Symposium 2001/29

11 July 2001


                                                                                                           English only

Symposium on Global Review of 2000 Round of

Population and Housing Censuses: 

Mid-Decade Assessment and Future Prospects

Statistics Division

Department of Economic and Social Affairs

United Nations Secretariat

New York, 7-10 August 2001











Census managers’ meeting 19-23 March 2001


South Africa*


















19-23 MARCH 2001








Jointly organised by UNSD, the SADC Secretariat and Statistics South Africa.






RAF/00/P03: UNFPA/SADC Census Support Project






April 2001





A.    Introduction. 4

B.    Opening remarks. 4

C.    Assessing the impact of training workshops. 5

1.     Census Management Workshops. 5

2.     Census Questionnaire Design Workshop. 6

3.     Census Mapping and Cartography Workshops. 7

4.     Sample Survey Design Workshop. 8

5.     Census Data Processing Workshop. 8

D.    Evaluation of census activities. 10

E.     United Nations Symposium on Global Review of 2000 Round of Population and Housing Censuses:  Mid-Decade Assessment and Future Prospects. 11

F.     Country reports. 12

Zambia. 12

Botswana. 13

Namibia. 14

South Africa. 15

United Republic of Tanzania. 16

Zimbabwe. 17

Malawi 18

Seychelles. 19

Mauritius. 19

Swaziland. 20

Lesotho. 21

Mozambique. 22

G.    Meeting recommendations. 23

H.    Analysis of census data. 23

I.      Review of questionnaires. 24

J.     Way forward. 24

K.    Closing. 25


A.      Introduction

1.         The meeting of the Southern African Development Community (SADC) Census Managers was held in Pretoria, South Africa, from 19-23 March 2001. The objective of the meeting was to critically review and discuss a report, particularly the findings and recommendations, by a consultant who evaluated the United Nations Population Fund (UNFPA)/SADC Census Support Project. The report focused on the preparedness of SADC countries to undertake the 2000 round of censuses of population and housing. Eight of the SADC member states are undertaking censuses during 2000-2002. Other member states will undertake large demographic surveys during the period. All the SADC member states were represented at the meeting except Angola and the Democratic Republic of Congo. Also in attendance were two representatives of the United Nations Statistics Division (UNSD). The SADC Statistician and the Project Manager for the Regional Statistical Training Programme represented the SADC Secretariat. A list of participants and an abridged work programme are in Annex 1 and Annex 2, respectively.


B.      Opening remarks

2.         The Statistician General of Statistics South Africa (SGSSA), Mr. Pali Lehohla, officially opened the meeting in his capacity as focal point of Census and Population statistics in SADC. In his remarks he underscored the fact that SADC member states had benefited from the UNFPA-funded Census Support Project through training workshops and exchange of experiences among member states. He expressed satisfaction at the level of participation and enthusiasm demonstrated by participants during the workshops and meetings organized under the project. He highlighted the achievements of the project, which included the agreement by SADC countries on core census questions and the willingness among the member states to exchange documents on census undertakings.


3.         He stated that other collaborative efforts had been initiated as a direct result of the project. Notable among these is the Census Analysis Project spearheaded by the University of Pennsylvania, which will assist SADC member states to build capacity on census analysis. The SGSSA disclosed that concerned parties to the project had agreed in July 2000 to evaluate activities carried out under the UNFPA/SADC Census Support Project and accordingly sanctioned for a consultancy to undertake an evaluation.


4.         In conclusion he expressed his gratitude to the UNSD and UNFPA for executing and funding the project respectively and also requested the participants to discuss issues with transparency and objectivity during the meeting.


5.         The representative of SADC Secretariat, Mr. Elliott Odirile, in his remarks concurred with the sentiments of the SGSSA and added that the meeting needed to map out a way forward for the process initiated through the project. He emphasized that the process should be undertaken with the cooperation and collaboration of all SADC member states with Statistics South Africa as the focal point so that ownership of the project remains with SADC as a region and UNFPA and UNSD as cooperating partners.


6.         He urged the participants to critically review the consultant’s report and concluded by appealing for frankness and openness in sharing experiences.


7.         In his remarks, the representative of UNSD, Dr. J. P. Banda congratulated Mr. Pali Lehohla on his appointment as SGSSA on behalf of UNSD. He also thanked the SGSSA for agreeing to host the meeting at very short notice in the midst of a pilot census.


8.         Dr. Banda pointed out that while UNSD was not a funding agency, it supports population and housing census programmes through production and issuance of census methodological handbooks and provision of requisite technical assistance. In the execution of the UNFPA/SADC Census Support Project UNSD conducted a number of census workshops, which were found useful by member countries in the region. He also informed the meeting that the UNSD is planning to produce a special issue of the Demographic Year Book to focus on census-related issues as its contribution to the 2000 round of census activities.


9.         Dr. Banda also took the opportunity to announce the United Nations Symposium, which will take place in August 2001 in New York, on “Global Review of 2000 Round of Population and Housing Censuses: Mid-decade Assessment and Future Prospects”. In conclusion he requested the meeting to come up with a short statement reflecting on experiences and the context of the SADC region with regard to the six themes of the symposium.

C.      Assessing the impact of training workshops

10.       The consultant, Mrs. Celestina Kabalu, presented the first part of her report on evaluation of census activities in the SADC region within the context of the UNFPA-funded Census Support Project.  The terms of reference for the consultancy were as follows:


·                    To assess the impact of training on the planning and implementation of census programmes;

·                    To evaluate census activities in light of the laid-out census plans


She visited six countries, namely, Botswana, Namibia, South Africa, United Republic of Tanzania, Zambia and Zimbabwe. The workshops she evaluated were on Census Management, Census Questionnaire Design, Census Mapping and Cartography, Sample Survey Design and Census Data Processing. In collaboration with the SADC Secretariat a questionnaire was designed for data collection and evaluation. The consultant also interviewed some participants of the various training workshops during the visits to the countries.


1.         Census Management Workshops


11.       Two workshops were conducted on Census Management with the main objectives to enhance the capacity of national statistics/census offices in successfully conducting population and housing censuses and intercensal demographic surveys. The meeting was informed that during the time of the missions most countries were busy with census preparations and some of the participants who attended the workshops were in the field. Comments of the participants who were interviewed included the following:


§                     Census Management workshops were very useful in that some census managers found the exchange of country experiences  very beneficial and assisted some in avoiding problems they could otherwise not have anticipated during the preparations and conduct of the censuses in their countries;

§                     The United Nations Census Management handbook should be customized to suit the prevailing situation in the SADC region. It was argued that some concepts and definitions were unique to the SADC especially with regard to economic activity;

§                     More time should have been spent on discussing census budgeting;

§                     Inadequate time was allocated to the discussion of country experiences.




12.       In the discussions, which ensued, some participants indicated that the basis for budgeting was adequately covered. It was felt that examples used in the workshop should have been drawn from the SADC countries rather than from the United States. Questions were raised about the ideal census management structure. It was pointed out that countries might have different structures depending on their legal set up.  What is, however, important is to establish effective networking arrangements with various organizations and persons that are crucial to the success of the census.  It was suggested that since SADC was trying to harmonize its statistical series it may be necessary to modify the United Nations recommendations, which provide general guidelines, to suit the conditions obtaining in the region. It was observed that concepts of households, dwellings and so forth needed to be carefully reviewed within the regional context.


2.         Census Questionnaire Design Workshop


13.       The main objective of the workshop was to harmonize concepts and introduce common core census questions for the SADC region. The consultant reported that most countries in the region had included the agreed-upon core questions. Comments from participants of the above workshop included the following:


§                     Questions discussed at the workshop had already been decided at an earlier meeting of senior officers;

§                     Some participants felt that they should have been taught how to design a questionnaire.




14.       The idea of SADC adopting core questions was applauded. It was, however, pointed out that beyond the core questions countries should have the liberty to develop additional questions. A question was raised whether there were mechanisms in the offing to streamline and develop core questions for censuses and Poverty/ Living Conditions surveys in the region.


15.       Some participants expressed disappointment with SADC for not taking the initiative to urge countries in this regard as it was recommended in an earlier meeting held in Namibia. The SADC Statistics Committee should have approved such a recommendation.  The meeting noted that the way forward is for member states to implement agreed-upon recommendations without necessarily waiting for the intervention of the SADC Secretariat.


16.       It was also suggested that there was a need for a manual for guiding countries in collecting data using core questions and that conclusions and recommendations of SADC Statistics Committee meetings should be transmitted to the relevant substantive technical staff in the national statistical offices for implementation purposes. Web sites could be used for disseminating proceedings of such meetings.


17.       It was clarified in the meeting that participants were not taught how to design a questionnaire because countries in the region were adopting different data-capturing approaches. The emphasis was therefore on content.


3.         Census Mapping and Cartography Workshops


18.       Two census workshops were conducted whose objectives were to train participants in concepts of census cartography and Geographic Information Systems (GIS) and encourage participants from SADC member states to establish contact and share professional experience. The workshops were considered very useful in terms of fostering cooperation in the area of cartography in the SADC region. Countries appreciated the assistance from other member countries that have the requisite expertise.


19.       While countries visited by the consultant were at different levels in the use of GIS, they nonetheless expressed confidence that cartographic work would be completed in time. The consultant was assured by the countries that population maps would be ready when census results are disseminated. Some participants were happy with the workshop to the effect that they developed their own cartographic manuals based on the notes and experiences of other countries although GIS necessitated acquiring expensive equipment.




20.       One country indicated that despite the problems of equipment and fuel, mapping would be completed although the exercise started late. Other countries had increased teams of mappers in order to be on target. A protracted discussion ensued on the merits and demerits of aerial mapping. It was pointed out that there was a danger of basing enumeration area demarcations exclusively on aerial photographs. It was agreed that it was always important to ensure quality control in all forms of mapping and in some cases field visits are essential for verification purposes even if aerial photography is the mode of mapping.


21.       Countries were advised to take particular care when considering outsourcing for services because some companies may not be conversant with or specialize in census mapping. It is always advisable to ensure that local staff is trained in order to ensure that the project is sustained.


22.       The meeting noted that sharing of cartographic expertise in the region should be promoted further through the establishment of a fund to support such exchange of staff between countries.  It was also underscored that census mapping should be expeditiously done because it is a basis for creating master sample frames.


4.         Sample Survey Design Workshop


23.       The workshop covered basic principles of survey design, whose objective was to upgrade skills in the use of sampling techniques in censuses. Comments from the participants who attended the workshop included the following:


·                    The workshop was found to be a very good refresher course;

·                    Some of the software prescribed at the workshop should be made available to member states;

·                    The workshop, which covered laboratory exercises, was very practical and useful;

·                    More workshops on sampling should be conducted and if possible replicated in SADC member countries.




24.       Most participants felt that designing of sample surveys should be strengthened in national statistical offices. It was stated that lack of proper documentation was the main problem in national statistical offices and not lack of sampling statisticians. It was emphasized that the workshop should be replicated at the national level in member countries. The meeting was, however, informed that replication might be possible if those trained at the subregional level could conduct the training workshop at the national level. One country plans to replicate the entire workshop, which has been conducted under the SADC project. Suggestions on how the training could be replicated at the country level were invited from the participants.


5.         Census Data Processing Workshop


25.       The workshop introduced participants to new technologies for processing population census data. The workshop had two parts. One catered for data processors and the other introduced Census Managers to the fundamentals of data processing. IMPS (Integrated Microcomputer Processing System) was the software used during the training. It was pointed out that the software was good for editing and dissemination. The following included the comments from participants who took part in the workshop:


·                    Time allocated was too short;

·                    Other software packages, other than IMPS, should have been introduced at the workshop;

·                    There was a need to replicate the training at the national level in all the member states;

·                    For countries to get maximum benefit from using IMPS, a UNFPA expert with a background in IMPS should be based at the Country Support Team in Harare, Zimbabwe;

·                    Some participants felt that they did not have the necessary programming background to derive maximum benefit from the workshop.




26.       Although most countries were apprehensive or ambivalent about the use of scanning, owing to high costs and capacity vis-ą-vis subsequent use of the equipment, it was generally agreed that scanning was the way forward mainly because the process would facilitate the speedy release of census results. The meeting learned that for scanning to yield desired results meticulous planning should be undertaken and all the necessary precautions should be taken to follow the stipulations for its meaningful application. It was also stated that collaborative institutional arrangements, in the case of one country in the region, went a long way to maximize the successful application of scanning. The adoption of the scanning process should, however, be gradual, starting with small surveys and progressively building to large-scale surveys including censuses.




27.       The recommendations of the meeting after reviewing and discussing the findings of the evaluation of the workshops by the consultant include the following:


§                     There is a need for an adapted census manual for the SADC region to cater for the unique circumstances obtaining in the region;

§                     A mechanism should be put in place within the region to upgrade census budgeting skills;

§                     Regular budget monitoring should be a part and parcel of census activities;

§                     Communication within national statistical/census organizations should be improved to facilitate the sharing of information on SADC census activities and decisions;

§                     The workshop on sample survey design should be replicated in member states;

§                     Development of statistics in SADC should be proactive by implementing approved decisions of the SADC Statistics Committee to ensure ownership;

§                     Member states should adopt new technologies to enhance data-processing activities and in particular scanning should eventually be introduced in SADC after extensive testing with small data sets;

§                     National statistical offices should send qualified officers to SADC workshops to enable them to get maximum benefit from workshops and ensure that the officers train others at the country level.

D.      Evaluation of census activities

28.       The consultant indicated that all countries had prepared work plans but actual dates of implementing various census activities were not documented. This implies that for activities which had been postponed the revised dates could not be obtained. In most cases census activities did not start in time due mainly to lack of funds. It was observed that governments’ commitment to fund censuses was low compared to other activities such as elections.


29.       Generally donor funds pledged for censuses are released after government financial commitment is obtained and the lack of commitment is manifested in the postponement of censuses by a number of countries in the region. High turnover of staff is another constraint in some countries, which has adversely affected census preparations. The consultant was of the opinion that assistance to countries, which are facing financial constraints, should continue to be obtained from multilateral and bilateral donors.


30.       The consultant informed the meeting that census mapping would be complete, in most countries, before census enumeration.  Lack of adequate equipment was, however, a problem in some countries. It was reported that most countries had developed census structures. In one country the census programme was divided into sub-programmes, headed by different people. The various groups in the census structures generally held regular meetings to monitor progress being achieved and instituting corrective measures as and when necessary.




31.       From the discussions it was clear that one of the countries was facing serious financial and resource constraints to conduct the census by the year 2002.   Donors are not forthcoming to provide resources for the census and the competing demands in the country for funds seem to relegate the census to the bottom of priorities. It was suggested that there was a need to intensify lobbying at the highest possible levels particularly at the political level. More effort and time should be spent in working out ways and means of lobbying for donor support. Countries were advised to explore the possible assistance from UNFPA from country funds. There is a need to make a case for censuses as a source of management information.


32.       Countries were advised to be cautious in conducting censuses during an election year, because governments generally give priority to elections when it comes to funding. Innovative ways of seeking funds and resources to support censuses may include approaching the private sector and perhaps having face-to-face discussions with the PARIS 21 group who have set up a Trust fund for assisting developing countries’ advocacy in statistics. It was also pointed out that in soliciting for funds from donors, there was a need to specify the type of assistance required from donors. In view of the resource constraints some member states were facing, some participants wanted to know what role the SADC Secretariat could play in mobilization of resources for censuses for member countries. The representative of the SADC Secretariat indicated that in order for them to assist in lobbying for census resources, it was imperative for member countries to submit work plans and budgets and other requisite information. Management information should be provided to the SADC Secretariat on a regular basis so that they are kept abreast of the situation in each member country.  At the moment such information was not available at the SADC Secretariat.




33.       The following recommendations were made:


§                     Countries should be submitting relevant information to the SADC Secretariat at the end of June and December of each year, to enable the Secretariat to monitor progress of the 2000 Round of censuses of population and housing in the region. On the basis of such information the SADC Secretariat would explore ways of assisting needy countries;

§                     Non-traditional donors for censuses should be approached to support census programmes including the private-sector companies and some non-governmental organizations.

E.      United Nations Symposium on Global Review of 2000 Round of Population and Housing Censuses:  Mid-Decade Assessment and Future Prospects.

34.       The UNSD representative indicated that the main purpose of the symposium is to address the question: “When and under what circumstances do censuses succeed?” In so doing the symposium would identify issues and problems inherent in the current round of censuses and seek solutions that may guide census planning during the next round through six themes relating to:


·                    Strategies of involving stakeholders in census activities;

·                    Strategies for choosing among data collection methods as sources of demographic and social statistics;

·                    Adapting new technologies to census operations;

·                    Maintaining census-related activities during intercensal years;

·                    Identifying and resolving problems of census mapping; and

·                    Post-enumeration surveys: are they worth it or not?


Participants were requested to draft a short statement on issues and reactions to the above themes in the context of the SADC experience in undertaking censuses.  The meeting was informed that details, main papers and statements in relation to the symposium would be posted on the UNSD web site.  


F.      Country reports



35.       Zambia carried out its fourth comprehensive census in October 2000 since the attainment of the country’s political independence. The census night was 16 October 2000, and the census was conducted on a de facto and de jure basis. Previous censuses were carried out in 1969, 1980 and 1990.  The main objectives of the Zambian 2000 census of population and housing were to achieve a high population coverage; generate relevant data at the lowest level of geographical area of governance and planning, ensuring conformity with international standards of the United Nations and adherence to the SADC recommendations; and also ensuring the timely release of data.


36.       The census was originally planned to be conducted in August 2000.  Actual planning and budgeting of the census started in 1997. The meeting was informed that the approval of the census plan by Cabinet, which included the management structure, was obtained  late, by  March 2000.  The census management structures included policy and technical committees.  The estimated budget for the census was US$ 20 million. In carrying out the census the National Statistical Office established bonds with collaborating partners who contributed in various ways to the success of the census.


37.       Cartographic work started in September 1998. The mapping systems, used for statistical work, elections and local government, were harmonized. Sketch maps based on quick counts were used in areas where detailed maps had not yet been completed by the time of conducting the census. Publicity was mounted using various means such as radio, television, print media, billboards, T-shirts etc.


38.       Zambia captured its 2000 census data by scanning using OMR (optical mark readers). Scanning was used in order to produce timely census results. Existing facilities at the Examination Council of Zambia were used to capture the data. At the time of the meeting 90 per cent of the questionnaires had been scanned with a 5 per cent rate of rejection.  The PES was conducted in February 2001 and its results will be used to evaluate the census.




39.       Participants noted that it was important to establish effective communication during the early stages with government, policy makers and decision makers, including political and opinion leaders to facilitate approval and commitment to the census programme.  It was observed that although Zambia missed most critical dates of the census, the census was carried out in the final analysis. The meeting was informed that Zambia administered its questionnaires on a 100-per-cent basis.




·                    Resource mobilization was difficult owing to strict fiscal policies of government;

·                    Printing of questionnaires was delayed, owing to the delay in obtaining approval for tender application, necessitating the postponement of the census enumeration from August to October 2000;

·                    Pledges from donors and support were delayed until government commitment to the Census was known.




40.       Botswana is poised to conduct a census in August 2001. According to the report, preparatory work is on schedule. The Botswana census is mandated by the constitution; therefore, there is strong government commitment to the census programme exemplified by provision, at the right time, of adequate resources for census activities. Mapping and listing of housing units is expected to be completed by the end of March 2001. The growth point areas will be revisited in order to update the lists. Eviction and demolition areas will also be revisited. In May 2001 census officials will carry out a final country inspection of the quality of mapping and housing.


41.       A pilot census was conducted in August 2000 to test various operational aspects of the census. This was based on a purposive sample of 61 enumeration areas (EAs) out of 2,600 (1991).  The findings from the pilot were taken into account in the preparation of the census. For example, changes were made to the questionnaire, especially items on economic activity.


42.       Various organizational structures have been established in preparation of the census including district and census committees. The meeting was informed that the census office was established in 1999. Training of Technical Officers and District Census officer was conducted from 12-17 March 2001. Trainers will undergo training towards the end of March 2001. Supervisors will be trained in April 2001 and enumerators in August 2001. The meeting was informed that a PES will not be conducted, as adequate quality assurance mechanisms will be put in place and the mapping and housing listings will be used to check coverage.


Botswana has in place an elaborate publicity campaign scaled up from March and reaching the peak in July 2001. Publicity includes messages through school notes, packaging materials, census film, radio drama, billboards, booklets and other materials. Procurement of other census materials will start in April 2001 and end by early June 2001. In order to ensure that the census period is given maximum attention, other activities such as competitions, political rallies and by-elections have been planned to be conducted. The census office will mobilize about 600 vehicles to be used by field staff. The census is expected to cost about 40,000,000 Pula (about US$7.2 million).


43.       It was decided to outsource data processing in order to produce a data-entry system using IMPS. Unfortunately, the consultant absconded after completing the first phase of the project. However, necessary corrective measures have been put in place to address the problem.




44.       It was observed that Botswana had one of the highest per capita costs of the census in the SADC region. It was reported that the terrain in some parts of Botswana was so rough that enumerators will have standby vehicles, especially in desert areas of the Kalahari. This will entail that more vehicles will be required. It was also pointed out that the suspension of other activities was only during the enumeration period and organizers of such other activities were requested well in advance to align their fixtures/functions in advance.


45.       The meeting was informed that radios will be used to explain questions to the public. Member states were urged to take extra precautions when outsourcing census activities. There was a general feeling that the questionnaire was too big and would be cumbersome for enumerators to use in the field. It was taking about 30-45 minutes on average to enumerate one household. The meeting was assured that the questionnaire did not pose major problems during the pilot. The role of publicity by enumerators was also discussed.




·                    Outsourcing of data processing has created a bottleneck in data  processing.

·                    To ensure that there are no slippages in all the preparatory activities so that the census is conducted as scheduled and with minimal or no operational problems.




46.       The preparation of the census, on an ad hoc basis, started in 1999, although the Census project budget was only approved in February 2001.  The Director General of Planning is at the top of the census organization. The law empowers the National Planning Commission to collect statistics and conduct population and housing censuses.  A high-level committee of Permanent Secretaries of some line ministries has been formed. In addition technical committees of people from both the public and private sectors have been formed. Furthermore, publicity and evaluation and monitoring committees were formed. Plans are under way to form regional census committees.  The Census office was established at the beginning of 2001 and consists of a Census Manager assisted by five Assistant Census Managers to manage the day-to-day planning of census activities. Teachers will be used as supervisors and students as enumerators during the census enumeration scheduled for two weeks from 27 August 2001.  


47.       Cartographic work started towards the end of 1999 at a slow pace mainly due to transport problems. Plans are under way to increase mapping teams. The GIS was outsourced to a private company. The quality of work is, however, unsatisfactory. The training of mappers was not satisfactory and the work done in four regions was not of acceptable quality. The deadline for completing the work is the end of June 2001 and it is uncertain that the deadline will be met. A member of CTS–Harare assisted in training staff in cartographic work. The office has now a cartographer from the Central Statistical Office of Zambia assisting with cartographic work.


48.       The pilot census was completed although some EAs were big and took three weeks to complete. For the pilot census enumerators were recruited through regional offices. Some of the field staff were not suitable for the work.  Census messages will be on radio, television, print media and so forth.


49.       With regard to data processing the programmer is available and is familiar with IMPS. However, he may need some training in CSPro. A decision has not yet been taken regarding whether to use keying or scanning for census data capture. It is planned that dissemination of census results should be through both the hard and electronic copies. There is optimism that sufficient funds will be available for the census from the government and some donor agencies.




50.       In the discussions, which followed, it was emphasized that there was a need to take extra care in selecting contractors for outsourcing. There was a need to institute thorough investigations into the profiles of the companies to be contracted before engaging them and to ensure that the contracts explicitly state that any data or information the contractors come across or use during the time they are engaged remains the property of the statistical office.




·                    Problems of outsourcing;

·                    Mobilization of funds;

·                    Inadequate staff in relevant census areas such as cartography and data processing;

·                    Shortage of time between now and the census date; need to hasten preparatory work, particularly cartographic work; and

·                    Need to take a decision, as soon as possible, whether to capture data by keying or scanning .


            SOUTH AFRICA


51.       South Africa conducts censuses every five years. The last census was conducted in 1996 and the next one is scheduled to be conducted in October 2001. Planning for the forthcoming census started in October 1999 with cartographic work, which is still not completed. Statistics South Africa has shifted from sketch maps to aerial photography as a basis for delineating of EA boundaries. It is planned to complete census mapping by June 2001. The work has been outsourced to a private company but staff for Statistics South Africa is attached to the project.


52.       Before designing a questionnaire, nine user workshops were organized. Discussion of the questionnaire was held within the Census Advisory Committee, but the Census Cabinet Committee has not yet ratified the questionnaire. There is pressure to increase the number of questions, of which there were 78 in 1996. A pilot census was carried out in March 2001. The ratio of supervisors to enumerators will be increased from 1:10 to 1:5. An effective publicity strategy is being developed to ensure maximum cooperation from respondents who are usually uncooperative, particularly commercial farmers.


53.       South Africa conducted a pilot PES in March 2001 and it is also intended to carry out a PES after the census in order to evaluate the coverage of the census. In 1996 the undercount was estimated at 10 per cent. PES results will be used to adjust census figures. The dissemination strategy will focus mainly on electronic products. The official population count will be announced in October 2003. An elaborate system for capturing the data using scanning has been set up.




54.       Some participants wondered whether South Africa should not change to a decennial census programme. The current Statistics Act stipulates that Censuses should be conducted every five years. Discussions are under way to review this issue. Regarding outsourcing of various census activities, it was strongly felt that Statistics South Africa should own the process by managing the activities. The meeting was informed that the contracts were carefully drawn such that the statistics office was building capacities from outsourcing.




·                    Keeping pace with evolving census geography;

·                    Development of proficiency in usage of aerial maps;

·                    Capacity-building and retention of core staff for the next census;

·                    Increasing response rates in subgroup populations that traditionally do not readily cooperate during censuses.




55.       The United Republic of Tanzania is planning to conduct its next census of population and housing in 2002.  The last census was conducted in 1988. The census, which was supposed to have been conducted in 1998, was postponed to 1999 and later to 2002 because cartographic work had not been completed.  Initially donors were reluctant to pledge funds for the census because they perceived that there was little commitment to the census by the government. The latter had not earmarked adequate resources towards the census. The census budget is estimated at US$32 million. The census activities have now been revamped with the release of government funds and substantive pledges and contribution of resources by donors.  Cartographic work is in progress with 60 per cent of the country having already been mapped. However, some of these areas have to be revisited for updating.


56.       The SADC core questions will be included in the questionnaire. Discussions are under way in the United Republic of Tanzania to reduce the number of questions in the census questionnaire to fit the budget and to do away with the conduct of the PES. With regard to data capturing, a decision has not yet been taken regarding the method to use. In the meantime a mission was sent to Zambia to observe and discuss the scanning process. GIS may be adopted; however, a separate budget will be prepared and submitted to donors.




57.       Participants wanted to know whether the reduction in the number of questions affected the SADC core questions.  It was reported that the core questions were retained and that questions on gender will be eliminated because they demand probing. It was further reported that teachers would be employed as supervisors and enumerators. The United Republic of Tanzania will follow a cascading approach in training field staff.




·                    Quick decision on method of data capturing is essential since this has implications on questionnaire design and quality of paper; and

·                    Progress in mapping to be closely monitored in order to meet the deadline.




58.       Zimbabwe carried out its last census in 1992. The next census is scheduled for 2002.  The census activities are generally behind schedule by six months. The major constraints include delays in recruitment of staff, fuel crisis, budgetary constraints, inadequate equipment, inaccessibility of some areas owing to heavy rainfall and inadequate staff at headquarters to monitor field operations. There are plans to limit the questionnaire to 40 questions. Stakeholders will be involved through circulation of proposals to those on the Central Statistical Office mailing list, i.e., individuals and institutions both in public and private sectors. Nationally acceptable definitions should be used in the census.


59.       The meeting was informed that Zimbabwe will stick to the technologies already in the country that can be accessed with minimum need for foreign exchange. They will, however, introduce GPS (Global Positioning System) for mapping. The mapping exercise is behind schedule due to delays in purchasing equipment. They will not carry out the PES because they will put in place effective quality control measures at every stage of the census activity. The census budget is estimated at US$19 million.




60.       It was suggested that efforts should be intensified to mobilize resources for the census. Some participants felt that it was risky to carry out a census during an election year because the latter is usually given priority in funding.




·                    Acute resource constraints;

·                    Some preparatory activities are behind schedule, e.g., mapping.




61.       Malawi conducted its census in 1998. Data processing was completed in 2000. The census was delayed from 1997 to 1998 because of lack of funds. Mapping started in 1996. At that time GIS was not in the picture. Plans are under way to introduce GIS in the intercensal period. Censuses in Malawi are usually conducted during the month of September when schools are closed because the census uses teachers and students as supervisors and enumerators respectively. In 1998 the school calendar changed and therefore the census could not use the services of teachers and students. The census could not be postponed to November because it is the rainy season. High school leavers were instead used for enumeration.


62.       Problems were encountered during training due to the fact that funds for the purpose could not be released as the government was implementing a cash budget system where funds are released in accordance with the amount of funds collected through a revenue collection mechanism. Special consideration was requested from donors to pay for training. There were problems at the time of completion of the enumeration because enumerators could not release census materials before being paid. Census materials were finally collected in December 1998 after securing funds to pay the enumerators. There were problems in data processing; however, the US Bureau of the Census provided assistance in this area. Malawi did not conduct a PES but used available funds for data processing. Some stakeholders felt that the census result of 9.9 million people in Malawi was low and considered 11 million as a better estimate. The Census office in Malawi, however, stuck to its estimate of 9.9 million people arguing that fertility decline, outmigration and rising mortality justified the census figure.




63.       Some participants wanted to know whether Malawi had conducted a large-scale survey. It was reported that a Demographic and Health Survey was conducted in 2000. Some participants wanted to know the source of the estimate of 11 million people in Malawi. The meeting was informed that it was based on projections based on the mortality, fertility and migration assumptions. Coding posed a problem in the Malawi census. It was stated that they were advised not to do any editing although they continued with minimal editing. With respect to the issue of outmigration it was advised that it would have been prudent to check if there was a large influx of people from Malawi into a neighboring country during that year.




·                    There were serious financial constraints leading to the postponement of the census and delay in retrieving completed schedules from enumerators.

·                    Need to assess coverage.




64.       Seychelles carried out its last Population and Housing census in 1994. Its objectives included the preparation for the revision of the electoral district boundaries, validation of the National Population Database (NPD), provision of a primary source of population data for administration and planning and a frame for sample surveys.


65.       The updating of maps was done manually. The main topics included in the questionnaire covered housing, population, agriculture, employment, education and fertility. Seychelles is not conducting a population census during the 2000 round. Plans are, however, under way to carry out a housing census in August 2002. The population register provides most of the data generated from the population census. In this census new technology will be used in mapping. Aerial maps are available; therefore, cartographic work should take less time to complete.




66.       It was noted that Seychelles was one of the few countries in Africa that has a comprehensive population registration system.  The system is partly sustained by rewards and sanctions. The former include social security and unemployment benefits, while the latter is a requirement for registration records at burials, schools and health facilities. It was pointed out that this was a good example of a country that uses population registers in conjunction with census data.




·                    To ensure that the mapping exercise is completed in good time and maps updated.




67.       Mauritius conducted the last Housing and Population censuses in 2000. The former was conducted from 7 February to 18 June 2000 and the latter from 19 June to 16 July 2000. The Census night was the night of 2 July 2000. By September 2000, the preliminary results of the Housing Census were published and the final results were released in November 2000. Results of the Population Census will only be available in September 2001.


68.       The census budget was US$3 million. Stakeholders were requested to submit their requirements with regard to items to be included in the questionnaire. SADC core questions were included except for mortality. About 1,000 supervisors and 5,000 enumerators were engaged for the Housing and Population censuses respectively.  Only one district out of 10 was digitized. Considerable delays were experienced in the printing of census questionnaires undertaken by the Government Printer. A PES was not carried out because it was felt that the methodology used in the censuses allowed for checks of undercoverage and duplication.




69.       Some participants asked why there was a time lag between the housing and population censuses. It was argued that the matching of the two data series could pose problems. It was pointed out that the identification codes and migration would make the linkage possible. The mortality question was omitted from the questionnaire because there is a good vital registration system in Mauritius.




·                    Printing problems

·                    Staff shortage




70.       The last population census for Swaziland was carried out in April-May 1997. The census adopted a de facto method of enumeration. Census committees were set up to support the census. Students were enumerators while teachers were supervisors. Teachers trained enumerators. Due to a teachers’ strike in 1996 the census had to be postponed. The delay caused budgetary cost overruns because questionnaires had to be reprinted and contracts of temporary enumerators extended. Data processing was undertaken by the government computing centre. The programmer assigned to the census project developed the data entry, editing and tabulation programmes in-house using a computer programming language called NATURAL ADABAS. Preliminary results were released in November 1997 and the final results were released after two years.


71.       A PES was conducted in order to evaluate coverage and content errors of the census; however, the analysis of the data has not yet been undertaken because of lack of staff. Once analysed the results of the PES will, however, be used to adjust the population projections figures and not the actual census results. A limited number of staff can be employed at the National Statistical Office due to lack of adequate office space, especially in data processing. It was stated that there were financial constraints.


72.       There are plans to conduct a demographic sample survey in 2002.  Planning for this survey is still in a preliminary stage. However, a request for funding of about 2.5 million South African Rands has been forwarded to the government. They are currently drawing a project proposal for submission to UNFPA. Hardware and software have already been bought for the computerization of mapping.




73.       The plans to conduct a large-scale survey were welcome.  A concern was expressed that the PES had taken so long to analyse that it may not be useful due to lapse of time. It was, however, reported that the results will be used for projecting population figures and not for adjustment of the census results. A tentative date of 2007 has been set for the conduct of the next census in Swaziland. Other participants urged that the census dates in the SADC region be synchronized.




·                    Financial constraints;

·                    Shortage of office space;

·                    Lack of capacity in data processing.




74.       The last population census in Lesotho was conducted in 1996 at a total cost of M6 million, out of which M3 million was UNFPA contribution. Over and above this, UNFPA donated 10 vehicles and 7 computers. The major constraint of the census was that funds for data collection were released very late. As a consequence, the enumerators went on strike and this delayed the data-collection activity.


75.       The Demographic Survey should have been conducted in 1992 but was not because of lack of funds and the fact that the Bureau of Statistics (BOS) in Lesotho was still struggling to release the 1986 population census results. The preparations for the 2001 Demographic Survey started in 1999. The cost for this survey was estimated at about M684,000, out of which M410,000 was supposed to be UNFPA’s contribution. However, in 2000 UNFPA indicated to BOS that it was experiencing financial problems and cut down the contribution to M7,000, which was never released for survey activities. No other donor has been forthcoming. The 1996 population census covered 3,059 Enumeration areas, of which 130 have been selected for the 2001 Demographic survey.  The major constraint facing BOS is that it has no qualified cartographers. The Cartographer from CST, Harare, has only trained the officers engaged in cartographic activities. Another constraint is lack of transport.


76.       All SADC core questions have been included in the demographic survey questionnaire. The training of supervisors and enumerators for the demographic survey was scheduled for the week of 19-23 March 2001, but did not take place due to lack of funds. The survey is planned for 8 April 2001. However, with no funds readily available, the above activity may be postponed.




77.       The Lesotho situation was considered serious. Participants, therefore, advised that with immediate effect high-level government officials should hold discussions with the UNDP and UNFPA representatives in the country to request funds for the country allocation. With respect to cartographers the meeting felt that it was important to have career development for cartographers.  There is also a need to have qualified cartographers. Other participants had suggested that SADC countries should come to Lesotho’s aid.




·                    Resource constraints;

·                    Lack of vehicles;

·                    No qualified cartographers;

·                    Serious time constraint, as the Demographic and Health Survey is supposed to start on 6 April 2001.




78.       The last census was carried out in 1997.  The final report was released in April 1999. Census products included provincial reports, master sample frame, population projections and thematic maps using GIS. Mozambique has very extensive intercensal survey under the integrated system of household surveys. The surveys provide data on a broad range of topics including poverty, education, labour, maternal mortality, impact of floods, reproductive behavior and so forth. The generic composition of field staff is as follows: one team per province with each team having five enumerators, one supervisor, one cartographer and one driver.


79.       There is a scanner for data entry with three computers. There are two data entry clerks and one programmer. Data processing is done concurrently with data collection.  There is a fund of about US$1.5 million to cater for the surveys.  The work programme indicates that Mozambique is planning for a census in 2007. A pilot census will be conducted in 2007.




80.       Some participants wondered whether Mozambique was adhering to the stipulation of SADC for the need to conduct a large-scale demographic survey during the intercensal period.  It was pointed out that most of the surveys had a demographic component. Participants cautioned that by adapting variables in the Core Welfare Indicator Questionnaire (CWIQ), care should be taken to maintain consistency in the data during the analysis.  Considering the culture in this part of the region, was it easy to get responses to questions on sex experience? The meeting was informed that since it was the first time the survey was being conducted it was not easy to assess how easy it was to obtain good responses to questions on sexual experiences from the youth. Well-trained enumerators will be deployed for collecting data in this survey. In addition a pilot is planned for the survey.


81.       It was suggested that countries should share some of the expertise, such as in  cartography. For example, Mozambique has 12 cartographers who seem to be well qualified in the field. Some advised that there was a need to come up with a directory of specialists in the SADC region. 




·                    The need to sustain a high standard of surveys if and when donors have pulled out.        

G.      Meeting recommendations

82.       The meeting noted that SADC member states had prepared for the 2000 round of censuses. Out of the eight member states scheduled to undertake censuses between 2000 and 2002, only Botswana and South Africa seem not to have financial constraints. Namibia was still looking for funding for some activities. Zambia had completed the census. The meeting also noted that experiences of Botswana and Namibia of outsourcing some of the activities did not produce the desired results. Botswana and Namibia had outsourced data processing and cartographic activities, respectively, to the private sector. In view of the foregoing, the meeting made the following recommendations:


·                    Major stakeholders including government, political and opinion leaders should be involved at an early stage during the planning stages of censuses;


·                    Outsourcing of census activities to the private sector should be undertaken with great care to ensure good value for money;


·                    National Statistical Offices should be represented on the management of outsourced activities in order to ensure that, among other things, confidentiality of data is assured.


·                    A training workshop on Organization and Management of Statistical Offices should be conducted for SADC member states.


·                    SADC countries should work towards a common year for conducting the next round of censuses.


·                     Member states should avoid conducting censuses during an election year in order to avoid competing for limited financial resources during the same year, considering that both programmes demand huge sums of money.

H.      Analysis of census data

83.       This agenda item focused on the path beyond census data collection. The SADC Millennium project (Phase II) on analysis and evaluation was presented as a draft proposal.  Training various specialists on analytical skills will be the focus of the project but should also include research activities as well. The training will include planners and policy makers.


84.       The proposal to establish a work group for dissemination of census and household information in the SADC region is intended to facilitate dissemination of statistics. Countries will be expected to submit selected data on selected variables to Statistics South Africa, which will initially coordinate the project.


85.       The meeting agreed that more work is required to focus and refine the proposals. Statistics South Africa was requested to take into consideration various comments made at the meeting when revising the draft proposal. Member states were requested to send the names of contact persons to Statistics South Africa by 30 March 2001. The contact persons from member states should ensure that relevant comments are forwarded to Statistics South Africa.

I.       Review of questionnaires

86.       The meeting reviewed census questionnaires to determine the extent to which member countries included or will include the core questions in the questionnaires. The meeting observed that member countries included most of the core questions in their questionnaires except in a few cases where minor modifications had to be incorporated to accommodate specific situations. The table below indicates countries which did not include some of the core questions.






§                Current economic activity (Modified slightly)

§                Fertility- Date of birth of last birth


§                Usual economic activity

§                Place of residence one and five years ago

§                Mortality

§                Refuse collection

 South Africa

§                Usual economic activity

§                Type of materials for walls and roofs of housing units


§                Mortality- deaths in last 12 months

§                Parental survival


J.       Way forward

87.       In view of the progress achieved by the member states and taking note of the constraints encountered by some of the member states, the meeting proposed and agreed, inter alia, that follow-up activities to the project should include the following:


(a)                development of a web page to facilitate sharing of information on Censuses and other related activities among member states;

(b)               development of analytical capacity in Census data analysis and dissemination;

(c)                formulation of a strategy for determination of a common SADC Census year;

(d)               depositing of Census questionnaires for SADC member states with the SADC Secretariat and Statistics South Africa as focal point for Census and Population issues; and

(e)                dissemination of  information through CD-ROM

K.      Closing

88.       Mr. Motale Phirwa, Census Manager, Statistics South Africa officially closed the meeting.  He emphasized the need to ensure that the recommendations of the meeting are implemented in the SADC member states and apologized to the participants for any shortcomings experienced during their stay in Pretoria. He commended the consultant for her work in evaluating the preparedness of the SADC member countries to conduct the 2000 round of Censuses of Population and Housing.


89.       He concluded by paying tribute to UNSD and UNFPA for their contribution to the success of the meeting and to the SADC Secretariat for effectively facilitating and guiding the meeting.

















22 213 5602


22 213 5601

Head Social & Demographic Statistics Dept. -

Box 796 Dar-es-Salaam Tanzania






248 383 169


248 225 339

Director of Statistics

PO Box 206 Victoria Seychelles


Banda, Jeremiah P.


1212 963 8338


1212 963 1940

Statistician –

Two UN plaza 1007 New York


Buthali, Dabilani


267 588 500


267 588 610

Census Manager   P/Bag BO 335 Gaborone



Dudu V.


268 404 1538/7539


268 404 3300

Head of Demographic and Vital Stats - Box 456 Mbabane



Dzekedzeke, Kumbutso


260 125 25 75


260 125 25 75

Census Manager  Box 31908, Lusaka Zambia





267 351 863


267 372 848

Project Manager SADC

P/Bag 0095 Gaborone Botswana

Kabalu, Celestina L.C.

 260 1 233377

/096 760 372




Box  39196 Lusaka,  Zambia





264 61 283 4111


264-61-239 376

Assistant Census Manager – P/Bag 13356 Windhoek Namibia





27 12 310 8075


27 12 320 0719

Statistician General – P/BagX44 Pretoria 0001


Mapeta, Washington T.


263-4 -703 971


263-4-728 529

Census Manager-

Box CY342 Causeway






258 1 490 384

09 258 8 231 275


258 1 492 114

Av.Ahmed Sekou Toure 21 C.P. 493 Maputo Mozambique


Ndawala, Jameson


265 524 377


265 525130

Assistant Commissioner  -PO Box 333 Zomba Malawi







267 351 863



267 372 848



P/Bag 0095

Gaborone, Botswana


Phirwa,  Motale


27 12 310 8108


27 12 310 8201

Census Manager   P/BagX44 Pretoria 0001


Suet,  L. F Cheung Kai


230 212 2316


230 211 4150

Survey Statistician Lic Bldg John Kennedy St. P.L Mauritius


Suliman, Sirageldin


1212  963 4375




1212 963 1940

Inter country Advisor -

Two UN plaza NY


Tsietsi, Matsotang


266 323 852


266 310 177

Principal Statistician

PO. Box 398 Maseru Lesotho










19th March, 2001


Ž                Registration of Participants

Ž                Administrative Matters

Ž                Assessment of Impact of Training workshops:

§         Census Management

§         Census Questionnaire Design

§         Census Mapping

§         Household Survey Sample Design

§         Census Data Processing

§         Findings and Recommendations




20th March, 2001



Ž                Evaluation of Census Activities

Ž                United Nations Symposium on 2000 Censuses of Population and Housing

Ž                Country Reports




21st March, 2001



Ž                Country Reports (continued)



22nd March, 2001


Ž                Country Reports (continued)

§                     Findings and Recommendations

Ž                Analysis of Census Data




23rd March, 2001


Ž                Review of Country Questionnaires

Ž                The Way Forward

Ž                Closing



*       This document was reproduced without formal editing.

**     Southern African Development Community, Botswana. The views expressed in this report are those of the authors and do not imply the expression of any opinion on the part of the United Nations Secretariat.