Background of the African Census Analysis Project (ACAP)

Introduction
Brief History of Census-Taking in Africa
Availability of the Census
African Census Analysis Project
The Future of ACAP

<< Availability of the Census

African Census Analysis Project

In recognition of the need to preserve African census data in order to avoid perpetual loss due to poor storage and also the need to encourage and enhance further analysis, dissemination, and utilization of the massive census data, the African Census Analysis Project (ACAP) was undertaken as a joint initiative of the Population Studies Center, University of Pennsylvania and African research and governmental institutions. This was to allow for collaboration with various African governments and research institutions at archiving and analyzing African census data, both at national and sub-national levels, in order to inform appropriate policy interventions on the continent.

At the onset , ACAP pursued three main objectives:

• Development of the Pan-African Census Explorer (PACE) for the purpose of archiving and analyzing African census micro-data .
• Demographic capacity-strengthening in Africa.
• Collaboration with African researchers.

ACCOMPLISHMENTS PHASE I

PACE development and data recovery efforts: Alongside the development of the Pan-African Census Explorer (PACE), one of the key accomplishments and attractions of the initial phase of ACAP is the creation of a unique data bank that helped avoid the destruction of many of the 1970 and 1980 rounds of data. We continue to be the repository of data for several countries as they experience the “disappearance” of their national censuses. We usually proceed by trying to read the data tapes in the countries and if not possible, they are brought over to the United States for further attempts. The recovery and transfer of past censuses constitutes an important accomplishment of the first phase as most of these censuses had been classified as irretrievable and unusable. ACAP has been quite successful in its role of negotiating and recovering census micro-data from various African nations. These data are then processed (i.e., made machine-readable, cleaned, and harmonized) for integration into PACE.

The difficulty of this recovery process usually starts with attempts to locate the census tapes on which the past censuses are stored. Based on our observation and past experience, it is often the case that once a new census is conducted the preceding census is relegated to the background. It not uncommon to arrive in a country to find that no one even recalls where the tapes containing the previous censuses were stored let alone talk of reading them. However, our multiple trips across the continent and untiring efforts to collaborate with the censuses offices in the countries have been quite rewarding. We managed to locate most of the old tapes of past censuses and have been particularly successful in recovering the data stored on such tapes and transferring them to modern computer storage media. Examples of recent successful recovery efforts include the Ghana 1984 census, Benin 1979 census, Cameroon 1976 census, and Malawi 1977 census .

For an idea of the challenges, time, and resources involved in this recovery exercise, it is worthwhile to describe t he most recent case of successful recovery of the 1977 Malawi census. A total of fourteen 9-track magnetic tapes that held this census (originally created on ICL 1900 series mainframe computer) have been in our keeping for over five years. During this period the tapes have made several rounds to commercial data recovery firms with little success at recovering and transferring the data to modern media. Most of the data recovery firms declared these tapes in bad shape and we had little hope of ever reading these census data. For instance, one of the data recovery services upon receipt of the tapes sent us this message along with the quote for services:

…The tapes are in pretty bad shape and will require treatment for striction to loosen up emulsion …and cleaning so they can be read by our drives….

But after treatment and diagnosis, our little hopes were dashed when we received the following from same firm:

I have researched the job sent to us from the U. of Pennsylvania to copy tapes. The tapes were in bad shape and could not be copied. They are returned to the U. and invoiced for….

Following similar results from other attempts we were reluctant to continue spending on tapes with little hope of success. But somehow we did not write this off and one of the eventual attempts was fruitful when the tapes’ contents were converted into sets of bitstream files which were subsequently converted to ASCII characters. However, we were not able to confirm that these were the Malawi data. The information in these ASCII characters did not match the census data reports, and they suggested that compressed data was embedded on the tape. In effect, some sort of memory-saving compression routine had been used to store the data, the details of which had been lost.² As part of the continuous efforts to recover these data, our systems specialist had concurrently posted on the internet a request for information on the ICL 1900 series machines. In the course of numerous exchanges with various experts about these machines, we identified a British programmer who had spent his entire career in application and systems programming on these machines. With his assistance, in November 2004 a program was finally written that successfully produced data that matched what was expected given the published reports from the original analysis of the data at the time of the census in the 1970s.

As a result of such relentless efforts and continuous resourcefulness we have made tremendous progress in helping avoid the destruction of many of the 1970 and 1980 rounds of African censuses. The expanding PACE collection currently consists of 55 census micro-data sets from 26 African nations . ACAP has the largest collection of complete African census data in the world (see Table 3).

Research capacity-strengthening: ACAP has promoted the use of census micro-data for a better understanding of African health and society. ACAP has helped to strengthen the research capacity of over 100 African researchers in several ways: granting support to students for training in Africa; collaborating with and supporting selected population- and health-oriented scholars working in Africa; bringing over 50 African students and researchers to the University of Pennsylvania for Masters and PhD degree training, or postdoctoral research; organizing over ten workshops for scholars using census data in their research; and co-organizing or regularly participating in seminars, conferences, and census-related training workshops or activities in Africa.

A major part of the training support covered graduate studies in demography at the University of Pennsylvania where most students are encouraged to work on the archived census data as part of ACAP research activities. Under this training support program, ACAP has supported the education of over 20 PhD and Masters graduates during the ten years of its existence. The most recent PhD graduates who have completed their studies and successfully defended their dissertations over the last five years are from Ghana, Malawi, Sierra Leone, South Africa, Tanzania, and Uganda. ACAP has also hosted about ten postdoctoral fellows from Ghana, Sierra Leone, South Africa, Nigeria, Tanzania, and Zimbabwe who worked on various aspects of African demography as part of ACAP research activities using the archived census data.

Table 3: African Census Data Collection (ACAP Archive) to date

* Former South Africa Homelands

The recently completed PhD dissertations based on the archived census data at Penn include:

• Differentials in infant and child mortality in East Africa (Kenya, Tanzania, Uganda)
• Health, well-being, and mortality in Africa ( Botswana, Lesotho, Zambia)
• Mortality in twentieth-century Malawi
• The demography of the migrant population in South Africa
• Grandparents and the mortality of their grandchildren in the Gambia: Findings from the 1993 census
• Mortality in the era of HIV/AIDS in Tanzania
• Poverty in post-apartheid South Africa: Measurement, trends, and the demography of the poor

Upon completion of their studies most of these students returned to their home countries and are currently working with their respective institutions or universities in Africa. They also retain the status of ACAP research collaborators for Africa. The return of most of these students is a demonstration of ACAP’s commitment to demographic capacity-strengthening in Africa. Besides these ACAP-supported African students who recently completed their doctoral studies, ACAP is currently supporting three African students from Kenya, Ghana, and Cameroon to pursue graduate/doctoral studies in demography at the University of Pennsylvania, one of whom is currently at the dissertation stage. ACAP has also supported several other African students for short internship programs at the University of Pennsylvania. Some of the training support was provided to students based at African institutions.

During the last five years, ACAP has been hosting on average 8–10 visiting scholars per year, most of them from African countries. The countries represented by these visits include Benin, Burkina Faso, Cameroon, Guinea, Mozambique, Ghana, Kenya, South Africa, Swaziland, and Uganda. The estimated number of such visitors and the person-months duration of stay are presented in Table 4.

Table 4: African scholars hosted at ACAP from 1998-2005

Publications: One of the main aims of ACAP is to contribute toward making census data more useful and relevant to scholarly pursuits, policy formulation, and development planning in Africa. This is accomplished in part through promoting extensive research using census micro-data for a better understanding of African health and society. In line with this, we now have in progress a multivolume series with two edited volumes already published. Over the past 8 years our research results have been presented at various international conferences and seminars, and subsequently published in peer review journals. The topics covered by the recent journal publications using the ACAP archived data include but are not limited to (see Appendixes 1 and 2 for the exact titles):

• Racial classification and the modern census in South Africa
• Polygyny and fertility
• Marriage influences on individual behavior in urban Africa
• Assessing the reliability of census data
• Migrants and the labor market

• Influence of individual and household characteristics on schooling attrition

• Living standards of the elderly
• Parental survival, living arrangements, and school enrollment of children in the era of HIV/AIDS
• The effect of living standards on childhood mortality;
• Socioeconomic status and child mortality: analysis using housing and household characteristics from African census data
• Disability in general and disability among the children of migrants
• Impact of external causes on expectation of life
• Childbearing and schooling
• The HIV/AIDS epidemic, kin relations, living arrangements, and the elderly

Collaborative Efforts at Using Census Data at the Regional Level

In collaboration with various regional and national institutions in the countries, ACAP has initiated a number of research programs aimed at enhancing the use of census data in development planning. Among such efforts are four major ones worthy of mention: a collaborative project with the Southern African Development Community (SADC) Statistical Committee program on the further analysis of the 2000 round of censuses; a collaborative program of research (with the International Network of field sites with continuous Demographic Evaluation of Populations and Their Health in developing countries—INDEPTH) on population and health in Africa, the program to strengthen the information base for proactive development planning and research program on the situation of orphans in African Societies.

SADC Program on Analysis and Utilization of Census Data

The SADC countries, like most African countries, have each conducted at least two population censuses. However, data from these censuses have been grossly underutilized and have not adequately informed socioeconomic development in the SADC region. In a bid to address the need for policy-oriented demographic and spatial analysis in the region, ACAP spearheaded this project that was approved by the SADC Council of Ministers. The main aim of the Analysis and Utilization of Census Data Project was to facilitate strengthening the capacity of SADC member-countries in demographic, social, and spatial analysis in order to promote the understanding and utilization of census data by government policymakers and planners.

ACAP–INDEPTH Research Program on Population and Health in Africa

Recognizing that African censuses archived by ACAP and the Demographic Surveillance Systems (DSS) coordinated by the INDEPTH Network produce vast and complementary kinds of demographic data for Africa and offer an exciting potential to examine African population and health, a collaboration between the two institutions was consolidated.The main objective of this collaboration was to pool the resources and expertise of ACAP and INDEPTH to undertake high-quality joint research projects in order to inform demographic and health policy in Africa. The first step of the collaboration started with a few INDEPTH sites that had the capacity to contribute the basic data needed for investigating common research topics. Based on mutual understanding between INDEPTH and ACAP, Burkina Faso, Ghana, Mozambique, Senegal, South Africa, and Tanzania were selected for this initial phase and serve as demonstrative examples of the usefulness of the collaboration. The second phase of the collaboration is expected to involve the majority of the INDEPTH sites and ACAP collaborators.

Data Needs for Proactive Development Planning

The aim of this collaborative activity is to enhance the process of planning and the implementation of democratically based proactive development programs at the subnational level by making census data easily accessible and available at the district level for analysis. This activity is the practical demonstration of the usefulness of PACE in enhancing the capacity to manage, analyze and disseminate census data. The pilot phase of this program was planned for Uganda in 2004 in collaboration with the Uganda Bureau of Statistics (UBOS) with the hope of eventual replication elsewhere. Indeed, the District Planning Units (DPUs) are the focus of this program. With decentralization of the planning function, the DPUs need to guide subcounties on how to develop their subcounty plans of action. This guidance has to be supported by the relevant data. Hence the DPUs need the ability to provide information for planning disaggregated by sub-county and lower levels.

Situation of Orphans in African Societies: Case of the Sahel

This collaborative effort with the Université Laval, Quebec (which coordinates the activities of the demography network of the francophone universities agency ( Réseau Démographie— Agence Universitaire de la Francophonie RD-AUF)is not only intended as a way to expand the geographical scope and coverage of the project. It was elaborated to enable the francophone community in Africa to equally benefit from our activities. It is in this light that our joint research project to examine the Situation of Orphans in African Societies focuses essentially on the Sahel countries. Orphanhood is not only a common phenomenon in sub-Saharan Africa, but recent evidence suggests a dramatic increase. While this is quickly attributed to the AIDS pandemic coupled with the persistence of other numerous pandemics and weaknesses and crises of the health/social system, there is little understanding of the support mechanisms available to this vulnerable group and the likely effect on the orphans’ life chances. This project proposes to examine the impact of parental loss in Africa. By exploring the different sources of data available on orphans, the proposed research will focus on three aspects of the lives of children—education, economic insertion, and entry into adulthood—that are affected by the death of one or both parents. A number of West/Central Africa countries have been selected for the current phase of the program. This research will result in a volume for the General Demography of Africa series.

² As explained earlier in the paper, it would have been helpful if there had been some metadata indicating clearly how these data were stored or compressed.

The Future of ACAP >>