Contents - Previous - Next
1. Davies B. Corbishley P. Evans J. Kenrick C. Integrating methodologies: if the intellectual relations don't get you, then the social will. In Burgess RG. Strategies in Educational Research: Qualitative Methods. London: The Falmer Press, 1985; 289-321.
2. Max Drake H. Research method or culture-bound technique? Pitfalls of survey research in Africa. In: O'Barr WM, Spain DH and Tessler MA, eds. Survey Research in Africa: Its Applications and Limits. Evanston: Northwestern University Press, 1973; 58-69.
3. Hall B. Breaking the monopoly of knowledge: research methods, participation and development. In: Hall B. Gillette A, Tandon R. eds. Creating Knowledge: A Monopoly? Toronto: International Council for Adult Education, 1982: 1-13.
4. Bernard HR, Killworth PD, Kronenfeld D, Sailer L. The problem of informant accuracy: the validity of retrospective data. Ann Rev Anthropol 1984; 13: 495-517.
5. Evans-Pritchard EE. The Nuer. Oxford: Clarendon Press, 1940.
6. Salamone FA. The methodological significance of the lying informant. Anthropol Quart 1984; 50 (3): 117-124.
7. Bleek W. Lying informants: a fieldwork experience from Ghana. Pop Dev Rev 1987; 13(2): 314-322.
8. Husband R. Foster W. Understanding qualitative research: a strategic approach to qualitative methodology. J Human Educ & Dev 1987; 26(2): 51-63.
9. Van den Berghe PL. Social science in Africa: epistemological problems. In: O'Barr WM et al Eds. Survey Research in Africa: Its Applications and Limits. Evanston: Northwestern University Press, 1973; 25-35.
10. Trow M. Comment on participant observation and interviewing: a comparison. Human Organization 1957; 16: 33-35.
11. Cook TD, Reichardt CS, eds. Qualitative and quantitative methods in evaluation research. Beverly Hills, CA: Sage, 1979.
12. Rossman GB, Wilson BL. Numbers and words: combining quantitative and qualitative methods in a single large scale evaluation study. Evaluation Review, 1985; 9(5): 627-643.
13. Collins R. Statistics versus methods. In: Collins R. ed. Sociological Theory. San Francisco: Jossey-Bass, 1984; 329-362.
14. Denzin NK. The research act. New York: McGraw-Hill, 1978.
15. Jick TD. Mixing qualitative and quantitative methods: triangulation in action. In: Van Maanen J, ed. Qualitative Methodology. Beverly Hills, CA: Sage Pub. 1983; 117-134.
16. Chambers R. Shortcut methods of gathering social information for rural development projects. In: Cernea MM, ed. Putting People First: Sociological Analysis in Rural development (Published for the World Bank). New York: Oxford University Press, 1985; 399-416.
17. Ratcliffe JW, Gonzales-del-Valle A. Rigor in health-related research: toward an expanded conceptualization. Int J Health Serv, 1988; 18(3): 361-392.
It would be useful to add another set of actors in the context of using RAP. In addition to the researchers, the methodological specialists, and in addition to the people themselves as key participants, we should not forget decision makers in the developing countries we are working with. It is those people who make decisions on priorities and resource allocations within their own countries. They may be from governmental or non-governmental institutions, who are providing services and health services to the rural client. They are making the decisions on the basis of whatever information is available to them at the time. It is important that the methodological specialists remember to work with them. First, to define what information they need, when and for what types of operational and political decisions they will have to make.
We should talk to them about the choice of methods in relation to the information the decision makers need. We need to show then that the tool kit is much wider than they may have supposed. We need to show them that RAP may be valuable to them, in complement to other types of methods. But the choice of method should be as a function of what they need and what can generate the information that they can effectively use.
The development agencies have an important role to play in helping decision makers in this area, and in showing the researchers and decision makers how to use RAP for themselves wherever appropriate. There is a need to show them how RAP can be not quick and dirty, but rapid and useful. This will help us put our discussions in a broader context. It is the decision makers that should be foremost in our minds as we consider the choice of problems and questions that will asked. It is consideration of the decision makers also that will help us at the other end of the research process to tell us what approaches to use to communicate the information we gain clearly and effectively to the decision making process. Continued
RAP does have a disciplinary face and there are pitfalls to it that are not self evident to the untrained. This points out the need for training.
Is it possible that we may face a danger of moving from macro-economics being the dominant form of data for planning to a stage where qualitative information is dominant and quantitative data is pushed under the carpet?
Despite the value and increased use of qualitative data and RAP, there is no looming risk to quantitative data becoming predominant on the agenda of the World Bank.
Dr. Pedersen's categories of purist, eclectic and pragmatist were extremely interesting and useful. Toward the end of his presentation he was also pointing to a false stereotype in the area of democratization of knowledge and the dichotomy of qualitative vs quantitative methods. Many of us have been brainwashed by our professional training, education and activities into thinking we are the only people who can count. We have tended to conclude that rural people are experts on their culture, beliefs and subjective experiences, but are not good at counting or estimating. This tends to obscure what anthropologists have known on and off: that rural people both literate and illiterate have a good capacity to count, to estimate, to recall quantities, to estimate trends, to rank, and to score. However, there are important preconditions to strong data gathering in this area. First is the critical need for an ability to establish rapport. Without this the value of participatory quantification work is very limited. Also if you wish rural people to quantify and estimate, you must develop a locally appropriate and relevant set of physical materials, such as seeds for counting, sticks broken into various lengths, stones for different seasons which people can quantify against. This area of participatory quantification is a frontier and this is extremely interesting.
theory is not unimportant to RAP and those who carry out
rapid appraisals. While random sampling is generally not
used, there is a strong need for anthropologists and
other RAP users to consider some of the issues of
sampling theory. They need to understand sampling theory
in order to determine what part of the population they
will interview and whose knowledge and opinions they want
their appraisal to reflect. RAP should avoid a simple use
of convenience sampling, and purposefully select criteria
for their sample of interviews, etc., and explain these
criteria in the methodology discussion of their
Nutritional surveillance: Objectives, principles, and lessons
The role of qualitative methodologies in nutritional surveillance
By David L. Pelletier
David L. Pelletier is affiliated with the Cornell Food and Nutrition Policy Program, Cornell University.
As a chapter also published in a new manual, Methodologies for Nutrition Surveillance, this paper raises two related issues which were both focal points of the conference. The first is that qualitative methods like RAP are essential investigatory tools for the follow-up of quantitative studies that search for explanations and solutions of identified nutritional problems. They are important to bring needed data to inform strategic decisions in the programme planning, evaluation, and adjustment loop. The second is the issue of institutionalizing RAP through professional training inside each country where it will be used. The translation of information gained through quantitative tools of nutritional surveillance into programme policy and strategies is seen here as being effectively and essentially mediated by the national ability to use methods such as RAP. - Eds.
EXCERPTED from a section of a manual on Methodologies for
Nutritional Surveillance, was prepared under the auspices of the
Cornell Nutritional Surveillance Programme and drawn from
experience gained under the Cooperative Agreement between the
Office of Nutrition, Bureau of Science and Technology of USAID
and the Colleges of Human Ecology and Agricultural and Life
Sciences at Cornell University.
A discussion on the issue of whether and how qualitative methodologies, whether rapid or not, can contribute to the field of nutritional surveillance is long overdue. Nutritional surveillance is one of the many planning-related fields that has suffered from what might be termed a "quantitative bias" in its approach. It is encouraging to see a growing receptivity - at least in certain quarters of some major donors - to complementing the existing set of planning and evaluation tools with a variety of qualitative approaches. One of the purposes of this paper is to examine the ways in which qualitative approaches could significantly strengthen nutritional surveillance activities.
At a higher
level of analysis this paper also raises the question of whether
and how qualitative approaches can be successfully transferred to
developing countries with the requisite levels of training,
sophistication of method, and validity. Indeed, there is a need
not only to transfer these, but also to institutionalize
them in the policy-making, planning and management of programmes
in developing countries. Although this is a desirable goal to
pursue, a number of questions exist concerning how to achieve it
in an acceptable manner. There are important lessons from other
fields - such as nutritional surveillance itself, previous
multisectoral nutrition planning, and farming systems research -
in which theoretically attractive approaches to planning were
developed but that encountered a number of difficulties in
practice. This is illustrated below and the implications of these
lessons for extending RAP or other qualitative approaches to
developing countries are examined.
The immediate impetus for launching nutritional surveillance as an ostensibly new set of activities to assist in the alleviation of world hunger and malnutrition was the World Food Conference of 1974. The concept, based on the model of infectious disease surveillance, sounded straightforward in principle, even if a bit on the ambitious side: To establish an ongoing system for generating information on the current and future magnitude, distribution and causes of malnutrition in populations - with emphasis on protein-energy malnutrition - in order to assist governments and international agencies in policy formulation, programme planning, management and evaluation. Attention was to be given to situations of acute food shortages as well as chronic malnutrition.
These broad objectives of nutritional surveillance and some guiding principles were published in 1976 by WHO, based on expert consultations, and subsequently elaborated upon by the Cornell Nutritional Surveillance Programme (CNSP) in 1984. CNSP itself was established in 1980 by a Cooperative Agreement with USAID in order to develop and disseminate the principles and methodologies of nutritional surveillance and to gain case-study experience. On the basis of CNSP's 1984 publication, and a number of experiences since that time , a number of principles have been articulated as described below.
As stressed in the 1984 publication , no one surveillance system can satisfy all the information needs implied in the original statement of objectives suggested by the WHO expert group. Instead, it appears that four distinct approaches should be recognized, based on the purpose for which the information is to be used [41. The four uses of nutritional surveillance are:
Problem Identification and Political Sensitization
Policy Formulation and Planning
Programme Management and Evaluation, and
Timely Warning and Intervention 
The rationale for distinguishing these four types is that each has its own requirements in terms of the types of data, periodicity, the need for predictive versus reflective indicators, the modes of data collection and analysis, the clients for that information, and so on . In recognition of this, it was proposed that the development of nutritional surveillance activities in a country be preceded by an initial assessment of some three to six months, during which time decisions should be reached concerning the type of systems needed, the specific users and their information needs, institutional arrangements, and a work plan for designing a system to meet those needs.
One of the immediate lessons for RAP lies in the observation that, despite these distinctions having been documented and widely disseminated through publications, training courses, etc., the vast majority of countries embarking on nutritional surveillance did so without having performed the prior steps of identifying decision-makers, the decisions requiring information and, thus, the type of surveillance systems required. Instead, the most common form of nutritional surveillance attempted - and with limited success - would be more appropriately considered "nutrition monitoring" . That is, the aggregation of data on the nutritional status of the population over time, usually in the form of weights and/or heights of clinic attenders or school children. This is illustrated below, followed by a discussion of how it relates to RAP.
In the output from one such surveillance system the prevalence of malnutrition among clinic attenders was plotted over time for each clinic along with the overall rate for all clinics combined. While such analysis disaggregates the trends by clinic, similar disaggregation could be and has been done at higher administrative levels as well, such as districts, regions and provinces in various countries. Apart from the question of validity that has never been properly answered for three kinds of data, a major limitation of such outputs is the lack of ancillary information on the factors responsible for the observed trends. It is therefore difficult to imagine what decisions and immediate action could be taken in response to information of this type.
Even when ancillary data are available from other administrative sources, they are generally adequate only for confirming that locations with a high prevalence of protein-energy malnutrition (PEM) also have a high prevalence of socioeconomic deprivation according to multiple indicators. However, these ancillary indicators are not suitable for pinpointing the specific factors requiring intervention. This is because of the multicollinearity among them, the inability to conduct the analysis at the individual or household level, the generally crude, proxy nature of the indicators, and the fact that information is typically available on only a few of the potential causes or intervention points. Thus, outputs such as these may be useful for political sensitization and geographic targeting, but do not indicate to planners or policy-makers what types of interventions should thereby be targeted to high-prevalence areas.
One of the attractions of these data is that they are often available through administrative sources, and merely need to be captured in an on-going fashion to turn them into a "nutritional surveillance system." However, the tendency to use those data that are readily available rather than those actually needed to support different categories of decisions may be one of the important reasons for the non-utilization of such data in decision-making. As elaborated in a recent publication  the extensive experience with this type of surveillance system in the countries of Central America and Panama have revealed far fewer tangible results than one would have hoped for, and this disappointing experience is not limited to that region.
The difficulties described above in the case of nutritional surveillance may have parallels with RAP. The diversity of applications for which RAP is relevant is clearly a potential strength rather than a weakness of the approach. However, the diversity of approaches being grouped under the term "RAP" does require serious examination. In particular, it would be useful to develop a topology of applications for which RAP might be appropriate and examine the implications of this for how RAP should be conducted. For example, one might distinguish between the use of RAP to enhance local participation in development planning versus its use as a technique for simply collecting information for managers, planners, and policy-makers. Such a distinction has clear implications for such things as who does the RAP, how rapidly it can/should be done, who should be the informants or participants at the local level, the degree and type of information bias to be expected and how it might be assessed and taken into account, what types of reports are required, etc.  In contrast to the case of nutritional surveillance in which the choice of methods and data sources was driven by consideration of what data were already available (i.e., the importance of the topology has generally not been appreciated), RAP has the potential to mold itself more closely to the problem at hand. Thus, the development of a topology of RAP applications and corresponding RAP guidelines could make a significant improvement in the use of RAP.
A second principle based on experience - which creates one of the rationales for linking qualitative methodologies to nutritional surveillance - is that the infectious disease model of surveillance is usually not the most appropriate one for nutritional surveillance. Under this model, which is best exemplified by CDC's system of weekly morbidity and mortality reports, a continuous system of reporting is put in place that will alert public health authorities to the location and magnitude of outbreaks of disease, so that well-defined methods of containment and control can be activated. Thus, an ongoing system of information leads to decisions and action in a rapid and continuous fashion.
With the exception of nutritional surveillance for timely warning and intervention, this model has, if anything, misled the design of nutritional surveillance systems. The reasons have to do in part with: 1) the complex etiology of chronic PEM, involving two proximate causes (inadequate nutrient intake and disease) and a myriad of context-specific contributing causes; 2) the equally complex decision-making processes at the policy level on a myriad of policy matters that bear on PEM; 3) the institutional responsibility for monitoring and acting upon PEM is not nearly as clear-cut as with infectious diseases; and 4) in the case of chronic PEM it is clear that changes in population nutritional status do not occur with nearly the rapidity seen in the disease outbreak model. In principle this should remove the rationale for a continuous surveillance system, and replace it with one that reports on a periodic basis of years rather than weeks or months.
translation of nutrition information into decisions and action
has more demanding information requirements, involves more
obscure decision-making processes, and is more protracted in time
than in the acute infectious disease model. Seen in this light it
is not surprising that nutritional surveillance of the simple,
nutrition monitoring type is hard-pressed to show evidence of
positive impact on decisions, action or nutritional status,
outside of its important role in political sensitization. It is
of interest to note that efforts to develop principles for
surveillance of chronic diseases are encountering similar
difficulties because these diseases possess many of the
properties described above for PEM.
Timely warning and intervention systems (TWIS)
Timely warning and intervention systems, abbreviated TWIS, are relevant in those situations in which natural or man-made events create a threat to household food security on a recurrent basis, beyond that seen in response to seasonal fluctuations in normal years. In such cases it makes sense to develop a system that can alert decision-makers to a possible food crisis with sufficient lead-time to permit decisions to be taken and interventions to be mobilized to ward off a disaster. The minimal requirements for designing a TWIS, therefore, are:
1. A system of simple, predictive indicators that can be collected locally and transmitted to appropriate levels in a timely fashion. Note that this typically does not include nutrition status indicators (see point 4 below).
2. Agreement between technicians and decision-makers concerning the choice of those indicators, their interpretation and the cutoff levels to define action.
3. Pre-identification of intervention options suitable for a variety of circumstances, that can be rapidly mobilized in response to the information system.
4. A set of fail-safe indicators that provides at least late indication of those areas in need of intervention but which were missed by the earlier indicators for some reason. This may include nutritional status indicators, among others.
5. A clear decision-making algorithm with tight bureaucratic integration with the information system.
In designing and operating a drought-related TWIS in Central Lombok District of Indonesia qualitative methodologies have been employed at several stages [10, 11]. For instance:
1. They were used to decide which of the existing agro-meteorological (ag-met) indicators collected by extension workers and others might be the best predictive indicator of food crises. In order to identify the best predictor, each of the potential ag-met indicators was analyzed statistically in relation to the occurrence of food crises based on historical experience.
2. The fail-safe indicators used in Indonesia were also developed through qualitative interviews with local respondents. In this case the indicators were based on household food consumption behaviours, namely the number of days in which rice was not consumed or in which wild roots were consumed over the previous five days.
3. A third use of qualitative information in the Indonesian case is seen when the district officials are alerted to an impending food crisis in several areas based on either the early or the late indicators. The first response is typically to dispatch a team from the district level to the problem areas to confirm the existence and the context of the crisis, in order to decide among several intervention options.
4. Finally, qualitative approaches have more recently been applied in Central Lombok to evaluate the extent to which this system has actually triggered decision and action in recent years .
Nutritional surveillance for policy and planning
Nutritional surveillance for policy and planning is the most difficult type to implement well and, the most difficult type to demonstrate the impact of information on decisions and action. In this section two examples of outputs from surveillance systems in this category are presented to illustrate how qualitative methodologies have been applied in the past and where considerably more work is needed in the future.
For a national sample survey conducted in Costa Rica, results of the prevalence of low weight-for-age in children was stratified by the father's occupation. In this study there was a clear gradient in prevalence rates from top to bottom, with labourers in sugar-cane and banana plantations showing the highest rates of malnutrition. In response to these results a three-month, focussed, ethnographic investigation was launched to identify the specific behavioural or environmental factors responsible for the high rates on these plantations [2, p66]. This study implicated poor water and high food prices as the likely reasons for this finding, and led to legislative action.
Results from a national sample survey of selected provinces in Kenya in 1978-1979  showed the prevalence of stunting to be 50% greater among children in households growing hybrid maize among those households cultivating less than 1.5 hectares. No such association was found among households cultivating more than 1.5 hectares.
Since results such as these from Kenya are typical of the surveillance outputs based on sample survey data it is worthwhile to examine in detail the possible policy recommendations from such outputs. In a well-reasoned discussion of these results the authors stated:
One possibility is that the local maize varieties are resistant to drought and yields are less variable, even if lower on average, than the hybrid varieties. Early adopters of the new varieties may not in fact have achieved the expected increases in yields due to unfamiliarity with the techniques or unavailability of other inputs required, or to failure to purchase new hybrid seeds for subsequent plantings, or they may have reduced acreage planted to maize by a greater proportion than the increase in yield and thus had less of their own produce to consume ... This somewhat surprising association, if causal and if confirmed by later surveys, would have important implications for crop policy. Hybrid maize varieties are being promoted as higher-yielding and (presumably) more profitable for farmers, and it is worrying if they are in fact leading to a deterioration in child nutrition. [13, p.306].
primary conclusion in this report, based on this and related
findings, was that further investigation was required in order to
better understand the reasons for the observed differences in
nutritional status among different agro-ecological classes of
farmers. Such a conclusion is actually the only responsible one,
based on the available information. However, the general
experience has been that, contrary to the Costa Rica example,
such second-stage investigations are seldom launched in practice
and the institutional capacity for doing so is often weak or
non-existent . This may be one of the reasons why the loop
connecting information, decisions and action has not been
completed, since planners and policy-makers are too often left
with a series of suggestive associations and no clear,
justifiable recommendations for action. This represents a major
area in which RAP-type investigations could make a significant
contribution to nutritional surveillance, in providing a timely
and feasible mechanism for answering the context-specific
questions raised by surveillance outputs.
1. It is for these reasons that the title of this paper and the preferred term as used in the text is qualitative methodologies rather than RAP. The former is considered a more generalized concept in its methods and applications, and is better suited to describe the qualitative approaches discussed here in connection with nutrition surveillance.
1. WHO 1976 Methodology of Nutritional Surveillance: Report of a Joint FAD/UNICEF/ WHO Expert Committee. WHO Technical Report Series No. 593, WHO, Geneva.
2. Mason JB, Habicht J-P, Tabatabai H. Valverde V. Nutritional Surveillance. WHO, Geneva., 1984: 66.
3. Tucker K, Pelletier D, Rasmussen K, Habicht J-P, Pinstrup-Andersen P. Roche F. 1989 Advances in Nutritional Surveillance: The Cornell Nutritional Surveillance Programme 1981-1987. Cornell Food and Nutrition Policy Program, Monograph 892, Ithaca, NY.
4. Habicht J-P, Pelletier, DL. The importance of context in choosing nutritional indicators. J Nutrit 1990;120(supp 11): 1519-1524.
5. Pelletier DL, Msukwa LAH. The role of information systems in decision-making following disasters: lessons from the mealy bug disaster in northern Malawi. Human Organiz 1990; 49(3): 245-254.
6. Habicht J-P. Nutritional surveillance for policy and planning. Pew/Cornell Lecture Series, Cornell Food and Nutrition Policy Programme, 1990.
7. Habicht J-P, Lane JM, McDowell AJ. National nutritional surveillance. Federation Proceedings 1978; 37(5): 1181- 1187.
8. Arnauld J, Alarcon JA, Immink MDC. Food security and nutritional surveillance in Central America: the need for functional approaches. Food and Nutr Bull 1990; 12(1): 26-33.
9. Thacker SB, Berkelman RL. Public health surveillance in the United States. Epidemiologic Reviews 1988; 10: 164-190.
10. Brooks RM, Abunain D, Karyadi D, Surnamo D, Williamson D, Latham MC, Habicht J-P. A timely warning and intervention system for preventing food crises in Indonesia: applying guidelines for nutritional surveillance. Food and Nutr Bull 1985; 11(2): 37-43.
11. Brooks RM, Habicht J-P, Williamson DF. Timely warning and intervention systems (TWIS) for periodic food consumption shortages: Experience from Indonesia. Cornell International Nutrition Monograph Series, No. 22, Division of Nutritional Sciences, Cornell University, Ithaca, NY, 1990.
12. Ralston K, Lubis D. A timely warning and intervention system for food crisis prevention: An update on the Indonesian experience. Mimeo, Cornell Food and Nutrition Policy Programme, 1989.
13. Haaga J, Mason J, Omoro FZ, Quinn V, Rafferty A, Test K, Wasonga L. Child malnutrition in rural Kenya: a geographic and agricultural classification. Ecology of Food and Nutr 1986; 18: 297-307.
14. Pelletier DL, Msukwa LAH n.d. The use of national sample surveys in nutritional surveillance: lessons from Malawi's National Sample Survey of Agriculture. Social Science and Medicine 1991; 32(2): 887-898.
Nutrition covers many sectors. We need to find the emic point of view. At the regional and local levels - how do we institutionalize nutritional surveillance.
The most useful process for educating decision makers had to be carefully identified and the choice of leaders is critical. Many researchers are more hesitant than necessary in revealing data. People who are encountering problems are eager to receive what we are developing. It is important to keep the information available and visible along the way.
The development of RAP requires sensitivity to decision-making.
paper is welcome and it reflects expertise in nutrition
surveillance that seems to reside in Northern countries.
It will be important to better explore and where
necessary strengthen national capacities for national
surveillance in developing countries.
The questionnaire syndrome
Case 1: Limitations of revealing relations (Philippine storage case)
Case 2: The leap frog problem (Mantaro Valley Project, Peru)
Case 3: Eliciting quantified data without the questionnaire (Potatoes in Nepal)
Case 4: Meeting a specific need (Identifying fallow periods in Bhutan)
Keeping the questionnaire in perspective
The cafeteria is now open: A diverse selection of methods
By Robert E. Rhoades
Robert E. Rhoades is Chairman at the Department of Anthropology at the University of Georgia.
This paper describes the author's progressive disillusionment with quantitative questionnaire methods used to evaluate rural agricultural programmes, and goes on to describe a diverse selection of participatory, qualitative methods that offer promise of more realistic and useful information about what farmers believe, their practices and their reasons for their behaviour.
The expense, time and invasiveness of quantitative surveys are referred to in a number of other papers in this volume, but not as specifically documented as here. Too often, when surveys have been done the reports are largely descriptive with too little emphasis on prescribing how the interventions could be improved. Such experience led to experiments in the health field with various qualitative approaches. One result was an UNU-UNICEF initiative that led to the publication of RAP Guidelines and to the many RAP applications in the health sector as presented in this volume. - Eds.
IN 1980 SENIOR ECONOMISTS and agronomists representing the international agricultural research centres gathered in Mexico at CIMMYT to discuss methods for conducting agro-economic, farm-level research aimed at generating appropriate crop technologies for small farmers1,2. The productive and very lively meeting focused largely on the tried and true, traditional research methods available to agricultural scientists concerned with linking technology with farmer needs. These research methods included questionnaires, farm budgets, on-farm trials, experiments, cost benefit analysis, yield extrapolations, various statistical tests and sampling techniques. As an anthropologist, a rare breed in the international centres back in those days, I was allotted five minutes to speak on anthropological or what the workshop organizers called "informal methods. " My brief paper, "Notes on the art of the informal survey," was met with a mixed reaction .
The primary concern of even anthropologically sympathetic agronomists and economists was how can such informal methods produce any degree of reliability, statistical proof, or generalizability? These were the very charges I feared most (as an uninitiated postdoc): that I was not a SCIENTIST! I had nervously anticipated this to a degree, and for that reason tried to cover myself by calling my method an "art" and not a "science."
I still basically endorse that distinction, my motive a decade
ago was in a great part to ward off the attacks of a group of
very prestigious but doubting economists and agronomists. Little
did I know that during these same years, many other experienced
workers in rural and agricultural development were beginning to
question the appropriateness of the traditional research tool
kit. Their voices were few and their methods or research tools
were described mainly in the grey literature (what
self-respecting journal would jeopardize their reputation by
publishing The Art of the Informal Survey!)3 The past
ten years, however, have brought what I see as a coming
revolution in methods from agricultural and rural development
research. In this paper, I want to reflect on what this new era
will bring and how agricultural researchers can benefit from the
changes. I should make it clear at the outset that this is not a
dialogue about how the earlier pioneers had gone wrong and are
about to be saved by the new guard. My purpose is much simpler:
explore the old in light of the new alternatives in methods and
point to common ground for practical action in solving problems
with rural people.
Contents - Previous - Next