Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Methodological issues in nutritional
anthropology
Report of a Workshop
INTRODUCTION
On 17-20 November 1980 a workshop was held at the Massachusetts Institute of Technology, Cambridge, Mass., USA, for the purpose of evaluating methodological issues in nutritional anthropology. The meeting represented the collaborative efforts of the Committee on Nutritional Anthropology (Committee Il/8) of the International Union of Nutritional Sciences (IUNS), the Western Hemisphere division of the International Commission on the Anthropology of Food (ICAF), and the World Hunger Programme of the United Nations University (UNU).* The 24 participants from 14 countries represented a broad spectrum of experience and disciplinary background, including cultural and physical anthropologists, nutritionists, and nutrition and health care planners and administrators.
Background for the Workshop
During the 1970s many nutritional scientists became interested in multidisciplinary approaches to understanding problems of malnutrition. At the same time, there was growing recognition among social scientists of their potential role in contributing to understanding food and nutrition problems in the modern world. Increasingly, social scientists (including anthropologists) have become involved in both research and applied programmes on nutrition. From such collaboration, new directions of research are emerging. Among them the development of "nutritional anthropology" is beginning to generate a body of data and theory on the relationships of nutrition to socio-cultural and ecological processes.
The scope of research interests in nutritional anthropology is broad. Some investigators focus on the cultural context, seeking to understand the meaning of food in cultural and symbolic terms. Others are interested primarily in identifying the linkages between local conditions and national and international political and economic forces.
Biological anthropologists emphasize the interactions of nutrition-related physiological processes with environmental and population characteristics. Common to much of the work in nutritional anthropology is the emphasis on a "holistic," multi-factored approach, seeking to examine complex interrelationships of food use or nutritional status in the context of specific human communities.
The new research styles that emerge from changing theoretical perspectives and practical experiences pose serious methodological problems for any field of scientific inquiry. Problems of research design and data collection are likely to be particularly acute during the initial phases of interdisciplinary growth. This workshop was convened in response to the felt need, expressed by anthropologists and others, to carry out a "methodological stocktaking" for nutritional anthropology in order to assess its strengths, as well as its weaknesses, as a step in the process of developing more meaningful research. It was not intended to be an examination of the contributions that anthropological theories can make to the understanding of food and nutrition problems. However, during the course of the workshop, the intimate relationship between theory and method frequently led discussion into the realm of "what is important to know?" as contrasted with the methodological questions of "how do we find out what we think we need to know?"
Assumptions
The planning and organization of the workshop was based on these assumptions:
1. The people who have the primary responsibility for methodological development and the primary need for better methodology in nutritional anthropology are anthropologists. On the one hand, it is useful to recognize that many anthropological techniques are effectively employed by other types of researchers, and it is important to work at "demystifying" the anthropological research process. However, it is also important to recognize the significant contribution that the perspective, framework, knowledge base, and skills engendered during the course of anthropological training and research experience make to the anthropologist's effectiveness as a researcher.
2. While the main focus is the improvement and facilitation of anthropological research, the process leading to improvement requires the interaction of anthropologists with nutritionists, as well as other social and health science disciplines. Furthermore, many of the methodological concerns of anthropology are closely related to issues in other social sciences and research concerns involved in social-biological interfaces, including epidemiology, social psychology, communications research, and sociology.
3. While the review of methodological approaches should be as broad and as comprehensive as possible, the evaluation of problems and analysis of priorities for future development should be based on the application of research to problems of malnutrition and hunger.
Basis for Workshop Discussion
Like anthropology as a whole, the work of nutritional anthropologists covers a broad spectrum of theoretical perspectives, utilizing a wide range of research skills. To bring the workshop to manageable proportions required selecting certain aspects from the full spectrum of relevant anthropological work.
One criterion used to narrow the focus of discussion was to exclude those aspects of methodology that have been extensively examined by nutritional scientists, including anthropometry and diet intake assessment. Although anthropologists have been working actively in these areas, the basis for excluding these topics from the workshop was that a good deal of intensive analysis of methodological problems in this area has already been carried out and will continue to be. The effect of the decision was to exclude much of the work in biological anthropology from the workshop and to shift the focus primarily to methodological problems in the socio-cultural side of research, an emphasis that is justified by the relative lack of attention these problems have received.
A second criterion by which the scope of discussion was narrowed was to exclude socio-cultural issues and concepts that are utilized in a wide variety of research not limited to food and nutritional studies. For example, problems of measuring "socio-economic status" or "level of modernization" would not be addressed, even though they are often of major importance in nutritional-anthropological research. Because these concepts are widely used by socio-cultural anthropologists, more attention has been given to methodological concerns attendant on their use.
A third criterion was that special attention should be given to the problems of transforming qualitative, descriptive knowledge (e.g. about food-use patterns, cultural beliefs, and social factors) into quantifiable observations, that can be used in statistical analyses. The rationale for this criterion involves the assumption that one of the important strengths of the anthropological contribution to nutrition research is the description of cultural processes. However, to be credible and useful in practicable application, these descriptions often have to be specifiable at the level of individual or household units. Such data should also involve variables that relate to processes at regional and national levels.
Finally, a fourth criterion was the emphasis on problems of research design and data analysis. These have tended to be less salient in social and cultural anthropology, a feature that becomes particularly critical in multidisciplinary research.
As preparation for the workshop, background papers prepared by individuals with special familiarity and expertise in the topic under consideration were sent to the participants in advance. These will be further edited and enlarged for publication in book form.
THE THEORETICAL FRAMEWORK
Throughout the workshop, sometimes as the direct focus of discussion, sometimes as counterpoint to other topics, an underlying theme of discourse was the defining of a general theoretical framework in terms of which the methodological issues were analysed. While all of the anthropological participants took parts of the framework for granted, there was no clear consensus among them regarding the components of the framework. Other participants in the workshop were familiar with the framework to varying degrees. As in all complex multidisciplinary groups, there were many points at which fundamental differences in perspective affected the group's ability to explore an issue fully. The following points outline major features of a framework that characterizes much of the work in nutritional anthropology.
Focus on Communities and Regions
Traditionally, the community has been the primary focus of anthropological research, and, typically, the community is the point of entry for the anthropological researcher. The community base continues to be a major feature in nutritional anthropology, even where the objective involves nutrition, health, or food programmes that originate at higher governmental levels, for such projects are seen as having their locus of differential effects in communities and households. Furthermore, differences in local political conditions, resources, administrative structures, and socio-cultural characteristics have powerful effects on the ways in which macro-level processes (national and international political and economic forces) affect people at the community level.
Linkages between Macro-, Intermediate-, and Micro-level Contexts
As noted above, events and conditions at the micro-level (of community and household) must be conceptualized as being strongly affected by economic, political, and other forces at macro-levels (regional, national, and international economic and political processes). In some nutritional-anthropological research, data-gathering includes regional and national statistics, as well as data gathered through interviewing and observing at higher administrative and commercial levels. Frequently, however, local level features are studied in such a manner as to include data about macro-level linkages as they are manifest at the local level.
Households and Individuals as Data Units
Households and individuals are primary units of data collection. The household (defined as the smallest economically co-operating, food distributing, and consuming unit) is assumed, by nutritional anthropologists, to be of primary theoretical and practical importance in most societies. Data on land-holding, food animals, cash income, general socio-economic status, ethnic group identity, food-use patterns, and so on are often conceptualized as characteristics of households, even when multi-household kinship units are also important economic and social units.
Within households, nutritional status, health condition, and food-intake data must be identified for individuals. At the same time, it must be recognized that individuals in households embody social roles that represent generation differences, sex differences, and other social organizational factors. Also, cultural characteristics, such as attitudes and knowledge, food preferences and avoidances, may be seen as individual qualities that can vary from person to person and from household to household.
Ecological Theory
Research in nutritional anthropology often involves a more-or-less explicit ecological theoretical framework, although this is not the only useful orientation that anthropologists apply to research related to food and nutrition. This theoretical framework places emphasis on features of the environment (social as well as physical and biological) as systems of resources and constraints within which people act and adapt (in obtaining, distributing, and consuming food).
The ecological framework also includes a broad focus on "natural communities," which includes a wide spectrum of characteristics of relevance to food and nutrition of the population. This contrasts with a research focus that is narrowly defined in terms of a few characteristics of interest to the researcher. This includes, inter alia, the emphasis on a multivariable and holistic description of relevant features, including economic resources, social structures, ideational patterns, and so on.
A major strength of the ecological perspective is to focus attention on the factors that can produce large differences between regions or ecological zones in food production and distribution, as well as differences in the impact of national food programmes. On the other hand, researchers have sometimes placed too much emphasis on local-level social and cultural processes and have failed to develop methodological tools for linking individual and household adaptation processes to regional and national economic and political structures. Without systematic relationships to such macro-level systems, ecological approaches fail to generate insights and data relevant to national-level development and social system reform.
The ecological framework includes the analysis of differential power relationships at all levels, including the study of local-level food use and nutrition. In much social science research, the most usual ways in which differential power relationships are recognized are in the widespread recognition of socio-economic status and ethnic membership variables. However, the consideration of differential economic resources and power can go far beyond those particular indicators.
Other theoretical approaches or paradigms also play a considerable role in nutritional anthropology. Some materialist-oriented research, for example, is especially concerned with processes of production and resource allocation at national and international levels. Symbolic analysis and most psychological approaches, on the other hand, deal with micro-levels of food and diet. In the course of the workshop it was clear that a comprehensive nutritional anthropology, now and in the future, will consist of a variety of perspectives suited to different kinds of theoretical and practical questions.
Culture and Belief Structures Are Only Part of the Framework
Contrary to some widely held stereotypes about anthropologists, most nutritional anthropological researchers do not regard "culture" (ideational systems) as the principal causal forces in food use. The cultural features of beliefs, attitudes, and values are generally seen as interacting with economic, demographic, social-structural, biological, and other elements.
THE METHODOLOGICAL PRIORITIES
While not all nutritional anthropologists are agreed on all aspects of the preceding research paradigm, the workshop generated a set of priorities for methodological development that build from this general framework.
Need for New Approaches to Operationalizing Micro-level/Macro-level Linkages
This is perhaps the foremost priority to emerge from the workshop and aroused wide-ranging and exciting discussion. Most research in food and nutrition (whether carried out by anthropologists, nutritionists, economists, or other types of researchers) is generally either large-scale, macro-level analysis or local-level study, without serious attempt to link issues to regional, national, and international processes. Members of the workshop expressed the view that this problem is endemic in much food and nutrition research and calls for new methodological and theoretical strategies. Some possible directions for operationalizing this aspect of research include:
- Unpacking or disaggregating (and aggregating} regional and national statistics on food production and distribution, along with data on economic expenditures and census data on populations, in order to examine the statistical specifics as they apply to given sub-populations. Unfortunately, many nations do not have sufficient flexibility in statistical reporting to make it possible to aggregate or disaggregate systematically for such purposes.
- Descriptive exploration of political and economic organizational structures that reach from national levels to intermediate and local levels. For example, new methods must be devised to quantify the differential impacts of commercial food merchandising and food production in different regions, reaching down to community levels. In a similar vein, national food-production programmes typically have areas of concentration and areas of low intensity, apparently because of differential administrative structures, differential monetary support, and variations in political structures.
- Certain key macro-level influences may be operationalized systematically at local community levels through study of local distributors, political brokers, and fiscal agents. These features have sometimes been descriptively included in food-use studies but are seldom systematically (and quantitatively} linked to local-level differentials in food production and food availability.
- Nested sampling techniques should be developed, in which local community samples {e.g. food-use interviews) are statistically relatable to regional surveys. Thus, there is a need to develop methodologies in which statistical analyses of community populations (e.g. percentage of children suffering from specified levels of malnutrition} can be directly tied to regional survey data, with accompanying assessment of comparable background variables.
Statistical and Mathematical Techniques for Multivariate Analysis
Holistic, ecological modes of research often involve dozens of variables per household or individual "case." The available statistical methods-including factor analysis, multiple regression, "MANOVA," and others-require development of more rigorous measurement than is common in anthropological field work. At the same time, the use of these complex statistical techniques requires careful design of research, from sample selection and scale development to the precise steps of statistical (or mathematical) analysis. More work is needed to refine the most effective combinations of statistical analysis for particular types of research problems. The development of appropriate quantitative strategies is a high priority.
Appropriate Mixture of Quantitative and Qualitative Analysis
The complex patterns of analysis involved in the study of nutrition problems and the sequelae of malnutrition clearly require multiple statistical procedures, but they also require effective qualitative descriptive modes of research, which is a prerequisite for the development of relevant, local-level numerical data-gathering and for the interpretation of statistical results. The qualitative descriptive modes of research need to be more closely articulated to the numerical data at every step of the research process.
Linkages with Other Disciplines
The workshop recognized that the general, eclectic style of research in nutritional anthropology also requires systematic interrelationships with other disciplines and that such interrelationships require more effective methodological tools. Particular attention was focused on the points of linkage between socio-cultural and biological/ clinical research, and it was noted that the main point of articulation between the two is in the individual person -embodying simultaneously the biological organism and the social/cultural actor who participates in the broader behavioural and ecological system. Linkages with other disciplines can be fostered if:
a. research collaboration occurs in the context of the same communities and populations;
b. congruent variables and terminologies are agreed on;
c. congruence of statistical methods is achieved;
d. data for different disciplines and different theoretical purposes is, where practical, gathered from the same individuals with the same overall sampling frameworks (e.g. small, specialized samples for intensive clinical analysis can be developed by random sampling from within larger survey samples in given regions and communities).
Main Issues and Priorities at the Micro-level of Research
1. Emic perspectives. The need for greater precision in identifying and giving quantitative expression to people's own views, their definitions of food and appropriate food-use behaviour, was stressed throughout the workshop. A recurrent theme was that effective analysis of nutrition problems and the development of effective solutions requires an analysis of people's viewpoints (the "emic" or actor's, perspective). However, despite some two decades of anthropological research utilizing the concepts of emic observations, there is need for far greater development of a comprehensive emic methodology and a need for comparative research to identify which emic methods produce the most reliable and valid results. The development of emic methodologies by indigenous researchers should be given high priority.
2. Cultural description and intra-cultural diversity. Certain emic methodologies produce data that portray relatively homogeneous cultural patterns (e.g. definitions of food categories, appropriate food combinations). These descriptions may be valuable, but they should be elaborated to permit systematic inclusion of intra-group variations in cultural definitions.
3. Levels and combinations in cultural patterns. Much research in nutritional anthropology focuses attention on food items and lists of food items (as well as taxonomies of food), without sufficient attention to combinations or culturally expected "eating events," "meals," and other second-order patterns. Methodological priorities include the development of ways to depict and to quantify such combinations in more meaningful ways.
4. Schedules and cycles of food behaviour. Closely related to the item above is the recognition that food-use patterns also involve chronological sequences. Here methodological priorities include the need to clearly separate trends and changes (e.g. "modernization") from short-term and longer-term food cycles (e.g. seasonal variations and responses to economic fluctuations).
5. Histories and traditions of food use. Most of the
currently utilized techniques for examining food-use beliefs,
behaviour, and related variables use a relatively flat,
synchronic perspective. New methods are needed for examining the
longer-term patterns of change and development, the growth of new
"traditions," the gradual "sacralization"
(enhancement of ritual importance)
of some foods, or the "de-sacralization" of food-use
patterns.
6. Relationship of "real" behaviour to reported behaviour. Very few field studies of food use and nutrition provide data on actual behaviour. Behaviour-oriented studies (e.g. 24-hour recalls or food diaries) provide materials on people's recall or their presumed behaviour. Frequently an individual's statements about recalled behaviour differ substantially from beliefs and conditions concerning "expected behaviour." An important research priority includes further comparisons and establishment of the relative utilities of these logically different, but related, data domains. Members of the workshop pointed out that direct observation of behaviour may appear to be the ideal methodology if one is to achieve accurate data about people's real behaviour, but direct observation has some shortcomings: (i) observation may have effects directly on the behaviour (the Heisenberg principle); and (ii) direct observation by trained researchers is slow and costly.
The Need to Define and Specify the Role of General Ethnography
In this discussion ethnography may be regarded as all methods of field-based data-gathering, at all stages of research, that are carried out for the purpose of providing general, descriptive (qualitative and quantitative) information in a community or other social unit. Survey interviewing may be part of the ethnographic process. Ethnography is broadly inclusive of all anthropological field research except the methodologically rigorous collection of quantified data for testing a specific hypothesis.
Anthropologists have often assumed that several months of general ethnographic, descriptive research is an essential prelude to focused, quantitative data-gathering for testing of hypotheses and related theoretical purposes. Such intensive ethnographic field work is usually seen by other researchers as too costly or simply time-wasting.
- A high priority is to define more precisely appropriate and needed aspects of ethnographic field research, in order that such "background" descriptive materials be achieved as quickly as possible and that participant observation and interviewing be focused on those aspects of culture, social organization, and economic systems that are directly relevant to specific theoretical and practical issues. One suggestion is that ethnographic field research must be opportunistic and open-ended, ever prepared to explore new areas of data that become relevant to specific food use and nutrition issues.
- Another high priority is to develop models of ethnographic research for different major ecological areas, within which certain sectors of descriptive work will have high priority and high relevance, especially for the refinement of locally appropriate variables. For example, in many intensive agricultural regions (much of India, South-East Asia, parts of Latin America), modes of land-ownership and cropping patterns are practically always relevant and crucial for understanding local differences in food availability. Religious beliefs and practices may be relatively peripheral in some areas, while for some regions they may play a large role in affecting food use patterns.
- In testing and developing more refined ethnographic strategies, the relative efficacies of direct observations as against key-informant interviewing in relation to particular kinds of data should be compared and assessed (recognizing that the effectiveness of particular ethnographic methods is context-specific and can vary greatly among different communities and regions).
General Problems of Precision and Estimation of Error
The workshop participants recognized that a major problem in all field work is the control of error in measurements. Problems of precision and error are endemic and serious in both biological and socio-cultural research, but it appears that estimations of the magnitudes and effects of imprecisions have been given more attention in biological field research.
A first step to management of error factors is to recognize their differential presence and differential effects in various kinds of research. Methodological exploration is much needed in the estimation of error effects. We need to know when random measurement fluctuations seriously affect the utility of research interpretations. More serious than random error is the presence of systematic biases in measurement. These two different problems both require methodological attention. In some instances refined statistical analysis can be an aid in sorting out and controlling for certain kinds of biases. Ultimately, however, no amount of secondary manipulation of data can overcome basic weaknesses in data collection.
Thus, a methodological priority is to develop some "first approximations" of error estimates in different types of research contexts as a step to improving precision of measurement. At the same time, it is recognized that precision and reliability should not be maximized at the expense of validity and theoretical utility of observations.
The Need for Longitudinal Data Strategies
A major weakness in much field research concerning food use, nutrition problems, food production economic systems, and related elements is the short chronological scope of such research. Most community nutrition studies (whether by nutritionists, anthropologists, sociologists, or others) fail to provide data on seasonal fluctuations, and most studies do not provide clear measurement of temporal changes except by somewhat questionable extrapolations.
Previous in-depth studies in particular regions provide one kind of baseline for re-study to establish patterns of change. Re-studies have serious problems of comparability if the original data-gathering methods are not entirely replicable, for whatever reason. Haphazard sampling procedures are generally more difficult to replicate than randomizing procedures. Therefore, random sampling in research sites is one way to provide for better future studies of patterns of change.
In addition, nutritional anthropologists are urged to develop new methodologies for recovering some aspects of past food systems. For example, old aerial photographs and maps sometimes contain much valuable information that can be difficult to transform from visual displays into numerical form.
FURTHER METHODOLOGICAL CONCERNS
The following methodological issues were the subject of discussion at the workshop but are not as readily characterizable as priorities for methodological development. They are, rather, issues and problems that need to be recognized, or suggestions for new types of research activities.
Presentation of Primary Ethnographic Data
It was recognized in the workshop that ethnographic descriptive materials are often presented in the form of declarative statements about cultural, social, and other features, without incorporation of evidence or data sources. More rigour and explicitness is needed in order for data-users to be able to assess the relative credibility of different ethnographic statements. (For example, the statement that "Most of the people here eat eggs regularly" may be the opinion of one key informant, it may be based on more or less systematic observations by a field worker, or it may be the result of a carefully conducted survey.) The differential accuracies of these variations in evidence require that ethnographies make data sources explicit where possible.
Appropriate Levels of Data and Methodological Rigour
Whereas the testing of key hypotheses and central issues in research generally requires as much precision and rigour of data collection as can reasonably attained, different aspects of research projects will have different levels of methodological requirements. In certain kinds of situations, less rigour with greater descriptive elaboration and contextual "flavour" may be more useful for given research objectives.
Levels of Systems and Their Elaboration
Members of the workshop recognized that a comprehensive analysis of food-use systems and of the related problems of malnutrition involves differential focus let different stages) on: (i) individual and household systems of data; (ii) communities as food distribution and energy systems; and (iii) local to regional to macro-level systems.
Identification of these systems and measurement of processes at different checkpoints in the systems require different strategies at the different levels. Analysis of a community-as-energy flow system, for example, requires fairly accurate estimates of both food intake and energy expenditure for different types of individuals at different points in the system. More methodological research is needed to establish the ways in which costly and complicated community-based energy models can be approximated and modified for analysis of problems in other kinds of ecological settings. There is also a need to refine and elaborate ways to test the effects of macro-level policies and processes as they impact and change community-level systems.
Computer Simulation and Other Re-analysis of Existing Data
Members of the workshop commented at several points that certain bodies of already existing data could be reanalysed productively and that certain types of research questions might be refined from existing data before costly new field research is undertaken.
Computer simulations based on already available data plus experimentation with changing parameters and assumptions can provide one such mode of re-analysis. Also, in the past, impressive bodies of field data have been analysed with relatively low-level and naive statistical methods, offering the possibility that more complex numerical analysis could generate important new theoretical insights without doing violence to the original data. Two main types of analysis could profitably be re-analysed:
a. research in which data have been aggregated and "tested" at community levels-which could, therefore, be disaggregated to examine intra-community processes;
b. research in which only pair-wise relationships have been examined-in which multivariate analysis could identify important new dimensions of interaction among key variables.
Certain types of numerical data that have been analysed statistically are fully amenable to mathematical exploration, including study of optimizing strategies of individuals, stochastic change processes, and perhaps cost-benefit analysis.
The workshop does not intend that researchers should over-analyse numerical data as an end in itself, and it is recognized that in some statistical analyses the data are "cooked" into practically unrecognizable forms. However, the costs-in terms of funding and the impositions on the research communities-of new field research suggest that already gathered data should be explored for new theoretical insights whenever practicable.
THE SOCIAL CONTEXT OF RESEARCH
A considerable amount of discussion in this seminar was devoted to the social contexts surrounding nutritional anthropological research. At issue was an "auto-cultural" analysis of the practice of nutritional anthropology, a consciousness that we, like all scientific researchers, are affected by social-cultural and environmental forces in our professional work. Especially problematic is the historical fact that nutritional anthropology has developed in the US academic context. Had it been developed in third-world contexts [either socialist or non-socialist), its shape and content might be quite different today.
This recognition led organizers of this workshop to invite as much third-world criticism and participation in discussion groups as possible-the goal being to share this new academic development with third-world researchers and to get them into the intellectual discourse and practice involved in further development of the field of nutritional anthropology.
The workshop could be viewed as the beginning of a broader, international practice of nutritional anthropology in all parts of the world, especially in third-world countries suffering the most serious nutrition problems. This wider and more varied context of research will have methodological as well as theoretical and practical implications to which we must be open. In practice, this means that researchers from the third world not only must be involved but must take leading positions in the definition and conduct of a generous measure of future research and publication.
This recognition should neither automatically illegitimize US and European scholars from doing research in the third world nor automatically legitimize all third-world researchers. Ideally, we will go beyond this "either/or" mentality so that skilful international teams will collaborate in research on the many urgent food and nutrition problems that are imbedded in international social, economic, and political systems. These systems are so wide that the social backgrounds, interests, and perspectives of very different researchers become complementary.
Participants in the workshop recognized that the present development of nutritional anthropology has been strongly influenced by the social context of its US academic origins. The workshop, however, has initiated the process of opening the field to the international research community. In practice, this should lead to much more nutritional anthropology research that is directed by, or involved within, third-world social contexts. Thus, not only will new methods, theories, and problems on food and nutrition emerge, but also a new broadened social context must guide the future definition of nutritional anthropology.
GENERAL OBSERVATIONS
Much of the discussion during the workshop focused on the need to refine and improve existing methodology. A central theme was that refinement of research methods must include recognition of the social and financial constraints within which actual field research takes place. In the next stage of development, nutritional anthropologists will continue to draw on the strengths of the considerable inventory of already available tools and techniques, while they work to refine them and create new methods. Research has to be context-sensitive and flexible in relation to the needs of programme development, implementation, and evaluation. However, as was repeatedly stressed during the workshop, it is possible to be rigorous despite the special needs and constraints of different national and ecological settings. Field research is, in part, the "science of the possible," seeking for the most rigorous and credible data systems within the context of responsibility to central issues of research directed to helping solve problems of malnutrition.
RECOMMENDATIONS CONCERNING PRIORITY AREAS
1. Research Design
a. There is need for greater utilization of multiple research focuses within the same project, involving household, community, regional, and national- and international-level analysis.
b. There is need for better integration of qualitative, descriptive data with quantitative measurement.
c. Nested sampling procedures should be developed so that data from intensive research on a small sample can be effectively related to large-scale survey results and national statistics.
d. There is need for increased utilization of longitudinal and prospective research designs.
e. There is need to extend methods to the analysis of social change processes broader than household or individual adaptation. This will require refinement of macro-level variables reflecting differential political and economic power.
2. Measurement
a. There is need for better utilization of emic research techniques to measure food beliefs and knowledge and perception of nutritional status.
b. Research is required on the effects of the observer on food-intake behaviour as compared with the error introduced through recall methods.
c. There is a need for identifying significant social and environmental schedules, conditions, and cycles that affect all research variables and for the development of time-sampling frames that take these schedules into account.
d. Research is needed to assess the precision and reliability of the measurements and techniques utilized by nutritional anthropologists.
3. Analysis
a. There is need for new techniques to analyse and quantitatively describe cultural patterns of food use.
b. There is great need for appropriate multivariate analysis. Also, work is needed to refine research design in relation to existing statistical techniques.
4. General
There is a need to ensure that rigour in methodology is maintained as new theoretical orientations are applied in nutritional anthropology.