Beyond Data Collection: Strategy for Assisting
Countries to Focus on Relevant Population Analysis

by:

Patience W. Stephens

Population Division

New York

1995

THE CONTEXT

Following the International Conference on Population and Development held in Cairo in September 1994, increased attention is now being drawn to the need to make demographic data and analysis relevant to development. Paragraph 12.2(a) of the Programme of Action recommends the establishment of "a factual basis for understanding and anticipating the interrelationship of population and socio-economic ... variables and for improving programme development, implementation monitoring and evaluation" as an objective of data collection and analysis. An important aspect of the Cairo Programme of Action is its clarity in expressing population goals within the context of overall development. All recommended actions are discussed within the context of the promotion of economic growth and sustainable development. The principles of the Programme of Action also distinctly outline intermediate targets in the promotion of development. These principles include:

a)
the improvement of standard of living
b)
gender sensitivity
c)
protection of the environment
d)
poverty eradication
e)
health promotion
f)
strengthening of the environment
g)
education
h)
protection of children and
i)
the proper treatment of migrants, refugees and indigenous people.

The principles reflect the primary objectives of national governments and they are the objectives most often stated in national development plans. To effectively address the multiplicity of issues that are represented under each of the 9 areas listed, analysis needs to produce not just estimates of the sizes and distribution and other statistical parameters of the subject population. It must also result in an understanding of the methods of analyzing data and producing results that are relevant to these broad areas of activity.

Earlier in this workshop, a number of presentations by UNFPA highlighted new emphasis on specific areas of activity in a manner that is consistent with a shift from routine demographic data collection to data collection that is programme-relevant. Specifically the utility of continued funding of routine census data collection in countries where this effort may already be self-sustaining was questioned. The need for integrated data bases that guide development planning has also been raised. There has even been a call for shifts in skills from those of demography to those of epidemiology in order to be more responsive to current research interests in major development concerns such as the global burden of disease.

What do these discussions imply for data collection and analysis? Do we need to throw our demographic skills out of the window and don new hats in order to meet "state of the art" demands or are there adaptations within the context of demography that can enable us to respond readily to the analytical requirements of development, whether they focus on family planning, reproductive health, international migration, education of women or infant mortality? How can the current technical support arrangements established by the UNFPA be strengthened to ensure better responsiveness of demographic analysis at the national level to country development needs?

CHANGING RELEVANCE OF APPROACHES

Proposed shifts in emphasis in the data collection effort are by no means indictments of the efforts of past decades. Past emphasis on largely demographic objectives in data collection played an important role in providing much-needed baseline information on individual countries. Many of these countries rose from levels of relative demographic data poverty in the 1950s and 1960s, to points where they can now provide reliable estimates and some recent trends in basic demographic indices. Many now have a cadre of well-trained demographers and the technological capabilities of statistical offices have been enhanced through the supply of computers and other equipment. For some countries there is now sufficient local capacity to plan and manage, without international technical assistance, the routine data collection efforts. The improvement in the availability and accuracy of demographic data over the past decades has also clearly been aided by advances in the development and application of methods of data collection, a number of which have been discussed in this workshop.

Although the past efforts have resulted in the generation of baseline data as input into planning at the national level, demographic data can hardly be said to have been integrated into development planning. Yet a number of technical assistance efforts have focussed specifically on promoting this integration of demography into broader intersectoral planning. At best, however, demographic data have been considered in the allocation of national resources. At worst -- and unfortunately so in many countries-- demographic data are ignored in planning, even for closely related sectors such as education, health and provision of basic infrastructure.

FACTORS ACCOUNTING FOR CURRENT APPROACHES TO DATA
COLLECTION AND ANALYSIS

Ironically, the persistence of the current approach to data collection and its processing in developing countries is a legacy of the intensive efforts of the past to encourage national governments to collect basic demographic data. Assistance to governments in the 1960s focussed sharply an estimation of vital rates through well-defined and well-taught indirect and direct estimation techniques. In the 1960s and 70s key demographic estimation manuals were published, software developed and training provided to students of demography and personnel of statistics offices in the application of these methods. To date there is no demographic manual (with associated software) that surpasses the United Nations Manuals 4 and 10 in their popularity with analysts in developing countries. This fact is perhaps a reflection of the most important message successfully transferred to demographers in developing countries, and thus represents local perceptions of "state of the art" approach to population data. Our success in communicating the methods of cross-sectoral research analysis have been much more limited.

A further reason for the neglect of analysis is the isolation of much data collection exercises from analysis efforts. In this regard, censuses and other internationally organized survey efforts are particularly culpable. The two methods of data collection often disregard the traditional instruction given to the student of introductory the research methods that field inquisitions should be developed only after clear analytical plans are defined. As a result, census enumeration in many countries has become an end in itself and volumes of data that could be used to generate policy-relevant analysis are never processed or tabulated because there were never plans at the national level to analyze them. Similar problems relate to large international survey efforts such as the World Fertility Surveys and the Demographic and Health Surveys, even though their focus on the collection of data on specific areas -- eg. fertility or health -- makes them more amenable to ex poste analytical designs.

It is important to note that even as volumes of yet unanalysed data accumulate in the archives of developing countries, at the international level, state of the art research and analysis of these same data virtually flourishes.

OPPORTUNITIES FOR REDIRECTING DEMOGRAPHIC ANALYSIS
USING TSS SYSTEM

In order to be responsive to the broader demands of development, countries need to be assisted to become more conversant with population research analysis which can be tailored to the investigation of specific issues of national concern. Given that a substantial part of demographic enquiry in most developing countries is driven by resources from the international donor community, it will take the redirection of efforts of the donor community to help countries to move beyond routine data collection and estimation, and to equip them with the expertise and tools to consider analysis issues that are more relevant to development. This is a role which the TSS system can play effectively, especially as CST advisors are at the helm of national efforts in data collection and analysis. It is proposed that the TSS system be used to foster the transfer of methodologies for population research analysis to countries. Advice to countries in areas of data collection, as well as analysis, should seek to reduce (though not eliminate) the emphasis on demographic estimation and in its place provide guidance in the collection and analysis of socio-demographic data.

There are several requirements for effective transfer and absorption of these new directions in demographic data use. First, there is a need to review and synthesize the most useful methodologies in population research analysis. Second, is a need to bring TSS advisors and specialists up-to-date on "state of the art" methodological approaches and issues and to discuss regional peculiarities that need to be considered in applying these methods. It is proposed that in order to achieve this objective a workshop be held to bring together TSS advisors and specialists in data collection and analysis, as well as in areas such as family planning, IEC, gender and development to discuss these methodologies, especially their application in specific country situations. An objective of the workshop will also be to design a programme for providing assistance at the country levels in the application of these methods of analysis.

SOME APPLICATIONS OF POPULATION RESEARCH ANALYSIS

Population research analysis involves the application of a wide array of research techniques in the investigation of the relationship of population and non-population issues. This array ranges from simple multi-way tabulation of data to investigate crude relationships among variables, to multivariate regression analysis of various forms. Depending on the nature of available data, various forms of survival and event history analysis may be involved. The range in the possible approaches to population analysis reflects the rather eclectic approach that is required in demographic analysis. Unlike demographic estimation for which there are more often than not, very specific methodologies for specific situations, in population research analysis the choice of which method to use may depend on researcher preference or competence, the nature of the dependent variable, the distribution of outcomes or even the subject matter being investigated.

Given the range of approaches to population research that exist, a synthesis will not be attempted here. Instead the reader's attention is drawn to valuable compilations of approaches presented in the volumes of the UNFPA (1993) publication - Readings in Population Research Methodology. The selection of articles in these volumes appropriately reflect the diversity in approaches that can be taken to population research. The methodologies covered in the volumes (especially volumes 6, 7 and 8) are an invaluable set of tools for development-relevant population research.

Three recent publications are used here to illustrate the possible applications of population research that can provide important information in the planning process.

CASE 1: Relationship between a Literacy Programme and Child Mortality and Health.

Sandiford, Cassel, Montenegro and Sanchez (1995) have recently contributed to an understanding of one of the factors that may account for lower observed child mortality among educated mothers. The study, which uses data on Nicaragua, established that the mortality and malnutrition risk of the children of mothers who had successfully participated in a national literacy programme were significantly lower than the risk among illiterates. The study also established specifically that there were sharp reductions, following the literacy campaign, in child mortality among the newly literate women but not among other groups of women.

The study uses both traditional demographic estimation and multivariate research techniques and thus demonstrates the potential to apply skills already acquired in countries with newer ones.

CASE 2: Implication of Changing Household structures on Family Welfare

A study by Lloyd and Gage Brandon (1993) in Ghana provides the policy maker with an appreciation of the nature and changes in family structure overtime, while analyzing the implications of the nature of the household on expenditure. This demonstrates the important role played by women in supporting the functioning of the household. The authors note, for example that in the roughly 30 percent of household headed by women the head typically contributes the majority of "market working hours" to the household. Perhaps the most important inference which the authors draw is that contrary to popular thinking, "the increasing proportion of households reported as female-headed in Ghana does not indicate a growing concentration of poverty among women".

This study used data from the Living Standards Measurement Survey, a rich multisectoral data source that exists for over a dozen developing countries but which has hardly been explored by country-level researchers. As with the previously cited work on literacy, the primary analytic tool used in the analysis was multivariate regression.

CASE 3. The Relative roles of Economic Development and Family Planning Programs in Reducing Fertility

In this article Gertler and Molyneaux (1994) examine the contributions of family planning programs, economic development and women's status to Indonesian fertility decline from 1982 to 1987 by combining demographic and economic frameworks to produce a unified framework. This article, which demonstrates the use of the combined demographic and economic frameworks in programme evaluation examines the factors accounting for Indonesia's fertility decline between 1970 and 1987. The study demonstrated that Indonesia's fertility decline resulted from increases in contraceptive use. Improvements in wages and in female educational attainment acted through contraceptive use to generate 45 to 60 percent of the decline in fertility. A key conclusion and particularly relevant policy lesson from this study is that 87 percent of the increase in contraceptive use was attributed to an intermediate factor ie. the change in education and wages.

This study uses data from the 1987 Indonesia Contraceptive Prevalence Survey, whose design is comparable to other fertility and demographic and health surveys conducted in developing countries in the 1970s and 1980s. The study's methodology is a good example of the eclectic approach to analysis as mentioned earlier. In addition to multivariate regression, it applied Bongaarts framework for decomposing the proximate determinants of fertility and reconciles apparently conflicting findings from different approaches.

A number of other studies focus on similar approaches to address development issues from a demographic perspective. The above examples, however, give an idea of what is possible even with already existing national data. With the emergence of new areas of emphasis such as reproductive health there will, no doubt be a need for some new data. However, whether the data are collected at the national level or community level, and whether they relate to reproductive health or other emergent areas, the population research approach will remain relevant.

References

Gertler, P. J. 1994. "How Economic Development and Family Planning Programs Combined to Reduce Indonesian Fertility". Demography (31) 1, 33-63

Lloyd, C. B. and Gage-Brandon A. J. 1993. Women's Role in Maintaining Households: Family Welfare and Sexual Inequality in Ghana". Population Studies (47) 1, 115-132

Sandiford, P., Cassel, J., Montenegro, M. and Sanchez, G. (1995) The Impact of Women's Literacy on Child Health and its Interaction with Access to Health Services. Population Studies (49) 1 5-18.

United Nations. Department for Economic and Social Information and Policy Analysis. 1995. "Population and Development. Volume 1. Programme of Action adopted at the International Conference on Population and Development, Cairo 5-13 September 1994." (United Nations publication, Sales No. ST/ESA/SER.A/149)

United Nations. Department of International Economic and Social Affairs. 1983. Manual X: "Indirect Techniques for Demographic Estimation". (United Nations publication, Sales No. 82.XIII.2)

United Nations. Department of International Economic and Social Affairs. 1967. Manual IV: "Methods of Estimating Basic Demographic Measures from Incomplete Data". (United Nations publication, Sales No. 67.XIII.2)

United Nations Population Fund. 1993. Readings in Population Research Methodology. (6 Volumes) UNFPA. New York