Previous - Next
Section introduction
1. Re-tooling in applied social investigation for development. Planning: Some methodological issues
2. Adaptation of anthropological methodologies to rapid assessment of nutrition and primary health care
3. Qualitative and quantitative: Two styles of viewing the world or two categories of reality?
4. The role of qualitative methodologies in nutritional surveillance
5. The coming revolution in methods for rural development research
This section deals with
those papers which opened the RAP Conference and were intended to
define the parameters of the methodological approach and set a
stage for further discussion. The role of qualitative research in
international development is defined and analyzed by Cernea and
Pedersen and more specifically in the field of nutrition by
Scrimshaw and Pelletier. Rhoades' paper, which although not
presented at the conference was specially requested by the
editors for inclusion in this volume, adds additional fuel to the
growing debate on the usefulness and cost effectiveness of much
of the past and current studies which are intended to inform the
development process.
These
papers provide a review of the foundations upon which RAP and RRA
rest. They reflect the excitement of an emerging field as well as
the need for caution against arrogance and the need to maintain
professionalism and scholarship as guiding beacons for further
developments in these qualitative fields.
Twin changes: In planning of projects and in social research
Twin epistemological risks in RAPs
The place of RAP within broader research strategies
Endnote
References
Additional reading
By Michael M. Cernea
Michael M. Cernea is the Senior Advisor of Social Policy/Sociology of the World Bank.
This paper opened the RAP conference, providing an overview the major issues for which the conference was convened. Methods and techniques are multiplying, as are their application to a broadening range of issues. The author groups Rapid Assessment Procedures (RAP) with Rapid Rural Appraisal (RRA) providing a scenario in which the former is viewed as an outgrowth of the latter. As the conference progressed, the more popular view was that the two groups of qualitative methodologies, in fact, had independent origins often based on different programme problems and needs. Regardless, RAP/RRA is viewed here as a new force changing the development planning process and development activities at community level. As a RAP practitioner, the author bolsters his points with examples from his own experience and from colleagues' to show both the power and risks of RAP. This paper also cautions that these approaches, while providing both useful information and a significant change to development perspectives and activities, require great care in terms of professionalism, training, and quality of use. - Eds.
THE EXPLOSIVE GROWTH and diversification of rapid assessment procedures (RAP) over the last eight to 10 years has opened up new avenues for social investigation in the service of development work. If we try to take stock and synthesize what has happened in the 1980s in RAPs1, we can distinguish at least four main processes and trends:
First, fast repertoire enrichment: new and imaginative procedures are invented and added to an already respectable inventory;
Second, application of RAP in new sectors and subsectors through content-adaptation and cross-fertilization. What initially was to serve syncretically all of "rural development" and was called "rapid rural appraisal" (work by IDS, especially Chambers), has lately been paralleled and reinforced by specialized systems of procedures designed for other sectors. Such other sectors are the primary health care and nutrition sectors (work by Scrimshaw and Hurtado), the social forestry sector (Molnar and others); irrigation projects (Chambers and others); micro-ecosystem assessment (e.g., work on rapid mapping by IIED), etc., I am sure that additional sectors will follow, or perhaps could already be listed. Cross-fertilization of experiences from different sectors gives birth to new procedures.
Third, geographic broadening in both the elaboration and application of RAPs. Work that started at Sussex in England, has been carried forward by the creativeness and organizational efforts of the Khon Kaen school in Thailand, has travelled on the wings of the UN University to Latin America and other places, was tried and enriched in Kenya by IIED and local researchers, and is right now gaining great momentum in India. Of course, these are not simply geographic expansions. They also are tests of cross-cultural adequacy, resulting in broader validation, refinement of methods and increased diversity.
Fourth, last but not least, there is a growing shift from technique to substance. When practiced correctly and creatively, these rapid techniques often develop the capacity to carry their practitioners further, to a new substantive direction: participatory data generation techniques increase the opportunities for participatory programmes; micro-ecosystem and natural resources assessment, done best by outsiders jointly with the users themselves, bring home sustainability goals and motivate action for better resource conservation; indigenous knowledge harvested thanks to RAPs gives unanticipated directions to programmes in agriculture or health care.
Reversing a known dictum, perhaps we could say that if rapid social assessment procedures are a medium, then, in this case, the medium changes the message. Quite often, the message is not only new information, but action itself. Incremental gains in awareness and knowledge, through participatory information gathering and direct rapport, are stimulating activities not envisaged otherwise.
The
international conference on RAP in health related programmes is a
testimony to the major trends in this field, as well as a working
laboratory for exchanging experiences and reflecting on
methodology. We need to take stock as we progress. At the
conference there is an opportunity to analyze critically our new
tools, their strength and weaknesses, with warm hearts and cool
heads. There are serious technical, epistemiological, and ethical
questions that demand answers if the RAP field is to
continue its growth and fully realize its potential impact. The
conference is designed to explore such questions. While the
specific emphasis of the conference is on health projects,
the methodological problems underlining this domain are largely
common to all sectors in which such procedures are used. Papers
by Susan Scrimshaw and Duncan Pedersen focus on the specifics of
health assessment procedures. Therefore, only several selected
issues of rather general methodological relevance are introduced
here.
The first change that I
want to point out is that a decade of work on RAP has yielded not
only piecemeal data findings, but something more important: a
compelling demonstration of RAP's potential for changing and
improving the planning of development. By cost-effectively
providing knowledge about the actors of development themselves,
RAPs can increase planners' ability to put people first in
development projects. And putting people first means often a
reversal in many development projects.
The second change, perhaps not less significant: a decade of RAP work has launched some social sciences on a path of methodological re-tooling.
To elaborate on these twin significant changes, it is useful to remind ourselves how the interest in rapid rural appraisal evolved. This interest was spawned, in fact, by a crisis-sized shortage of adequate social knowledge in development interventions. The expansion of development aid in the 1960s and 1970s multiplied programmes often conceived without considering people, and far distant from the places where these programmes were to be implemented. Social information tended to be incomplete, unreliable, superficial, subjective, wrong - thus misdirecting such projects in many ways. Time and again, project ill-match after project ill-match, failure after project failure, were traceable to the dearth of good information on the local society (even though this was not the only cause of failure).
I know firsthand about this acute need for social knowledge on project areas and populations because I work at the World Bank as a sociologist - and I still have to address the shortage and need for social investigation day in and day out (Cernea 1990). Yet, just prescribing more conventional social science research was largely inoperative, since practical project activities could hardly wait for research designed to last several years.
This was the gap into which the proponents of shortcut, rapid rural appraisal procedures stepped. Most of these proponents were social scientists with applied development orientation. They realized that business as usual in social investigation could not live up to the day's challenges. They wanted also to avoid the double impasse of either "quick and dirty" or "long and late" research.
Today, a decade or so later, the results of searching for shortcuts to knowledge are tangible. A broad arsenal of new techniques for data generation on the life, behaviour and production patterns of individuals, households and communities has been invented, tested, refined, and disseminated. Their summary listing would include: novel forms of direct observation and participant observation; researcher's participation in the studied activity; semi-structured interviews; group interviews; focus groups; mapping; aerial photographs; group walks; diagramming; quantifications; ranking; group reading of satellite imagery; simulation games and role playing; sondeo techniques and small team investigations; imaginative selection of key information as in chain-interviews; procedures for eliciting the subject's assessments; self-definition; etc. The list is far from exhaustive, yet it testifies to the creativity, intellectual excitement, and diversity that characterize the work in this area.
The effects of the creative search for new approaches on development work are manifold. Not only was additional knowledge generated for many development interventions, but the fallacious argument that a development intervention must proceed even without adequate knowledge - because, allegedly, "it couldn't obtain it anyway" - was voided. Rapid assessment procedures - certainly not alone - are apt to produce knowledge within reasonable time-spans and at costs lower than conventional procedures.
It is significant to note that some major development agencies - among them, the World Bank, USAID, ODA, etc. - have started to use RAP in their work, at least to a certain extent. For the World Bank it has been part and parcel of a broader and longer-term effort to introduce knowledge from non-economic social sciences in designing project strategies, and to promote the use of sociological/anthropological investigation methods for generating the social information needed in preparing, supervising and evaluating projects (Cernea end Tepping 1977, Cernea 1985,1989, Casley and Kumar 1988, Murphy 1988, Salmen 1989). The problem is often to correct, improve and enhance the spontaneous practices of development agency staff used in "quick and dirty" field assessments, with trained skills in using correctly tested RAPs.
For the social sciences - I refer primarily to sociology, cultural, medical, or economic anthropology, and their applied domains - the sets of rapid assessment procedures represent a consequential re-tooling process, a retooling process which does not reject or abandon their traditional methods and techniques, but complements and enriches them. RAP represents a new generation of flexible knowledge-producing instruments, which increase the capacity of social sciences for applied research. Social sciences are enriched also by the refinements brought to many pre-existing, time honored anthropological or sociological field techniques.
Economists have demonstrated that advances in natural science knowledge reduced the cost of technical change. It was also hypothesized that advances in social science knowledge would reduce the costs of social/ institutional change (Ruttan 1988). The recent progress in crafting RAPs as new and penetrating tools for understanding social processes is clearly one of the ways to reduce the costs of using social sciences in development projects.
I note also another outcome - which we could call a multiplier effect. It goes beyond the techniques of social data gathering and consists in the conceptual load that these techniques inevitably carry into project practice. A considerable body of notions, propositions, explanations, knowledge from general sociology, medical and general anthropology, political science, and other social science descriptions, knowledge in which such rapid techniques are embedded, are introduced into development work. Such substantive understandings improve work for inducing development. For those who argue that the design of development programmes should be guided not just by econocratic or technocratic approaches, but by better knowledge about people's social organization and culture, this method-linked entry point for social science knowledge is an important added value.
Sometimes, one vivid case can capture well an entire process, as we all know from RAP field work. This re-tooling process is illustrated with the story of a medical anthropologist (Edward C. Green) and of his intellectual journey from lengthy statistical sample surveys to rapid procedures, and then to a balanced combination of them. Green had received a two-year USAID contract to conduct survey research in Swaziland, within a waterborne disease control project. Once in the field, however, he soon developed doubts about the appropriateness of a lengthy sample survey for the project. "The information I was after," he wrote later, "related to sensitive areas such as toilet behaviour, personal hygiene and health beliefs ... (But) the impersonal, pre-coded questionnaire typical of survey research is notoriously deficient in eliciting information of this kind.... (Green 1986)." After a preliminary ethnographic reconnaissance, Green decided to modify the pre-established research strategy and use key-informant interviewing and participant-observation. Specifically, he purposely selected two special groups, not representative as a cross-sample of the population at large, but nevertheless highly relevant for his inquiry: first, he set his sights on Swaziland's traditional healers and carried out an informal study of health beliefs and behaviour among them and among another sub-population - the Rural Health Motivators (RHM) - expecting that as "community insiders" they would be likelier to give "candid and truthful replies to sensitive and even embarrassing questions" about health beliefs and behaviour. Indeed, within six months of his arrival Green was able to collect general socio-cultural information, plus specific health-related findings, to guide the components of the USAID disease control project.
Later, when
his contract called for another survey study, the researcher
decided that at that stage the survey was desirable. In
designing it, he used the insights gained from his prior rapid
investigations. The combined result was an ethnomedical study
that integrated - note the medical vocabulary! - "the
quantified bones of the survey with the qualitative flesh"
of the quick assessment studies. Green's full account (Green
1986) of his subsequent studies shows how a wise combination of
rapid and not-so-rapid but statistically-based studies led to
significant improvements in the public-health network. One
tangible result was a programme of cooperation between the
indigenous and biomedically trained health practitioners within
the government's disease-control campaigns against cholera and
diarrhoeal diseases.
By definition, shortcuts
are not beaten paths. They may be strewn wit obstacles, confront
slippery slopes, and hide methodological dangers as shortcuts to
social information. Those who practice, or recommend, rapid
assessment procedures must be aware of the epistemological risks
involved in their use or misuse and issue clear warnings.
The methodological problems or distortions that tend to occur in RAPs can be divided, in my view, into two categories. I suggest this grouping for discussion during our symposium as a way to illuminate the nature of the methodological problems we are facing and the kind of precautions that must be taken to minimize them.
The first category of risks results from the limitations intrinsic to the rapid procedures themselves as investigative tools (such limitations vary from one procedure to another) and from the specific biases that may develop during their use as individual procedures.
The second category of risks are extrinsic to these procedures as such: they typically tend to result from giving RAPs an improper contextual place or weight within the research strategy used for project planning and e valuation.
In the first category I would list primarily the problems of (a) accuracy; (b) representativeness; (c) cultural inappropriateness; (d) subjectivity.
It should never be forgotten that on the birth certificate of RAPs, a tradeoff was inscribed - the trade-off between research duration, on the one hand, and the quantity and accuracy of collected information, on the other hand. This trade-off is not an original sin. It rather is the original blessing with which these procedures have entered a world in great need of knowledge but always short of time. Yet speed of acquisition, even when proudly worn on the sleeve, has cognitive costs, not only benefits.
Robert Chambers has aptly conceptualized (1985) two principles these procedures must follow for rationalizing and keeping under control such cognitive costs. These two principles are "optimal ignorance" and "appropriate imprecision." Indeed these are, also in my view, acceptable and necessary principles, with an in-built balance coming from their paradoxical wording.
In practice, however, it is the investigator's judgement call as to what he can ignore and how much imprecision he can appropriately tolerate, without self-destroying penalties. How rapid isn't too rapid? The question to be asked may be: when is rapid not slow enough?
The shortcut methods for projects have evolved in an "intermediate" zone as a middle level methodological solution. They came to occupy the ground between the two extremes of a wide range of possible methods: the highly formalized (survey and other) research methods, rather solid but often disqualified in project work (for reasons of costs, length, and indigestibility), are at one extreme. At the other extreme are the highly informal, bordering on the irresponsible, amateurish manners in which many project appraisers, planners or supervisors used to patch a smattering of data and fugitive impressions, knitted together with their own ethnocentric assumptions credited as hard facts. The shortcut methods come in-between because they introduce formality in a reasonable degree, accuracy strictures, verification (triangulation, etc.) and represent an adaptation of legitimized social science procedures and codified personal experiences. If, however, the shortcut becomes too short (of time or of data), and the "rapid" slides into too rapid, then the trade-off is violated. The risk of inaccuracy, going beyond "optimal ignorance" and allowing in- "appropriate imprecision" to creep in, can snowball, compromise the good label of "rapid assessment procedures" and mislead the project they are supposed to guide.
Mutatis mutandis, sacrificing the representativeness of statistical samples in favour of purposive selection of informants is also a temptation not devoid of perils. Determining the quantity of information required to minimize "sampling" error in non-random informant selection is largely left to the responsible judgement of the RAP users.
In other words, rapid assessment procedures run the risk of sliding into little more than the quick and unreliable amateurish manner of misgathering social information, that they wanted to replace in the first place. It is not an abstract risk: I have seen it at work, wreaking havoc. And I have seen it lurking in the pages of some glossy consultant firms' field reports, marketed now under the newly fashionable RAP label. This is why it is important to maintain a sharp awareness of the methodological hurdles present on the shortcut path.
We could ask: is there also, on the other hand, the twin risk of RAP work sliding towards the opposite extreme, taking on overly complex procedures and losing flexibility and promptness?
Again, this is the sort of question that requires a mutual professional answer. My personal answer is - yes, there is this risk too, but it is not the principal one.
Actual time allocation for a rapid investigation cannot but vary with the variables - more or less complex - that need to be investigated. No rigid or universal recipes for time-quotas can be given. What is adequately rapid in one case, may not be slow enough, or conversely not rapid enough, in another. The healthy way of cutting time is not watching your wrist watch or calendar, though, but putting (initial) time in thoughtfully picking the critical variables, in preparing the most knowledge-producing questions, in discovering proxies.
For instance, after much research on poverty, it was realized that fairly accurate shortcut poverty-level assessments for a certain area or population and for the dynamic of the actual incomes of the poor can be obtained by focusing on two basic variables: the hourly/daily wages for labour and the free access to social services (like primary education, basic health care, and family planning). Other researchers may add one or two other basic variables (e.g., terms of trade for agricultural products) - but it all remains in the domain of relatively easier feasibility (and without pretending to replace in-depth, specialized poverty-measurement studies).
Another example can be given in connection with the need to assess rapidly the self-organization of communities, with their networks of social interaction, propensity or non-propensity for volunteer work, local authority systems, family kinship systems, rights and obligations, etc. Lengthy studies are easier to imagine than finding shortcuts for getting to the core of such social systems, but the latter are the ones imperatively needed.
If I can relate some experience from my own field work, I found one simple question that infallibly opened many doors to understanding various community systems of social organization. More importantly, this question is easy to ask, easy to understand, and is always quickly and eagerly answered. I ask: "When a family's hut and crops happen to catch fire and burn down, how does the village help out that family?" The answers to this question usually erupt and flow freely: I learn whose traditional obligation is to offer shelter (kin or neighbours), whose is to supply food, for how long, how the village leaders, healers, priests, intervene, who is to contribute wood, grasses, or free labour for quickly rebuilding the burned hut, who buries the dead, etc. The social structures, kin obligations, and the mutual-help-networks unfold like on a real-life story-screen. From this question it is not difficult then to move to other themes: for instance, to offer reflection to my informants the question - whether or not similar cooperation would be forthcoming if the village were to build (or repair) not a hut, but a small irrigation system? Or a health room? My questions, I assure you, worked in many countries, helped me in different cultures - and I invite others to test them out for themselves.
Another methodological risk with rapid procedures stems from the culturally coloured differences in the perception of the same thing by different respondents. This may often occur, for instance, in rapid health assessments of morbidity rates. Answers to questions like "Were you ill during the month (or year, etc.), and if so what was the kind and length of each illness?" will depend on the respondent's perception of illness. It was observed (Srinivasan 1989) that in a socioeconomic context where the poor, or women, or some other groups, do not perceive or are culturally conditioned not to admit an illness unless it is sufficiently serious - while the rich tend to be hypochondriacs - one cannot rely on an intergroup comparability of morbidity rate assessments derived from such responses; and there are going to be substantial differences between morbidity estimates through informal interviews and morbidity estimates rated through frequent clinical examination of the same respondents.
To fight off such kinds of limitations in RAPs, we need to address the paramount problem of professional training for the practitioners of RAPs. Briefly stated:
• first, that the use of RAP requires training, training in participatory use of RAPs. The advantages of RAPs can be invalidated if the users mistake RAPs' informal nature for an unbounded permission to "play it by ear."
• second, that social scientists - anthropologists, sociologists, political scientists, social psychologists, etc. - already trained in their disciplines' basic field methods, have a comparative advantage in using RAP and guarding against methodological distortions.
• third, that the methodological re-tooling of applied social research for development projects through RAPs is aiming precisely at broadening the use of social investigation by a larger community of professionals than the trained social scientists. It is crucial that planners, economic analysts and technical specialists who design development projects recognize the potential and accessibility of this broadened social research and absorb it gradually in their tool kit. This, however, is a long term goal, and presupposes changes in many university curricula.
Before concluding, a few
words about the extrinsic circumstances that may distort
the use and benefits from RAPs in development projects. These
circumstances refer to the place and weight
attributed to RAPs in overall strategies for project planning or
evaluation.
It is incumbent upon RAP practitioners, I believe, to warn that RAPs are not, and cannot tee, a universal cure for all the gaps of social information and that rapid appraisals are not a substitute for long term basic research methods and procedures. It has to be said explicitly and loudly, that in many cases the design and strategy of development projects cannot be sound without the benefit of long term, non-shortcut, longitudinal, academic, old-fashioned types of social research.
One example: starting from 1975-76, ICRISAT has developed a series of village surveys in India that provide time-series information on up to ten years in three agroclimatic regions of India's semi-arid tropics; it collected longitudinal information approximately every three weeks on all household transactions (consumption, production, investment). This data set may be the only one in existence that allows the measurement of net farm profits over many years and therefore of farm profit risk (H. P. B. Binswanger and M. R. Rosenzweig, personal communication, 1990). Of course, no shortcut procedures are able to or should substitute for this kind of study and cannot supply similar knowledge. Any institution or research strategy that would depend on RAP for similar depth or precision on RAP would entertain gravely misplaced expectations.
Social researchers would, in turn, be methodologically ill-advised to resort to RAP when assisting the planning of certain projects or project components which need full sets of data on the senture population affected by a project.
Take for instance, project planning for involuntary population displacement from a dam site, and its relocation. I would be more than a little worried if such planning would be expedited through low-cost, informal, shortcut procedures. To be done right, it requires full censuses, house by house inventory of lost assets, determination of joint ownership on specific natural resources, etc. Unfortunately, certain dam planners display unexpected enthusiasm for shortcut procedures in just such cases, preferring imprecision when imprecision is simply not tolerable. The social researchers practicing RAP are obligated to reject the misuse of RAPs in such projects and in research strategies that demand alternative approaches. This does not mean, however, that a combination of procedures cannot be constructed, including RAP as a complementary approach.
Another example, from medical anthropology this time, refers to the study of AIDS. There are recent valuable attempts to put the research on behaviour patterns that may lead to AIDS on the path of RAPs. If epistemological strictures are observed, these attempts may produce useful and urgently needed results. But many things cannot be learned this way. Take for instance, the case of a recent ethnographic synthesis on practices of male circumcision among 409 ethnic groups of African populations, which also analyzed the relationship between circumcision practices and the rates of HIV seroprevalence based on recent statistics. The authors of this study (Bongaarts, Reining et al. 1989), rightfully hold that such analyses are beyond the claims of RAP. But complementarily among the two approaches is compatible and desirable.
To conclude, I would emphasize that I see great promise for strengthening the cognitive power and contribution of RAPs and that avoiding the misuse of RAPs is possible too. These are two facets of the same process of methodological retooling in applied social research for development.
Towards
this re-tooling process in social investigation, this
international conference on RAP will make a very valuable
contribution.
The author expresses his
gratitude in particular to Robert Chambers, whose "letters
from the field" are bringing an unending stream of ideas and
information about rural rapid appraisal procedures in-the-making;
the academic community - in particular Susan Scrimshaw, Scott
Guggenheim, Augusta Molnar - engaged in a continuous effort to
enhance the use of rapid field-assessment methods in the service
of development projects. Thanks also are due to Gracie Ochieng
who rapidly processed this paper at odd hours.