Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
BEATRICE LORGE ROGERS AND NINA P. SCHLOSSMAN
The underlying goal of economic development is to improve the welfare of disadvantaged people. The individual is the ultimate target of development efforts, be they undertaken at the household, community, or national level. A successful programme or policy improves human welfare. It benefits the intended target group and achieves the desired result without any unforeseen negative side-effects on other groups. Understanding current patterns of resource distribution and the factors which determine them is necessary in order to design programmes, effectively, as new projects or policies may alter the determining factors and thereby change the patterns of allocation.
The effects of projects on intra-household dynamics may be subtle and complex, but they are absolutely central to the successful outcome, and even to the successful implementation of development interventions. This is because development projects are based on certain assumptions about how households will behave in the face of the change they intend to bring about.
For example, a new technology which increases productivity may change the value of an individual's time, with consequences for the time spent on other tasks he or she performs. An immunization programme which alters the probabilities of child survival may change the strategies by which households decide to invest in certain children. A paved road may open new opportunities for wage employment, altering the opportunity cost of time devoted to unpaid household production.
Understanding intra-household decision-making is important in order to predict who in the household is likely to gain and lose (in both the short and long run) as a result of an intervention. Projects which place unacceptable burdens on some individuals without compensation may find themselves compromised by lack of participation. Projects which attempt to target consumption goods to particular household members often find that the household's own preferences in allocation supersede those of the programme planners.
The papers in this volume describe the wide variety of linkages between programme interventions and individual outcomes. They spell out a set of issues which, in turn, suggest the need for obtaining information on some specific variables, including house hold composition, income and assets, task allocation and time use, and individual consumption and welfare. Neither income, capital, nor the time of given members is necessarily interchangeable with that of other household members, so that these variables need to be measured for each individual.
The papers discuss alternative approaches to collecting data on the internal processes by which households allocate resources and responsibilities among their members. The authors make it clear that both information gleaned from secondary sources and primary data obtained using a wide range of qualitative and quantitative methods are needed to investigate intra-household processes. None of these methods is novel, although their application to household dynamics is relatively untried. The important message is that the combination of methods is stronger than any single methodological approach, because analysis of the household draws on the concepts of a variety of disciplines.
Many measurement problems have not been satisfactorily resolved. The household and its internal dynamics have only relatively recently become a focus for research, and one result of this work has been to identify new areas of investigation, requiring the development of new measurement techniques.
The search for inexpensive and quick methods of measurement has not always been successful, and in many cases short-cuts simply cannot be taken. Collecting data on individuals is necessarily more time-consuming than collecting aggregate data. It makes sense to seek proxy variables and easily accessible indicators for some of the more difficult-to-measure concepts in household-level studies, but it will not always be possible to find them.
These papers taken together clearly demonstrate that information on intra-household allocation and its determinants is central to effective development planning. Including this level of analysis in the planning process certainly requires more resources, both in time and money. But the return on these additional costs should be more than compensated by the greater likelihood of project success.
INTRODUCTION
These tables present, in summary form, an approach for incorporating intra-household analysis into development planning. They are taken from a methodological guideline prepared for the United States Agency for International Development and other donor agencies which implement development projects abroad. The guideline, Incorporating the Intrahousehold Dimension into Development Projects: A Guide (Rogers, 1988), can be obtained from USAID/Bureau of Policy and Program Coordination/ Human Resources Division, in Washington, D.C.
The papers in this volume underscore the importance of including intra-household issues in the design, monitoring, and assessment of projects and programmes, The tables include a list of key variables, methods of measurement, and the comparative advantages of various approaches, as they pertain to each stage of project implementation.
REFERENCE
Rogers, B.L. 1988. Incorporating the Household Dimension into Development Projects: A Guide. Office of Policy Development and Program Review, Bureau of Policy and Program Coordination, US Agency for International Development, Washington, D.C.
Table A. Data to be collected: variables, uses, methods
| Type of data | Variables | Uses of the information | Methods |
| Household composition, structure, and function | |||
| Members living under one roof or in one compound (co-residential unit) | a. Number of members, age and sex | Measure level of need in relation to resources (vars. a, e, f) | Secondary data is often available on common household structures |
| b. Common household structures (nuclear, extended, multi-generational, other) | Assess possibilities for task- sharing within household (vars. a, b, e, f) | Key informant interviews can cover common household structures | |
| c. Number of unrelated individuals | |||
| d. Number of members tied by blood, by marriage, to house-hold head | Indicate vulnerable groups at risk of low (relative) levels of consumption (vars. c, g) | Small-scale surveys of households should start with a listing of all members (in table form) including age, sex, relation to household head, educational level, occupation(s) for each | |
| e. Ratio of children to adults (age depends on local definition) | Identify possible sources of resistance to change (var. b) | ||
| f. Ratio of non-working to working members (depending ratio) | |||
| g. Sex of household head | Seasonal variation may be addressed by questions on the house-hold list ("Is this person usually present all year? During what season is she absent?") or by covering several seasons. Servants or other unrelated individuals may be defined as members for some purposes and not others, e.g. if they eat with the household they may be members; if they are paid by the household rather than contributing to it, they may be considered separate | ||
| h. Seasonal changes in co-residential household size and composition due to in- and outmigration | |||
| If servants are considered a separate unit, any study of households should include them in the sample | |||
| Group eating from a common food supply (commensal unit) | a. Degree of overlap of this group with the co-residential unit | Identify
possibility for leakage of benefits |
Secondary data on social organization may include information on sharing and gift-giving |
| b. Frequency of food gifts sent and received | Identify possible paths for dispersion of benefits | ||
| c. Rules governing exchanges of food | Direct observation of households at mealtimes to determine whether members are usually absent or guests are present | ||
| d. Frequency of members eating away from home | Define level of need. Co-residential unit may not accurately define need if much sharing occurs | In small-scale surveys, a question in the household list may be added for some or all members: "How many meals are taken at home/ away from home?" | |
| Group pooling its income and resources for common support (income-pooling unit) | a. Degree of overlap of this group with co-residential and commensal units | Identify possibility for leakage of benefits | Information on income/resource pooling is very difficult to obtain in a short-cut manner. Food-sharing may be a proxy in some cases |
| Identify possible paths for dispersion of benefits | |||
| b. Degree of pooling by different members: male head, female head, young adult children, elderly relatives, unrelated members | |||
| Secondary data on pooling is not commonly available | |||
| Define level of need | |||
| c. Degree of pooling with persons not living in the household: relatives living elsewhere, foster children | Direct observation of pooling is not possible | ||
| Informal questioning of key informants may give idealized rather than actual picture but should indicate how to pose survey questions | |||
| d. Frequency and source of gifts in cash and kind | In small-scale surveys, may ask individual members: "What categories of expenditure do you spend your income on?" "How frequently do you receive (give) gifts of cash, of goods?" "Does anyone outside the household pend on your income?" "How much of your income is reserved for your personal (as opposed to household) use?" | ||
| Labour-sharing unit | a. Degree of overlap with co-residential, commensal, or income-pooling group | Identify possible conflicts with time and labour requirements resuiting from development projects or policies, and possible means of accommodating to them | Studies on labour obligations may exist in the anthropological or sociological literature |
| b. Nature of labour obligation (type of work, whether mutual or one-way) | Informal direct questioning of local informants can reveal rules for labour-sharing and the nature of the shared tasks | ||
| c. Whether labour obligations are determined by blood, affinal, or other ties | Identify possible sources of resistance to change, or barriers to individuals taking advantage of new programmes or policies | ||
| d. Seasonality of obligations | |||
| e. Whether there are several different labour-sharing units with different obligations and tasks | |||
| Identify possible paths for dispersion of benefits, especially of productive assets and training | |||
| f. Degree of overlap among the different labour-sharing units | Identify possible detrimental effects from disruption of the labour-sharing units or possible shifts in membership (due to introduction of new technology, for example) | ||
| Income | a. Agriculture | Estimate level of income adequacy and security of households | Information on the nature of the economy (types of employment, types of production) will certainly be available from either published studies or internal government or donor agency reports |
| —Degree of dependence on agricultural wage labour, subsistence farming, and farming for sale (proportion of households which earn income from each source; proportion of each household's income from each source) | |||
| Estimate returns to different kinds of human capital (wage rates) based on education, age, sex, to predict incentives for investment in particular individuals; to estimate relative value of individuals' time in home and market activities | |||
| Direct observation of workplaces (fields, markets, factories) can in dicate by whom certain jobs are done and the types of work per formed | |||
| — Are certain types of agricultural labour performed by women, men, children (whether for pay or for own production)? | Identify likely degree of control (by individuals) of income as a whole; of income from different sources | ||
| Key informant interviews can proice information on: | |||
| — Who in the household | - association of crops, tasks with certain individuals; | Predict changes in returns to different kinds of human capital; possibly predict changes in household's investment in different individuals | |
| - owns land; | |||
| - controls the uses of its products; | - seasonality of employment; | ||
| - labour shortage/surplus; | |||
| - markets the products? | - types of work available | ||
| —Are certain crops or types of crops the responsibility of certain members (women, men)? | Focus group methods may be used to obtain information on: | ||
| Predict changes in returns to physical assets and possible consequences for access | |||
| - general pattern of income earning (number of earners, their sex, age, type of jobs); | |||
| —Seasonality of income from different crops and labour | |||
| Predict changes in individuals' incomes (amount, reliability, frequency) | |||
| - perceived association of individuals' income with specific expenditures; | |||
| —Rates of pay for different kinds of work | |||
| —Form of pay (cash, inkind) | Predict changes in household income (amount, reliability, frequency) | - perceived association of individual income with control over income uses | |
| b. Formal sector employment: | |||
| —Types of jobs available | Small-scale surveys can identify: | ||
| —Full-time, part-time, seasonal | Predict possible changes in control over assets and income, and in their uses reliability of income; | ||
| - individual income streams within households: approximate amount, frequency, | |||
| —Skill or educational level required | |||
| —Rate of pay | |||
| Identify possible sources of resistance to change | |||
| —Period of pay (piece, day, week, etc. ) | - categories of expenditure associated with individual in come streams (subjective perception of respondents) | ||
| — Form of pay (cash, inkind) | Identify possible change in how food is acquired (purchased, home-grown), and possible consequences for food adequacy and security | ||
| —Are the jobs for men, women, children, or no restriction? | |||
| —Is labour in surplus, or scarce in different kinds of jobs? | |||
| Predict who might gain and lose from altered employment opportunities | |||
| — Seasonality of labour demand | |||
| c. Informal sector employment: | |||
| — Types of jobs available | |||
| —Level and reliability of income | |||
| —Done by men, women, children | |||
| —Seasonality | |||
| Assets and wealth | |||
| Productive assets | a. What major productive assets are owned by households? (Proportion type of household) | Predict changes in ownership or access to the use of resources | Existing studies |
| Key informant interviews | |||
| b. Is ownership joint or individual? | Identify possible sources of resistance to change | Small-scale surveys of households, asking about ownership and use of resources (e.g. a checklist format covering, for each listed asset, ownership, use, how obtained) | |
| c. How is access obtained? Distinguish use rights (rights to the product) from ownership (right to allocate). Are rights obtained by purchase, in heritance, through blood or marriage ties, etc.? | |||
| Predict who may be displaced from use of assets | Previous studies might exist in the anthropological literature | Predict who will and will not benefit from changes in the productivity of assets | |
| Publicly owned assets | a. What resources are freely available to all? | Identify possible changes in availability of free goods (e.g. food, water, fuel, other goods) | Direct observation |
| Key informant interviews | |||
| Predict consequences for consumption and time use | |||
| Ownership of consumption goods | a. Quality of housing (roof, walls. floor) | Estimate general economic level of households | Direct observation |
| b. Utilities (electricity. water, waste disposal available to household) | Identify vulnerable population | In small-scale surveys, include questions about ownership of resources | |
| c. Ownership of goods indicating wealth (e.g. bicycles, auto-mobiles, radios, televisions, cows, goats) | Predict project or policy effects which may vary depending on total resource level of household | Note that choice of which specific goods are accurate indicators of wealth depends on local knowledge | |
| Task allocation and time use | |||
| Inventory of tasks | a. Major taks of | Use to collect and organize information from subsequent sections | Existing studies (including anthropological and sociological research) |
| - household maintenance; | |||
| - home-board production; | |||
| - work outside home | Key informant interviews | ||
| b. Range of time required for each task | |||
| Direct observation of tasks that can be publicly viewed | |||
| Organization of tasks | a. Which tasks can be done together | Predict possible conflicts with new tasks required as result of a policy change | Existing studies |
| Key informant interviews | |||
| b. Which tasks must be done together or in a fixed sequence | |||
| c. Time restrictions on tasks (e.g. done only at certain times of day, week, year) | Predict shifts in the time spent on certain tasks or their frequency | Direct observation | |
| d. Location restrictions on tasks | Focus groups | ||
| Task allocation | a. Which tasks are performed by individuals of particular age, sex, status, what proportion of the time | Predict possible conflicts of current tasks with new tasks required as a result of a project or a policy change | Existing studies may be available for some tasks |
| Key informants may be useful but may provide idealized rather than accurate information | |||
| Predict which tasks are likely to shift and from/to which individuals | |||
| Direct observation of the performance of tasks is necessary to determine actual distribution of tasks | |||
| Predict consequences for quality of the work performed | |||
| Predict changes in total work burden of individuals | |||
| Social norms regarding work | a. Restrictions on types of work or place of work based on sex, age, status, religion | Predict which individuals will take advantage of changing work and income-earning opportunities | Previous studies |
| Key informants most useful for information on norms | |||
| b. Degree to which these restrictions are observed | Identify possible sources of resistance to changes resulting from project | ||
| c. Social norms governing earnings by age, sex | Focus groups may be useful to determine which restrictions are followed in practice, especially for activities which are difficult for an outsider to observe | ||
| Identify possible sources of social stress, family disruption, and violence as a result of changes due to a project or policy | |||
| Time burden of individuals | a. Amount of time spent in each of various tasks identified by methods in previous four sections | Identify possible conflicts with project-related tasks | Previous studies of time-use patterns may exist in a few cases only; check UCLA database of time allocation studies (see Johnson's chapter in this volume) |
| Predict possible changes in the performance of tasks (who does them; how well; how much time is spent) | |||
| b. Time constraints on individuals (available leisure; amount of time spent sleeping; working; in recreation; flexibility in allocation of time) | |||
| Information from direct observation of tasks, the time they requite, how they are organized, and their allocation among individuals can be combined with in formation on household size and composition to estimate work burden on particular individuals | |||
| Assess whether time will be shifted, reduced, increased, for a given task as a result of the time requirements of the project; or whether the task will be displaced to another person altogether | |||
| c. Work burden of individual | |||
| Predict changes in the work burden of certain individuals; consequences for their welfare; the welfare of children | Small-scale surveys can ask questions on: | ||
| - the frequency of performance of certain tasks; | |||
| Assess available leisure, amount of sleep, as a measure of welfare of individuals | - the frequency of available help, or number of helpers, for the task; | ||
| - age and sex of helpers, to estimate work load; | |||
| - the range of time different activities take; | |||
| - how tasks are organized (sequence steps; are they done alone or always in conjunction with other tasks?) | |||
| 24-hour activity recall or spot checks | |||
| Direct observation of time use of individuals | |||
| Consumption | |||
| Food | a. Growth outcomes, growth rates of children by age and sex | Identify vulnerable groups, groups at risk of inadequate food intake | Growth outcomes can be measured in a small-scale survey measuring height, weight, and age of children and comparing height/ age and weight/height with a standard. This the best measure of adequacy of food consumption |
| b. Specific foods or types of foods allocated to certain individuals (by age, sex, work or pregnancy status, kinship status in household) | |||
| Predict who will benefit in food availability at household level | |||
| c. Specific foods or type of foods allocated or withheld during illness, pregnancy, lactation | Anticipate possible changes in access to food if consumption pattern changes | Key informant interviews and focus groups can indicate whether specific foods or types of food are preferentially given to certain types of individuals, and what foods are given or withheld in illness, pregnancy, etc. | |
| d. Meal patterns of individuals: frequency of formal meals at home, away, and informal consumption (wild food, street food, snacks at neighbours'); illness | Assess degree to which household food availability is a proxy for food available to each member | ||
| In a survey, questions may be included on allocation of foods to individuals and on allocation in sickness. For example, a checklist of local foods may be presented with questions like "Is this food mainly given to children? babies? boys? girls? adults? men? women?" "If your child is sick, do you increase feeding of any foods? Which? Decrease? Which?" | |||
| e. Food intake (quantity) of specific individuals | |||
| A food-frequency questionnaire may be administered to certain individuals in a sample of households; principal caretaker may answer for younger children | |||
| Meal pattern information may be obtained from local informants and from direct observation inside households | |||
| Note that household food consumption cannot be used as a proxy for adequacy of consumption by individuals. | |||
| Health care | a. Morbidity of individuals by age, sex | Identify vulnerable groups | Previous studies and government or agency reports may exist on available types of health care (but reliability may be questionable) |
| b. Infant and child mortality by age and sex | Predict likely pattern of use if available services are changed | ||
| c. Frequency of use of different categories of health-care services | Predict who is likely to benefit first from changed services; who in long-term | Direct observation at health service locations can indicate who uses the services; how much time is required; and what services, personnel, and supplies are available | |
| d. Time and cash costs of services | |||
| e. Number of hours services are open; who staffs them during which hours (doctor; nurse; lay health worker); sex of staff members | For what kinds of problems are services likely to be used | ||
| For which members | Focus groups and local informants can provide information on what services are used for what complaints; who is responsible for providing the care; time costs and other constraints to use. Information on sex bias will probably not emerge from this method | ||
| f. What medicines/vitamins are dispensed; under which circumstances | In a small-scale survey, questions may be included on: | ||
| - morbidity of children and adults (accurate retrospection probably limited to 2-4 weeks; may distinguish diarrhoea, fever, respiratory problems); | |||
| - use of services: accuracy is probably better if questions refer to the last illness episode of the individual rather than "usual practice"; it is helpful to distinguish first source of care, second source of care (if applicable), etc. | |||
| Education | a. Educational levels of house-hold members by age, sex | Indicate preference for investment in certain individuals | Information may be available in government or other agency reposts |
| b. Current school enrolment of individuals by age, sex | |||
| Identify groups not receiving services | |||
| c. Proportion of girls and boys in school by age/grade level | Predict who will benefit first from changes in availability of services or access to them | Direct observation of schools can indicate relative attendance of boys and girls, members of different ethnic classes | |
| Predict who will benefit from a change in the returns to education of specific individuals | Direct observation of communities can suggest degree of non-attendance (in some circumstances) | ||
| Focus groups can address questions of who is sent to school and why, and what barriers to attendance exist | |||
| In a small-scale survey, questions on educational level, literacy, and current enrolment can be included in the household listing | |||
Table B. Intrahousehold analysis procedures
| Steps | Outputs | Approximate time required |
| Review of project idea or plan | ||
| 1. Project or policy idea is presented | 1-2 weeks | |
| 2. Objectives of the project or policy change are specified | ||
| 3. Linkages between inputs of project or policy and expected individual outcomes are spelled out in detail, using the following questions regarding intra-household issues. A detailed scheme or model of these links between the project inputs or policy induced changes and individual outcomes is prepared | A flow chart or other framework specifying the linkages between project (or policy) inputs and expected outcomes | |
| 4.
Missing information needed to complete the model is
identified - Who will participate in project activities? - Will the project require or cause a change in household structure, composition, or function? - Will the project change any person s access to productive resources, or any person's control over what is produced (including control over income from his/her labour)? - Will the project affect any person's wage rate (returns to labour) or the rate of return to assets under any person's control? - Will the project require changes in the inventory of tasks performed by household members, or in the organization of tasks? - Will the project change the allocation of tasks among members or the time use of members? - Will the project change any person's access to consumption goods (food, health care, education, etc.) which affect welfare? |
Written
specification of a model of expected effects on
individual income, command over resources, task
performance, time burden, and consumption Written identification of gaps in knowledge, necessary to assess these effects |
|
| Integrating information from existing sources | 1-6 weeks | |
| 5. Published and unpublished literature is reviewed to fill in missing information on projects' or policies' effect on individuals | ||
| 6. Literature review is used to identify people who have worked in the area of the proposed project or policy | ||
| 7. These people are contacted for the information they can provide. Additional written information and personal contacts may be identified | ||
| 8. The written model for project effects on individuals is updated, and the remaining areas of missing information are identified | Updated
written specification of model of expected effects Updated written indication of missing information |
|
| Planning for field-work | 1-2 weeks | |
| 9. These are used to prepare a set of topic guidelines (questions appropriate to direct observation and to different categories of informants) to be followed during on-site data collection | Topic
guides Data collection plan |
|
| Preparations are made for on-site data collection | 1 month | |
| 10. Geographic area(s) for on-site data collection are identified; government and agency concurrence obtained if necessary | List of geographic areas for data collection, with reasons for selection | |
| 11. Identify skill areas required for data collection; identify person or team to conduct the field-work | List of people, qualifications and availability | |
| 12. Identify persons in-country who may help with data collection effort | ||
| 13. Select team; prepare contracts; deploy in field | Administrative paperwork required to field team | |
| Field-work | 1-2 weeks | |
| 14.
Contact with local sources of information: team visits
knowledgeable social scientists, programme
administrators, government of ficials to obtain opinions
and information required to complete intra-household
analysis of proposed project. Hiring of additional team
members 15. Field data collection - Team travels to area of proposed
project (repeated in several areas if appropri ate to
nature of project). - Data collected using topic guides, updated as needed. - Direct observation of representative sample of households, schools, clinics, markets. shops, or other locations as appropriate to the proposed project or policy. Observations used to obtain data and to verify the accuracy of verbal reports - Informal interviews with local informants chosen to represent varying perspectives and points of view - Focus groups of local population conducted as appropriate - Model of project is updated and remaining knowledge gaps identified |
Updated written specification of model of expected project or policy effects | Minimum 4 weeks in most cases |
| 16. Survey research: if necessary, a small-scale household survey may be undertaken, using the results of the previous data-collection effects as a basis for designing a survey instrument | Minimum 2-4 additional weeks in most cases | |
| 17.
Integration of data into project plan - Proposed project or policy plan is reviewed and elaborated or modified as needed - Completed model of expected effects is prepared - A plan for the timing and date requirements for monitoring of the intra-household effects of the project is developed |
Completed
written model of predieted effects Project or policy implementation plan Written project-monitoring plan |
1-3 weeks |
Similolu Afonja
University of Ife
Department of Sociology and
Anthropology
PO Box 1052
Ile-lfe, Nigeria
Jere Behrman
Economics Department
University of Pennsylvania
3718 Locust Walk
Philadelphia, PA 19104-6297, USA
Patrice Engle
Psychology and Human Development
Department
California Polytechnic State University
San Luis Obispo, CA 93407, USA
David Franklin
Sigma One Corporation
PO Box 12425
Raleigh, NC 27605, USA
Marito Garcia
International Food Policy Research
Institute
1776 Massachusetts Avenue, NW
Suite 400
Washington, DC 20036, USA
Peter Heywood
Papua New Guinea Institute of Medical
Research
PO Box 378
Madang, Papua New Guinea
Elizabeth Jelín
Centro de Estudios de Estado y
Sociedad (CEDES)
Avenida Pueyrredon 510-Fe
1032 Buenos Aires
Argentina
Allen Johnson, Chairman
Department of Anthropology
University of California
Los Angeles, CA 90024, USA
Shubh Kumar
International Food Policy Research
Institute
1776 Massachusetts Avenue, NW
Washington, DC 20036, USA
Judith McGuire
The World Bank
1818 H Street NW
Washington, DC 20433, USA
Ellen Messer
Alan Shawn Feinstein World Hunger
Program
Brown University
Providence, RI 02912, USA
Isabel Nieves
Instituto Nutricional de Centro
America y Panama (INCAP)
Guatemala City, Guatemala
Patricia O'Brien-Place
US Agency for International
Development Africa Bureau
Washington, DC 20520, USA
Per Pinstrup-Andersen
Director of Cornell Food and Nutrition
Policy Project
Division of Nutrition Sciences
Savage Hall
Cornell University
Ithaca, NY 14853, USA
Ellen Piwoz
Pan American Health Organization
525 Twenty-third Street, NW
Washington, DC 20037, USA
Najma Rizvi
International Centre for Diarrhoea)
Disease Research (ICDDR)
GPO Box 128
Dhaka-2
Bangladesh
Beatrice Lorge Rogers
Tufts University School of Nutrition
126 Curtis Street
Medford, MA 02155, USA
Mark Rosenzweig
Department of Economics
1035 Business Administration
271 19th Avenue South
University of Minnesota
St Paul, MN 55455, USA
Constantina Safilios-Rothschild
Population Council
1 Dag Hammerskjold Plaza
New York, NY 10017 USA
Susan Scrimshaw
School of Public Health
University of California
Los Angeles, CA 90024, USA
John Strauss
Economics and Statistics Department
RAND Corporation
1700 Main Street
Santa Monica, CA 90406, USA