervisors. Moreover, in some districts, supervisors had to be elevated to the position of trainer to arrest the problem of shortage of trainers in those districts. Such supervisors needed to be assisted to prepare them to discharge the role of trainers.
34. Technical Officers and DCOs conducted the training o 4 – 5 August in cities and towns, and on 8 – 9 in other census districts.
10.8 Training of Enumerators
35. Training of enumerators took place over five days from 11 - 15 August 2001. Participants arrived at the centres on 10 August 2001. It was compulsory for all trainee enumerators, supervisors, trainers, TOs, and DCOs, who came from outside the host locality to stay at the accommodation provided at the training centres. This was necessary to ensure that a situation does not arise, where TOs and DCOs are provided with better hotel accommodation, while other census personnel are left to endure the modest accommodation provided at the training centres, lest they would boycott the accommodation and demand hotel
99 accommodation as well, which government would not afford for a number of people as many as over 5,000.
36. There were between 4,700 – 5,000 enumerators trained around the country. The first four days of training were devoted to the presentation of concepts and procedures of enumeration as outlined in the enumerator’s manual. Examples and mock interviews were discussed in class. Profiles were given out as homework for trainees to do and discuss among themselves in the evenings. The fifth day was for supervised practical interviews, discussion of the experiences on the practical, assessment on practical interviews, written test, and marking. Trainers and supervisors did the marking. Grading, team allocation and deployment of enumerators and supervisors, distribution of census enumeration materials, and payments were done on the sixth and last day.
37. Generally, the training period was considered to have been insufficient.
10.9 Security During Payment and Enumeration
38. Society is sophisticated these days, thus each census district had made its own arrangements for security of cash from the bank or revenue office to the training centre where field personnel were paid. Central and local police officers provided the security.
39. In urban areas and large villages measures were in place to have police patrols in the evenings during enumeration, to ensure security of enumerators.
10.10 Census Enumeration
40. Census enumeration was scheduled for 17 – 26 August 2001. These dates were so chosen to include two weekends so that the working population could be captured at their homes on weekends.
41. In most parts of the country enumeration started on 17 August as scheduled while in a small number of cases it started the next day on 18 August. This was mainly due to late departures from the training centres to areas of assignments. Late departures, where they occurred, were caused mainly by delayed payments. Enumeration was completed as scheduled in all parts of the country, except in Gaborone and Mogoditshane where, an extension by two days was necessary.
42. Central Transport Organization mechanics were placed in strategic places around the country, during enumeration, to ensure that breakdowns were attended to immediately.
100 43. It turned out though, at the end of census, that people were not counted in a few small localities in lands and cattle posts areas, which localities had been missed by the cartographic exercise. Although this was anticipated, but to a very minimal level, efforts had been made to ensure that such localities are identified before the census, so as to make arrangements for enumeration of people living in them. Cartographic inspections by TOs and familiarization tours by DCOs and TOs were the two main quality assurance measures in place. In addition, appeals had been made to parliamentarians and councillors, who are supposed to know their areas better, to report to district administration or the Census Office, any localities they found with dwellings that have no census numbers.
10.11 Transport Arrangement
44. Transport was arranged centrally with CTO headquarters in Gaborone. Vehicles were pooled from central and local government departments through coordination with CTO depots around the country, and distributed to district administration centres.
45. Initially 600 vehicles were requested from CTO based on estimation from 1991 census. But after EA delineations it turned out that about 1,200 vehicles would be needed for the whole census operation. This number of vehicles required turned out to be too big for CTO to meet, even when there were supplementary vehicles from Councils and Land boards. To this end, authority was sought to allow the use of own vehicles by TOs, DCOs, and supervisors, but only in cities, towns, and urban villages. This was based on an expected maximum distance of 200 kilometres per vehicle per day.
10.12 Fuel Arrangements
46. Central Transport Organization fuel points were opened every day of the census enumeration from 0730 to 1600hrs for government vehicles. Council and Land board vehicles used their usual fuel stations but were reimbursed later even though this took longer than necessary because of shortage of funds. The request for supplementary funding to meet the over-expenditure took long to process.
47. Special arrangement was made by the Census Office to provide fuel dumps carried in drums at strategic areas around the country, to minimise travelling distance for refuelling by central government vehicles. Household Surveys Unit staff that had received some training on fuel accounts at CTO headquarters manned the fuel dumps, located at the following places:
101 DISTRICT LOCALITY
Southern Pitsane Molopo Kweneng East Lephephe Kweneng West Takatokwane Kgatleng Olifant Drift Kgatleng Mosomane Central Serowe Mmashoro Central Boteti Tsienyane Central Tutume Mokubilo Ngamiland West Seronga Gantsi Lone Tree Kgalagadi South Middlepits Kgalagadi South Hereford
48. Ghanzi district made special arrangement for mobile fuel to be carried in a truck for Central Kgalagadi Game Reserve.
10.13 Recommendations
49. Cartographic work needs to be planned in such a way that the latest date for completion of EA delineation is end of February of the census year. This will allow for more correct figures to be determined regarding field personnel and number of vehicles to be requested from CTO.
50. Strategies must be in place during cartographic exercise, to ensure that no locality is missed. One such possibility could be that cartographic inspection starts as early as possible and should be in such a way that members of Census Standing Committee are assigned their districts permanently from beginning to the end of cartographic exercise. They should take trips to conduct inspections in their respective districts regularly and more frequently, following the cartographic team schedule until they are satisfied that no lands and/or cattle post areas have been missed.
51. District census committees should be given the responsibility to take inspection trips to cover the whole district once the cartographic work completed in that district. They should report to the Census Office any missed localities they find, so that remedial action is taken before finalisation of maps and control lists.
52. The recruitment of University of Botswana students for census enumeration is a matter that needs a lot of careful consideration since the University opens during the course of enumeration. Those recruited abandon after training, creating serious manpower problems.
102 11 Quality and Security Measures
11.1 Introduction
1. The quality of census results depends on the interplay of many factors in each phases of the project. It is therefore important to built in quality assurance mechanisms in each of the stage and make optimum use any medium that each phase offers. In the following sections an attempt is made to articulate the mechanism that were put in place.
11.2 The Design of the Questionnaire
2. Questionnaire design offers an opportunity through which quality control mechanisms could be developed. In that respect, quality control measures of the census started with the design of the questionnaire. The flow of the questions is an important aspect of quality control. The questionnaire started with the listing of all household members who slept or spent the census night with the household. After identifying the head of household, the questions proceeded in order of complexity regarding the possible answers to the questions. There followed classification of the population into specific age and sex segments: all ages; ages 2 years and above; ages 12 years and above; females aged 12 years and above and females aged 12 to 49 years.
3. After dealing with all individuals in the household, the questionnaire then dealt with specific individuals; absentees, the disabled and those that have died. The questionnaire then moved to the household as a unit before eventually dealing with housing. By dividing the questionnaire into units in this way, there was a more logical flow to the questions and this helped both the enumerator and the respondent.
4. Since some errors introduced to a survey or a census may be impossible to correct, it is therefore important to design the questionnaire in such a way as to minimise the possibility of such errors.
11.3 Design of EA Books
5. Without doubt the loss of a source document containing collected data, in a data collection exercise, is a major loss. In a census it is even more so, as the document contains information about individuals, which if the information falls into wrong hands can be misused. Hence the need for elaborate security arrangements against the loss of completed EA Books. Arrangements were therefore made to account
103 for each enumeration area book and each page in the book. It is against such background that importance is attached to accounting for individual sheets in the EA Book, as discussed in the preceding paragraphs.
11.4 Prelisting Form and Control List
6. Listing of dwellings before enumeration is an important census activity since the assurance of complete coverage of enumeration, depends on that to a large extent. It is essential that all localities are listed and identified on the ground and furthermore that all the dwellings are uniquely identified and listed. The locality and dwelling listings for the last three censuses were carried out using the Prelisting Form. This form is very similar to the Control List. The column headings of the Housing List and the Control List consist of:
Prelisting Form Control List
Serial number Serial number Lolwapa/House number Lolwapa/House Number Use of House/Lolwapa Name of head of lolwapa/household Name of head of lolwapa/household Number of persons Status (of enumeration) Enumeration Details Households and persons enumerated Comments Comments
11.5 The Control List
7. After the enumeration questionnaire, the Control List (shown as Appendix 10) is perhaps the second most important document in the census. The Control List is an address list giving the dwelling number and the name of the heads of household of all the dwellings to be visited and enumerated by the enumerator. This list is meant to contain the list of all dwellings in a locality. The names of occupants listed on the Control List were only a guide since at the time of the enumeration a dwelling may have had different occupants.
8. The Control List is designed to monitor the progress of enumeration in each locality - the number of dwellings enumerated, the number to be revisited, the dwellings in which the enumerator has been refused access, etc. The summaries entered into the Control List by the enumerator at the conclusion of enumeration of each dwelling provide the information required for compiling census preliminary returns.
9. The importance of the Control List for the accuracy of the census cannot be over- emphasized. If dwellings are not properly and completely listed, the accuracy of the enumeration becomes dependent on the thoroughness of the enumerators. A
104 poorly motivated enumerator may enumerate only the dwellings on his list if the listing was not complete. Indeed if the listing is not exhaustive, complete coverage of the EA cannot be ensured.
10. Difficulties experienced in the 1991 Census, with regard of completion of a Control List for multiple households necessitated a modification of the form used in the 2001 census. The new form made it possible to account for all households within one dwelling and also allowed easy correction of the mistakes made by the enumerators.
11.6 Enumeration Check Control
11. Another quality control measure independent of the Control List was the Enumeration Check Control. A panel was provided on the front cover of the EA Book (see Appendix 12) for this purpose, this was the Enumeration Check Control. The enumerator was required to account for the use of every single page in the 25 – page EA Book and in the process giving summaries of enumeration for each household. An account of each of the 25 pages in each book was given on each line. Additional provisions were made for necessary summations. The first column of the panel dealt with serial number corresponding to the numbering of each page of the EA Book. The next three columns were designed to give village code, locality code and lolwapa/plot number respectively. The last two columns were designed to give the summary of enumeration in each household i.e. number of persons enumerated (summary of Part A of the questionnaire) in column 6, and Batswana citizen absent (summary of Part B) in column 7.
12. Instructions were given to enumerators as to how to enter these figures under varying circumstances. The summation of entries in the Enumeration Check Control in respect of a locality should, if instructions had been followed properly, equal the summation for that particular locality in the Control List. Thus there were three independent sources, the Enumeration Forms, the Enumeration Check Control and the Control List from where to obtain the results of enumeration in each dwelling. Of the three forms of controls, the Enumeration Check Control was the one the enumerators found most difficult to understand in the early days of enumeration. However, as enumeration progressed they were able to correct their own mistakes.
11.7 Supervisor's Manual
13. To avoid a complete loss of information in case an EA Book was lost or destroyed, supervisors were instructed to record the summary of enumeration carried out each day by each enumerator. This was recorded in the space provided for this purpose in the Supervisor's Manual whenever the supervisor met the enumerator. Sometimes a supervisor may not meet an enumerator under his/her
105 supervision for several days. At the conclusion of enumeration, the supervisor was required to complete the Supervisor's Summary List. Consequently all supervisors’ manuals were surrendered to the DCO for onward forwarding to the Census Office.
11.8 District Census Officer's Manual
14. When the supervisors had completed their summaries, the DCO for the Census District further compiled the results in the Districts Census Officer's Summary Sheet, which formed part of the DCO Manual. For each Enumeration Area the DCO was required to identify and record the name of the largest locality. The DCO was in addition required to record the total number of households, persons present and persons outside Botswana; the total number of un-enumerated households and un-enumerated persons for each EA in Census District. The aggregates reflected in the DCO's summary had to be equal to the sum of all the supervisors' summaries in that district (See DCO Manual which is not part of this report). Similarly the DCO manuals were surrendered to the Census Office. DCO and Supervisor summaries formed the basis for census preliminary results.
11.9 The Satchel
15. Packing for enumerators was on a satchel-by-satchel basis, i.e. all the documents and material an enumerator was going to use were packed into a satchel and that satchel formed the unit of account. A record of EA Books given to each enumerator was also kept. An extra supply of EA Books was given to each supervisor and DCO. The books allocated to each enumerator, supervisor and DCO had to be accounted for.
11.10 Document Control Form
16. A transaction document (see Appendix 16) showing the contents of enumeration documents and material handed over by the DCO to the Supervisor and from the latter to the enumerator in the first hand, and returned by enumerator to the Supervisor and from the latter to the DCO on the other, was prepared. Since Technical Officers collected materials in the first instance, they had to account for each item before handing over to the DCO.
17. The back of the document was divided into two halves. On one half, the supervisor was instructed to list items that might be missing from the enumeration documents and materials the DCO was supposed to have given to the supervisor. On the other half of the back of the transaction document, the DCO was also expected to list missing items from the enumeration documents and material that
106 the supervisor was expected to have returned to the DCO. Both the supervisor and the DCO knew before hand the list of things they should get from one another.
11.11 Security Arrangements at the Census Office
18. When the census documents were returned to the Census Office, they were carefully checked against the records that were prepared when the documents were going out and any discrepancies explained and reconciled. Only one incident was reported of a completed EA book that was lost. Fortunately this was discovered the following day and thus enabling re-enumeration of the affected areas. While some satchels were reported missing or misplaced, however, the centre of attention was the EA Books.
19. Prior to the enumeration, an office had been set-up as an archive to receive and keep all the enumeration books in a library fashion. In view of the volume of materials and documents, space was secured at a private building (Beta House) for dispatch and receipt. A tracking system that was developed by the Data Processing Manager was quite useful for locating and accounting for materials. All the books used for an enumeration area were tied together. The EA was the unit for taking out or returning EA Books needed for data processing purposes. An archiving officer was designated for issuing out and receiving EA Books that went out for either coding or data capture and at later stages data checking and reconciliation. For security reasons, the census archive, where the EA Books were kept was declared out of bounds for persons who did not have any direct link with data processing.
11.12 Cartographic Inspection and Familiarisation Tours
20. In addition to inbuilt checks on census documents, a lot of effort was made to check completeness of mapping and house listing. On realising that staffing capacity of the Cartography Unit could not enable full supervision of mapping teams, a decision was made to deploy technical officers to all districts to assess the quality of cartographic work – listing, dwelling number sequencing, missed localities, etc. Three such tours were conducted before the commencement of census enumeration. On completion of each tour, the Technical Officer wrote a report on the inspection. On the basis of such reports follow-ups were made to normalise the situation before enumeration.
21. Familiarisation tours by Technical Officers and DCOs just before enumeration took place were very useful quality assurance measures. Although these were conducted almost on the eleventh hour, and thus not allowing corrective action regarding cartographic work, such familiarisation proved very useful.
107 22. In order to get a full picture on the subject, it is recommended that this chapter be read in conjunction with Chapter 13, which deals with other quality control measures during data capture and computer editing.
108
12 Post Census Enumeration Activities
12.1 Introduction
1. Post census enumeration activities were to start immediately on completion of enumeration, scheduled for 26 August. Among the first activities undertaken was the reconciliation of population counts in the field. This was followed by receipt of materials in the field, preparation of summaries, receipt of materials in the office, preliminary counts, as well as coding and editing. While these activities are addressed in turn, a full coverage of coding, editing and data entry is on Chapter 13.
2. Although census enumeration was scheduled to end on 26 August, enumeration was extended in many areas thereby delaying the commencement of post enumeration activities.
12.2 EA Books and Control List
3. Before accounting for materials, enumerators had to reconcile the enumerated population and households in the EA book against the Enumeration Check Control and the Control List. This proved to be a challenging task for enumerators, depending on their level of numeracy as well as adherence to instructions. In some instances this task took longer than the extra day allowed for administrative matters.
4. Experience from preparation of preliminary results proved that documents were in most cases manipulated to balance. Even though the Enumeration Check Control seemed to balance when compared with contents of the EA book, it turned out that the figures were manipulated according to what was expected. Suffice it to mention that once again, like in the 1991 census, this form was the most difficult to most enumerators. Out of zealousness by some supervisors, some books were cancelled and re-entered on a fresh EA books because of cancellations on the front cover. Apart from the Enumeration Check Control, the Control list was a real problem to most enumerators. Once again30 a number of control lists had to be re-written and properly filled in the office.
12.3 Receipt of Materials in the field
5. Once the figures had been reconciled (or supposedly so as it turned out to be) the EA satchel, which was a unit of accounting was handed over to the supervisor. A
30 Just like in the 1991 census
109 form was designed for use by the supervisor when issuing materials to enumerators. This form was used to account for returnable materials (EA Books, Institutional Questionnaire and EA File). It was optional to return the rest of materials issued at the beginning of enumeration. Upon being satisfied that the EA books had been reconciled and that all returnable materials had been accounted for, the supervisor received the satchel and made a recommendation that the enumerator is paid his/her allowance.
12.4 Receipt of Supervisor Documents in the field
6. Supervisors were given two days to compile summary totals and reconcile Control List and EA book totals before accounting for materials to DCOs. This was done once the supervisor had checked and corrected the enumerator’s work. Supervisor summary sheets were prepared and were part of the Supervisors’ Manual. Therefore the manuals had to be returned, unlike the enumerators’ manual.
7. Summaries for supervisors were from two sources - EA Book total and Control List. The instruction of what was to be done was very clear. But given the difficulty with which the supervisors completed their summaries, it seem either some had not exercise due care from the onset or were simply not very numerate for the exercise. The fact that some EA checks control totals were manipulated and Control list wrongly completed meant that the work of some supervisors was below expectation.
8. Only after all the documents had been checked and corrected was the supervisor expected to hand in all such to the DCO. Accounting forms were again used to check for all returnable materials. Indeed in most cases this was not observed, to the extent that some of the returns were submitted uncorrected and un-reconciled. This was made possible through the manner in which the DCO and TO conducted their checks.
12.5 District Census Officer Manual
9. When the documents had been reconciled and corrected and supervisor summaries prepared, DCO compiled district summaries. These summaries followed the same pattern as for supervisors in that there were summaries from the EA books and those from the Control List. This was done on an EA-by EA basis such that totals from the EA book and those from the Control List had to balance. These summaries gave the district total which formed the basis for the compilation of preliminary results.
10. Having noted the lack of care in the reconciliation of the figures, there was indeed a cumulative effect in that one set of wrong results led to wrong summaries. The
110 fact that DCO summaries were not checked at the time of handing over materials meant that anomalies could only be detected at the time of compiling preliminary results.
12.6 Receipt of Materials in the office
11. In anticipation of the volume of documents and heavy traffic, space was secured in a private building, which was hitherto used by another government department. This was a useful plan in that it provided needed space for batching of materials (a batch being an EA) and ease of archiving later on at the Census Officer after accounting of documents and materials. Moreover, most of the non-returnable materials were returned thus increasing the volume. However a set back was that all usable materials had to be transported to the Census Office, which was about three kilometres away from the batching site.
12. One week was initially earmarked for receipt and accounting of materials, starting from 27 August 2002. It was also thought that arrival of district teams was going to be uniformly spread across the week. However, DCOs and TOs spent more time than allocated for field accounting and preparation of summaries. After spending days awaiting the arrival of district teams, most teams arrived almost at the same time and thus making accounting a slow and cumbersome exercise. Since accounting of materials was on a district-by-district basis, teams had to wait until late in the night for their turn to handover materials. In fact some DCOs spent many days in Gaborone because it took a very long time for large districts to account for their materials. This was further compounded by the fact that some extra questionnaires supplied to the district were used during training after experiencing shortage of loose sheets. While this was justified, instructions were issued to the effect that the front cover of such EA books must be kept safe for accounting purposes. It would seem this was not adhered to in a number of instances.
13. With many teams arriving at the same time, additional material receiving teams were formed within a very short time. Combined with the volume of materials, the quality of accounting was somewhat compromised to the extent that some used questionnaires were filed alongside blank ones. Fortunately at the time of data processing the cleaned data was compared with the summaries on an EA-by- EA basis thus enabling identification of missing books. Upon such discoveries, it was then decided to check all the supposedly blank questionnaires in case some completed books were inadvertently put in wrong piles.
14. Batching was done by way of tying together used EA book together with the EA file. Also included in the batch were cancelled books since experience has shown that there are usually transcription errors in moving information from the cancelled book to a clean book. Having EA books and EA file bound together
111 was very useful not only during both manual and online editing but also when compiling preliminary results.
12.7 Preliminary Results
15. Immediately upon completion of receipt and accounting of documents and materials, compilation of preliminary results started. Four teams of three members, each headed by a senior statistician, were charged with the responsibility of producing preliminary figures. To the extent that most of the district summaries were incorrect there was need to refer to the EA books more often than had been anticipated. Preparation of preliminary results took place at the time when questionnaires were being moved from Beta House to the Census Office for archiving. This made it very difficult to locate batches for district that had by then not been archived.
16. Summaries were prepared such that disaggregation was possible at various levels. The hierarchy as built into geographic identification system was employed at this stage of the preparation of summaries. In towns and villages the smallest unit was the EA whereas for the rest of the country it was the locality within an EA. After checking through all these units, everything was added to the census district and national level. For each district there were sub-totals for each EA, locality and village.
17. Preparation of district and national summaries culminated in the first report titled “2001 Population and Housing Census Preliminary Census Brief”. This publication was released on 19 September 2001. This report presented results by census as well as by administrative districts, thereby making comparison with the 1991 and projected figures easier. These results were very close to the projection and the finally published results.
12.8 Storage of EA Books and Control List
18. Storage of census questionnaires and other related materials proved to be a problem. The office that was designated for storage was just too small to accommodate all the materials and furthermore the shelves initially acquired were not strong enough for the bulky questionnaires. It would seem the decision to half the number of pages for an EA book, supposedly to allow ease of checking of questionnaires was not well thought out. Contrary to the thought that this did not change the volume of questionnaires, this was not necessarily the case since the questionnaire cover was very think and heavy. These questionnaires were so heavy that the shelves collapsed resulting in injury of one of the data entry clerks. Stronger shelves were acquired while data processing was in progress. The speed with which the Central Tender Board dealt with the matter is well appreciated.
112 19. After printing, all materials were stored at the temporary space that had been acquired. Thus, it could not be established that the archiving room could not accommodate all the questionnaires. Faced with the problem, the Census Office Conference was used as temporary storage space pending completion of editing and coding. This arrangement made dispatch and receipt of materials for processing a very difficult task. In addition, accounting for materials was quite a daunting task. On completion of data processing questionnaires were moved from the Arching Room and the Conference Room to a bigger room that was hitherto used for editing and coding.
12.9 Constraints and Recommendations
20. Once again reconciliation of EA books, Control List and EA Check Control was poorly done. It is clear that field personnel did not follow instructions on the completion of the documents in reference. Although the Control List was modified to make it easier to complete as per the recommendation made after the 1991 census, this did not help enumerators and supervisors. Halving the number of pages per questionnaire resulted with number of lines on the Check Control reduced to one column. This did not help either.
21. Most Control Lists were wrought with cancellations, a factor that made reconciliation a daunting if not an impossible task. Worst still, for EAs made up of many localities, enumerators often entered details of a household in a wrong Control List page thus making supervisor summaries derived from the Control List impossible to reconcile. As regards the contents of Control List, it does seem some enumerators were again not clear about completion of the various columns. Un-enumerated households were obviously not clearly understood since it turned out that in some districts there were more dwellings than people prompting one to suspect that empty dwellings were treated as un-enumerated households.
22. EA Check Controls were so poorly completed that there were many cancellations that photocopies of EA book front covers were pasted on some EA books. In some instances, the front covers were so messy that data on the EA books were transcribed to clean books thus resulting in transcription errors. While enumerators and supervisors had their own fair share in the matter, poor checking by the DCOs and Technical Officers ensured that field staff were paid and released even when their documents were not in good order. It is recommended that some mechanism be established to ensure that DCOs and TOs are held accountable for unsatisfactory work. Suffice it to note that some team spent almost two weeks producing summaries, without spotting the anomalies in their district results.
23. In spite of being fairly junior , the Archiving Officer discharged her obligation with great care ensuring that almost all the questionnaires and materials were easily tracked. This was indeed a tall order for her moreover that she was just
113 alone. It is recommended that in the next census a unit be established which could also help in the procurement of materials since experience has shown that Supplies Unit is not capable of fully handling procurement of materials.
114 13 Census Data Processing
13.1 Introduction
1. Census data processing is the set of operations that translates the raw information collected from the field into a meaningful, useful and timely set of statistical tables and reports. A national Population and Housing Census is invariably the largest and most costly data processing operation a country undertakes. Furthermore, the time period between two censuses, normally ten years, is so long that the transfer of knowledge and techniques from one census to another is minimal, particularly by virtue of the changes in technology, which take place in the interval. In effect, the processing of each national census must be treated as an extremely large, first-time operation. The increased availability and use of microcomputers, operating either in stand-alone mode or in a networked environment, has made a significant impact on the design and implementation of data processing activities.
2. The Central Statistics Office, having already processed the 1991 Census of Population and Housing using the then-version of the IMPS software (Integrated Microcomputer Processing System) and having now adopted the software as a standard for processing censuses and surveys, naturally opted for IMPS as the development platform for the 2001 census. Like other software packages, however, IMPS had also undergone major transformations since 1991 and all programs and procedures had to be devised afresh.
3. IMPS is designed and developed by the International Programs Centre (IPC) of the US Bureau of Census and is specifically geared towards census and survey data processing. Indeed, IMPS performs the major tasks involved in such processing: data entry, data editing, tabulation, data dissemination, statistical analysis and document tracking.
4. As early as in 1999, being given that it was facing a serious exodus of staff and realising that it would not be able to employ a person with the required background in IMPS, CSO took the decision to outsource the “Provision of Census Data Processing Software”. A tender was launched in late 1999, inviting prospective tenderers to submit proposals for (i) Development of computer application software, and (ii) Services for managing and implementing the application at the CSO Census Office. The period of assignment was July 2000 to April 2001 to enable coverage of both the pilot Census (sample basis) and the actual Census (100 percent enumeration).
5. In the meantime, the Census Office started to build up the necessary data processing capability in terms of equipment as well as human resources (to assist
115 and supplement the eventual private consultants) so that by July 2000, the data processing unit was operational.
6. During the pre-Pilot Census period, Botswana hosted a three-week training program in IMPS, organised by SADC for the benefit of member states in the context of the 2000 round of censuses. A visiting Consultant from the US Bureau of Census conducted the training. Three Officers from the Computer Services Unit of CSO and two from the Statistician cadre attended the course. However, the three among these Officers who had been earmarked for the Census subsequently quit Government service, thus leaving no trained Officer at the Census Office for supporting and overseeing the data processing component of the Census project.
7. The contract for Census data processing was duly awarded in June 2000, but the consultant deserted in January 2001 after completing the processing of the Pilot Census for which field enumeration had been done in August 2000.
8. Another tendering exercise had, therefore, to be resorted to for the Main Census, and that was done in March 2001. The new contract was duly allotted in late June 2001 to DCDM Botswana (Pty) Ltd. An intermediate Implementation Manager took office in mid-July. The substantive Implementation Manager took over at the beginning of September and, starting October 2001, he has been backed by an Assistant Implementation Manager.
13.2 The IMPS Package
9. As mentioned earlier, the IMPS software is composed of different program modules, which can be used together as a complete processing system or as independent applications. All modules, however, share a common data dictionary which has to be present whenever any other module is to be used.
10. The version of IMPS used for processing the 2001 Census was IMPS 3.1 for Data Entry, Data Editing and certain tabulations, and IMPS 4.1 for most Frequency and Cross-Tabulation runs. IMPS 3.1 is the latest, and, indeed, the last DOS version of the software and forms an integral part of the distribution package of IMPS 4.1. The latter version is the second Windows™ version of IMPS, but does not have its own Data Entry, Data Editing and Publication Tabulation modules.
11. The different IMPS modules used were as follows:
• IMPS 4.1 Data Dictionary for defining data file layouts and field values. The dictionaries were later saved as IMPS 3.1 versions for use by the other IMPS 3.1 components of the system. • CENTRY for developing data entry screens and entering data from questionnaires.
116 • CONCOR for developing and implementing computer edits. CONCOR was used both independently (batch editing) and in conjunction with CENTRY online editing).
• IMPS 4.1 Cross Tabulation for producing frequency distributions on all questionnaires items and direct cross-tabulations of up to 3 dimensions. (As a matter of fact, the CROSSTAB module can cross-tabulate up to four items, generate tabulations across different record types, select items with multiple occurrences and can tabulate multiple data files in a single run).
• CENTS for developing and producing more complex tabulations (e.g. on disability fields and on encoded items) is a package for tabulating, summarizing, and displaying statistical tables for publication.
12. The major advantages of IMPS are:
• It is free. • It is PC-based and can be used over standard networks. • Support on technical problems is provided promptly and freely by the US Bureau of Census through its International Programs Centre • It is a standard in national statistics offices of many developing countries around the world, and especially in the SADC region, thus promoting collaboration, exchange and mutual support on processing issues. • Data files are in standard ASCII format, thus facilitating transfer to an external platform for further analysis and other processing.
13.3 Computer Equipment
13. In the context of the census, the Census Office acquired and set up the following equipment, starting September 2000:
• 1 Compaq Proliant Server with the following configuration:
Windows NT 4.0 Operating System 1024 MB of RAM Disk Array (RAID5) with a total of 72 GB storage capacity Cartridge tape backup (DAT24 x 6 autoloader) with a global archiving capacity of 300 GB, with Back-Up Exec software Server anti-virus software 52X CD-ROM drive Floppy diskette drive 15-inch screen monitor
• 22 MECER Pentium III PCs with the following configuration:
Windows 98 SE Operating System
117 128 MB of RAM 10 GB hard disk capacity 52X CD-ROM drive Floppy diskette drive 15-inch screen monitor
• 3 HP LaserJet printers with printing capability of 16 pages per minute • 1 heavy-duty HP LaserJet printer with printing capability of 16 pages per minute • 1 heavy-duty Manessman Tally line printer with printing capability of 1000 lines per minute
14. Some 10 existing PCs were also redeployed for census-related tasks (particularly online editing) on different occasions during the period of data processing.
15. The server was power-protected by a 700 VA UPS supported by auto-shutdown software. Similarly, all PCs were protected against power failure by 500 VA UPS with built-in surge protection.
16. All PCs were connected to the file server through a Local Area Network (LAN), and this local file server was itself linked to the server at CSO headquarters through the Government Data Network (GDN).
17. The heavy-duty LaserJet printer and the line printer were meant to be network printers (i.e. connected to the server and thus accessible directly from any client PC), but the set-up could not be made to be operational that way. Instead, each of these printers was connected to a client PC. Printing jobs had to be launched from the latter PCs only after print documents had been sent to the server.
18. The other LaserJet printers were connected locally to three client PCs used by Census Officer, Implementation Manager and Principal Statistician.
19. Twenty of the installed PCs were used exclusively for data entry and verification until mid-March 2002 when the exercise ended. They were then used for online editing of questionnaires (which were done, in part, by the same Data Entry personnel).
13.4 The Census Questionnaire
20. Two types of questionnaires were used for the 2001 Census, namely the household (main) questionnaire and the institutional questionnaire.
118 a) Household Questionnaire
21. The Household questionnaire contained five distinct sections, apart from the common leading section comprising geographical and control data. The sections were as follows:
22. Common Section: This section contains a combination of six fields that uniquely identify a household, i.e.
District Code EA Number Village Code Locality Code Lolwapa/Dwelling Number Household Number and three other fields for EA book identification and control purposes, viz: Book Number Book Check Digit Household Check Digit
23. The Book Check Digit is a prelist total of the number of households enumerated in that particular book and serves to compute any discrepancy when matched against the actual number of households captured at Data Entry stage.
24. Similarly, the Household Check Digit indicates the number of persons present in one household and serves to check the accuracy of the number of persons captured for that household.
25. The fields in the Common Section add up to a total of 29 characters or bytes.
26. Section A – Population Present: There were 37 individual data items in this section making up a total of 81 bytes.
27. Section B – Population Outside Botswana: There were 13 data items in this section making up a total of 23 bytes.
28. Section C – Disability: There were 2 data items in this section. The disability field is sub-divided into 7 slots to allow entry of 7 different disability types. The total number of bytes was 16.
29. Section D – Deaths: There were 5 data items in this section making up a total of 14 bytes.
30. Section E – Housing Unit and Household Characteristics: There were 22 data items in this section making up a total of 45 bytes.
119 31. Basically, one questionnaire was designed to record information on up to 10 persons in Section A, and up to 3 persons each in Sections B, C and D (since the unit of enumeration was the household, there could exist one and only one set of entries in Section E). All “overflowing” persons had to be entered on a continuation questionnaire, with proper annotation in the comments box of each questionnaire used. As a rule, Section E was filled in only for the last questionnaire of the household.
32. The maximum number of persons in Section A was set at 50 (the figure was initially put at 30, but after encountering several “genuine” cases of higher numbers on questionnaires, the figure was raised). The range of acceptable number of records for Sections A to E are as follows:
Minimum Maximum Section A 1 50 Section B 0 15 Section C 0 10 Section D 0 10 Section E 1 1
b) Institutional Questionnaire
33. The Institutional questionnaire has only one section, apart from the common leading section comprising geographical and control data. The common section is the same as for Household Questionnaire, except that Household Number and Household Check Digit are replaced by Institution Number and Institution Check Digit, respectively.
34. The main section is identified as Section F and serves to collect information on all persons present on Census night in Institutions such as prisons, barracks, hospitals, hotels, etc.
35. Section F has 34 fields, making up 86 bytes. The disability field is sub-divided into 7 slots to allow entry of 7 different disability types.
36. One Institutional questionnaire catered for a maximum of 15 persons, so any additional entry beyond this number had to be entered on a second and subsequent questionnaires. The maximum number of persons in an institution was set to 999.
37. The questions asked to persons in Institutions and to persons in Households are the same, except for the following:
• Relation to Head of Household, being Not Applicable, is not present in Section F • Residency Status is present only in Section F
120 • Disability is present in Section F. For Households, the approach is different: disability questions are asked only after Section A is completed for all members of the household and responses are entered in Section C. • Questions “Number of Children Living with Mother” and “Number of Children Living Elsewhere” are not present in Section F.
c) Organisation of Questionnaires
38. All questionnaires were bound in book form, each book having a front cover and a back cover. The District Number, EA Number, Book Number and Book Check Digit figured prominently on the front cover among a variety of control and monitoring information relating to the book. The back cover contained Enumerator appointment and revisit records.
39. The household questionnaire book was initially meant to contain exactly 25 questionnaires serially numbered 1 to 25, but books with a different number of questionnaires have not been uncommon; in an extreme case, one book had 32 questionnaires. The institutional questionnaire book also theoretically contained 25 questionnaires.
40. In all, about 30,000 household books and 2,000 institution books were used for the entire census enumeration. The issue and return of books to office were strictly controlled and monitored. Even unused and cancelled books had to be accounted for by field staff.
13.5 The Tracking System
41. The unit of control being an EA, all completed books, both household and institutional, for an EA, were attached together with the EA file and transmitted in this fashion throughout all processing phases.
42. The Tracking system used was an EXCEL spreadsheet having one row per EA. There were columns provided for the following entries: (i) EA number (ii) range of household Book Numbers used (iii) number of household books (iv) range of institutional Book Numbers used (v) number of institutional books (vi) number of households (vii) date checked in from field (viii) date manual editing in (ix) date manual editing out (x) date coding in (xi) date coding out (xii) date data entry in
121 (xiii) date data entry out (xiv) date computer editing in (xv) date computer editing out (xvi) status (stage at).
43. The sheet was designed in such a way as to allow “grouping” and “ungrouping” of EAs by district; the EAs for a district are visible only on a press of the “ungroup” button. This feature facilitates navigation between and within districts.
44. The EA numbers were recorded by the Check-In officer as and when books were being received from the field after enumeration. She also entered the figures relating to the book numbers; the column for number of households was not filled in.
45. Indeed, recording onto the Tracking system never went beyond column 5. Each functional unit devised its own method for controlling the flow of EA books in and out of the unit, and for monitoring progress. The Archives unit, for example, maintained a register of EAs (together with the number of corresponding books) issued and returned to the unit.
13.6 Data Entry Process
a) Preamble
46. As mentioned earlier, twenty PC workstations were dedicated for data entry and verification. Provision was also made for two spare PCs to cater for emergencies. All the PCs were linked via a Local Area Network (LAN) to a Windows NT file server where all data files were kept. The data entry software used was CENTRY, the data entry module of IMPS Version 3.1.
47. The unit of data entry was the Enumeration Area (EA), uniquely identified by the combination District – EA number. This combination is also used to name the EA batch data file.
48. Data entry was done EA-wise by district. An EA consisted of one or more books of questionnaires, but, in no event, did a book contain information on more than one EA. Besides, the data entry system was set in such a way that district and EA numbers were automatically replicated for all questionnaires in a particular EA.
b) Data Entry Set-Up
49. CENTRY provides for three hierarchical levels of data entry screens, namely:
• Batch Screen (all entries therein are replicated in the whole batch)
122 • Questionnaire Screen (entries therein are replicated for all records in the questionnaire) • Record Screen (entries specific to that record)
50. The Batch screen was made up of the combination District - Enumeration Area (EA) number, which uniquely identifies an EA.
51. The Questionnaire screen was made up of the other common fields (see paragraph 22 above), i.e. Book Number, Book Check Digit, Household Check Digit, Village Code, Locality Code, Lolwapa Number and Household Number.
52. There were five distinct Record screens for the household questionnaire, one each for Sections A, B, C, D and E, and one Record screen for the institutional questionnaire.
53. It has already been mentioned that the unit of data entry was the EA (identified by the District-EA number combination). This combination was also used to name the EA batch file (the institution data file name was preceded by an “I”); the file extension .BCH is imposed by CENTRY. As all other operations (especially archiving) were to be done by district, it was decided to work district-wise, i.e. to complete one district before proceeding to the next.
54. The data entry system, besides the application programs and the CENTRY component of IMPS, consisted of a series of script files (extension .bat) which automated several procedures and made the task of both the data entry operators and the supervisors easier. Thus, at the time of invoking the application from the Windows desktop, the system was automatically set to the required server folder and executed the password-control program (developed in dBase). Thus, the operators did not need to take any technical decision and could fully concentrate on the keying exercise, proper.
55. Data entry and verification were performed directly on the file server in folder “POP” and sub-folder “ADD”. On completion of a district, all data entry files, comprising the data files (.BCH), pointer files (.BOP) and operator statistics files (.LOG) were reconciled by the Implementation Manager and moved to “VERIFIED” sub-folder, to which operators had no access. c) The Data Entry Programs
56. Two separate data entry programs were developed to cater for household questionnaires and institutional questionnaires.
57. Range checks, i.e. checks for valid values, were automatically enforced. It was a matter of interest at the start of the data entry exercise whether to incorporate a few consistency checks (e.g. identify two different field values in conflict with each other) at the data entry stage. Taking into consideration that this might delay
123 the exercise, especially as the Editing and Coding section could not afford to assign one or two officers to the Data Entry Section on a permanent basis, it was resolved to postpone such checks for the online editing stage.
58. Another matter for debate was whether to allow operators to force values of fields (CENTRY may be set to accept only values present in the data dictionary or the “forcing” F2 key may be enabled to allow forcing of “out of range” values). The same rationale as above prompted the decision that they be allowed to force values, but, at each such occurrence, the data entry supervisors would have to be alerted for necessary liaison with Editing and Coding Section.
59. The automatic skipping of all ‘Not Applicable’ fields was programmed into the data entry applications. For example, for a person less than 12 years old, the cursor, upon completion of the entries relevant to that age group in Section A, would automatically move to the end-of-record field.
60. The fields on the Data Entry batch screen (which determined the name of the physical data file, on the one hand, and which were replicated on all subsequent questionnaires and records, on the other hand) were key fields, which, if entered incorrectly, would cause serious problems in EA reconciliation and in later processing. It was, therefore, essential that these fields (District and EA) be correctly entered. For this purpose, the Batch ID Check feature was implemented. Basically, this implies having an external reference file (the Batch ID file) which contains the correct values. During Data Entry, this file is accessed online to ensure that the codes keyed in by data entry Operators are valid and that there is no duplication of batches. 61. The Batch ID file was initially created using the District-EA combinations from the EA Tracking System, but updates were necessary from time to time, especially when the system indicated “Batch ID not found in ID Check file” for a valid District-EA combination. This would be the case when the original EA entry in the Tracking System was erroneous. 62. Furthermore, the Data Entry System updates the Batch ID file with the date of Entry so that, if an attempt is made to enter the same batch at a later date, the system responds by the message “Batch ID already entered”.
63. Another feature implemented was password control. IMPS, by itself, does not have any scheme for password-protected access to data entry screens. However, the US Bureau of Census does provide the source code of a small dBase program designed to provide a rudimentary password control and some degree of automation for running data entry programs. This program was customised to suit the local set-up.
124 13.7 Personnel and Training
64. The data entry staff was composed of forty operators, two supervisors and two assistant supervisors. The staff was equally distributed between a morning and an evening shift, working seven hours each on a five-day-week basis. The morning shift hours were from 7.30 a.m. to 3.20 p.m. and the evening shift hours were from 3.30 p.m. to 11.20 p.m. The teams alternated shifts every week.
65. Overall supervision was the responsibility of the Implementation Manager.
66. The mainstream training of operators for the main census started in July 2001 with the arrival of the intermediate Implementation Manager from the application- supplier company, with questionnaires and application program from the pilot census being used as practice material. The training included initiation to microcomputer concepts, Windows operating system, mastering of keyboard skills, use of mouse, file naming conventions, CENTRY system and familiarisation with different function keys.
67. After consultation with the Census Officer and senior staff from the Computing Services Unit of CSO, the substantive Implementation Manager reviewed the data entry application to incorporate amendments brought about by changes in questionnaire layout. A front-end script to launch the process via an icon on the Windows desktop and to enable password-protected access to the data entry application was also introduced.
68. The training on the new system lasted about two weeks until end of September 2001 during which some five hundred dummy questionnaires (filled in by enumerators during their own training) were used to test the system. This period also allowed the testing of system integration, more particularly the communication between PC workstations and file server.
69. A comprehensive set of data entry instructions, developed earlier on, was thereafter consolidated and integrated in the Data Entry Operator Manual which was distributed to each operator.
70. The data entry for the main census started full swing on 05 October 2001 and was completed on 13 March 2002.
71. All data entry staff, except the two supervisors, were temporary staff recruited specifically for the census. This often resulted in staff leaving employment at short notice (e.g. for permanent jobs elsewhere or to pursue higher studies) and the process of recruitment and training had to be started again. It was estimated that a data entry operator joining the team without prior keyboard skills would be proficient only after two months in the job.
125 72. It is also to be noted that, following persistent shortage in the supply of edited and coded EA books to the Data Entry Section after the New Year holiday period, 15 data entry operators were redeployed to the Data Editing and Coding Section to help clear the bottleneck.
13.8 Control of Flow of Questionnaires
73. A manual system was set up to control the flow of edited and coded questionnaires from the Archives Unit to the Data Entry Section and within the latter section. Since all questionnaires were transmitted in batches corresponding to an EA, EA identification codes (District + EA) were used at all stages.
74. As mentioned earlier, the Archives Unit maintained a tracking system whereby all EAs received from the field, along with the number of books and the book serial numbers for each EA were recorded. On the first issue of EA books for a district to Data Entry Section, the Archives Unit also provided a printout of the tracking records for that district by EA. The Unit also kept a register of EAs being issued and had it countersigned by the data entry supervisor.
75. At the same time, the data entry supervisor recorded receipt of the EAs on the “Data Entry Control Sheet” by inserting the EA number, the number of books in the EA and the date. The supervisor, when assigning work to the operators, kept a record of the EA given away for entry or verification, the date and the name of the operator. Similarly, when he/she took back the keyed-in batches, the date of completion was recorded. Additionally, to enforce control and monitoring at the operator level on a daily basis, the data entry supervisor also kept a record of all assigned EA books operator-wise.
76. By keeping and maintaining accurate records, the supervisor was able to know the whereabouts of any given EA batch under his/her custody at any time. Furthermore, these control lists were, from time to time, checked with the master list of EAs from the tracking system to avoid duplication and also to assess progress of work.
13.9 Monitoring of Operators’ Work
77. The first level of monitoring the output and quality of data entry operators’ work was done by the data entry supervisors. Mention has been made above that a record of EA book assignments was kept per operator so as to monitor the operator’s output on a session-by-session basis.
78. As all operators were called upon to do both entry and verification, it had to be ensured that no one was allowed to verify his own file. Also, care was taken to avoid situations whereby two operators systematically verified each other’s work.
126
79. It must be noted here that 100% verification was done only for the first 7 districts entered. Two separate CONCOR validation runs, once on entered data only and the other on verified data for the same district, did not show any marked improvement on the verified counterpart.
80. For subsequent districts, the proportion of verified EAs varied from 10 to 35%.
81. The supervisors had to be aware of the skills and the accuracy level of each of the operators so that work could be assigned accordingly. Special attention was given to new operators and those whose performance was not at the expected level. The choice of EAs to be verified was often dictated by the output quality of the original operator.
82. Moreover, being responsible for the whole operation, supervisors were expected to report any serious problems, such as hardware breakdown, program or system problems, rate of absenteeism, etc, which could jeopardise completion of the work on schedule.
83. Performance Reports on operator output were produced after completion of a district. To this end, the operator statistics produced as a by-product of CENTRY were widely used. The EA statistics were sorted and consolidated by shift and by operator, and reports and charts were produced. Charts showing the comparative performance among operators and the progress of an operator over time were posted in the Data Entry Section for all to see. It is felt that this may have served as a motivating factor, which eventually saw the completion of the work before the originally set target date. (Production of comparative charts was suspended as from January 2002, as EA books were in short supply and a number of operators had to be redeployed to other sections).
84. The average number of keystrokes per hour was around 10,000 towards the end of the exercise, with a range extending from 6,000 to 16,000. This average is a “sustained” average computed over a period of at least two weeks. Short bursts of some 20,000 keystrokes per hour were not uncommon among the high- performing operators.
85. It is to be noted that two operators were dismissed from work because of their persistent poor output, even after four to five months in the job.
86. The total number of EAs entered for the whole of Botswana is 4,166, distributed among the 28 census districts. The number of EAs per district varied from 2 for Central Kgalagadi Game Reserve (CKGR) to 481 for Gaborone. A total of 377 EAs had institutions, including 14 which had institutions only.
87. All having been said, it is estimated that a grand total of some 180 million keystrokes have been hit by data entry operators for the whole census.
127
13.10 Monitoring of Data Entry Progress at Management Level
88. The overall output of the Data Entry Section was closely monitored at management level. A performance report was produced every week to analyse variations from targeted figures and to ensure that corrective measures were taken promptly.
89. Taking the preliminary population counts as the base figure and running a short program to tally the number of persons already captured, the number of persons outstanding could be directly computed. Dividing this number by the number of weeks left up to the targeted completion date gave the following week’s target in terms of entries to be made. Any variation from that target had to be justified.
90. The weekly target figure was, of course, a dynamic one, changing in value by virtue of the actual number of entries made during the previous week.
13.11 Backup of data files
91. All data files were backed up on a daily basis on the tape backup system at the local Census Office server. Backup was scheduled to run on all weekdays at 11.30 p.m., after the end of the evening Data Entry shift. Folder “POP” and all associated sub-folders were systematically backed up.
92. A backup on hard disk was also done by the Implementation Manager onto his own PC every morning.
93. The above was further supplemented by a tape backup, at more staggered intervals, at the CSO Headquarters server through the Government Data Network facility.
13.12 Computer Data Editing
a) Data Correction Strategies
94. In the ‘traditional’ census data processing system, the data are first entered into batches with minimal or no checking. After data entry, these batches are combined into a large data file, which is sent through an editing program to produce a listing of errors. The listings are then given to subject-matter personnel who mark the appropriate corrections on them. The corrections are then effected on the data entry records by data entry staff using the same data entry program (in “modify” mode) and the editing program is run again. This cycle of
128 correction/editing run continues, with the number of rejects becoming smaller and smaller until there are none.
95. The ‘traditional’ process is time-consuming and lengthy because it requires many iterations. It is error-prone due to the various steps that involve tracking of documents and data, and the transcription of corrections. Especially when the editing rules are complex, a change to the data which corrects one error may trigger another error from a different edit rule. The traditional process involves both subject-matter and data entry personnel.
96. A simpler manual correction system may be implemented by activating an editing program online during data ‘modification’ phase of a data entry program. As in the ‘traditional system’, data entry operators key the census data in batches. However, instead of performing batch edits, printing error listings and making corrections on paper, the edits and corrections can both be done at the same time at a computer terminal by subject-matter personnel, saving time and paper. This is the “online editing” method.
97. Thus, online editing can be done by using CONCOR in conjunction with CENTRY. The online editor enters into CENTRY and tells the system to run CONCOR interactively against the batch of data. CONCOR will process each questionnaire until an error is found. The system will stop at this point with both the questionnaire and the CONCOR messages displayed on the screen. The online editor can then easily make the necessary corrections. The system will immediately indicate if the amended questionnaire still contains any errors. In this way, ideally, a batch of data may be cleaned in a single session by a single person.
b) The Census 2001 Approach
98. A mix of both strategies was adopted for the editing of data pertaining to the 2001 census.
99. When all EA batches for a district were completed, the individual data files were combined into a district file. The CONCOR program was run ‘offline’ on this district file to produce an error listing for the whole district. The listing was further split into smaller listings, each such listing constituting the workload of an online editor. The online editor made the corrections at his computer workstation on the individual batches of data stored on the file server.
100. When the first round of online editing for the district was completed, the updated data batches were again combined into a new version of the district file and batch editing was repeated. In a perfect world, the output listing should indicate no errors. However, in practice, errors were still present, sometimes different from the original ones.
129 101. The online/offline cycle was repeated until there were no errors or “acceptable” errors only, as certified by the census statisticians.
c) The CONCOR programs
102. The CONCOR Validation and Editing programs were developed on the basis of edit specifications prepared by the Census Officer. However, being given the very nature of census enumeration covering each and every person in the country and coming across many unforeseen situations, the programs were continually being updated to cater for the latest encounters. As a matter of fact, even the latest versions cannot be said to be exhaustive in terms of checks and edits, and, as a result, complete “cleanliness” of the data file could not be attained.
103. It had been resolved since the very beginning that no recourse would be made to systematic imputation of data (such as using the “cold deck” and “hot deck” methods) which guarantees complete consistency of the data, or, at least makes the data “look” good. The adopted procedure of having two levels of computer-generated errors (EI errors to be mandatorily corrected, and EII errors to be checked for genuineness and not necessarily entailing correction) proved to be very strenuous and time-consuming, but it definitely ensures better translation of the reality on the field.
104. Furthermore, after about two months of online editing, it was realised that, if we were to meet the target for producing basic tables, we could not afford, in the first instance, to be checking and/or correcting all errors. It was, therefore, resolved to prioritize the errors in two groups, (notwithstanding the fact that we already had EI and EII errors), top priority being given to those errors (e.g. two heads in a household) likely to affect the population count.
105. The guiding principle was that no amendment would be allowed on the global population count by locality, village and district, once the figures are published in the basic tables. Subsequent correction of no-priority errors could, however, still change population distribution when tabulated against other variables.
106. The full versions of the programs repeated and extended all the range checks enforced during data entry. Thus, any value that might have been “forced” at data entry was trapped and thrown out for correction.
107. The CONCOR programs also made reference to a master Locality List file to check the accuracy of codes entered for village and locality.
d) Set-Up for the CONCOR Runs
108. As mentioned earlier, data captured in the Data Entry Section were transferred to \POP\VERIFIED folder after verification. These data were kept
130 intact on that folder throughout the data processing duration, so that if reference had to be made to “raw” entered data, that would be the place to look at.
109. When the online editing team was ready to take on a new district, all individual data batches (.BCH files) for that district were reconciled and copied to \POP\CLEAN folder on which the “editors” user group had full control. Access to all other folders was barred to online editors.
110. “Cleaned” data files for districts were stored in \POP\EDITED folder, each district having its own sub-folder. Access to this folder was restricted to the implementation management team.
13.13 Quality Assurance Measures
a) Preamble
111. Reports produced by the data processing system to implement Quality Assurance checks may be classified under four broad categories, namely:
• Duplicate Questionnaire Checks • EA Counts • Locality Counts • Frequency Counts
b) Duplicate Questionnaire Checks
112. Two questionnaires were considered to be in duplicate when the Lolwapa number taken in conjunction with Household number, District, Village and Locality were the same for both questionnaires.
113. An occurrence of duplicate questionnaires did not necessarily mean that the same household information had been filled in twice or that the particular questionnaire had been entered twice by error. It might simply mean that two different households had wrongly been allocated the same lolwapa or household number, or that the data entry operator had wrongly entered one of these numbers.
114. The objective of the duplicate questionnaire check was, precisely, to identify and separate the “genuine” duplicates from the false ones. The first type of duplicate implied canceling one of the questionnaires while the second type implied correcting any data entry error or allocating a new number to the household.
131 c) Comparative EA Counts
115. Here, the computed population counts by EA for a district were compared to the preliminary counts by EA compiled from Supervisor Manuals. Indeed, one panel in the Manual was meant for entering population counts per enumeration book for all books used in an EA. The grand totals for the EA were obtained by summing up the book totals.
116. All cases of discrepancy exceeding 8 were systematically investigated, while the lesser levels of discrepancy were checked on a sample basis. The checking procedure often involved going back to the physical Supervisor Manuals, to Enumeration Area files and, more often than not, going to the EA books and questionnaires.
117. The common causes of discrepancy noted were :
(i) missing book or book not entered for an EA, (ii) questionnaire for a household not entered, (iii) one or more persons in a household not entered, (iv) wrong entry of enumeration book population in Supervisor Manual, (v) wrong summation of enumeration book totals to give grand total for the EA in Supervisor Manual, and (vi) enumeration books mixed up during original batching or during data entry.
118. In an extreme case, there was a household belonging to one EA which had been enumerated within a book (in a sheet in the middle of the book) for a different EA. d) Comparative Locality Counts
119. The checks at locality level are similar to the EA count checks.
120. A separate panel in the Supervisor Manual provided for entering the different localities enumerated within an EA together with their respective population counts. The calculated grand total for localities would, of course, match the grand total calculated from book counts.
121. The threshold in discrepancies warranting investigation was slightly different from EA counts check in that, for small localities, a difference of 1 or 2 persons might mean affecting the percentage count by 50 or, even, 100%. Special attention was paid to localities reportedly having no population.
132 122. For certain localities known to the Census Officer, the figures were compared to 1991 census figures, especially in instances where there was a perceptible decrease in population.
123. Some of the factors causing discrepancies in locality counts were the same as for discrepancies in EA counts (see previous paragraph), but, additionally, the following causes were identified:
(i) Same code given to different localities in locality coding list, (ii) Different codes given to same locality in locality list, (iii) Insertion of incorrect/wrong locality code in locality box of questionnaire, and (iv) Entry of incorrect locality code at data entry.
124. (If the entered code did not exist in locality list, it was rejected by CONCOR edit program, otherwise the population had been lumped with the population of the ‘other’ locality).
e) Frequency Counts
125. Before declaring the data for a census as being “statistically correct”, it is necessary to produce frequency counts for all pertinent variables. Such frequencies often reveal the insufficiencies of existing editing programs and procedures, and, indeed, serve to refine the latter by introducing more rigorous checks.
126. Frequency runs were effected for each district after the data had passed through all other “cleaning” phases. These runs served to introduce new consistency checks in the CONCOR edit program (for example, when it was “discovered” that households in CKGR were disposing of their refuse through “roadside dumping” or that households in Orapa were owning their housing units). On the other hand, the frequency runs also served to relax on certain strict edits (for example, contrary to popular belief, some households in remote areas did genuinely have “water piped indoors”).
127. The frequency runs were supplemented by a few simple cross-tabulation runs which achieved the same objectives set out above. One of these cross- tabulated disability type against occupation to ensure, for example (but this is really an extreme example), that a blind person is not entered as being an airplane pilot!
133 13.14 Consolidation and Harmonization of Household and Institutional Data
a) The Rationale
128. It had been a matter of concern that, for all the phases in the census data processing suite, be it data entry, editing or production of counts, there was the need to have two separate sets of programs and files, one for processing household questionnaires and the other for processing institutional questionnaires.
129. It happened on several occasions, especially during the processing of the first districts, that, through an oversight, the institution counterpart had been overlooked when producing certain counts for an EA or for a district. Furthermore, additional resources, in terms of manpower and file storage space, had to be put in to develop and maintain two sets of programs, store two separate sets of data files, each with their own data dictionaries, and keep track of their whereabouts.
130. All the above pleaded in favour of combining, for further tabulation, analysis and dissemination, the household data and the institutional data into a consolidated file having a harmonized file layout and, thus, a single data dictionary. An “origin flag” would still distinguish the household data from the institutional one.
131. District data files were thus combined after all “cleaning” was completed..
b) The Harmonized File Layout
132. One major change was the ‘merging’ of the Person record in household questionnaire (Section A) with Person record of institutional questionnaire (Section F). An indicator flag was introduced (in fact it was incorporated within the new Person Serial Number) to distinguish one type from the other.
133. The other major change was the integration of the Disability record within the Person record in a restructured ‘decoded’ format.
134. The new record types were, therefore, the following:
o Record Type A : Person record (‘de facto’ population) o Record Type B : Outside record (citizens outside Botswana) o Record Type D : Deaths record (persons deceased since 1 year) o Record Type E : Housing Unit and Household record.
135. The following are the other changes brought to original file layouts:
134 o Person Serial Number was increased to 4 digits (as compared to 2 digits in household questionnaire and 3 digits in institutional questionnaire). The first digit served to distinguish between the two originating types: 0 for household and 1 for institution. Taken as a whole, serial numbers 0001 to 0999 represented households and 1001 to 1999 represented institutions. o Relationship to Head, which did not exist in institutional questionnaire, was present on all ’A’ records. The field was left blank for institutions. o Residency Status, which did not exist in household questionnaire, was present on all ‘A’ records. The field was left blank for households. o Fields “Number of children living with mother” and “Number of children living elsewhere” (both male and female), which did not exist on institutional questionnaire, were present on all “A” records. The fields were left blank for institutions. o Computed (or derived) fields were added to facilitate cross-tabulation. These fields are: Total Children Ever Born (i.e. Male + Female), Total Children Living with Mother, Total Children Living Elsewhere. Total Children Died, Total Children Born since 1 Year and Total Children Born Last Year and still Alive. o All multiple response items in the original Section E records were now decoded.
13.15 Data Tabulation
136. All frequencies and cross-tabulations produced for checking data consistency were run on the combined and consolidated district data files. The basic tables, “Population of Towns, Villages and Associated Localities”, were also produced from these files.
137. The combined files were in ASCII fixed length format and could thus be used directly for further tabulation and analysis by IMPS and other packages supporting this format. However, it was decided to use the SPSS package for production of topical tables and there arose the need, therefore, to reformat the files to facilitate export to that platform.
138. A special procedure was thus developed to generate comma-delimited (CSV) versions of the combined data files. These were the files that were used for producing all remaining tables in the census tabulation program.
135 13.16 Issues and Recommendations
a) Questionnaire Layout
• Questionnaire size is considered too big. It posed handling problems to data entry operators. Given the limited workspace, some operators had to place the EA books on their laps or under the keyboard (the more so, as folded up questionnaires kept falling back on the one being entered). There was also the risk of parallax errors as the data was entered row-wise (mixing up data on adjacent rows). • It is felt that reading codes from a dark background might be tiring for the eyes in the long run. It is suggested that the boxes containing codes should have a light background, preferably white. • It is suggested that whenever a section is not applicable, it be crossed out diagonally by field and/or editing staff. For one thing, much time is spent at source entering dashes, and also, it takes more time for the data entry operator to figure out which sections are to be entered and which are to be to be skipped. Another advantage is that the operator will have no pretext for omitting a section. • To speed up data entry, it would be a good idea to have mandatory sections (section E, in this case) placed on top, followed by optional sections. The operator would then have the possibility to release a questionnaire earlier, instead of having to navigate through the intermediate sections before reaching the one to be entered. (Of course, this would encourage malingering but strict supervision and control would be a deterrent). • The indication of a continuation questionnaire for a more-than-10-member household overflowing into the next questionnaire has not always been obvious, especially as the rule of filling section E on the last questionnaire has not been respected in some cases. Indeed, the annotation to that effect in the comments box is not sufficiently highlighted. A special continuation box (to be ticked, if applicable) placed at the end of the last “person row” might prove useful.
b) Questionnaire Fields
• Presetting the first digit of person serial number to 9 (Section B) restricted the number of entries in that section to 9 (91 to 99), whereas there have been instances of higher occurrences. Duplicated serial numbers, i.e. restarting with 91 for the tenth person, had to be resorted to. It is suggested that there be no presetting of first digit and that the first acceptable serial number in that section be 71 (it is unlikely that a household size will be over 70 persons). • The disability field was entered differently in section B and F. The terminator for disability codes in section B was a blank slot whereas that in section F was 00. this led to some confusion among data entry operators, especially at the
136 beginning. It is suggested that similar fields be entered in a uniform and harmonized fashion. • Only six slots in section B and only one slot in section F are provided for disability where there could actually be up to seven disability types entered (it had been discovered at a later date that up to nine entries were possible). The data file layout was thus different from the view on the questionnaire. • In field E2, the number of box positions provided was insufficient to hold the four-digit code and the latter was often squeezed into the box, thus hampering readability. (On the other extreme, field E6 could have been 2 digits long instead of the 3 provided for it on the questionnaire). It is suggested that the pilot census be on the lookout for such cases and that the final questionnaire incorporate the findings. • For institutions, the questionnaire check digit was 2 characters long, whereas the number of persons could be well over 99. The relevant consistency check had to be relaxed in such cases. The length of the check digit field should be increased accordingly. • The questionnaire check digit is a pre-list total meant to countercheck the number of persons in a household or in an institution. Similarly, at the next higher level, the purpose of the book check digit is to countercheck the number of households or institutions in the book. It was realised, when performing quality assurance checks and finding out that books had been missed out or that cancelled books had been entered, that there should have been a still higher level of check digit: that at EA level. It is, therefore, recommended that the EA check digit field be introduced to indicate the number of actual books used for the EA.
c) General • Provision should be made well in advance to equip the Data Entry Section with solid metal racks capable of supporting the weight of questionnaires. The racks are necessary for keeping the different stacks of EA books (to be entered, being entered, to be verified, to be returned, etc) in an organised way to facilitate identification, retrieval and tracking. • The data entry process was often halted when incomplete or ‘out of range’ entries were encountered. This often adversely affected the work schedule. It is recommended that one or two officers from the Editing and Coding Section be posted permanently to the Data Entry Section to clear such problems, especially at the start of the whole exercise. This secondment of staff would also serve to highlight the deficiencies of editing and coding, and thus help fine-tuning fresh staff briefings. • Data entry staff spent a lot of time going downstairs to the Archives unit to collect and return EA books. Apart from the fact that operators kept complaining about body aches due to the heavy loads carried, this also affected the work schedule. It is recommended that another grade of staff be recruited specifically to convey EA
137 batches to and from the Data Entry Section. It would also help if the Archives unit were on the same floor as the Data Entry Section. • There were too frequent disruptions in the supply of coded books to the Data Entry Section, especially after the end-of-year holidays (so much so that several data entry operators were redeployed to help out with editing and coding). Such disruptions break the momentum and have a demoralising effect on staff. To avoid any recurrence in a situation where deadlines are so strict, it is recommended that editing and coding should start well in advance of data entry. • Institutional information was often entered on a household questionnaire in the middle of an EA book. Instructions were given to data entry operators that such occurrences be reported to supervisors who would do the necessary work to have the information transcribed onto institutional books. Despite these instructions, the risk was high that, in the heat of the action, the questionnaire would be missed. While it is understood that such a situation cannot be avoided on the field, it is suggested that the DCO (or TO) keep a record of such instances. The list could be transmitted to the office during check-in of books for immediate transcription onto institutional books. • The piece-meal update of master Locality file has been a strenuous and time- consuming exercise during the online editing phase. It is suggested that field staff should report all new localities and villages, and indeed all changes in locality status, at the moment of checking-in questionnaires to the office. This would allow prior update of the locality file before running the CONCOR validation program and would thus dramatically reduce the number of rejected village and locality codes. • Numerous cases of locality codes wrongly transcribed onto questionnaires and same locality codes assigned to different localities have been encountered. A special emphasis should be laid on the necessity to check these entries at source on the field, as undetected cases would negatively impact on the basic tables. • Locating and retrieving a missing book for an EA from the piles of reportedly blank or cancelled books in the storeroom has not been easy, as the books were haphazardly placed. It is desirable, for tracking purposes, that cancelled books for an EA be physically bundled along with the valid books for that EA and be clearly marked “CANCELLED”.
138 14 Conclusions and Recommendation
14.1 Conclusions
1. Unlike the previous censuses, the 2001 Population and Housing Census was funded by the Botswana Government. This was expected, since the contribution from development partners, the United Nation Fund for Population Activities (UNFPA) in particular, had been diminishing with successive censuses. This was mainly because of the country’s per capita GDP that has been growing over the years.
2. Notwithstanding the foregoing, some technical assistance was offered through missions conducted by the UNFPA CST on cartography. These missions were very useful in giving direction as well as assessing the overall quality of the cartographic preparatory work. While these missions added value to the overall quality, this was somewhat diluted by the inadequate staffing of the Cartography Unit, a factor which made effective monitoring of fieldwork almost impossible. Also worth noting is the assistance offered in the area of economic activities by a reputable labour statistician31. This enabled first-hand advice from an expert, not only with a wealth of knowledge on the subject, but who had prepared relevant UN Manuals on the subject.
3. An important point with regard to capacity building, in readiness for the census round, was the SADC 2000 Census Project. Collaboration between SADC member states, as part of the project, enabled harmonisation of concepts and definitions, synchronised data collection and training of census personnel. Botswana benefited immensely from the project.
4. Although in-depth analysis of census results is still awaited, the 2001 census is considered a success. Census results serve a diverse number of users whose interests are also very different. To the extent that the population total was almost the same as the projections, and the fact that very few incidents of missed persons were reported, the project could be considered a success. Suffice it to note that the success of an exercise of such magnitude should be measured by the extent to which the needs of most general users are met. The hurdles and hiccups experienced during the exercise are part of a normal process.
5. The continued collaboration between the Ministry of Finance and Development Planning and that of Local Government should be hailed for contributing to the success of the census. The selfless service rendered by District Census Officers, drawn from the Ministry of Local Government, contributed immensely to the smooth operation of the exercise.
31 Reg Gilbert, who succumbed to ill-health later in South Africa
139
6. Preliminary census results were released within two-and-half weeks of the end of fieldwork while the 2001 basic census counts were released within ten months of completion of fieldwork. This was a result of a harmonious working relation between the data processing consulting firm32 and the Census Office. Mutual understanding between CSO units enabled ease of deployment of staff in response to census work demand.
7. There was a major setback during the tabulation, arising from the computer handling of the census data. A decision had been made to tabulate the census data using the Statistical Package for the Social Scientist (SPSS), a package the CSO had adopted for all data analysis. SPSS creates numerous temporary files during processing and this results in a computer running out on memory and eventually freezing particularly when handling huge datasets. This seriously delayed the release of the national statistical tables.
14.2 Recommendations
a) Census Office Staffing
8. There have been repeated appeals to CSO Management to be responsiveness to the staffing needs of the Census Office. Such responsiveness should be demonstrated even during the inter-census period, not just during peak period. The fact that CSO is plagued by staff exodus should not make management loose focus on the importance of an adequately staffed Census Office.
9. Although the census was conducted successfully, many administrative bottlenecks that are linked to the grading of the Census Officer post must be addressed in future censuses. For effective management of a project of the magnitude of the census, the officer in charge must have authority. Authority is needed when issuing instructions and directives, as dictated by the course of events. With very limited authority, instructions and directives are reduced to requests and most certainly this is not a very conducive environment. The initial organisational structure that the Census Officer reports to the Government Statistician guarantees a somewhat smooth flow of operation and thus it would be ideal in practice if such arrangement carries with it the desired authority. Against such background, it is recommended that the Census Officer post be upgraded to a level of deputy Government Statistician.
b) Role of Census Office
10. Mass exodus of demographers to greener pastures is likely to affect the extent to which the census data could be analysed. In the light of that, it is recommended
32 DCDM Botswana Ltd, a subsidiary of a Mauritius firm.
140 that Census Office staff is given thorough grounding on demography and related issues. This will enable second stage analysis beyond the census analytical report. The Census Office must be responsive to the changing needs of users, who are more educated and very articulate. Products that have traditionally been produced in past censuses may not adequately address the needs of such users.
11. In order to maintain the Census Office functional during the inter-census period, it is recommended that the office undertake the following:
o Conduct in-depth analysis on contemporary issues (AIDS, Orphan hood, housing, disability, economic activity, etc) and publish monographs on such analysis o Conduct consultancy on population and related issues (a lot could be learnt from the US Bureau of the Census hence a mission could be arranged – to the Bureau or to the Census Office). o Familiarisation to sound census methodology, taking into account contemporary issues, development in technology (scanning, etc) o Conduct research on appropriate training methodology (mediums, durations, etc) o Explore ways of obtaining quick and reliable preliminary (re-visit or review summaries on supervisors manual, control list and check controls) o Conduct research on question phrasing and responses – this will enable an assessment of quality of census data at question-level o Explore ways of ensuring efficient institutional structures in the midst of changing government structures – role of DCO, role of District Administration, sources of field personnel, etc o Explore innovative enumeration methods as well as innovative data processing and dissemination methods o Assess the data needs of contemporary issues o Collaborate with Demography Unit in production of population projections o Provide training on data processing to Census Office and provide the necessary platform for perfecting such skills through surveys o Produce regular population reports c) Cartographic Work
12. Notwithstanding the success of the 2001 census, it must be noted that there were many deficiencies in cartographic work. This was due largely to inadequate staffing of the Cartographic Unit. Attention must, therefore, be paid to the resources requirement and staffing situation in the unit.
13. The ongoing EA digitisation is a very important preparatory exercise for the 2011 census hence must be closely monitored. Intermediate level staff must be equipped with appropriate training on digitisation and other geographic information system related matters. For the digitisation to succeed, it is also
141 important that coordinates are collected, using Global Positioning System (GPS), and therefore the staffing needs of the unit at that level must be addressed.
14. One area of concern is the identification of ward boundaries. This is a very emotive, sensitive and often politicised matter. Notwithstanding that, the Census Office has been inundated with requests for data at that level. It is important that ward or planning area boundaries are identified, in spite of the difficulties alluded to above. Strategies of how to deal with the problem must be devised. This will be a signal of the CSO’s responsive to user demands
15. The move to encourage government and quasi-government organs to standardise on CSO codes is a very welcome development. This calls for maintenance and monitoring of CSO codes thereby ensuring that changes to locality codes are minimal. Collaboration with National and Civil Registration in matching the codes used by the three departments dictates that once CSO codes are completely matched with those of the National Registration, there should be no changes in the locality codes. Due care is needed in the handling of codes of localities that attain village status with time. A clearly defined framework is needed regarding procedures in the handling of the CSO coding system – not just the localities codes but also the entire set of codes. d) General Data Quality
16. While the overall data quality of the census results could be considered good, there were a number of issues that might somewhat compromise the integrity of the results. Among these is the length of enumeration period. With a ten-day enumeration period and a moving reference night, in the midst of a highly mobile population, there is a high likelihood for omissions. Indeed it could be argued that such omissions could be counter-balanced by multiple enumerations, but then experience has always shown that the net result is an undercount. It is therefore recommended that ways of reducing the enumeration period be explored in preparation for the next census.
17. Another area of concern relates to the collection of data for absentee households members but within the country at the time of enumeration. While there is no doubt that such persons are canvassed elsewhere, the fact that essential details about such people cannot be linked to their households seriously limits the resulting data. The same sentiments were expressed after the 1991 census and recommendations were made regarding capturing essential data for such persons. Not much attention was paid to the recommendations when preparing for the 2001 census. It should be stressed once more that due attention be paid to the collection of limited data on such persons in future censuses. The Census Office should, during the inter-census period, explore possibilities of collecting such information, drawing parallel to the section on absentee citizens outside the country (Section B).
142 e) Management and Administrative Procedures
18. While the merits of bureaucracy are clearly understood, often times the red tapes seems to be over-stretched rather unjustifiably. In particular, one could note administrative bottlenecks connected with protocol relating to the granting of various authorities. For instance, if there is shortage of funds (arising from factors beyond the control of census team) during the census and the authority to overspend is granted a month later, then such inordinate delay shows that there is something seriously wrong within the system.
19. The role of planning officers in a project as huge and complex as the census is very important for the exercise to succeed. It is therefore recommended that such role should not be relegated to newly appointed officers who have virtually no idea about the direction of the project course. The fact that two project memoranda were not processed after presentation at the project review meeting, and furthermore the handling of funds to offset the “over expenditure”, leaves too much to be desired. Such inefficiencies do not augur well for a project as complex as the census. It is therefore recommended that in future censuses such role should be assigned to more answerable officers.
20. Without doubt the CSO management team provided the desired direction and guidance to ensure that the project succeeds. However, the handling of some issues brings into question the resoluteness of the management team in taking critical decisions. Lack of resoluteness was clearly demonstrated in the way one of the earlier consultant was conducting himself during the tenure of the consultancy. Indeed there were a number of areas where it was clear that the operation was that of management by crisis. It is therefore recommended that management is both firm and resolute in taking critical decisions, if the census is to succeed. Apart from census publicity campaign, there was very little done to counter false reports on the census. f) Communication During Census Enumeration
21. That communication is critical in any modern census cannot be overemphasized. Developments in technology have brought in a variety of mediums that have improved communications tremendously. Such mediums include cellular phones. Much as these could be deemed expensive, perhaps in future government should review its policies relating to the use of such mediums, bearing in mind the value added of such projects. Quantification (in money terms) of the worthiness of the project must be brought to bear rather than subject the project to the mercy of blanket policy decisions. g) Administration and Supplies Sections of Census
22. The adage that “a chain is as strong as its weakest link” aptly exemplifies the census process, especially as regards administration and supplies. The state of our
143 ledgers during the 2001 census left too much to be desired. For a census to succeed there must be a dedicated administration officer for the census. The arrangement where such responsibility is relegated to fairly junior and inexperienced administrators, with the most senior administrator operating from a remote site does not work and therefore should be discouraged. Perhaps this arrangement could only work if the Census Office and the CSO are on the same site. Apart from general administration, huge projects with complex operations require a qualified accountant dedicated to the project. The on-the-job-learning approach with inexperienced staff has serious limitations and therefore must be avoided in future censuses.
144 Appendix 1: Timetable of Census Activities
Activity Starting Date (Duration)
1. PRE-CENSUS ENUMERATION ACTIVITIES
Scheduled Actual
A. Administrative
(a) General
1. Project Document Preparation 8/99 9/99 (2ms) (1.5ms) 2. Establishment of Census Office 7/99 10/99 (1m) (1m) 3. Establishment of Census Committees: - Central Census Committee 11/99 11/99 - Technical Advisory Committee 11/99 11/99 - National Communications Committee: 11/99 5/00
(b) Recruitment of Personnel
5. Census Officer 3/99 4/99 6. Cartographic Officer and Draughtsman 7/99 7/99 7. Administration Officer 8. Supply Officer 7/99 7/99
B. Cartographic Work
1. Inventory of the Map 3/99 3/99 (7ms) (7ms) 2. Recruitment and Training of Field Mappers: 9/99 9/99 (2ms) (2ms) 3. Acquisition of Maps, Equipment and Supplies: 10/99 6/00 (10ms) (10ms) 4. Field Map Updating 10/99 10/99 (20ms) (20) 5. Fair Drawing of Maps 1/01 1/01 (17ms) (17ms) 6. EA Delineation: 02/01 02/01 (6ms) (6ms) 7. Delineation of Supervision Areas 6/01 6/01 (2ms) (2ms)
145 8. Transcription of Control Forms 2/01 2/01 (7ms) (7ms) 8. Reproduction of EA & SA Maps 2/01 2/01 (7ms) (7ms)
C. Census Data Processing
1. Delivery of Desktop Computers 5/00 11/00 and Ancillary Equipment 2. Recruitment and Training of 7/01 7/01 Data Entry Operators (3ms) (3ms)
D. Preparation of Census Documents
1. Census Questionnaires: 9/99 11/00 (13ms) (13ms) 2. Enumerators’ Manual: 9/99 5/00 (13ms) (12ms) 3. Supervisors’ Manual 4/00 10/00 (11ms) (8ms) 5. District Census Officers’ Manual 3/00 5/00 (1m) (1m) 6. Technical Officers’ Manual 5/00 5/00 (12ms) (14ms) 7. Manuals for Training of enumerators and 6/00 5/00 Supervisors (12ms) (12ms) 8. Occupation and Industry Coding Manual 6/00 6/00 (12ms) (12ms) 9. Coding and Editing Instructions 6/00 6/00 (12ms) (12ms) 10. Publicity Material 01/00 5/00 (20ms) (20ms) E. Other Preparatory Activities
1. The Pilot Census 8/00 8/00 (1wk) (1wk) 2. Preparation of Tabulation and 1/00 7/00 Publication Programmes (14ms) (20ms) 3. Printing of Census Enumeration Documents 1/01 1/01 (6ms) (7ms) 4. Assembly, Packing and Distribution of Census Enumeration Documents and Related Materials 6/01 6/01 (2ms) (3ms) G. Training
1. Training of DCOs 5/00 5/00
146 (2wks) (2wks) 2. Central Statistics Office Staff 3/00 7/00 (1wk) (1wk) 3. Training Workshop on Use of 7/00 7/00 IMPS and other Packages (2wks) (2wks)
4. Trainers of Field Staff - Training of Trainers 3/01 3/01 (5days) (5days) - Training and Selection of Supervisors 4/01 4/01 (5days) (5days) - Training and Selection of Enumerators 8/01 8/01 (5days) (5days)
2. CENSUS ENUMERATION ACTIVITIES
1. Main Census Enumeration 8/01 8/01 (10days) (12days) 2. Return of Completed Questionnaires and other enumeration Materials to the Census Office 8/01 8/01 (5days) (10days) 3. POST – ENUMERATION ACTIVITIES
(a) Manual Editing and Data Processing
1. Development of Tracking System 8/01 8/01 (1m) (1m) 2. Manual Checking and Coding of 10/01 10/01 Census Questionnaires (6ms) (6ms)
3. Data Entry 10/01 10/01 (6ms) (7ms) 4. Editing and Validation of Census Data 11/01 12/01 (6ms) (6ms) 5. Tabulation of Census Data 11/01 11/01 (8ms) (8ms) 8. Collection of Supplementary Data * * for “Guide to the Villages of Botswana” (*) (*)
b) Preparation of Census Reports
1. Census Stake holder’s Workshop 5/02 5/02 (2days) (2days) 3. Evaluations and Basic Analysis 5/02 5/02 of the Census Data (3ms) (3ms)
147
4. Preparation of Census Publication Maps 4/02 4/02 (4ms) (4ms) 5. Preparation of Detailed Account of 9/01 9/1 Census Operations (Costs included) (6ms) (6ms)
(c ) Printing and Distribution of Census Publications:
- Census Preliminary Report 9/01 9/01 (2m) (2wks) - Population of Towns, Villages and 11/01 11/01 Associated Localities (6ms) (6ms)
- Administrative Report and National 1/02 9/01 Statistical Tables (12ms) (15ms)
- Guide to the Villages of Botswana 1/02 9/01 (12ms) (6ms) - Analytical Reports 8/03 8/03 (12ms) * - Dissemination Seminars/Workshops 1/94 7/03 (2ms) (1m) - Evaluation of the Census Project 8/03 * (2ms) * - Summary Statistics on Small Areas 9/02 * (12ms) * KEY: * Activities still pending as at time of printing. m(s) Month (s) wk(s) week (s)
148
Appendix 2: Census Order and Regulations
Statutory Instrument No.26 of 2001
CENSUS ACT (Cap. 17:02)
CENSUS ORDER, 2001 (Published on 22 June, 2001)
IN EXERCISE of the powers conferred by section 2 of the Census Act, the Minister of Finance and Development Planning hereby makes the following order –
1. This Order may be cited as the Census Order 2001 2. A census shall be taken in Botswana beginning on the 17th August to the 26th August, 2001.
MADE this 15th day of June 2001.
B. GAOLATHE, Minister of Finance and Development Planning
149
Statutory Instrument No. 29 of 2001
CENSUS ACT (Cap. 17:02)
CENSUS (AMENDMENT) REGULATIONS, 2001 (Published on 6th July, 2001) ARRANGEMENT OF REGULATIONS
REGULATIONS 1. Citation 2. Amendment of Regulation 6 3. Amendment of Schedule
In exercise of the powers of conferred on the Minister of Finance and Development Planning by section 11 of the Census Act, the following Regulations are hereby made – 1. These Regulations may be cited as the Census (Amendment) Citation Regulations, 2001. 2. The Census Regulations (hereinafter referred to as “the Regulations” are Amended in Regulation 6, by submitting - Amendment Regulation 6 (a) for the figure “P100” the figure “P500”; and (b) for the words “3 months” the words “6 months”. 4. The Schedule to the Regulations is amended by submitting for Amendment the Schedule thereto, the following new Schedule - of the Schedule
CENSUS REGULATIONS, 2001
(Incorporating Amendments by Statutory Instrument No. 29 of 2001 Published on 6 July 2001)
REGULATION
2. The duties of officers appointed to be Census Officers shall be-
(a) to enumerate persons within Botswana during the period the census is to be taken at such times, other than the periods between 10.00 pm and 6.00 am, and at such places as may be expedient;
150
(b) to ask such questions of any persons so enumerated as may be necessary in order to obtain, in respect of that person and, where expedient, in respect of any other person such as is referred to in regulation 4, the information specified in the Schedule hereto; and
(c) to map inhabited areas of Botswana and to include in the maps, among other things, all arable land and cattle post areas.
3. Every person questioned by a census officer in or within the precincts of any house or dwelling shall, if so required by such census officer,-
(a) give to the best his knowledge and belief the information referred to in the Schedule in respect of any member of the household who is absent from the house or other dwelling at the time of such questioning;
(b) give such information as he may possess as to how long any member of the household who is absent from Botswana has been so absent; and
(c) give to the best of his knowledge and belief, the information referred to in the Schedule in respect of any other person who spend the preceding night at the house or other dwelling.
4. A census officer may require any person who has attained the age or apparent age of 12 years to give any information required by or under these Regulations.
5. Forms to be used for purposes of section 6,7 and 9 of the Act shall be such forms as may be approved therefore by the Minister.
6. (1) Any person who, without reasonable cause,-
(a) fails to give any information which he is required by or under these Regulations to give;
(b) for purposes of the census, gives to any census officer any information which he knows to be false or does not believe to be true; or
(c) wilfully hinders or obstructs any census officer in the exercise of the powers or the performance of the duties conferred or imposed on him by the Act or these Regulations, shall be guilty of an offence and liable to a fine of P500 and to imprisonment for 6 months.
“SCHEDULE (reg 2 (b))
151 INFORMATION TO BE OBTAINED BY A CENSUS OFFICER IN RESPECT OF ANY PERSON
Name Relationship to head of household Sex Age Citizenship Place of birth Place of usual residence Parental Survival Language Education Training (institutions of higher learning or trade schools) Religion Marital status Economic activity Occupation Industry Number of Children born alive Births 12 months prior to the census Household members outside Botswana Any disability of member of household Death in the household 12 months prior to the census Ownership of Livestock Agricultural activity and ownership/access to/of agricultural land Source of household income Type of housing unit Method of acquisition of housing unit Number of rooms in housing unit Material of construction of housing unit Water supply to the household Type of toilet facility of the household Energy source for lighting, cooking and heating Refuse disposal Ownership of durable goods Home-based care”
MADE this 25th day of June 2001.
B. GAOLATHE, Minister of Finance and Development Planning
152
Appendix 3: Index of Maps
SOUTHERN 10
LOCALITY NAME LOCALITY CODE SCALE 1.Southern (3 Sheets) 10 1:100,000 2.Kanye 10-01-100 Approx 1:6000 3.Ranaka 10-02-100 Sketch 4.Gasita 10-05-100 1:10,000 5.Lorolwane 10-06-100 Sketch 6.Tsonyane 10-08-100 Sketch 7.Kgomokasitwa 10-09-100 Sketch 8.Pitseng 10-10-100 Sketch 9.Mokhomba 10-11-100 1:10,000 10.Lekgolobotlo 10-12-100 1:10,000 11.Seherelela 10-13-100 Sketch 12.Lotlhakane 10-14-100 1:10,000 13.Sese 10-16-100 Sketch 14.Sesung 10-17-100 Sketch 15.Mogotlhwane 10-18-100 1:10,000 16.Segwagwa 10-19-100 Sketch 17.Manyana 10-20-100 Sketch 18.Maokane 10-21-100 Sketch 19.Morojane 10-23-100 Sketch 20.Molapowabojang 10-26-100 1:10,000 21.Moshaneng 10-28-100 Sketch 22.Moshupa 10-29-100 1:10,000 23.Ntlhantlhe 10-30-100 1:10,000 24.Samane 10-32-100 1:10,000 25.Selokolela 10-35-100 1:10,000 26.Mogonye 10-38-100 Sketch 27.Sebohelo 10-43-100 Sketch 28.Lotlhakane west 10-44-100 Sketch
BOROLONG 11 LOCALITY NAME LOCALITY CODE SCALE 1.Borolong (2 sheets) 11 1:100,000 2.Pitsane 11-01-100 1:10,000 3.Tlhareseleele 11-02-100 1:10,000 4.Pitsane-Potokwe 11-03-100 1:10,000 5.Rakhuna 11-04-100 1:10,000 6.Malokaganyane 11-05-100 Sketch 7.Bethel 11-06-100 1:10,000 8.Dinatshana 11-07-100 1:10,000
153 9.Ngwatsau 11-08-100 Sketch 10.Ramatlabama 11-09-100 1:10,000 11.Goodhope 11-10-100 1:10,000 12.Mokatako 11-11-100 1:10,000 13.Tswaaneng 11-12-100 1:10,000 14.Metlojane 11-13-100 1:10,000 15.Borobadilepe 11-14-100 Sketch 16.Hebron 11-15-100 1:10,000 17.Logagane 11-16-100 Sketch 18.Lotlhaje 11-17-100 Sketch 19.Makokwe 11-18-100 Sketch 20.Marojane 11-19-100 Sketch 21.Papatlo 11-20-100 1:10,000 22.Pihitswane 11-21-100 Sketch 23.Molete 11-22-100 Sketch 24.Ditlharapa 11-23-100 1:10,000 25.Madingwana 11-24-100 Sketch 26.Kgoro 11-25-100 1:10,000 27.Sheepfarm 11-26-100 Sketch 28.Lokalana 11-27-100 Sketch 29.Mogwalale 11-28-100 Sketch 30.Gathwane 11-29-100 Sketch 31.Digawana 11-30-100 1:5,000 32.Magoriapitse 11-31-100 Sketch 33.Lejwana 11-32-100 Sketch 34.Mogojogojo 11-33-100 Sketch 35.Mmathethe 11-34-100 Sketch 36.Mokgomane 11-35-100 Sketch 37.Pitsane-Molopo 11-36-100 1:10,000 38.Sedibeng 11-37-100 1:10,000 39.Gamajaalela 11-40-100 Sketch 40.Dikhukhung 11-41-100 Sketch 41.Leporung 11-42-100 Sketch 42.Mmakgori 11-43-100 Sketch 43.Mabule 11-44-100 Sketch 44.Tshidilamolomo 11-45-100 Sketch 45.Metlobo 11-47-100 Sketch 46.Lorwana 11-48-100 Sketch 47.Kangwe 11-51-100 Sketch 48.Sekhutlane 11-52-100 Sketch
NGWAKETSE WEST 12
LOCALITY NAME LOCALITY CODE SCALE 1.Ngwaketse west (2 sheets) 12 1:100,00
154 2.Mabutsane 12-01-100 1:10,000 3.Morwamosu 12-02-100 1:10,000 4.Sekoma 12-03-100 Sketch 5.Khonkhwa 12-04-100 Sketch 6.Keng 12-05-100 1:10,000 7.Khakhea 12-06-100 1:10,000 8.Kokong 12-07-100 1:10,000 9.Kanaku 12-08-100 Sketch 10.Mahotshwane 12-09-100 Sketch 11.Ithokole 12-10-100 Sketch
SOUTH EAST 20
LOCALITY NAME LOCALITY CODE SCALE 1.South East (1 Sheet) 20 1:100,000 2.Otse 20-01-100 1:10,000 3.St Josephs College 20-02-100 Sketch 4.Ramotswa Station 20-03-100 1:10,000 5.Ramotswa 20-04-100 1:10,000 6.Mogobane 20-05-100 1:10,000 7.Tlokweng 20-06-100 1:10,000
KWENENG SOUTH 30
LOCALITY NAME LOCALITY CODE SCALE 1.Kweneng south (2 Sheets) 30 1:100,000 2.Molepolole 30-01-100 1:10,000 3.Lentsweletau 30-02-100 Sketch 4.Mahetlwe 30-03-100 1:10,000 5.Gakgatla 30-04-100 1:10,000 6.Medie 30-05-100 1:10,000 7.Gabane 30-07-100 1:10,000 8.Hatsalatladi 30-09-100 1:10,000 9.Kopong 30-12-100 Sketch 10.Kumakwane 30-14-100 1:10,000 11.Lephepe 30-15-100 1:10,000 12.Boatlaname 30-18-100 1:10,000 13.Monwane 30-20-100 Sketch 14.Metsimotlhabe 30-21-100 1:10,000 15.Mmankgodi 30-22-100 1:10,000 16.Mmopane 30-23-100 1:10,000 17.Mogoditshane 30-24-100 1:10,000 18.Shadishadi 30-29-100 1:10,000 19.Sojwe 30-30-100 1:10,000
155 20. Thamaga 30-32-100 1:10,000 21.Mmanoko 30-35-100 1:10,000 22.Kubung 30-36-100 1:10,000 23.Kgope 30-39-100 Sketch 24.Mogonono 30-40-100 Sketch 25.Sorilatholo 30-41-100 Sketch
KWENENG WEST 31
LOCALITY NAME LOCALITY CODE SCALE 1.Kweneng West (1Sheet) 31 1:100,00 2.Letlhakeng 31-01-100 1:10,000 3.Botlhapatlou 31-02-100 1:10,000 4.Ditshegwane 31-03-100 1:10,000 5.Dutlwe 31-04-100 1:10,000 6.Moshaweng 31-05-100 Sketch 7.Khudumelapye 31-06-100 1:10,000 8.Mantshwabise 31-08-100 Sketch 9.Motokwe 31-09-100 Sketch 10.Ngware 31-10-100 Sketch 11.Salajwe 31-11-100 1:10,000 12.Takatokwane 31-12-100 1:10,000 13.Tsetseng 31-13-100 1:10,000 14.Tshwaane 31-14-100 Sketch 15.Malwelwe 31-16-100 Sketch 16.Maboane 31-17-100 Sketch 17.Sesung 31-18-100 Sketch 18.Kaudwane 31-22-100 Sketch
KGATLENG 40
LOCALITY NAME LOCALITY CODE SCALE 1.Kgatleng (2 Sheets) 40 1:100,000 2.Mochudi 40-01-100 1:10,000 3.Pilane 40-02-100 1:10,000 4.Rasesa 40-03-100 1:10,000 5.Bokaa 40-04-100 1:5000 6.Morwa 40-05-100 1:10,000 7.Matebeleng 40-06-100 1:10,000 8.Oodi 40-07-100 1:10,000 9.Modipane 40-08-100 1:8,000 10.Mabalane 40-09-100 1:10,000
156 11.Sikwane 40-10-100 1:10,000 12.Mmathubudukwane 40-11-100 Sketch 13.Ramonaka 40-12-100 1:10,000 14.Malolwane 40-13-100 1:10,000 15.Olifants Drift 40-14-100 Sketch 16.Atersia/Mosomane 40-15-100 1:10,000 17.Malotwana 40-16-100 1:10,000 18.Leshibitse 40-17-100 1:10,000 19.Ramotlabaki 40-18-100 1:10,000 20.Kgomodiatshaba 40-19-100 Sketch 21.Dikgonnye 40-20-100 Sketch 22.Khurutshe 40-21-100 Sketch 23.Dikwididi 40-22-100 Sketch
CENTRAL SEROWE 50
LOCALITY NAME LOCALITY NAME SCALE 1.Central Serowe (6 Sheets) 50 1:100,000 2.Serowe 50-01-100 1:5,000 3.Palapye 50-02-100 1:10,000 4.Lecheng 50-03-100 1:10,000 5.Moremi 50-04-100 Sketch 6.Malaka 50-05-100 Sketch 7.Mogapi 50-06-100 1:10,000 8.Mogapinyana 50-07-100 1:10,000 9.Kgagodi 50-08-100 1:10,000 10.Maunatlala 50-09-100 1:10,000 11.Tamasane 50-10-100 1:10,000 12.Diloro 50-11-100 Sketch 13.Lesenepole 50-12-100 Sketch 14.Mosweu 50-13-100 Sketch 15.Mokokwana 50-14-100 Sketch 16.Seolwane 50-15-100 1:10,000 17.Lerala 50-16-100 Sketch 18.Majwaneng 50-17-100 1:10,000 19.Ratholo 50-18-100 1:10,000 20.Moeng 50-19-100 Sketch 21.Gootau 50-20-100 1:10,000 22.Goo-Sekgweng 50-21-100 Sketch 23.Matlhakola 50-22-100 Sketch 24.Topisi 50-23-100 1:10,000 25.Paje 50-24-100 1:10,000 26.Mabeleapudi 50-25-100 1:10,000 27.Tshimoyapula 50-26-100 1:10,000
157 28.Mmashoro 50-27-100 Sketch 29.Mogorosi 50-28-100 1:10,000 30.Thabala 50-29-100 1:10,000 31.Moiyabana 50-30-100 Sketch 32.Gamabuo 50-31-100 Sketch 33.Radisele 50-32-100 Sketch 34.Mogome 50-33-100 Sketch 35.Mokgware 50-34-100 Sketch 36.Motshegaletau 50-36-100 Sketch 37.Malatswai 50-37-100 Sketch 38.Mokhungwana 50-38-100 Sketch 39.Serule 50-40-100 Sketch 40.Moreomabele 50-41-100 Sketch 41.Gojwane 50-42-100 Sketch 42.Manaledi 50-43-100 Sketch 43.Dimajwe 50-44-100 Sketch 44.Majwanaadipitse 50-45-100 Sketch
CENTRAL MAHALAPYE 51
LOCALITY NAME LOCALITY CODE SCALE 1.Central Mahalapye (5 sheets) 51 1:100,000 2.Mahalapye 51-01-100 1:10,000 3.Palla Road/Dinokwe 51-02-100 Sketch 4.Mookane 51-03-100 1:10,000 5.Mmaphashalala 51-04-100 1:10,000 6.Taupye 51-05-100 1:10,000 7.Dibete 51-06-100 1:10,000 8.Dovedale 51-07-100 1:10,000 9.Kudumatse 51-08-100 1:10,000 10.Makwate 51-09-100 Sketch 11.Shakwe 51-10-100 1:10,000 12.Moshopha 51-11-100 1:10,000 13.Machaneng 51-12-100 1:10,000 14.Sefhare 51-13-100 1:10,000 15.Chadibe 51-14-100 1:10,000 16.Bonwapitse 51-15-100 Sketch 17.Mmutlane 51-16-100 1:10,000 18.Mokobeng 51-17-100 Sketch 19.Ngwapa 51-18-100 Sketch 20.Tumasera/Seleka 51-19-100 1:10,000 21.Ramokgonami 51-20-100 1:10,000 22.Maape 51-21-100 Sketch 23.Pilikwe 51-22-100 Sketch
158 24.Mhalapitsa 51-23-100 Sketch 25.Borotsi 51-24-100 Sketch 26.Tewane 51-25-100 Sketch 27.Kalamare 51-26-100 1:10,000 28.Mokgenene 51-27-100 Sketch 29.Shoshong 51-28-100 1:10,000 30.Moralane 51-29-100 Sketch 31.Mosolotshane 51-30-100 1:10,000 32.Kodibeleng 51-31-100 Sketch 33.Ikongwe 51-32-100 Sketch 34.Poloka 51-33-100 Sketch 35.Otse 51-34-100 Sketch 36.Matlhako 51-35-100 Sketch 37.Mokoswane 51-36-100 Sketch
CENTRAL BOBONONG 52
LOCALITY NAME LOCALITY CODE SCALE 1.Central Bobonong (4 sheets) 52 1: 100,000 2.Bobonong 52-01-100 1:10,000 3.Tsetsebjwe 52-02-100 1:10,000 4.Tshokwe 52-03-100 Sketch 5.Mathathane 52-04-100 1:10,000 6.Lentswelemoriti 52-05-100 Sketch 7.Semolale 52-06-100 1:10,000 8.Gobajango 52-07-100 1:10,000 9.Lepokole 52-08-100 Sketch 10.Molalatau 52-09-100 Sketch 11.Tobane 52-10-100 1:10,000 12.Sefophe 52-11-100 1:10,000 13.Mabolwe 52-12-100 Sketch 14.Mmadinare 52-13-100 1:10,000 15.Robelela 52-14-100 Sketch 16.Damochujenaa 52-15-100 Sketch 17.Motlhabaneng 52-17-100 Sketch 18.Moletemane 52-39-100 Sketch
CENTRAL BOTETI 53
LOCALITY NAME LOCALITY CODE SCALE 1.Central Boteti (1 Sheets) 53 1:250,000 2.Letlhakane 53-01-100 1:10,000 3.Mosu 53-02-100 Sketch 4.Mmatshumo 53-03-100 Sketch
159 5.Mopipi 53-04-100 Sketch 6.Xhumo 53-05-100 Sketch 7.Kedie 53-06-100 Sketch 8.Rakops/Tsienyane 53-07-100 1:10,000 9.Toromoja 53-08-100 Sketch 10.Moreomaoto 53-09-100 Sketch 11.Makalamabedi 53-10-100 Sketch 12.Khwee 53-11-100 Sketch 13.Kumaga 53-12-100 Sketch 14.Motopi 53-13-100 Sketch 15.Mmadikola 53-14-100 Sketch 16.Mokoboxane 53-15-100 Sketch
CENTRAL TUTUME 54
LOCALITY NANE LOCALITY CODE SCALE 1.Central Tutume (2 Sheets) 54 1:250,000 2.Maitengwe 54-01-100 1:10,000 3.Nata 54-02-100 1:10,000 4.Dagwi 54-03-100 Sketch 5.Nkange 54-04-100 1:10,000 6.Senete 54-05-100 1:10,000 7.Changate 54-06-100 1:10,000 8.Tutume 54-07-100 1:10,000 9.Matobo 54-08-100 1:10,000 10.Goshwe 54-09-100 1:10,000 11.Mosetse 54-10-100 1:10,000 12.Nswazwi 54-11-100 1:10,000 13.Nshakashokwe 54-12-100 1:10,000 14.Sebina 54-13-100 Sketch 15.Marobela 54-14-100 1:10,000 16.Mathangwane 54-15-100 1:10,000 17.Marapong 54-16-100 1:10,000 18.Chadibe 54-17-100 1:10,000 19.Matsitama 54-18-100 1:10,000 20.Borolong 54-19-100 1:10,000 21.Dukwi 54-20-100 Sketch 22.Mokubilo 54-22-100 Sketch 23.Tonota 54-23-100 1:10,000 24.Borotsi 54-24-100 Sketch 25.Mandunyane 54-25-100 1:10,000 26.Natale 54-26-100 Sketch 27.Semotswane 54-27-100 Sketch
160 28.Shashemooke 54-29-100 1:10,000 29.Gweta 54-30-100 1:10,000 30.Zoroga 54-31-100 Sketch 31.Lepashe 54-32-100 Sketch 32.Mmanxotae 54-33-100 Sketch 33.Sepako 54-34-100 Sketch 34.Semitwe 54-35-100 Sketch 35.Makobo 54-36-100 Sketch 36.Mabesekwa 54-37-100 Sketch 37.Tshokotshaa 54-43-100 Sketch 38.Makuta 54-44-100 Sketch 39.Maposa 54-45-100 Sketch 40.Kutamogoree 54-46-100 Sketch 41.Mmeya 54-47-100 Sketch
NORTH EAST 60
LOCALITY NAME LOCALITYCODE SCALE 1.North East (2 Sheets)60 1:100,000 2.Mbalambi 60-01-100 Sketch 3.Gambule 60-03-100 1:10,000 4.Gulubane 60-04-100 1:10,000 6.Gungwe 60-05-100 Sketch 7.Jackalas1 60-06-100 1:10,000 8.Jackalas 2 60-07-100 Sketch 9.Kalakamati 60-08-100 Sketch 10.Letsholathebe 60-09-100 1:10,000 11.Makaleng 60-10-100 1:10,000 12.Mulambakwena 60-11-100 1:10,000 13.Mambo 60-12-100 1:10,000 14.Mapoka 60-13-100 1:10,000
15.Masingwaneng 60-14-100 Sketch 16.Masunga 60-15-100 1:10,000 17.Matsiloje 60-16-100 1:10,000 18.Moroka 60-17-100 1;10,000 19.Mosojane 60-18-100 1:10,000 20.Nlapkhwane 60-19-100 1:10,000 21.Ramokgwebana 60-20-100 1:10,000 22.Sechele 60-21-100 Sketch 23.Sekakangwe 60-22-100 1:10,000 24.Senyawe 60-23-100 1:10,000 25.Siviya 60-24-100 1:10,000 26.Tati Siding 60-25-100 1:10,000
161 27.Themashanga 60-26-100 Sketch 28.Tsamaya 60-27-100 1:10,000 29.Tshesebe 60-28-100 1:10.000 30.Zwenshambe 60-29-100 1:10,000 31.Vukwi 60-30-100 1:10,000 32.Butale 60-31-100 Sketch 33.Masukwane 60-32-100 1:10,000 34.Shashe Bridge 60-33-100 Sketch 35.Matshelagabedi 60-34-100 1:10,000 36.Ditladi 60-35-100 Sketch 37.Botalaote 60-36-100 Sketch 38.Toteng 60-37-100 Sketch 37.Matenge 60-38-100 Sketch 38.Pole 60-40-100 Sketch 39.Kgari 60-41-100 Sketch 40.Mowana 60-42-100 Sketch 41.Mabudzane 60-43-100 Sketch 42.Matopi 60-44-100 Sketch
NGAMILAND SOUTH 70
LOCALITY NAME LOCALITY CODE SCALE 1.Ngamiland South (1 Sheet) 70 1:350,000 2.Maun 70-01-100 1:10,000 3.Makalamabedi 70-02-100 Sketch 4.Shorobe 70-03-100 Sketch 5.Toteng 70-04-100 Sketch 6.Matlapana 70-19-100 Sketch 7.Phuduhudu 70-20-100 Sketch 8.Komana 70-21-100 Sketch 9.Chanoga 70-24-100 Sketch 10.Tsao 71-21-100? Sketch 11.Sehithwa 71-23-100? Sketch 12.Kareng 71-24-100? Sketch 13.Semboyo 71-30-100 Sketch 14.Habu 71-32-100 Sketch
NGAMILAND WEST 71
LOCALITY NAME LOCALITY NAME SCALE 1.Ngamiland West (1 Sheet) 71 1:350,000
162 2.Seronga 71-01-100 Sketch 3.Ngarange 71-02-100 Sketch 4.Beetsha 71-03-100 Sketch 5.Gunotsoga 71-04-100 Sketch 6.Mokgacha 71-05-100 Sketch 7.Shaowe 71-07-100 Sketch 8.Xadau 71-08-100 Sketch 9.Xangwa 71-09-100 Sketch 10.Xakao 71-10-100 Sketch 11.Kauxwhi 71-11-100 Sketch 12.Etsha 6 71-12-100 Sketch 13.Ikoga 71-13-100 Sketch 14.Sepopa 71-14-100 Sketch 15.Nxamasere 71-15-100 Sketch 16.Shakawe 71-16-100 Sketch 16.Nxaunxau 71-17-100 Sketch 17.Mohembo West 71-18-100 Sketch 18.Etsha 13 71-19-100 Sketch 19.Nokaneng 71-20-100 Sketch 20.Tubu 71-25-100 Sketch 21.Gumare 71-26-100 Sketch 22.Xaxa 71-27-100 Sketch 23.Bodibeng 71-28-100 Sketch 24.Kgakge 71-31-100 Sketch 25.MohemboEast ? Sketch
CHOBE 72
LOCALITY NAME LOCALITY CODE SCALE 1.Chobe (1 Sheet) 72 1:350,000 2.Kachikau 72-01-100 Sketch 3.Kasane 72-02-100 1:10,000 4.Kazungula 72-03-100 Sketch 5.Kavimba 72-04-100 Sketch 6.Lesoma 72-05-100 Sketch 7.Mabele/Muchenje 72-06-100 Sketch 8.Parakarungu 72-07-100 Sketch 9.Pandamatenga 72-08-100 Sketch 10.Satau 72-09-100 Sketch
NGAMILAND DELTA 73
LOCALITY NAME L0CALITY CODE SCALE
163 1.Ngamiland Delta (1 Sheet) 73 1:350,000 2.Xaxaba 73-01-100 No Map 3.Jao 73-02-100 No Map 4.Katamaha 73-06-100 No Map 5.Daonara 73-07-100 No Map 6.Ditshiping 73-09-100 No Map 7.Diadora 73-10-100 No Map
GHANZI 80
LOCALITY NAME LOCALITY CODE SCALE 1.Ghanzi District (1 Sheet)80 1:500,000 2.Ghanzi 80-01-100 1:5,000 3.Dekar 80-02-100 1:10,000 4.Tsootsha/Kalkfontein 80-03-100 Sketch 5.Karakubis 80-04-100 Sketch 6.Tshobokwane 80-05-100 Sketch 7.West Hanahai 80-06-100 Sketch 8.Charles Hill 80-07-100 Sketch 9.Makunda 80-09-100 1:10,000 10.Kule 80-10-100 1:10,000 11.Ncojane 80-11-100 Sketch 12.Groote Laagte 80-12-100 1:10,000 13.New Xanagas 80-13-100 1:10,000 14.East Hanahai 80-14-100 Sketch 15.Kacgae 80-15-100 Sketch 16.Bere 80-16-100 Sketch 17.New Xade 80-18-100 Sketch
KGALAGADI SOUTH 90
LOCALITY NAME LOCALITY CODE SCALE 1.KgalagadiSouth (1 Sheet) 90 1:500,000 2.Werda 90-01-100 1:10,000 3.Makopong 90-02-100 1:10,000 4.Kisa 90-03-100 1:10,000 5.Omaweneno 90-04-100 1:10,000 6.Tsabong 90-05-100 1:10,000 7.Kolonkwaneng 90-06-100 1:10,000 8.Bogogobo 90-07-100 Sketch 9.Middlepits 90-08-100 Sketch 10.Khuis 90-09-100 Sketch 11.Gakhibana 90-10-100 Sketch
164 12.Rapplespan 90-11-100 Sketch 13.Vaal Hoek 90-12-100 1:10,000 14.Boks Pits 90-13-100 1:10,000 15.Struinzendam 90-14-100 1:10,000 16.Bray 90-23-100 1:10,000 17.Draai Hoek 90-24-100 Sketch 18.Maubelo 90-25-100 Sketch 19.Khawa 90-26-100 1:10,000 20.Kokotsha 90-33-100 Sketch 21.Maralaleng 90-34-100 1:10,000 22.Maleshe 90-35-100 Sketch
KGALAGADI NORTH 91
LOCALITY NAME LOCALITY CODE SCALE 1.Kgalagadi North (1 Sheet) 91 1:500,000 2.Kang 91-15-100 1:10,000 3.Tshane 91-16-100 Sketch 4.Hukuntsi 91-17-100 10,000 5.Lehututu 91-18-100 1:10,000 6.Lokgwabe 91-19-100 1:10,000 7.Monong 91-20-100 Sketch 8.Ncaang 91-21-100 Sketch 9.Hunhukwe 91-26-100 Sketch 10.Zutshwa 91-27-100 Sketch 11.Ngwatle 91-28-100 Sketch 12Ukhwi. 91-29-100 Sketch 13.Make 91-30-100 Sketch 14.Inalegolo 91-31-100 Sketch 16.Phuduhudu 91-32-100 Sketch
165 Appendix 4. Members of Various Committees
4.0 Census Central Committee members S. S. G. Tumelo Permanent Secretary MFDP - Chairman L. P. Gakale Permanent Secretary MOA P. T. Ramatswi Permanent Secretary MOE B. K. Sentle Deputy Permanent Secretary MLG M. J. Moatshe Deputy Permanent Secretary - MLH M. L. Nchunga Executive Secretary, National Conservation Strategy M. Chakalisa Deputy Permanent Secretary MLHA R. Segodi Director of Towns and Regional Planning - MLG W. Manyeneng Assistant Director, Primary Healthcare - MOH G. M. Charumbira Government Statistician (Secretary) D. Buthali Census Officer (Invited Member - Recording)
4.1 Census Technical Advisory Committee
G. M. Charumbira Government Statistician - Chairman B. C. Molomo Director Strategic and Contingency Planning A. N. Majelantle Chief Statistician - CSO R. G. Majelantle Department of Demography - UB M. L. Mokgautsi Civil Registration - MLHA T. Masalila Principal Statistician - CSO D. Buthali Census Officer – Secretary A. A. Mokgare Principal Agriculture Economist - MOA R. Segodi Director, Town and Regional Planning R. Greener Statistician - BIDPA Dr. D. Hudson Statistician - Phaleng Consultancy M. P. Kerekang Principal Statistician – CSO S. Basiamang Director of Primary Education - MOE) H. Nthibe Principal Research Officer - MLG
166
4.2 Census Standing Committee Members
G. M. Charumbira Government Statistician (Chairman) M. Malepa Senior Statistician K. G. Johane Senior Statistician L. Busang Statistician I O. Silas Statistician II L. P. Mosele Senior Statistician D. Mello Statistician II A. N. Majelantle Chief Statistician M.P. Kerekang Principal Statistician I D. Buthali Census Officer (Secretary) T. Botana Principal Statistician I J. Katale Principal Statistician I V. K. Dwivedi Principal Statistician II M. Shatera Senior Statistician M. Phatshwane Senior Statistician T. Maruatona Principal Statistician II I. M. Mokubung Principal Statistician II K. Makhumalo Principal Statistician II K. Mogotsinyana Principal Statistician II G. N. Mosupi Principal Statistician II
4.3 Census Technical Officers (CSO Staff)
Name Designation
R. K. Mothibi Senior Statistician M. Phatshwane Senior Statistician P. Kelekwang Senior Statistician M. Malepa Senior Statistician K. G. Johane Senior Statistician S. Ramolemane Statistician I P. P. Gopolang Statistician I L. Busang Statistician I O. Silas Statistician II M. Gwafila Statistician II L. P. Mosele Senior Statistician N. Nko Statistician I V.T. Sebekedi Statistician I O.J Puleng Statistical Officer D. Gaobotse Statistician II D. N. Motlapele Senior Statistical Officer
167 O. Gaosekelwe Statistician I S. Modise Statistician I P. Labobedi Statistician I A. Kebonyemotse Statistician I K. Mbiganyi Statistician I O. Lebogang Statistician I A.S. Thuto Assistant Statistician D. Mello Statistician II K. M. Motlhabane Statistician II B.S. Tlhomelang Statistician II T. Basinyi Senior Statistical Officer M. Pheto Statistician II T. Laletsang Statistician II J. Tauetsile Statistician II J. M. Letebele Senior Statistical Officer H. K. Koboa Statistician II B. Dema Senior Statistical Officer S.R. Pheko Statistician II M. Masingwane Principal Statistical Officer T. Marapo Statistician II G. Ramotsoko Statistician II E. Onyadile Statistician I S. Sabone Statistician II M. Moreti Senior Statistical Officer A. N. Majelantle Chief Statistician M.P. Kerekang Principal Statistician I D. Buthali Census Officer T. Botana Principal Statistician I D. M. Lesetedi Principal Statistician II J. Katale Principal Statistician I V. K. Dwivedi Principal Statistician II (Research) M. Shatera Senior Statistician T. Maruatona Principal Statistician II I. M. Mokubung Principal Statistician II S.E. Kabaija Principal Statistician II K. Makhumalo Principal Statistician II K. Mogotsinyana Principal Statistician II K. Moalosi Principal Statistician II G. N. Mosupi Principal Statistician II G. M. Charumbira Government Statistician
168
4.4 District Census Officers
Place Name Designation Kanye/Moshupa Nkabo Kefhilwe District Officer Dev. O. Mosweu Physical Planner Goodhope Sub District P. Moalafi Assistant Physical Planner O. K. Ditedu District Agric. Officer
Sowa Town F. Ookeditswe Senior Education Officer
North East District K. Ntapu Senior Physical Planner G. Shatera District Officer Lands
Francistown C. D. Siviya Regional Election Officer G. Basupi Regional Civil Reg. Officer G. Lepang Senior Health Inspector
Ghanzi D. Khudu Senior Physical Planner M. G. Mochanang Senior Ass. District Officer P. Makambe Assistant District Officer
Kgalagadi M. Khuduego Senior Lands Officer E. Keabaitse Ag. Senior Physical Planner G.B Kenosi Ass. District Officer Lands M. L. Phiri Assistant District Officer E. Onneetse District Agric. Officer
Ngamiland U. M. Maswibilili Principal Lands Officer S. Aaron Assistant Physical Planner
Kasane M. Oageng District Officer (Lands)
Gumare B. J. Galekgorwe District Agric. Officer Keikitse Tladi Assistant Physical Planner
Mabutsane T. L. Moloko District Officer
Orapa J. Morwe Principal Adult Education Officer B. Gasha Assistant Physical Planner
Mahalapye L. Plaajie DOD Violet Kaunda PDOD
169
Bobonong K. Motlopi District Officer Lands G. Matsila Physical Planner
Tutume M. Macha District Officer Dev. C. S. Madwala District Agric. Officer
Jwaneng L. Mahape Teacher S. Oitsile Economic Planner
Lobatse O. Modisa District Officer Dev. M. Makope Principal Housing Officer
Mochudi K. Rutherford Principal Physical Planner H.T. Tapela District Officer Dev.
Molepolole M. Ramashaba District Officer Dev. Z. Letshoo Principal Admin. Officer K. N. Relaeng District Officer Admin.
Letlhakeng G.B. Smile Senior Education Officer S.A. Mosinyi Internal Auditor
South East A. Ramotsebe District Horticul. Officer M. Marata Land Board Secretary Selebi Phikwe Mpho Mathe Physical planner H. Mathumo Senior Education Officer
Gaborone B. Masole Sen. Social Welfare Officer M. Tiro Admin Officer R. Mosimenyana District Labour Officer T Balesamang Physical Planner K. Moalosi Senior Admin. Officer
170
Appendix 5: Training Schedules (Enumerators and Supervisors)
Training Schedule – Census Enumerators (11 – 15 August 2001)
DAY SESSION TOPIC/ACTIVITY
Administrative Logistics Session I (0800-1300) Definition of Terms and Concepts (Tea Break 1000 – 1030) Preparation for Enumeration
Day I 1300 – 1400 LUNCH
Enumeration Procedure Session II 1400 – 1700 Questionnaire (A1-A7) Questionnaire (A8 – A12) (Tea Break 1530 – 1545) Questionnaire (A13 – A20) Questionnaire (A21- A26)
Session I Questionnaire (A21- A26)
0800-1300 Questionnaire (A27 – A32) Questionnaire (Section B) (Tea Break 1000 – 1030) Questionnaire (Section C) Day 2 1300- 1400 LUNCH
Questionnaire (Section D) Session II 1400- 1700 Questionnaire (E1 – E7) Questionnaire (E8 – E15) (Tea Break 1530 – 1545) Questionnaire (E16 – E22)
Procedure After Household Enumeration Session I 0800 – 1300 Institutional Questionnaire (Tea Break 1000 – 1030) Day 3 1300 – 1400 LUNCH
171 Interview Demonstration (2@ 40 Min) Session II 1400 – 1700 Discussions (1 Hour) (Tea Break 1530 – 1545) Profile I as homework (not for discussion)
Supervised Practical Interviews (4 @ trainee) Session I 0800 – 1300 Discussion of Experiences (Tea Break 1000 – 1030)
Day 4 1300 – 1400 LUNCH
Supervised Practical Interviews (2@ trainee) Session II 1400 – 1700 Discussion of Experiences (Tea Break 1530 – 1545) Profile II as Homework (not for discussion)
Supervised Practical Interviews (2@ trainee) Session I (Practical Assessment) 0800 – 1300 Discussion of Experiences (Tea Break 1000 – 1030) Review of important concepts and procedures
Day 5 1300 – 1400 LUNCH
Assessment Test (Written) – Duration 1 Hour Session II 1400 – 1700 Administrative Logistics, Marking, etc End of Training
172
Training Schedule – Census Supervisors (23 – 27 April 2001)
DAY SESSION TOPIC/ACTIVITY
Administrative Logistics Session I (0800-1300) Definition of Terms and Concepts (Tea Break 1000 – 1030) Preparation for Enumeration
Day I 1300 – 1400 LUNCH
Enumeration Procedure Session II 1400 – 1700 Questionnaire (A1-A7) Questionnaire (A8 – A12) (Tea Break 1530 – 1545) Questionnaire (A13 – A20) Questionnaire (A21- A26)
Questionnaire (A21- A26) Session I 0800-1300 Questionnaire (A27 – A32) Questionnaire (Section B) (Tea Break 1000 – 1030) Questionnaire (Section C) Day 2 1300- 1400 LUNCH
Questionnaire (Section D) Session II 1400- 1700 Questionnaire (E1 – E7) Questionnaire (E8 – E15) (Tea Break 1530 – 1545) Questionnaire (E16 – E22)
Procedure After Household Enumeration Session I 0800 – 1300 Institutional Questionnaire (Tea Break 1000 – 1030) Day 3 1300 – 1400 LUNCH
173 Interview Demonstration (2@ 40 Min) Session II 1400 – 1700 Discussions (1 Hour) (Tea Break 1530 – 1545) Profile I as homework (not for discussion)
Supervised Practical Interviews (3 @ trainee) Session I 0800 – 1300 Discussion of Experiences (Tea Break 1000 – 1030)
Day 4 1300 – 1400 LUNCH
Supervised Practical Interviews (2@ trainee) Session II 1400 – 1700 Discussion of Experiences (Tea Break 1530 – 1545) Profile II as Homework (not for discussion)
Duties of Supervisors Session I 0800 – 1300 Detailed Checks (Tea Break 1000 – 1030) Supervisor Summaries Assessment Test (Written)
Day 5 1300 – 1400 LUNCH
Contingencies Session II 1400 – 1500 Administrative Logistics End of Training
174 Appendix 6: Guidelines to Training Officers
While the training follows the Enumerator’s Manual, some sections are fairly straightforward hence trainers may produce notes rather than read line by line. Some suggestions have been proposed to follow in making such notes. In some sections it is suggested that trainers read the manual line by line in order to retain the require detail. It is planned that trainees will be provided with the manuals in advance.
Administrative Issues
- Census Organisation - Objectives of census - Reference to Census Act
Concepts and Definitions
Read the section line by line. Trainees must understand the section fully hence trainers must devote more time to the section.
Preparation for Enumeration
Read the section line by line. Trainees must understand the section fully hence trainers must devote more time to the section.
Enumeration Procedure
As a guide, put emphasis on the following issues:
- Courtesy in dealing with members of the public - Establishing number of households - Establishing the household head - Whom to Enumerate - Cases of no one at home - Questionnaire General Points o Who fills the questionnaire o Shaded and un-shaded parts of the questionnaire lines o Number of digits for each answer o Cases of “not applicable”/Skips o Cancellation of responses o Answers not known o Avoid skipping/reserving pages o Use of comments box - Enumeration of those who slept with but have left the household - Revisits and revisits forms
175 - Structure of questionnaire
Census Questionnaire
The presentation should focus on the census questionnaire but with constant reference to the manual except where it is indicated otherwise.
A1 – A7
As a guide, emphasize the following issues:
- Order of listing persons - Birth and death during census night - Serial number and households with more than 10 members - No sex not known - Age to be estimated at all costs, use events calendar, ID, Clinic cards, etc - Age below 1 year or above 98 - Age in completed years - Age consistency with relationship - Writing of country name and code, country codes, locality codes - Place of birth, usual place of residence of mother - Name and code of locality of birth - Significance of place of birth
A8 - A12
- What is usual place of residence - Usual place of residence of residents, tourists, etc - Importance of parental survival/orphan hood - Biological rather than foster parents
A13 – A20 - Language and dialects - Age cut-off point for questions - Education level, use of 09, 19, 29, 69,etc - Education level for those still attending - Connotation attached to first digit of education level - Formal training of at least 3 months - Religion but not church denominations - Acceptance of religion as given - Distinction of marital status categories in brief
A21 – A26
- Go through questionnaire followed by reading section paragraph by paragraph - Answers of at least two words (occupation and industry)
176 - Make reference to a few probing examples for occupation and industry
A27 – A32
- Linking fertility to females rather than males - Treatment of still births - Consistency of A27 to A30 - Case of visiting mothers - Treatment of no births and unknown number of children for eligible females - Total fertility and current fertility, age restrictions - Consistency between A27 to A30 and A31 to A32
Section B
- Household members (CITIZENS) outside Botswana during census night - Cases of no absentees - Duration of absence since last visit - Cases of re-visits - Duration of absence (years and months, e.g. treatment of 18 months) - More than 3 absentees (treatment of Part A, C, D, E) - No sex unknown to be allowed - Age to be estimated at all costs, use events calendar, ID, Clinic cards, etc - Age below 1 year or above 98 - Age in completed years - Age consistency with relationship - Questions B11 – B14 also apply to persons aged below 2 years and those below 12 years - Education and Training (refer to presentation for A13 – A20)
Section C
- Persons disabled must be from B1 or C1 - Distinction between defect and inability to use - Brief description of responses
Section D
- Multiple reporting, need for probing - 14 days qualification - Deaths after 6 o’clock - Treatment of more than 3 deaths - Name and code of usual place of residence - Consistency between relationship and age
177 E1 – E7
- Exclusion of non-household members - Circling of codes and recording of responses not listed - Brief description of land acquisition modes - Importance of informal sector sources of income - Cash activities of members not enumerated with the household - Unearned income, remittances, maintenance (flow direction), etc
E8 - E15
- Distinction between housing unit and house - Brief description of types of housing units - What tenureship is - What qualifies to be a living room - What is main house - Materials of construction of main house - Many sources of water but one principal - Toilet facility used but necessarily owned - Distinction between Pit latrine, VIP and Enviro-Loo
E16 – E22
- Many sources (for lighting, cooking, heating) but just one principal - Space heating rather any other heating - Main mode of refuse disposal - Durables owned must be working - Durables owned must be in the housing unit - Critically ill rather than minor illnesses
After Household Enumeration
Read Part 5 of the Manual paragraph-by-paragraph
Institutional Questionnaire
- Draw attention to similarity and differences (compared with EA Book) o Sections B, D and E excluded o Hotels lodges forms, transcription of data into questionnaire o Residence question o Prisoners response category o Fertility Questions - Use of household questionnaire for institutions - Whom does the IQ cover - Identification Particulars for Institutions - Completing the Control List - Completing the IQ book summary
178 Duties of Supervisors (Supervisor’s Manual)
- Enumerator and supervisor materials - Public relations and courtesy - Enumerator monitoring sheet - Check completeness of enumeration (White and Orange labels) - Re-deployment of enumerators - Colour of pen to use
Detailed Checks (Supervisor’s Manual)
Read the relevant section of the Supervisor’s Manual paragraph by paragraph
Supervisor Summaries (Supervisor’s Manual)
Read the relevant section of the Supervisor’s Manual paragraph by paragraph, put emphasis on
- Source of details for EA Book totals - Source of details for Control List totals - Balance EA Book and Control List totals
179 Appendix 7: Training Centre Packaging Schedule
180 Appendix 8: Profile
Form CF2 -2000 REPBUBLIC OF BOTSWANA Profile I 2001 Population and Housing Census
District South East 2 0 Village Ramotswa 0 4
Locality Metsimaswaana 5 0 1 Enumeration Area 0 2 6 0
Serial Lolwapa/ Head of Lolwapa Status: No. Number Dwelling 1. Enumerated of Enumerated Households Number 2. Revisit hh Comments 3. Empty 4. Other (1) (2) (3) (4) P/ P/ P/ P/ P P P/ P/ p a p a / / p a p a 1 001 Phiri e Thamaga 1 1 3 0 ------2 002 De-Cruz 3 003 MmaSemausu
The De-cruz are four in the family - De-cruz who is the head, his wife Beauty, daughter Six-pence and son Sheleng. All members of the family are Batswana and they speak Setswana most often at home. Although Beauty and the children are Apostolic Faith Mission members, the head of the household claims to be a traditionalist hence believes in his ancestors. De-cruz was born 48 years ago in Ramotswa were both his parents still live. After completing Standard Seven, De-cruz obtained a plumbing certificate through apprenticeship and is now working as a plumber at Concor Construction Company. He has been working for that construction company since 1991. He usually lives in Gaborone and as such was living there a year ago as well as five years ago.
Both Beauty’s parents died 36 years ago when she was only 4 years old. Although she was born in Khakhea she was raised by her aunt at Metsimaswaana were she has always lived. She dropped out of school whilst doing Standard Seven. She attended a four- months tailoring course in Lobatse five years ago. Equipped with the Singer Sewing machine that she bought in October last year after a good harvest, she is now the “village tailor”. In the 1998/1999 agricultural season she harvested a lot of sorghum, melons and sweet reeds most of which she sold for cash. The newfound trade is so profitable that she did not plant anything during the last agricultural season. Apart from a stillbirth, she has had two children one of whom lives with her.
Sheleng was born on 3 August 1978 at Princes Marina Hospital in Gaborone. After completing Form 5 he proceeded to the University of Botswana where he is a second year mechanical engineering student. This time last year he was living in Gaborone while five
181 years ago he was living in Lobatse where he was attending secondary school. Since vacation in May he has been working temporarily at a local shop as a Shop Assistant.
Six-pence was born thirteen calendar years ago on 13 September and is currently a Standard Seven pupil at the local primary school. She was born in this village and has always lived here although he sometimes visits his father in Gaborone whilst on school holidays. Apart from attending primary school, she has never undergone any training nor does he do any work.
After acquiring a plot through the Tribal Authority, the De-cuz built a three-roomed house and two huts. One of the huts is used as a kitchen while Ausi Smallpox who lives alone rents the other. The main house is made of cement blocks, corrugated iron roof and cement floor. The household fetch water from a communal standpipe and have a Ventilated Improved Pit latrine. The household uses gas for cooking, firewood for space heating and paraffin for lighting. They have a pit in which they throw rubbish although they sometimes incinerate waste materials. The household owns a working donkey cart and a bicycle.
Other details of the household are as follows: - no member of the household outside the country, no one disabled except Sheleng who has two missing teeth. De-cruz, who was the respondent, told the enumerator that his girlfriend in Gaborone died in March this year but no one else should know. The family owns no livestock except a few scruffy chickens and a dog. No family member is critically ill.
182 Appendix 9: Vehicles Deployment List
183 Appendix 10: Control List
184 Appendix 11: Prelisting Form
PF1
District: ………………………………………. Code
Village: ………………………………………. Code
Locality: ……………………………………… Code Page
Enumeration Area: …………………………… Code of pages:
No. Lolwapa Use of house: Name of head of Number of Comments Number 1 = Occupied Lolwapa persons 2 = Empty 3 = Constr. 4 = Shop 5 = Others 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 … 25 26 Total this page …………
185 No. Lolwapa Use of house: Name of head of Number of Comments Number 1 = Occupied Lolwapa persons 2 = Empty 3 = Constr. 4 = Shop 5 = Others 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
Total this page………….
Mapper’s Name: Signature: Date:
Supervisor : Signature: Date:
186 Appendix 12: Enumeration Check Control
187 Appendix 13: Household Questionnaire
(See Back-page pouch)
188 Appendix 14: Revisit Record
189 Appendix 15: Institutional Questionnaire
(See Back-page pouch)
190 Appendix 16: Document Control Form (Main)
191 192 Appendix 17: School Essays Competition
ECONOMIC IMPLICATIONS OF POPULATION AND HOUSING CENSUS IN A DEVELOPING COUNTRY SUCH AS BOTSWANA
By Donald Molosiwa (Secondary School Position 1) – Molalatau CJSS
Population refers to the total number of people living in a particular country. Census refers to the official counting of people.
Population and housing census are conducted after every ten years in Botswana, but one may ask, why spend a lot of money and even time on such an expensive and laborious exercise, so many answers may be thrown to people.
Every developing country and developed countries as well conduct the population and housing census so as to use the statistics obtained for development planning. The statistics obtained will be used for guiding the Government on how much money, social amenities and services should be provided to a certain area, and as a result helps to avoid the problems of over budgeting or under budgeting.
The statistics also help the Government to plan for the future, for instance, if there are 5000 (five thousand) two year old children throughout Botswana, the Government will plan on how it will cater for them in the next five years by providing schools, health facilities and social services accurately.
Some donor agencies and programmes from other countries willing to help the developing countries will need population statistics so that they can provide necessary help based on the number of people who need help. People who are affected by poverty, illiteracy and housing problems will be able to get help if their number is known, suppose there are 5000 people affected by poverty, the Government will then calculate how much is needed basing on their number.
Information on housing will help planners with a good picture of the housing situation, and this will enable the Government to take the appropriate action. Unemployment and other issues, statistics will be taken in to consideration, therefore the Government may be able to offer help through some programmes and policies which will improve the standard of living of people. Investors from abroad will need the information on population statistics of a country so that they can establish their production enterprises, bearing in mind the population which would create a market for their produce.
193 The census statistics also help the Government to decide on the methods to use in order to achieve sustainable development, as the more people the country has, the more the effect they have on the environment. The Government may conduct research on how to manage the natural resources. The statistics will also help us as citizens, because it will provide a picture on whether the population is increasing or decreasing so as to monitor it.
194
PALO YA BATHO LE MATLO
Otsile Madu (Primary Position 1) – Kazungula Primary School
Palo ya batho le matlo go tewa go balwa ga batho le matlo mo eleng gore go tla sologela setshaba molemo, ka go tla dirwa ditlhabololo dingwe jaaka go aga dikolo tse di lekaneng.
Palo ya batho le matlo e simolotse ka ngwaga wa 1904, lefatshe la Botswana le ise le tsee boipuso. Ka dingwaga tseo lefatshe la rona le ne le bidiwa Bechuanaland Protectorate le le ka fa tlase ga tshiretso ya ga Mma-Mosadinyana, le okametswe ke lefatshe la Boritani. Mme palo ya batho le matlo ya tswelela ka go diragadiwa mo dingwageng dingwe le dingwe tse di lesome go fitlhela nako eno.
Palo ya batho le matlo e thusa ka go lemotsha setshaba gore a palo ya sone e a gola kgotsa e wela tlase kana e eme golo gole gongwe, le go thusa go bona gore a batho ba fudugetse kwa ntle ga lefatshe kgotsa ba tsene mo teng. E bile gape puso e ka kgona go bona gore a itsholelo ya lefatshe e gatetse pele. Mosola wa palo ya batho le matlo ke gore maduo a yone a tla tlhokafala mo go direng ditogamano tsa ditlhabololo tsa matshelo a batho le itsholelo ya bone.
Palo ya batho le matlo e mosola thata ka gore fa e sena go dirwa, maduo a yone a a kgobokanngwa mme morago go lebelelwe gore lefatshe le tlhoka eng go tlhabolola batho le itsholelo ya lone. Mme fa go dirwa jaana Goromente wa Botswana a tsenye letsogo ka go thusa ka tsa kago jaaka dikolo, dipatela le dikolelo le go rutuntsha badiri ba ba lekaneng. Maduo a yone a tlhokafala mo ditlhabololong tsa madirelo, thuto, ditlhopho, boitekanelo, temo thuo, matlo le tse dingwe.
Kgothatso yame ke gore mongwe le mongwe o tshwanetse go tlhomamisa gore oa balwa a dirisanye le babadi ba batho le matlo sentle ka gore ba dira tiro ee mosola thata mo go tlhabololeng lefatshe le itsholelo ya lone ka go diragatsa tse di tlhokegang ka nako eo, le ee tlang.
195 Appendix 18: Census Budget 2001/2002
Item # Description Expenditure (31/03/02) 63 International Comparison Project 23,061 66 Purchase and Installation of Computers 175,700 67 Consultancies 122,038 70 Cartographic Equipment 1,433 73 Office Furniture and Equipment 126,669 75 Camping and Field Equipment 100,976 76 Personal Emoluments 2,580,794 77 Allowances 19,729,314 78 Transport Cost 4,890,062 80 Training Facilities 2,367,892 82 Data Processing Consultancy 500,000 85 Telephone Charges 60,000 86 Census Film 426,102 88 Advertising and Publicity 1,430,581 Total 32,534,622
196 Appendix 19: Cost of 2001 Census Printed Jobs
Job Description: 27 500 Advocacy for the Census booklets printed in full colour.
Size: A5 (32 pages)
Cost/Price: P 36 200.00
Job Description: 25 000 Guide to the 2001 Population and Housing Census books printed in full colour.
Size: A4 (40 pages)
Cost/Price: P 38 000.00
Job Description: 18 000 Population and Housing Census Forms in green ink.
Size: A4
Cost/Price: P 8 100.00
Job Description: 1 000 Front Cover Poster in green ink.
Size: A1
Cost/Price: P 8 160.00
Job Description: 1 000 Back Cover Posters in green ink.
Size: A1
Cost/Price: P 15 660.00
Job Description: 6 000 Institutional Questionnaire books in green ink.
Size: A2
Cost/Price: P 65 200.00
197
Job Description: 100 000 Household Questionnaire loose sheets in green ink.
Size: A2
Cost/Price: P 23 206.00
Job Description: 3 000 Household Questionnaire Posters in green ink.
Size: A1
Cost/Price: P 9 040.00
Job Description: 1 000 Institutional Questionnaire Posters in green ink.
Size: A1
Cost/Price: P 8 160.00
Job Description: 300 District Census Officers’ Manual in black ink.
Size: A4 (24 pages)
Cost/Price: P 7 920.00
Job Description: 7 000 Events Calendar printed black text with cover in full colour.
Size: A4 (50 pages)
Cost/Price: P 23 200.00
Job Description: Enumerators’ Manual printed black text with cover in full colour.
Size: A4 (132 pages)
Cost/Price: P 110 500.00
198
Job Description: 1 500 Supervisors’ Manual printed in black ink.
Size: A4 (21 pages)
Cost/Price: P 12 900.00
Job Description: 6 800 Locality and Country Coding List books printed in full colour.
Size: A4 (300 pages)
Cost/Price: P 74 900.00
Job Description: 30 000 Household Questionnaire books printed in green ink.
Size: A2 (25 pages)
Cost/Price: P 397 000.00
Job Description: 235 000 ‘Be there to be counted’ brochure printed in full colour.
Size: A4 folded
Cost/Price: P 18 620.00
Job Description: 235 000 ‘Nna teng gore o balwe’ brochure printed in full colour.
Size: A4 folded
Cost/Price: P 18 620.00
Job Description: 60 000 Primary School Lesson Notes leaflets printed in full colour.
Size: 205 x 105
Cost/Price: P 13 340.00
Job Description: 80 000 Secondary School Lesson Notes leaflets printed full colour.
199
Size: 205 X 105
Cost/Price: P 17 010.00
THE TOTAL COST FOR ALL JOBS: P 904 236.00.
200