Proceedings of RRA Review
Part 1: Presented Papers
Part 2: Discussion
SUSTAINABLE AGRICULTURE PROGRAMME
Proceedings of 2nd Joint IDS/IIED RRA Review Workshop
Part 1: Presentations:
Diagrams: General Theory and
Profiles and Matrices
Diagrams and Cartoons
Diagrams for Training
2. Aerial Photographs
Aerial Photography and Household
Studies in Kenya
Aerial Photographs in Rapid Land
Resource Appraisal, Papua New Guinea
3. Interviews and Groups
RRA Interview Schedules within a
Collaborative Research Project in
Farmer Groups and ITK in Benin
Focussed Groups in Ethiopia
Some Difficulties in Training for
Wealth Ranking in Sudan
Preference and Direct Matrix Rankings,
Ranking of Carpentry Skills
Using Ranking in Training of Field
Rapid Appraisal Trial, Mbeya,
Susan Rifkin and
Rapid Appraisal South Sefton
(Merseyside) Health Authority
6. Participatory Approaches
From the Ground Up and Participatory
RRA in Kenya
Diagrams for Participatory RRA
Hugh Annett and
John Thompson and
7. Monitoring and Evaluation
PBME and Rapid Rural Appraisal
Using RRA to Evaluate NGO Projects
Part 2: Plenary Discussion Topics
Notes on objectives
Notes of aerial photo session
Notes on groups and group interviews
Discussion on health
Discussion on M&E
Notes on participation
Notes on ideology
Notes on the dangers of RRA
Notes on training in RRA
This special edition of RRA Notes comprises the proceedings of a
recent two day RRA review Workshop, jointly organised by the
Institute for Development Studies, Sussex, England and IIED, and
held at IDS in June 1989. This was the second in a series, the
first having been reported in RRA Notes 1. It was attended by
some 35 participants from NGOs, universities and development
The workshop was organised into two distinct sections. In the
first a number of presentations were made of recent findings and
new developments, covering the topics of Diagrams, Aerial
Photographs, Interviewing and Groups, Ranking, Health,
Participation as Theme, and Monitoring and Evaluation. Brief
summaries of these presentations are reproduced here in Part One
and readers wishing to receive full versions should write
directly to the authors. Each of the seven sections was followed
by plenary discussions: for the purposes of this report all have
been relocated in the appropriate discussions that came after all
the presentations. The second part of the workshop was conducted
entirely in plenary after groups of participants had identified a
list of issues crucial to the future of the field of RRA. The
full discussion is reported in Part Two. Several points should
be highlighted, particularly those relating to present
limitations and future dangers:
* we must avoid RRA becoming the new orthodoxy there is great
danger in the kind of statements implying that RRA 'should be
used wherever possible';
* we must not oversell RRA beware of excessive faith. RRA
could very quickly be killed by too much money and too rapid
* great care is required in training good RRA is not easy,
and requires close quality control and long follow-up;'
* RRA may be quick, but the development process is still long;
* self-criticism of methodologies used should be rewarded it
adds greatly to the credibility of the work;
* focus on principles, not labels. The very diversity of RRA
is its strength, it should not be pinned down, but be
developed separately and according to individual
institutional and locational needs and conditions. The
principles and rationale should always shine through.
We would welcome comments of any length on these Notes,
particularly on contentious points and key issues that may have
remained untouched by the two day review.
Jules N Pretty
Diagrams: General Theory and Practice
What is a Diagram?
A diagram is any simple, schematic device which presents is
information in a readily understandable form.
There are many types of useful diagrams, including maps,
transects, seasonal calendars, historical profiles, decision
trees, activity profiles, venn diagrams, histograms, graphs, bar
diagrams, decision trees etc.
Why Use Diagrams?
* Diagrams can capture and present information which would be
less precise, less clear, and much less succinct if expressed
* Diagrams are shared information which can be.checked,
discussed and amended, thus they create consensus and
facilitate communication between different people and
* The constructors of diagrams must continually be asking
questions during the process: these questions are more open-
ended than in formal surveys.
* The act of constructing
extremes in space (e.g.
community) and in time
between years). Helps to
a diagram forces exploration of
to the periphery of a village or
(e.g. unusual events within and
discover surprising and unexpected
* Helps in the development of interviews.
* Rural people understand, and can contribute to, diagrams.
* Illiterate people can understand diagrams.
* Sometimes a humbling experience when diagrams are shown to
rural people and are found to require many changes.
* Drawing diagrams is fun.
Where and When to Draw
best on the ground or on large pieces of
paper, so that the diagram is shared with
* In the field:
all; second best in notebook, because
information is private.
* In the workshop: on large pieces of paper or on overhead
How to Draw
* Drawing diagrams is neither art nor technical drawing.
* Do not use a ruler: it encourages excessive and unnecessary
* If the diagrams is too pretty then the constructors may be
disinclined to change it when new information comes to light:
never be afraid to scribble over something that is incorrect.
* Be bold, clear and quick.
Jules N Pretty
Profiles and Matrices
Resources and Benefits Profile
We know that communities are not homogenous groups of people.
Likewise, households cannot be seen to be undifferentiated units.
Technological innovation, or intervention causing change, will
affect the rich and poor, the male and female, the old and young
in different ways. Two key areas in which people and households
are affected by change are in their access and control over (a)
resources used in production, and (b) the benefits derived from
production. Access is defined as the use of a resource, control
is defined as the capacity ultimately to decide about the use of
that resource. Diagrams I and II depict a way of recording this
information. The activities noted in the left hand column were
taken from a previously compiled Seasonal Calendar. The codes in
the two right hand columns identify relationships of access and
control both within and between households.
Decision Making: Matrices and Trees
Decision making is a complex process. It can be thought of as
having three stages, (i) Initiation of Discussion, (ii)
Discussion, (iii) Execution of Decision. There may be one or
more actors involved at any one of these stages. Diagrams III
and IV illustrate how to record this information using a matrix.
FA = AMe WA \}
7T~ &e IV
)ec~&r HGI~ PNk4iV
zfision M, : h"e & ,nt P4eys
MA- Me tAM W
FA = ,w ,6 A
44 -- MA A d40
PA = a le 4Adti-
,t. H: Lj tidl
HL -- iee) (-rA
Once again these show how patterns within and between households
can be understood and documented. This understanding has proved
useful in uncovering some of the social and economic
relationships surrounding productive activities that have been
the focus of project work.
The 'three stage matrix' can also be used in conjunction with the
familiar Decision Tree diagram. This to date has only been used
to conceptualise flows of decisions or to present semi-aggregated
information in a report. Diagram I gives an outline example of a
model which identifies critical points at which decisions have to
be made about a range of choices. In surveys in which this tool
has been used the patterns of (a) paths that households take at
any point, and (b) the actors involved in the decisions at that
point have tended to vary according to the resource position of
These descriptions are very short. They do not describe original
tools, rather adaptations of the work of other development
practitioners. Additionally no comment is made on the power of
the tools or of their use and abuse. What is presented is
therefore a brief introduction to ways of recording information
and an indication of a conceptual framework which gives rise to
the questions of intra and inter household dynamics that the
Ruth Alsop From mid-October:
Economist Research Fellow
ITDG School of Devlopment Studies
Myson House University of East Anglia
Railway Terrace Norwich NRY 7TJ
Rugby CV21 3HT
Diagrams and Cartoons
See RRA Notes No. 6, pages 26-29. June 1989.
Diagrams for Training
We recently ran a course in project identification for local
government officers in Nigeria, using the RRA philosophy,
methodology and techniques. These techniques included diagrams
such as maps, seasonal calendars, transects, historical profiles
and impact diagrams.
Course participants were shown an example of each of these
diagrams, followed by a discussion of how to construct them and
their usefulness in terms of project identification. Course
participants divided into their respective village teams and
interview pairings and constructed these diagrams on the basis of
their knowledge of their own villages. Finally, participants
presented their diagrams (prepared with marker pens on large
cardboard sheets) in plenary and this occasion was filmed.
Unfortunately, we made the mistake of not asking them to tell us
how they could make use of these diagrams for project
identification purposes. Nevertheless, the participants
thoroughly enjoyed the experience and were so proud of their
diagrams that they took them home.
The uses of RRA techniques for project identification purposes
It was intended that the seasonal calendar should be used to
indicate seasonal trends in activities and problems so that
critical bottlenecks or times for project intervention could be
The transect was used because we hoped that the problems and
opportunities listed would correspond to possible agro-ecological
The reason for including the historical profile was to highlight
the nature, direction and pace of development in the village
concerned, so that gaps within this pattern of development could
Although the primary purpose of the course was to improve project
identification skills, the fieldwork was also used as an
opportunity to get feedback from the villagers about various
facilities and services provided there by local government and
others. The impact diagram was therefore selected as a useful
tool for structuring and presenting the analysis of these
facilities and services.
Diagrams were selected from those available in the IIED Notes
(1988) and the Khon Kaen Conference proceedings (1985), on the
basis of their expected usefulness for project identification
As far as I know, these techniques have not yet been discussed
specifically in terms of the project cycle before. I think that
a discussion of when these and other RRA techniques could best be
used in the project cycle and what kinds of operational
information could be expected from them would constitute a useful
development in the RRA literature or at least one which would
interest me particularly.
All the above mentioned diagrams took less than an hour to
complete in the field. One or two individuals from each of the
five village teams were assigned to complete one or two diagrams
(not necessarily the same ones which they had chosen during the
teaching week) with different groups of villagers which varied in
size from two to more than a dozen people.
The historical profiles were done with a group of people that
included some elders. Similarly, the transect was done with a
group of farmers, while the impact diagram was done with people
who were directly affected by the service or facility in
question. The seasonal calendars, finally, were not done with a
specialised group of villagers.
Although the trainers did not accompany the participants on these
tasks, the reports show that none of the participants had
understood the intended operational purpose of the diagrams.
In addition, many of the diagrams were skimpily filled out and
others were treated more as pretty pictures than useful schematic
devised. On the other hand, there was a wide range in terms of
quality amongst the 30 participants.
Development Administration Group
University of Birmingham
PO Box 363
2. AERIAL PHOTOGRAPHS
Aerial Photography and Household Studies in Kenya
I have been working in Murang'a District in Kenya on problems of
capital and labour utilization and their relationship to on-farm
tree growing. Some of my research depends on the collection of
land-use data for individual households. Aerial photography,
coupled with ground truthing, seemed the best way to get this
type of information. I also wanted to use photographs during
household interviews as the focus of discussions about land-use
practices and changes which have involved trees. I had been
thinking that low-level aerial photography would have been ideal
for providing high resolution images of smallholdings. Larger
format approaches, particularly high altitude photography and
satellite imagery, would have been ill-suited for household work
because of their relatively poor resolution. They are best
suited for identifying gross landform features and land-use
I had originally planned to use what I had thought was an
existing set of low-level aerial photographs of the study area
which had been taken in 1985. Long after I had made a commitment
to carry out research in Murang'a, it became clear that aerial
photos would not be available. Even so, I thought I should at
least try to get some photos taken of the study area. This note
is intended to offer some guidance for others who might try to do
the same thing. It was surprising how relatively inexpensive
aerial photography could be (if the costs of early mistakes are
excluded), although this may simply be a function of the Kenyan
context, and may not be applicable elsewhere. With a little
planning though, I would think that aerial photographs could be
taken and incorporated into many Rapid Appraisal types of
exercises at relatively low cost.
The study area, Murang'a District, covers a wide range of
agroecological zones. The altitude ranges from around 600 metres
in the east to over 3,000 metres in the west, and the topography
is made up of an extensive ridge/valley system. Rivers generally
run from west to east. Land-use practices are especially
variable, depending on the agroecological zone and altitude. The
population density is similarly widely variable: on average, it
is around 250 people per km but in some areas is as many as
1,000 per km
Taking the Photos
A primary concern was that the photographs could be used for
identifying specific and complete smallholdings on the ground. It
was also concluded that sampling within a narrower agroclimatic
range would be more appropriate than broad sampling across the
district. Using earlier survey work which had been carried out
in the area for guidance, the transitional tea/dairy and
coffee/tea zone was chosen as the area to be photographed. This
zone roughly follows the 1800 metre contour.
Topographic maps at a scale of 1:50,000 were used for planning
flight lines. The 1800 metre contour was first outlined on these
maps and then 15 east/west transects were drawn across the
contour, at 2 km north/south intervals. By limiting the
transects to 5 km in length (instead of the earlier 15 km), the
study area was logistically far more manageable.
A Nikon FG camera with a 50mm lens was clamp-mounted out the open
window of a Cessna 152 and pointed vertically. The clamp was
nothing special, and was picked up in a photo supply shop in the
U.S. for around $30.00. One had to make sure, of course, that it
was tightly fixed to the plane's window frame. I tied a safety
cord to the whole apparatus to keep from losing it all together.
Flying at an altitude of 2000 feet above ground level allowed a
photographic ground coverage of about 12 ha per photograph.
Flying at a speed of about 80 knots and taking exposures at 7
second intervals gave photographs at a scale of around 1:12,000
with around 30 percent overlap. Enough photographs could be
taken with a roll of 36 exposures to easily cover the 5 km
transect. In all, it took about 10 hours of flying time to get a
satisfactory set of photos.
There were a number of drawbacks. I tried to fly at a constant
altitude which, on average, was around 2,000 feet above the
ground. Without a means of determining the changing altitude
above ground level and because the landscape changed quite
dramatically over short stretches, the scale of the photographs
varied between 1:10,000 and 1:15,000. I will have eventually to
calibrate the photos from ground measurements. If some sort of
sample were to be set up, the varying scale would introduce
problems of error: the photographic coverage was greatest where
the air-to-ground altitude was the greatest, introducing a bias
toward the lower agroclimatic zones.
Other drawbacks were a function of the need to fly particularly
accurate transects. There are a few major landmarks in Murang'a
District to ease this task, and flying and photographing for
lengthy periods of time was extremely exhausting. Even using
1:50,000 topographic maps, the regularity of some features of the
landscape (particularly Murang'a's endless series of ridges and
valleys) made it very difficult to get proper bearings.
Once a full set of satisfactory slides had been taken, ground
locations were plotted on 1:50,000 topographic maps. The fine
detail of the maps, and the good resolution of the low level
photography enabled the plotting of ground locations to within
less than 100 metres. The number of each photograph was marked
at its corresponding location in the map. Working copies of the
original transparencies were prepared as colour print
enlargements (5"x7") a scale of about 1:2000.
Using the Photos
To begin with, I have found the photos especially useful for
thinking about different land-use processes, particularly those
which would not be immediately evident from the ground. The
spatial arrangement of the shamba becomes really clear from the
air, and regular land-use patterns which are common amongst
smallholdings stand out. For instance, most buildings are built
close to the road. Arable crops are planted on the flattest
land, while permanent crops are planted on more steeply sloping
fields. Trees in woodlots are generally planted on the most
steep sites. Particular trees are often left or planted in
fields. Many smallholdings are demarcated with trees and hedges.
Valley bottoms are riddled with small drainage canals, and these
areas are intensively cultivated as well. Most of this I suppose
seems to make a lot of intuitive sense, and that's where the
photos are a tremendous help in confirming what seems to make
The photos also tend to reduce spatial biases which even the most
intrepid field worker can encounter on the ground. There is a
tendency, for instance, to walk along the contours and along the
ridges where there are paths, rather than longitudinally from one
ridge into a valley, across the river and to the next ridge. The
photos enable one to identify specific features of land-use which
are of interest, and then to find them on the ground. In this
sense, the search through the photos for the extremes the
smallest holdings, the largest holdings, the most heavily tree-
covered holdings, the most barren holdings, and so on can be
especially enlightening when one has a chance to interview
farmers. By walking the photographed transect, one can prepare
sketch maps of changing land-uses, and the agro-climatic features
which have contributed to these changes. This process as well
helps to identify holdings where land-uses do not seem to be
intuitive (for instance, where someone has planted trees where
you would expect them to plant coffee), and can help to identify
how households respond to resource constraints.
In household interviews, people seem quite comfortable with
interpreting the photos. There is seldom a perceptual gap, and
the thought of looking down on a shamba seems quite natural.
People in the area are generally literate and in many cases have
seen aerial photos before. During the land tenure reform
programmes of the early 1960s, aerial photos were used to
identify and register consolidated land holdings. During
household interviews, I have been using transparent sheets held
over the photos, to mark out and record the boundaries of the
holding, and to make a note of specific land-use practices which
are of interest. Interviewing has been essential to identify
which boundaries belong to whom because, although they are often
well-marked, even since the early 1960s there has been some
fragmentation as a result of inheritance and sale, and it is
seldom evident from the photos whose boundaries are whose.
In some respects, the taking of aerial photographs poses a number
of contradictions for Rapid Appraisal Practitioners. It was not
exactly cheap. It took time to get a good set of photos. It is
about as hands-off as one could get. But coupled with a
reasonable field technique and interviewing practices, I have
found them to be invaluable; where field workers have the
resources and the time to acquire photos, I think they could be
Oxford Forestry Institute
South Parks Road
Aerial Photographs in Rapid Land Resource Appraisal,
Papua New Guinea
The Southern Highlands Rural Development Project (SHRDP) (1978-
85) was a World Bank funded IRDP in Papua New Guinea with
programme activities costed at over US$30 million. In one of its
study areas, Upper Mendi, the first part of a formal, FAO-type
land evaluation was carried out by SHRDP to identify land mapping
units (LMUs) as a basis for agricultural and rural development
planning in the area. The next stage, never carried through,
would have been to use this physical inventory as the basis for
classifying the LMUs according to their suitability for specific
types of land use.
My objective was to evaluate this technical land resource
inventory in terms of how farmers resident in the area used and
valued their land resources: in short, to identify the social
overlay on the physical resource base which governs who has
access to what land. The fundamental unit of social organisation
in the area is the clan. The study took 3 months in total
(including archival work), based on 6 local clans. It occurred
towards the end of the funding period of the SHRDP. In this
brief summary I hope to show how aerial photographs were used
alongside a range of both RRA and more conventional research
RRA methods used
The following methods were all carried out in the field by a team
* walking clan boundaries with clan elders, using aerial photos
in the field for boundary mapping
* semi-structured interviewing with groups of men and women,
* direct matrix ranking of land resource preferences for
particular types of land use (sweet potato cultivation,
vegetable cultivation, pig foraging, collection of minor
forest products e.g. karuka nuts), using both local names and
LMUs (i.e. did the difference in the category used itself
affect the ranking/criteria in any way?)
* diagramming of inter-clan linkages and land disputes
* story-telling and oral histories
* local calendars
Complementarities between RRA and 'conventional' methods
There is an overlap between some of these techniques as 'RRA' and
what are in effect more conventional anthropological approaches
based on investigation of inter-clan linkages (taking
anthropological literature based on local fieldwork as a point of
departure) and oral history. Other methods included an agro-
ecological survey using aerial photographs in a fairly
conventional mapping procedure.
The aerial photos were a major asset as a tool for use in the
field as well as for the more conventional mapping and analysis
stages in the field office. The photographs had been flown by
the Office of Forests in 1982, and were high quality B&W 10" x
10" prints at an approximate scale of 1:4000. Their two main
field uses can be summarised as follows:
* as aids for mapping clan boundaries. Clan elders and others
who took part in the extensive walks to map clan boundaries
and indicate disputed land had little difficulty in using the
hard copy B&W photographs themselves, orienting them as
necessary in order to get their bearings, to point out
features of interest. We used chinagraph pencils to draw
directly onto the photos, double-checking with clan members
as we went. Boundaries that separated continuous lands of
neighboring clans were walked twice, once each with groups
from the two clans.
* as a focus for discussion in interviews. When carrying out
direct matrix ranking of crop varieties, etc., RRA
practitioners have found it best to use actual seeds in the
ranking exercise. With land resources this is more
difficult! Although we had a 'topographical advantage' in
the rugged highlands of PNG, and ensured that all such
interviews were conducted on ridge-tops overlooking the areas
of land in question, we also needed something more immediate
for both interviewees and interviewers to use as a reference
in distinguishing between land types. The aerial photos
again proved to be invaluable for this purpose, as they did
in discussions about land disputes (see below).
Clan linkages and access to valued land resources
The formerly flexible practice of 'multilocal residence' in the
area has been steadily tightening up with the provision of rural
services and great incentives and government requirements for
people to stay in one place. Thus gardens are cropped much less
frequently in areas further afield than a few hours' walk than
they would have been a decade or two ago. This means that clans
are more constrained by the land resources they customarily own
in the immediate locality.
For clans without sufficient land of particularly valued types,
in relation to their existing endowments, the major form of
access to the land they need is through their ability to exploit
the linkages they may have with neighboring clans. These
linkages take different forms and are constantly changing. For
those clans without sufficiently strong linkages with nearby
clans, it is hypothesised, disputes over particular parcels of
land are more likely to occur. This hypothesis was formed during
the course of discussions with members of local clans, when the
nature of the linkages was triangulatedd' by asking each clan
about all the others.
The diagram below attempts to represent in stylised form the
linkages between the clans (the bolder the line, the stronger the
links) and the rate of incidence of land disputes between 1977
and 1985 (from the District Land Disputes Register and verified
in group discussion in the field). In the broad view it does
appear to show that the incidence of disputes is low between
clans that are more strongly linked, and high between clans with
relatively few links with neighboring clans (e.g. Samarip clan).
Thus the extent of land disputes may be used as a rough secondary
indicator of problems in access to valued land resources at the
Institute of Development Studies
University of Sussex
IjNTeZ--CLAl LIt KAGEz Z & LA DuTiS
BeTLuJ&Tl LeWOFZ AEo CGLANSi Pr&
-- \ r-CLA AGES (SBt LS UTWoCI' N }K)
LAtm DSPATS, (tUAgeR 91t7-%5)
3. INTERVIEWS AND GROUPS
The Use of RRA Interview Schedules within a Collaborative
Research Project in Northern Thailand
The research project described in these notes is entitled "The
Role of Rural Peoples' Organisations in Agricultural Development
in Northern Thailand". It is a collaborative research project
involving the AERDD, University of Reading and the Department of
Agricultural Extension, University of Chiang Mai (CMU). Phase I
was concerned with identifying the range and characteristics of
Rural Peoples' Organisations (RPOs) in Northern Thailand. The
research methodology of this phase of the work revolved round the
use of conventional survey questionnaires to provide a census of
RPOs in the nine Provinces of this part of Thailand. Phase II of
the research project centres round eight "case studies" designed
and implemented by four separate teams of staff. Each team
comprises both AERDD and CMU staff. The two case studies with
which I have been involved have been concerned with a comparative
analysis of "successful" and "unsuccessful" Paddy Farmer Groups
(PFGs) in the Provinces of Nan and Lampoon. The rest of these
notes explain how for this stage of the research an RRA approach
evolved and the coverage of the Interview Schedules. An
assessment of the use of these Interview Schedules will be given
I did not go to Northern Thailand with the conscious intention of
using RRA. However, time constraints, the nature of the
research, problems experienced in the use of the Phase I survey
questionnaires and the practical difficulties in achieving
collaborative research all resulted in the design of a research
approach that utilised key elements of an RRA approach.
There were a number of advantages in the use of an RRA approach.
First, it enabled a defined problem to be explored in a flexible
but structured way taking into account perspectives from a range
of farmers and officials. Second, it enabled a great deal of
work to be done during a relatively short field period in
Thailand important given the difficulties faced in achieving
effective "collaboration" in the research work. Third, the
approach enabled a balance between data collection,
interpretation and analysis to be achieved during the field visit
itself. Fourth, the approach led to and was able to incorporate a
variety of secondary data sources e.g., reports from the
Cooperative Inspection Office. Finally, the use of an RRA
approach identified and compensated for some the unreliability of
the Phase I data which had used conventional survey research
Three semi-structured interview schedules were designed for the
Phase II work with the Paddy Farmer Groups. First, an interview
schedule for the Paddy Group Committee. Second, an interview
schedule with the Kaset Tambon (village level extension worker).
Third, an interview schedule with the Kaset Amphur (district
extension officer). A major objective was to assess the relative
"success" or "failure" of the Paddy Farmer groups from these
three different perspectives. The fourth perspective, that of
the farmers themselves, is being assessed by the use of
questionnaires by CMU staff and is not discussed in these notes.
However, the design of this questionnaire was aided considerably
by the preceding RRA interview schedules.
1. Interview Schedule for the Paddy Group Committee
After two initial tables designed to obtain information on
the villages, committee members and farmers in these groups
this interview schedule divided into six main sections.
These were Agro-economic background to the village and tambon
(sub-district); Background History of the Paddy Group;
Activities of the Paddy Group; Participation by members;
Economic background to the Paddy Group; Factors affecting the
success/failure of the Paddy Group.
2. Interview Schedule for the Kaset Tambon
This included Length of service in the extension department
and in the tambon; Age and education background; Division of
work between extension and non-extension work; Most important
crops for the farmers; Main problems for the farmers and for
their extension work; Details of their work with each of the
different types of RPO; Links (if any) between the RPOs and
T&V extension; Instructions/training from the district
extension office regarding working with RPOs; which types of
RPOs the least successful in the tambon and why?
3. Interview Schedule with the Kaset Amphur, including Agro-
economic background to the District; Extension policy re:
Farmer Groups, Housewives Groups and Natural Groups;
Extension Policy regarding the Paddy Farmer Groups; Official
view regarding the relative "success" or "failure" of the
Some preliminary results from the Paddy Group Committee interview
schedule are shown in the attached chart. The chart shows data
for two successful paddy groups (success) and five unsuccessful
paddy groups (US). The RRA interviews showed that the great
majority of these farmer groups were not operating effectively,
were heavily indebted, had experienced a great turnover of
officials and were largely ignored by an extension service which
concentrates on the few successful PFGs or utilises other types of
group extension purposes.
J Howard M Jones
Agricultural Extension and Rural Development Department
University of Reading
Reading RG1 5AQ
SOME RESULTS FROM COMMITTEE IINI'ERVIEWS
Date formed 1976
No. villas~ s 13
No. members 446
o. agric. units 8
Con. members 6
CHM:n other groups 2F.A.
CM: turnover no changes
Group:started by KA
K.A. attends sometimes
K.A. meetings 5/6 a year
Planning annual meeting
Cash credit no
Fertilizer credit yes
Group income share Y
Group problems some default.
3 F + T
5/7 a year
4/5 a year
2 F + T/1 orchard
1 Teach/F, 2 farmers
five x 2
one x 3
1 a year
3 a year
Group solutions godown already OK
2 r + other
2 Agric. Coops
three x 3
two x 1
6 a year
2 a year
2 Agric. Coop
two x 2
three x 1
2 a year
1 a year
6 a year
1 a year
V.H. no interest
If K could
2 F T
two x 2
three x I
2 a year
start longon group
Farmer Groups and ITK in Benin
In most cases, young field researchers do not reach all their
goals, or have to modify goals and methodologies during their
field research, but very few try to expose clearly the
difficulties encountered so that other fieldworkers can benefit
from their experience.
I report here on field surveys which have been conducted in an
integrated rural development project in the south of Benin. The
aim of these surveys was to identify problem areas so that
appropriate subjects for research and extension activities for
the project could be formulated, as well as to find out about
indigenous technical knowledge (ITK) which could be developed
further. Emphasis was to be put on areas of ITK concerned with
management of soil fertility and sustainability. To reach a
holistic understanding of the major problems, factors and their
relationships without spending years collecting and processing
data, we agreed that discussions and the qualitative point of
view of farmers were to be preferred to comprehensive whole-farm
The team of agronomists, agricultural officers and extension
workers asked farmers encountered in the fields and villages to
participate in a discussion either on problems at the whole farm
level or on agronomic topics. Informal groups were built without
difficulty and farmers described their farms and discussed their
main problems. These groups focused on production and on
cropping systems, and produced information on resources,
activities, outputs together with some indicators on bush-fallow
performance. These were to be examined more systematically: the
main challenge farmers have to face is the development of more
intensive cropping and production systems which remain
sustainable, as time for bush fallowing is getting shorter.
This approach can be summarised through the device of the survey
team: "learn from the farmers and about them".
Difficulties of Assessment by Researchers
To elicit possible changes in the land use we stratified
eventually the province according to ecological impoverishment
patterns. Even then, the picture about the adjustments farmers
are practising was still confusing, mainly because we had not
conducted discussions with groups stratified on a socio-economic
If this survey was to be prepared again then:
- These broad discussions would be conducted with a smaller
number of groups, instead of aiming at an accurate
geographical coverage. Thereafter some more topical
discussions could take place.
Even if we got some insight into regional differences in land
use and cropping patterns, it would have been more efficient
if a few areas had been selected regarding the main
agroecological indicators and if we had surveyed the land use
patterns in relation to region (land scarcity, soil
fertility, climate, etc.) and to the socio-economic situation
of different kinds of farmers in these regions.
Moreover, we should also have let the farmers evaluate the
performance of their different cropping systems: for example
according to the frequency and period of scarce food and
cash, to labour peaks in the-calendar.
Finally we should have returned to the villages with the
conclusion on different cropping systems which we drew from
farmers' assessments about the performance of these systems
and with our own opinions concerning their sustainability.
On this basis we could have discussed with farmers about
further work on a sound basis.
Reflections about our approach to tapping ITK
In the first stage, the team developed a general attitude towards
farmers of learning from them and questioning what they are
doing. It was a healthy reaction to usual attitudes, where
farmers are considered as backward illiterates reluctant to
accept change. Emphasis was put on looking for rationality and
knowledge in what farmers are doing.
In fact we could only be amazed about farmers' sound knowledge
about their environment, soils and plants. Farmers have also
been very active in screening new cassava varieties and were more
efficient and quicker in finding CMV resistant, sweet and early
varieties than the research and extension programmes.
There is also some evidence that we, and other farmers, could
learn from some of the better farmers. These must be better
farmers because they succeed in allocating their own resources in
order to solve their problems of scarcity, or because they
develop better technical skills, or because they adjust faster
than other farmers to changes, and not because they have access
to more resources.
Yet a difficulty in assessing problems and solutions with farmers
is their location specific knowledge: they cannot be aware of
problems if they cannot compare situations where these problems
are occurring and not occurring. In addition farmers often only
mention problems for which they know potential solutions (other
topics are not problems but hazards). Asking farmers about
changes in composition of fallow showed 'that they have a very
sound knowledge about its dynamics, that they know plants which
they think are able to let soil recover faster than the others
and that some of these plants are multipurpose trees; but they
never thought about planting these trees as they do not foresee
that these species will disappear or, even if they do, thy feel
helpless as they never saw anybody planting that kind of tree and
would not know how to do it.
RRA is a very useful way of helping scientists to learn about
farmers' concerns, which are mainly problems of short term
scarcities. It should not prevent scientists from developing
their own appraisal on the systems according to their own
concerns, which are linked to long term sustainability. In
conclusion, RRA is a good instrument to discuss possible changes,
but the team would have gained more experience by coming back to
the villages for further discussions: later experiences show
that it leads to a sound consensus on further possible solutions
which can be experimented or tried. Some of these resolutions
could have been "on farm" experiments, some might have involved
the whole village for common decision on land management,
infrastructures, and so on. But then, a strong team is required
to conduct the follow-up work!
Institute fur Sozial- and Agraroekonomie der Tropen und Subtropen
7000 Stuttgart 70
Focussed Groups in Ethiopia
Group work is central to any community based research and
planning. This example comes from a Rapid Rural Appraisal
exercise carried out in Wollo in Ethiopia by the Ethiopian Red
Cross Society and the Ministry of Agriculture. The aim of the
exercise was to explore ways of local level planning for
natural resource management at the Peasant Association (PA)
level. Using RRA techniques local extension workers and
research staff can facilitate local planning by engaging
farmers in the development process.
The particular focus of this RRA was the management of
hillside closure areas. These are portions of land where
agriculture and grazing is restricted in order to allow
regeneration of natural vegetation; this may be assisted by tree
planting programmes. In the past, blanket restrictions have been
imposed and local people have not been centrally involved in
the management of their hillside vegetation resources. There
is an increasing realisation of the necessity to develop
local management plans that allow the benefits of the closures
to be received locally.
Different groups within the Peasant Association have
different views as to the potential role of the hillside
closures and the components necessary for local management. The
RRA team explored this diversity of views in a series of focus
group discussions with different sectors of the community.
This note describes what was done and some of the problems
associated with the application of this approach in
community level planning.
What was done
1. A group discussion was held among the RRA team to list the
possible interest groups within the PA. These included: PA
leaders, Producer's Cooperative members, individual farmers,
Women's groups, youth, old men, old women, closure guards,
livestock owners/non-owners, those living near/far from the
2. Discussion groups were set up with each of the different
interest groups. Between 3 and 15 locals and 3-4 team
members attended. The meetings lasted between 1 and 3 hours.
A short checklist of questions was produced by the RRA team,
but the discussions were allowed to flow freely; often being
led by members of the farmer group. Ranking games were used
to focus discussion around preferences and attitudes to
different options. The attitudes of the members of the group
were recorded as notes; these included verbatim quotes that
demonstrated particular local views.
3. A comparison of attitudes and plans for management was made
following the first series of discussions (see Table 1).
This matrix compares the attitudes of four of the groups
to hillside closures. This gave the team an idea of the full
range of views and an idea of the social/political position
of the different adherents.
4. A general workshop meeting was held where representatives
from each of the focus groups were invited. Since they had
each been party to the previous discussions, the debate in
the larger meeting (30+ people) was fluid and open; all
groups felt able to contribute. This provided a good forum
for a discussion of future plans and a consensus on what
action should be taken next was reached.
Problems and biases
1. Place for meetings the decision of where the group
meetings were held was made by PA officials; it was usually
the central meeting place. The 'officialdom' of this may
have introduced a bias into the discussions. It is
therefore important to complement group discussions with
individual interviews in other places (at peoples' homes,
in fields, at the closures etc.).
2. Contacts the people invited to the group discussions
tended to be the most accessible and often official
position holders in committees etc. The representativeness
of these individuals must also be cross-checked.
3. Groups the choice of groupings was made by the RRA team.
They had a lot of local knowledge of the area, but
incorporated their own perceptions in the choice. Only two
of all the groups included women. Does the choice of groups
represent effectively the socio-political reality of the
4. Topic the choice of a particular topic hillside
closures and woodland management may restrict the open
discussion of the issues that people think are really
pressing. The topical focus should not be adhered to too
strongly and the linkages with other components of the
system need to be fully explored and peoples own priorities
allowed to come out in the discussions.
5. Groups and implementation the involvement of different
interest groups in planning certainly provides important
insight for implementation (potential conflicts,
identification of key actors etc). However the mechanism
for continuing the participation of different groups
through research and planning into implementation has not
been fully addressed. If it is not, the institutions
involved in the top-down development of the past will
inevitably take control and the value of generating local
involvement in the early stages lost.
A full write-up of the Wollo RRA (Participatory RRA in
Wollo: Peasant Association planning for natural resource
management) will soon be available from ERCS, Addis Ababa and
- Shortage of land
- Definite benefits
- Fear for lives
- Lenient on poaching
- Do job because of
- Rights of use not
- Do not regard trees
as belonging to
- Cannot get access
for fuelwood, clay
- Extra labour in
- Wildlife pests
Plans for Management
- Thinning of bush and
pruning to increase
- Cut and Carry
- Controlled grazing?
- No New closures
- PA Level control
- More guards
- More PA support
- Supervised cut and
- Increase use-grazing
access, wood and bark
- Split closures to
'belong' to different
Local management and
- Alternative home
planting useless as
trees will be taken
- Open the areas for
Some Difficulties in Training for Interviewing
The core of our fieldwork in Nigeria consisted of a series of
community, group and household interviews. This represents one
aspect of triangulation. In addition, we had also envisaged a
specialised function for each type of interview.
The community interview was meant to generate information about
possible community development ideas (bearing in mind that nearly
half of our course participants were community development
officers). Similarly, the household interviews were meant to
generate possible agricultural project ideas in particular, since
most of the other course participants were agricultural officers.
Finally, the focus group interviews were meant to generate
information about villager's reactions to existing projects.
In practice, however, all three interviewing formats produced
information across the board (relating to health, education and
public works as well as to community development and
agriculture). They also tended to produce information about
perceived problems and needs in a very generalised way which made
it difficult to come up with particular project ideas and/or made
it impossible to judge adequately between very different types of
possible project solutions. For example, the DAG trainers
thought that the problems with existing facilities in health,
education, agriculture and so on demonstrated that the model of
development which they embodied was inappropriate to Nigerian
circumstances. Many of the participants, on the other hand,
believed it was more a question of improving these facilities
and/or of "enlightening" the people.
Unlike most practitioners of RRA, we decided to use semi-
structured questionnaires instead of a checklist of issues. The
reason for this was that the training team felt that the
preparation of such questionnaires (one for each of the different
types of interviews) would constitute a useful intellectually
rigorous exercise for our participants, which would force them
into thinking clearly about what it was they really wanted to
This approach was prompted by my impression that existing project
identification was based on very casual and occasional (and
elitist) conversations with villagers (to the extent it was based
on information from villagers at all).
Unfortunately, although it was a successful classroom exercise,
the questionnaires proved to be disappointing in the field,
despite our training emphasis on probing through the use of the
"six helpers" and the use of open ended questions and blank
sheets for taking notes.
The effect of such an approach was probably to impose a structure
which inevitably reflected our participant's perception of the
world, whereas a checklist approach would probably have allowed
more of the villager's perception of the world to emerge. In
methodological terms, it allowed less room for learning during
the fieldwork itself because it provided less flexibility to
follow up spontaneous new leads, as and when they emerged.
As far as the methodology of the interview formats themselves
were concerned, the participants were all given protocols on
household, group and community interviewing. However, we did not
attempt to judge their performances in the field against these
protocols and cannot therefore say how much they learnt and
applied from these, nor how their interviewing techniques
different from their previous performance.
It did seem to be however that course participants were most at
home with the community interview. There were also some
interesting variations which arose spontaneously. My own team,
for example, decided to allocate responsibility for each sectoral
section of their questionnaire to a different team member, while
I allocated the remaining team members to observation duties.
This set up worked very well at our community meeting, at which
about 150 people (exclusively men and children) were present.
Another team, however, decided to conduct their community
interview on foot while walking around the village. This suited
the circumstances of their particular village where the homes
were very scattered. None of our teams overcame the well know
problem of dominance by the village head and other prominent
villagers, nor did any of them really adopt any of the measures
suggested in the literature which were included in their
protocols. Similarly, none of the village teams applied the
protocols in relation to group interviews. In addition, many of
the participants seemed to be unclear about the nature and
purpose of group interviews.
As far as the household interviews were concerned, finally, less
can be said with confidence since the trainers were not present.
However, going by the reports presented and the discussions we
had, it seems likely that these interviews were (like the others)
characterized by a lack of probing. The lack of attention to
detail was striking, as was the uncritical acceptance of the
answers given. Equally, course participants also seemed to equate
project intentions with the likely future realisation of such
projects, a belief which is not borne out by past performance of
village projects in Nigeria.
The overall conclusion, therefore, has to be that interviewing
skills cannot be taken for granted and that communicating such
skills is a difficult and time consuming business which requires
more attention and practice than we were able to give during our
Robert Leurs, Development Administration Group, Birmingham
Wealth Ranking in Sudan
This technique uses the perceptions of informants to rank
households within a village or quarter of a village according to
overall wealth. Researchers very often feel reticent before
embarking upon wealth ranking. Wealth is a sensitive topic. But
this game ensures that any discussion of absolute wealth does not
take place with reference to specific households. Classes or
groups of households may be characterized as having certain
features; wealth as a whole may be discussed; but when it comes
to individual households these are only compared with eachother,
and the discussion remains solely of relative wealth/poverty.
Moreover, the game appears to be more successful if informants
who are known to the team do the ranking. They may be previous
interviewees, or have attended protocol meetings. Better still
they may have been met and talked to in a very informal setting,
such as in the evening.
This technique requires careful preparation: first the list of
households must be prepared; second the name of head of household
must be written onto separate pieces of card or paper; next the
informants identified; next the interview begun with a discussion
of the informant's perceptions of wealth; then the cards are
sorted by the informant into piles or wealth classes; these are
reviewed and changes made accordingly; and finally the informant
is asked to name the principal features of each household's
livelihood strategy. The ranking is cross-checked with several
rankings of the same list, and the final wealth classes computed
(See Grandin, B. 1988. Wealth Ranking. IT Publications, for
detailed discussion of procedure).
This example comes from an RRA conducted in a village in Sudan.
There were no lists of households available for Faki Hashim. The
team of investigators had hoped to use the sugar ration lists
held by shopkeepers. These contain all the households in the
immediate neighbourhood, and had' the apparent advantage that
people would have an incentive to be on the list, unlike tax or
census lists. But on the day that the team came to collect the
lists and elicit the help of a shopkeeper, the shop happened to
Instead a key informant, the supervisor of the government mango
scheme and a lifelong resident of the village, was asked to name
all the heads of households residing in the central part of the
village. He had previously been interviewed and by this time
knew the team well. From his list of about 70 a sample of 50 was
taken at random, and the name of each written on separate pieces
of paper. Although this may have produced a biased list through
selective recollection of the informant, he did indicate that he
was conducting a geographical sweep of the village to ensure none
were omitted. The wealth ranking was then conducted on these 50
The procedure of discussing terms for wealth and the placing of
these cards into separate piles was conducted with three
different informants. All three were in agreement over the
features of household livelihoods that characterized their level
of wealth. In general the most wealthy were thought to own
agricultural land, own livestock, own transport vehicles, be
involved in commercial activity or be receiving remittances from
overseas. Those of middling wealth were involved in farming, but
mainly as sharecroppers, and might own a few livestock; and the
poorest households were those relying solely upon agricultural
labouring as a source of income.
The first informant was wealthy. He began with five piles, but
during checking after allocating all 50 cards he divided the pile
for the richest into two. All changes then made were form richer
to poorer piles. Finally he created an extra category for the
very poorest, leaving seven piles in all. The second and third
informants were both poor. Neither changed the number of piles
from their starting five. Following the rankings the first
informant was asked to name the key components to the livelihood
of each household. This produced summaries such as merchant,
lorry owner, land owner in agricultural scheme, etc.
The results of the ranking are shown in Tables 1 and 2. Only 48
were eventually given aggregate scores as informants 2 and 3 were
not asked to rank themselves. Informant 1 was not contained on
the list. The aggregate scores for each household were then
broken into 5 classes: A for 2.5-3, B 2.0-2.49, C 1.5-1.99,
D 1.0-1.49, and E < 1.0. This exercise produced some very
1. Clearly the majority of households rely on non-farm income
sources. Very few rely solely upon farming. Those that do
so are mainly in Class E.
2. The large number of merchants and owners of transport
reflects the proximity and opportunities of Khartoum.
3. There are some interesting comparisons to be made between the
rankings of the 3 informants. Over some households they are
in close agreement: all of 17, 30, 34 and 47 are wealthy; and
all of 2, 8 and 13 are very poor. But there are also some
large disagreements, particularly between the rich and the
two poor informants: household 44 was placed in the top pile
by informant 1, but in the bottom piles by the other two.
Perhaps informants 2 and 3 did not know of the two migrant
sons sending remittances. Informant 1 may have had some
special information about household 23, or just believed he
was a successful pedlar.
Jules N Pretty
Table 1 Results of wealth ranking conducted by three
informants on 50 households of Faki Hashim
HOUSEHOLDS INFORMANTS' RANKINGS AGGREGATE WEALTH
1 2 3 SCORE CLASS
1 0.72 0.4 -
2 0.29 0.2 0.2 0.69 E
3 1.00 0.8 0.6 2.40 B
4 0.43 0.6 0.4 1.43 D
5 0.72 0.8 1.0 2.52 A
6 0.43 0.2 0.4 1.03 D
7 0.57 0.2 0.2 0.97 E
8 0.29 0.2 0.2 0.69 E
9 1.00 0.2 0.6 1.80 C
10 1.00 0.8 1.0 2.80 A
11 0.86 0.6 0.6 2.26 B
12 0.72 0.4 0.2 1.32 D
13 0.14 0.2 0.2 0.69 E
14 0.43 0.2 0.2 0.83 E
15 0.72 0.8 0.2 1.72 C
16 0.86 0.8 1.0 2.66 A
17 1.00 0.8 1.0 2.80 A
18 0.72 0.6 0.2 1.52 C
19 1.00 1.0 1.0 3.00 A
20 0.57 0.4 0.2 1.17 D
21 0.57 0.2. 0.2 0.97 E
22 0.29 0.2 0.8 1.29 D
23 0.86 0.2 0.2 1.26 D
24 1.00 0.6 0.8 2.40 B
25 0.43 0.8 0.2 1.43 D
26 0.72 0.4 1.0 2.12 B
27 0.72 0.4 0.4 1.37 D
28 0.43 0.2 0.2 0.83 E
29 0.72 0.2 0.4 1.32 D
30 1.00 0.8 1.0 2.80 A
31 0.72 0.8 0.6 2.12 B
32 1.00 0.4 0.4 1.80 C
33 1.00 0.6 1.0 2.60 A
34 1.00 1.0 0.8 2.80 A
35 0.57 0.4 0.2 1.17 D
36 0.43 0.2 0.2 0.83 E
37 0.57 0.6 0.4 1.57 C
38 0.43 0.2
39 0.43 0.2 1.0 1.63 C
40 0.43 0.6 0.8 1.83 C
41 0.57 0.6 0.2 1.37 D
42 1.00 0.6 1.0 2.60 A
43 1.00 0.6 1.0 2.60 A
44 1.00 0.2 0.2 1.40 D
45 0.86 0.6 1.0 2.46 B
46 0.57 0.6 1.0 2.17 B
47 1.00 1.0 1.0 3.00 A
48 1.00 1.0 0.8 2.80 A
49 0.86 0.6 0.2 1.66 C
50 0.43 0.4 0.4 1.23 D
No. Classes 7 5 5
Table 2 Major occupations of each household in the five
classes produced by the wealth ranking
Class A Wealthy
Principal components of household livelihood
10 Lorry; pick-up; tractor; shop
16 Flour mill; sorghum merchant
19 Farmer; brick maker; 2 lorries
30 Merchant; good agricultural land
33 Agricultural scheme
34 Army colonel
42 Big merchant
43 Army Major
7 Medical laboratory owner
48 Medical assistant
3 Big merchant
11 Butcher; lorry
26 Lawyer; 2 lorries
31 Son in Saudi Arabia
45 Merchant; good agricultural land; shop; trailer
46 Lorry and merchant
9 Supervisor of agricultural scheme, owns land
and 20 cows
15 Supervisor of scheme, owns land
18 Shop, tractor, owns land
32 Agricultural scheme, owns land (owner)
7 Lorry, merchant
39 Lorry and taxi
49 Lorry and taxi, son is a Doctor
4 Retired Army officer
6 Civil servant with agricultural land
12 Medical assistant
20 Farmer and several lorries
22 Cultivator in scheme; average farmer
23 Lorry and pedlar
27 Lorry, good agricultural land and sons
29 Taxi, shop
35 Fodder shop
41 Farmer, lorry, and 2 migrant sons
44 Agricultural land, farmer, 2 migrant sons
50 Lorry; official in University
Class E Poorest
2 Small farmer
7 Poor farmer
8 Old man; small farmer
13 Street sweeper
14 Government worker, some livestock
28 Not cultivating his agricultural land, migrant
Preference and Direct Matrix Rankings, Sudan
This note is based on the experience of using ranking
techniques to investigate local incentives to tree
management in a RRA exercise carried out near Khartoum by
members of the Institute for Environmental Studies
(University of Khartoum) and representatives from various NGOs
Two ranking techniques were used: pair-wise preference
ranking and direct matrix ranking. I will not discuss the
actual techniques (information on this can be found in RRA Notes
1), but will concentrate on a comparison of their uses and
some of the potential problems of their application.
Pair-wise preference ranking
This technique was used to compare preferences for
different tree species between individuals (men, women,
young, old, richer, poorer etc.) and between different
groups (settled residents and displaced immigrants). The
ranking highlights the differences in priorities (as
expressed by the ranking of the 6 species) and differences in
decision-criteria used (as expressed in the list of 'good'
and 'bad' properties of each tree).
Figures 1 and 2 illustrate the results from a comparison
between two distinct groups in the village of Faki Hashim, north
of Khartoum. These results came from two ranking exercises
conducted with a few people from each group (all men). The two
groups chose different trees as the most important 6; the
long-term residents choosing trees that are particularly
important for shade in their homes or are common along the flood
retreat farming land on the banks of the Nile. The migrants
chose trees significant in the common property grazing land
beyond the village.
The two groups came up with a set of criteria; some
elements were common and others quite distinct (Figure 2).
Sidir, Zizyphus spina-christi, was ranked highly by the
displaced migrants group (No 1), but lower (No 4) by the long-
term residents. The reasons for this can be found when the
criteria for choice are investigated. The migrants use Sidir
as an important component of their funeral ceremonies;
it also has a valuable fruit important in nutrition. The reason
the settled group do not rank Sidir highly is because of the
trouble caused by kids throwing stones over the household's
walls to dislodge the fruit.
Direct matrix ranking (DMR)
DMR starts with the criteria for choice and ranks each item
according to the different criteria. A discussion of the
criteria provides a good starting point for investigating
choices. DMR can be linked to pair-wise ranking by using the
list of criteria generated as the basis for the ranking.
Getting a full list of local criteria is a complex task in
the above example 31 different criteria were mentioned by the
two groups in their choices for trees. In the Sudan study both
techniques were used for ranking and a good level of
comparability between the rankings was found.
The results of a DMR based on criteria generated by a pair-wise
ranking exercise are shown in Figure 3.
An investigation of the local weighting of criteria can also
provide useful insight. A DMR can be followed by asking a
forced question: 'if you could choose only one of the trees
which would you choose'. Sometimes the top ranked item is not
chosen because one particular criterion outweighs the
IRosnulLu from two pLrefeLtnce ranking exercises conducted
in lF.'aki llianhim
A. Diltplaced mig anti(i (Wesitern Sudan)
Sa Su //
Si Si Si /
Sa Su T Si
D. Long-term residents of Faki Hasim
Su Su //,
H T Su
N N N
Si T Su
H T Su
17 4 Comparisons of criteria for tree uses given by two
different groups in Fakir Hashim
A. Criteria suggested only by displaced migrants
1. Bark critical component for washing bodies in funeral
2. Fruit pods for tanning leather
3. Aroma of smoke whilst cooking
4. Straightness of wood
5. Wood for mortars
B. Criteria suggested only by Faki Hashim residents
1. Regrowth following pruning
2. Grows from seedlings
3. Availability of seedlings
4. Ornamental and beauty value
5. Wood for boats
6. Perfume and skin colouring
7. Wood for writing tablets
8. Fruit pods used in marriage ceremony
9. Nuisance in compounds: attracts stone-throwing young boys
10. Windbreaks on field boundaries
11. Fuel for brickmaking and bakeries
C. Criteria common to both groups
1. Fruit edible to humans
2. Fruits/pod/flowers for fodder
3. Leaves for fodder
4. Fruit pods medicinal
5. Strength of fire
6. Susceptibility to termites
7. Wood strength
8. Produces gum valued ingredient in inks and for mixing with
sand for building purposes
9. Branches good for hedges
10. Shade tree
11. Wood for handtools
12. Wood for building walls
13. Wood for building roofs
14. Smoke anti-rheumatic
15. Wood for furniture, beds
E N Si
Fig 3: DIRECT MATRIX RANKING
Laot Sidir Heglig Sunut
1. Resist drought
2. Good for growing on
3. Fast growing
4. Best for fodder
5. Ease of getting pods
6. Leaves as fodder
7. Fruits as fodder
8. Unripe pods are bad
9. Low for grazing
10. Gum good
11. Best fuelwood
12. Best for burning
13. Smoke less
14. Best for slow
15. Best smell
16. Best for building
17. For making furniture
18. For making boats
19. Making saddles
20. For rosaries
21. For human food
22. For medicines
23. Does not attract
24. Providing good
25. Ability to regener-
ate from seed
Criteria derived from
2 6 3 4 5
S 1 2
2 6 1 5 4
2 5 1 3 6
pairwise ranking; Informant same
primary school teacher (age c. 40)
others. Another option can be to rank the criteria themselves
and assign some kind of weighting system.
Ranking techniques provide a useful way of investigating local
decision-making criteria and they provide general comparisons
between different priorities. Outstanding questions about the
use of ranking techniques include:
How useful are they as a quantitative method of
assessment? What are the dangers of combining/adding up
different rankings to come to a planning decision?
What is the potential for the development of ranking
techniques through criteria weighting etc. or should they be
treated simply as a game context for essentially
How appropriate are the games to local cultural
situations (eg forced comparisons etc.)? Are there
alternative adaptations of local games that could be used as a
focus for ranking discussions.
[The full report of the IES/IIED Rapid Rural Appraisal
exercise will soon by available from IIED, London. It is
entitled: Rapid Rural Appraisal for Economics: exploring the
incentives to tree management in Sudan.]
Ranking of Carpentry Skills
See RRA Notes No. 6, pages 4-12. June 1989.
Using Ranking in Training of Field Level Staff
These notes are developed from the training of agricultural
extension officers from India who are front-line workers in a
project aimed at training marginal and resource-poor farmers in a
balanced and economic use of fertilizer. They are employed by
the extension organisation for the Indo-British Fertiliser
Education Project (IBFEP), funded by the ODA, in six states of
North and Northeast India, which is currently nearing the end of
its second phase.
The training as a whole emphasises participative learning, and
aims to respond to participants' needs for acquiring skills to
reach rural women effectively. IBFEP has recently introduced an
element into the project which requires the extension officers to
actively involve women from target households in their
agricultural extension training. Their ability to learn what
women's existing activities, responsibilities and resources might
be, across six very different states, is therefore the starting
point in the training. How they monitor the progress of their
women's programmes is a second important aim. A technique such
as ranking was therefore identified as a useful learning process
which they in turn could use in the field with different groups
of women to gather first hand, relevant information.
Constructing the Direct Ranking Matrix
The technique was introduced in three stages. The first stage
demonstrated what the technique is and how it works by showing
participants a ranking of fertilizer types done in an IBFEP
village by Robert Chambers in April 1988 (RRA Notes 1, June
1988). This example provided a lot of discussion and excitement
and paved the way for stage two in which participants were
divided up into five groups of five members each and asked to
construct their own direct ranking matrix, as if they were
farmers: two groups on paddy varieties, two groups on vegetables
types and one group on fertilizer. The intention was to compare-
the two rankings of paddy, the two of vegetables and the
fertilizer ranking with the Chambers version.
Even though the participants were using their knowledge and
preferences as agronomists, it did not detract from the process
of experiencing the quality of discussion required to establish
the list of criteria and to score them. Reaching a consensus took
longer for some groups than for others. The learning process was
reinforced by the fact that each group's matrix was different.
The third stage of the exercise was to ask the participants
whether they would add any further criteria if they were women
farmers. Those groups ranking vegetables were able to respond to
this first, by adding criteria such as "easy to cook", "nice
taste" and "nutritious" and this encouraged the other groups to
consider similar criteria for paddy, the suitability for both
growing seasons and the market price. The fertilizer group had
problems and this led onto discussion of how socio-economic
criteria are as important to cultivating households as the
technical criteria. This awareness of different but equally
valid perspectives was judged to be an important outcome and one
which might not have been so effectively achieved if the
participants had all been asked to role play women from the
beginning of the exercise.
Again the impact might have been less if one of each set of
rankings had been role played as women for comparison in stage
two; during that stage trainees had to experience actually doing
the ranking, to accept that their colleagues' rankings were also
of value and to transpose the idea to their work situations where
using the technique with farmers might lead them to useful
information exchange. A final reinforcement was provided however
by comparing with Chambers' other rankings done in other IBFEP
villages which, by showing the variations and the similarities,
increased the participants' confidence in trying it out
This was the first time ranking had been used in the training
programme and it was received enthusiastically. For fieldworkers
working with groups it does offer a concrete way of collecting
data and reinforcing group solidarity. In a training situation
in the UK there is a limit to how much role play can
realistically be undertaken; a technique such as ranking is an
opportunity for participants such as these male extension
officers to begin to concern themselves with issues identified by
As far as training for gender awareness is concerned, techniques
which enable male extension workers to learn about rural women
directly are invaluable and those which provide the means for
women's knowledge and judgement to be the determining factors in
prioritising needs and information even more so.
Agricultural Extension and Rural Development Department
University of Reading
Report of Rapid Appraisal Trial, Mbeya, Tanzania
This is a report of a rapid appraisal (RA) of health problems in
Mbeya, Tanzania. The workshop was divided into three phases:
* introduction to the methodology
* data collection
* formulation of a plan of action.
Fieldwork was carried out in three wards chosen by the municipal
medical officer. All three wards were squatter areas. Two were
areas of high concentration of population. The third was a peri-
urban ward in which the main source of income was still farming.
In addition, this ward had key officials who lived some distance
from the area and thus had limited contact with the residents and
their problems. As a result, this ward presented problems not
encountered in the two more urbanised wards and posed different
challenges for the rapid appraisal methodology.
It was explained that:
* RA was based on three sources of information documents, key
informant interviews and observations.
* is undertaken by professionals in multidisciplinary teams in
order that various aspects of information about one subject
can be explored and experiences can be applied to judge the
importance of the information presented.
* RA is not merely a method for collecting data about the
health problems of the urban poor but, more importantly, a
process by which to make a plan of action to improve the
living conditions of the people based on the participation of
these urban residents in defining their own problems.
The idea of using a planning profile in the shape of a pyramid
as a means of identifying important areas for which information
needed to be collected was presented. Participants were told
that the blocks of information which made the pyramid were
collected from the three sources mentioned above. To illustrate
graphically this approach participants were asked to write
information collected from documents on yellow cards, from key
informants on pink cards and from observations on green cards.
1The methodology is described in the document "Guidelines for
Rapid Rural Appraisal to Assess Community Health Needs: A focus
on Health Improvements for Low Income Urban Areas".
COMMUNITY COMMUNITY ORGANISATION COMMUNITY
CO-POSITON i STRUCTURE CAPACITY
The recording of information on the cards and placing them in the
categories of blocks of the planning profile provided the basis
both for the recording of information and the analysis.
The planning profile was explored in detail. Participants were
divided into three teams composed of members from different
sectors. Each team brainstormed on questions which needed to be
asked to build the blocks of the pyramid. Using white cards,
they wrote the questions down and each was read, placed in the
appropriate block of the pyramid (which had been drawn on large
sheets of white paper and attached to a blank wall), then grouped
together around specific issues. These groupings provided the
basis for the categorisation of data and included health policy,
health and environmental issues, social services, physical
environment, socio-economic environment, disease profile,
community composition, community organizations and structure and
community capacity for self-help. Participants then identified
from which sources these questions might best be answered.
Based on the categorisation of data in the planning profile a
checklist of information for interviews and observations was
developed. The checklist reflected a choice of information based
on discussions about each item. The entire group also considered
which people might be key informants. A list was compiled and
written on white paper.
Finally participants looked at documents as a source of
information to answer some of the questions which had been
previously identified. Each participant was asked to bring from
his/her office documents which would help to identify health
problems. Each team was assigned to draw from the documents
general information on Mbeya.
In preparation for data collection by semi-structured interviews
and observation the checklist which was prepared on the second
day was recorded on the first pages of a notebook given to each
of the participants. On the front cover of the notebook
information was recorded which reminded participants about how to
open interviews, how to record notes, how to conduct semi-
structured interviews in teams and how to end interviews.
Data Collection and Analysis
The first field visit was by pre-arrangement to three wards
selected by the municipal medical officer. Each team went to the
wards, met the ward officials and presented the reasons for the
interviews. Assistance of the officials was asked for obtaining
interviews with other key informants. In the meeting with ward
officials, information was collected from them about health
problems. However, ward officials comprised less than one
quarter of all ward interviews.
Much time was given to key informant interviews. On returning
the teams identified the major health problems in the three
wards. Each ward had different problems analysed by these
professionals based on key informant interviews, observations,
documents and their own experience. Participants then reported
in plenary the answers to the following questions:
* What were the major health problems?
* Who told you about these problems?
* Did your observations confirm these problems?
* Do the documents suggest that these are the problems?
However, when the problems were analysed, it was realized that no
priority had been given to them. It was thus arranged for teams
to return to the field to ask key informants to rank the order of
priority of the problems they had identified.
Each key informant was given 8-10 cards with the name of one
identified problem on each card, then asked to rank these cards
according to the most important problems. Blank cards were
provided in case a problem was identified which had not earlier
been recorded. Health priorities were then compiled for each
ward. The teams then suggested solutions to these problems.
A matrix was introduced by which to rank the feasibility of the
recommendations in order to place priorities on which was to be
undertaken first. The method was illustrated by asking each team
to choose one possible recommendation for the ward it surveyed
and to judge its feasibility by the following criteria:
* health benefit (what was the overall health impact?)
* community capacity (how committed was the community to
solving the problem and what could they contribute to its
* sustainability (would the intervention be able to be
maintained and at what cost for maintenance?)
* equitability (which income groups were likely to benefit
* cost (what are the initial capital and manpower costs?)
* time for benefit (how long would it take before changes would
Each recommendation was scored in these categories by giving "+"
for low, "++" for medium, and "+++" for high. The highest score
was given the highest priority.
Assessment of Methodology: Views of the participants
In general, comments were positive and enthusiastic. Positive
* Discovering aspects of community life which were unknown to
each before the investigations,
* Working in multidisciplinary teams to contribute to and draw
upon experiences from other sectors,
* Using semi-structured interviews instead of questionnaires to
discover community problems.
There were however some difficulties with the approach. The
following needs were highlighted:
* The need to overcome the bias of the sample
Because of possible bias, it was noted that it was important
that ward official interviews comprised less than one-quarter
of total key informant interviews, that interviews were
undertaken both in focus group and in individual situations
and that interviewers be constantly aware of these possible
biases. It was felt however that bias could be limited if
key informants were carefully selected. It was suggested that
in the future more time be spent in identifying key
* The need to overcome the shortage of time
Time was not sufficient to complete the planning process. A
ten day workshop would be the minimum to come up with some
* The need to overcome problems of interviews
Participants felt that the lack of experience of how to do
semi-structured interviews was detrimental to the collection
of data. They suggested that a pilot interview be undertaken
to give them some experience in these methods.
It appears to us the approach developed in the guidelines
provides a solid basis for programme design and development. It
also has the advantages of gaining community dialogue at the very
early stages of programme planning to build a basis for
negotiation and partnership between the resource holders and
beneficiaries. As indicated, the use and validity of the
approach to improving the health of the urban poor will depend in
great part on the interest and commitment of the authorities to
deal with the complex problems in slum and squatter area.
Adaptations to individuals situations will of course have to be
made. However, the general approach appears to be both
acceptable and useful to municipal planners in their search to
deal with the problems of the urban poor.
Susan Rifkin and Hugh Annett
Department of International Community Health
Liverpool School of Tropical Medicine
Rapid Appraisal South Sefton (Merseyside) Health Authority
March June 1989
1. The introduction of Locality Management in South Sefton
raised questions about the assessment of need of the population
served by the District Health Authority. Traditionally needs
have been indirectly assessed through the use of health
indicators. However, with the advent of more consumer-orientated
thinking the qualitative approach has to be considered as a
crucial option in understanding user's perspectives on health and
Community development approaches, particularly in Health
Promotion, have focused on users' perspectives, but the methods
of work tend to be labour-intensive and small scale. The main
problem has been how grassroots ideas can be fed into the
planning and policy setting process. Furthermore, how can local
concerns be translated into action when organizations such as the
Health Service tend to be centrally controlled. Rapid Appraisal
(RA) provides some solutions to the above questions.
2. The choice of RA in South Sefton was informed by the decision
to "go local" in management and provision of service. This opens
up opportunities to involve local communities in diagnosis of
health priorities, and thus identifies for Locality Managers key
people who can participate on a longer term basis in the health
planning process. It also has benefits in creating a baseline
for regular follow-up RA exercises, assessing progress in service
development as seen by the users.
Building relationships with a community is important for managers
who are committed to bottom-up planning as they can create more
permanent forums for debate through RA. In South Sefton, RA is
seen as the first step to involve communities in planning and
evaluating health services. The managers carrying out RA are
capable of translating communities' views into workable policies
and practices, and through the RA mechanism will be accountable
to the community.
3. The RA exercise in South Sefton was based upon the WHO
document 'Guidelines for Rapid Appraisal to Assess Community
Needs: a Focus on Health Improvements for Low Income Urban
Areas'. The authors of the document planned a two-day workshop
with the Manager of Research and Development in Sefton, who had
secured participation of a variety of managers.
From the Health Authority:
The Director of Nursing Services, Community, Elderly and
The Director of Nursing Services, Mental Health;
The District Health Promotion Officer;
The Operational Planning Manager;
The Manager of Research and Development.
From other agencies:
A Principal Housing Officer;
The Deputy Administrator of the Family Practitioner
The Planning Officer of Social Services;
The Research Officer of Social Services.
The workshop took place on March 16th and 17th 1989. The first
day and a half was devoted to formulating questions relating to
the information pyramid as described in the document. The
managers in South Sefton could not be released for the required
ten days, and the interview programme had to accommodate this.
Instead, a timespan of eight weeks was agreed in which all the
interviews and preliminary analysis would be completed.
The District Health Promotion Officer and a Nursing Officer who
were both knowledgeable about the ward to be investigated
(Linacre ward in Bootle) drew up a list of names of key
informants. Three multi-disciplinary subgroups were then
allocated a mixed group of interviewees and arranged their own
interviews. One intermediate working meeting was planned to
The subgroups carried out almost all their interviews within the
set period and did preliminary analysis on the data they had
collected as a group. A one-day workshop was held on May 18th to
analyse the total data set, and clearly defined issues were
highlighted. These were "reduced" to one-line statements and
then ordered into the separate categories. Each statement was
typed into a colour-coded card to be presented to the informants
for placement into priority order. Subsequently a final meeting
was held to complete the analysis of the data and prepare a plan
4. The information pyramid was found to be equally relevant in
the Sefton situation as in the Third World. However, emphasis on
data collection should shift mainly to interviewing and
observation, rather than gathering documentary evidence. Health
planning in the developed world is heavily based on quantitive
data, but in contrast to underdeveloped countries, this data is
so abundant that planners are struggling with turning this into
information rather than having to search for data. This,
therefore, poses the problem of losing sight of what the
population for whom services are planned actually want
themselves. In using RA, the investigators became aware of the
need to re-focus their attention on collecting qualitative data
by listening to the community.
The interviews themselves went very well, because the cooperation
of all interviewees was good. The three teams were very
enthusiastic about talking directly with the community about
their perceptions of priority problems. As managers they had
only indirect contact with people who (potentially) use their
services and the RA interviews made them feel "in touch" again.
Furthermore, they gained new insights into the complexity of
causes of ill-health as they were explained to them in terms of
the socio-economic and cultural fabric of Linacre ward. On the
other hand, certain findings confirmed what was known already.
An additional benefit was the truly multi-disciplinary nature of
the investigative work itself. Rather than sitting together in
joint planning meetings, managers were out actually working
together, getting to know each other, sharing information and
It is too soon to evaluate the success of the exercise in leading
to joint action for health in the ward. However, this is being
monitored and a paper on the exercise is now being prepared for
Hugh Annett Pauline Ong
Liverpool School of Tropical South Sefton (Merseyside)
Medicine Health Authority
Pembroke Place Fazakerley Hospital
Liverpool L3 5QA Longmoor Lane
Liverpool L9 7AL
6. PARTICIPATORY APPROACHES
From the Ground Up and Participatory RRA in Kenya
From the Ground Up (FGU) is a collaborative effort of
institutions in Africa and North America committed to improving
natural resource management in Africa. The programme is
administered and coordinated by the Center for International
Development and Environment of World Resources Institute, in
Washington, DC, USA, together with assistance form Clark
University, Massachusetts, USA.
The objectives of FGU are threefold:
1. to develop a better understanding of the core elements and
key instititutional, managerial, and technological
relationships which contribute to successful natural resources
management at the community level;
2. to effectively disseminate the results, conclusions, and
implications of the investigations 'across' to other
communities and 'up' to the national policy apparatus and the
3. to assist communities to better identify long-term needs and
opportunities toward enhancement and sustainable use of the
local resource bases.
Over the past year, the National Environment Secretariat of the
Ministry of Environment and Natural Resources, Kenya, has been
conducting a series of village-level exercises on effective
natural resources management as part of this coordinated project.
NES has employed Participatory Rapid Rural Appraisal (PRRA)
methods along with conventional household surveys to develop case
studies of successful community efforts in resource management,
and to prepare Village Resource Management Plans (VRMPS) -
community-specific action plans for improved local management of
critical natural resources.
NES has met with promising results in the five Kenyan communities
in which PRRA methods have been used. The communities, with the
assistance of government technical extension officers, have
organised realistic VRMPs for developing, conserving and
sustaining local natural resources, which they are now in the
process of implementing.
The PRRA methodology proved especially useful in targeting
particular priority areas, such as the rehabilitation of domestic
water supplies and the improvement of horticultural crop
production and marketing programmes, which required external,
advice and assistance. NES has been able to enlist specialised
agencies, including the Kenya Water for Health Organisation
(KWAHO), and the Horticultural Crops Development Authority (HCDA)
to work with the communtieis to develop viable strategies and
implement appropriate activities to deal with these and other
In June of this year, NES, Egerton University/Kenya, Clark
University, and the Center jointly sponsored an intensive, month-
long training programme in the PRRA/VRMP methodology. The 24
participants included senior Government officials, Government
technical extension officers, representatives of Kenyan NGOs, as
well as officials from FGU lead organizations in Somalia, Ghana,
and Sierra Leone.
NES believes the PRRA/VRMP methodology has the potential to:
1. provide a systematic, but semi-structured approach to enable
a multisectoral team consisting of community leaders,
extension officers, and NGO staff to examine community-based
resource management concerns and carry out village
2. provide high quality information and clear pictures of
community problems, opportunities, and capabilities in a
relatively short amount of time, at a low cost, and without
the need for foreign experts;
3. establish locally developed and managed action plans to bring
together, on one hand, development needs defined and ranked
by community groups, and on the other, inputs and technical
skills of development specialists with expertise in water,
livestock agriculture, and forestry from government technical
services, donot agencies, and NGOs;
4. maximise local participation by gathering and analysing data
in group discussions, using mostly visual instruments, and to
facilitate community mobilisation, particularly of women, in
activities that are truly sustainable;
5. move beyond the conventional sectoral approach to project
design, monitoring, and evaluation, and offer a holistic
persepctive on the factors that impinge on a community's
6. bridge the gap between intended beneficiaries and the
professionals who manage development resources, and introduce
natural resource management practices that village leaders
and institutions have incentive to maintain.
However, a number of important issues have yet to be fully
addressed. These include:
1. the long-term application of the VRMPs (i.e. will the
communities continue to refer to plans and update them
periodically once the initial enthusiasm fades?);
2. the effectiveness of VRMPS in high potential areas (i.e. all
VRMPS prepared by NES have been in communities situated in
different semi-arid, low-potential sub-locations in Kenya):
3. the appropriate lead organisation to conduct PRRA/VRMP
exercises (i.e. should the government be primarily
responsible, or would it be more effective and efficient to
leave such work to specialised NGOs with the institutional
capacity and flexibility to orchestrate the appraisals?);
4. the effectiveness of institutionalising the PRRA/VRMP
methodology (i.e. what is the proper set of techniques to
internalise the procedures into local organizations and to
integrate them into the country's policies?).
National Environment Secretariat
Ministry of Environment and Natural Resources
PO Box 67839
(now at Graduate School of Geography, Clark University, 950 Main
Street, Worcester, MA 01610, USA)
Center for International Development and the Environment
1709 New York Avenue NW
Washington DC 20006
Diagrams for Participatory RRA
In an earlier article for RRA Notes (number 4) I described some
participatory RRA work with the Aga Khan Rural Support Programme
(India), (AKRSP(1)). The variety of diagrams which we used
during this work certainly seemed to be valuable in encouraging
us to involve some of the local people in the RRA. Here I will
highlight some of the advantages of using such diagrams, and some
of their limitations which I feel need to be overcome if the
diagrams are to be widely applicable in participatory RRA work.
* At an early stage of the RRA, updating an existing map of
the village with a group of farmers meant we spent several
hours together, discussing with them the purpose of our
visit, the nature of the RRA as well as the recent changes in
land use in the village.
* After the field investigations, presenting our findings in
the form of diagrams to a selected group of community leaders
and women allowed us to further cross-check the information,
correct any mistakes and discuss the issues represented in
* The leaders felt that the diagrams would help them to put
across the main findings of our work, by highlighting the key*
issues which had emerged, and so the leaders themselves
presented the diagrams at the village meeting scheduled at
the end of the RRA.
* The diagrams presented at this meeting served the purposes
describing the village (through the map and transect) so
discussions could begin on a common understanding of the
highlighting the key issues, and, by their very presence,
forcing these issues to be addressed. One leader,
resenting the fuelwood calendar only fleetingly, was told by
a farmer to hold it up again so this important problem could
showing community achievements and potential for
improvements (for example a crop calendar of salt-tolerant
varieties, some of which had been tried by farmers, and a
water quality calendar focussing on ways of extending the
period when the water is fresh),
making problems and opportunities explicit (particularly in
the transect) so they can be questioned and discussed openly,
encouraging individuals who might otherwise have remained
silent to speak up. For example, several women commented on
a fire-wood calendar saying that it was unrealistic the
real situation was much worse.
The way the diagrams were used in these village meetings also
taught us much about the social structures in the villages. For
example, an elderly Brahmin villager took much time to explain
the diagrams to the men and women and clearly had respect. He
was identified as an important contact for AKRSP (1) and indeed
became chairman of the village organisation which was formed
after the village meeting. In another village, the authoritative
domineering manner of the village chief became obvious as he took
control of the meeting, using diagrams to try and "blind by
science" the other villagers. A possible threat to his
leadership became apparent as a young man continually questioned
the chief's comments and criticised his presentation. It turned
out that this was an ongoing power struggle which we only learned
about at this late stage of the RRA.
* Some of the diagrams are too complicated to be widely
accessible. This is especially true of the transect. We
tried to avoid this problem by presenting the transect as a
dialogue, with an RRA team member asking one of the leaders
about the transect as it was held up eg "what particular
problems are there in the grazing land?", and the leader
would respond by reading out the problems listed in the
* Related to this point, the wording in the diagrams needs to
be minimised. We used some colour coding in the seasonal
calendars to represent the three seasons, rather than writing
out the months. Better ways need to be found for
representing issues pictorially.
* Even with clear and simple diagrams, their usefulness is
limited if the size of the meeting means that a proportion of
the audience cannot see them.
* The diagrams we used were drawn by ourselves, the RRA team,
and although frequently amended by the villagers still showed
signs of outsiders' perspectives. In particular, the map -
drawn conventionally with north to the top proved to be
upside-down as far as the leader who presented it was
concerned. He ended up turning it upside-down to avoid
confusion! Local people should have been involved in the
initial drawing of the diagrams.
Overall I think the use of the diagrams was a successful
experiment, especially as they were presented by the community
leaders. Many adaptations are needed, and I am sure there is
still much undiscovered potential for diagrams as tools for
participatory analysis and decision-making.
7. MONITORING AND EVALUATION
PBME and Rapid Rural Appraisal
A new discipline has emerged in recent years Monitoring and
Evaluation, usually known as M&E.
M&E usually involves analysis of three aspects of development
Project Impact or Benefit Monitoring.
The last of these has its own acronym, PBME (Project Benefit
Monitoring and Evaluation). PBME aims to identify the extent to
which intended project beneficiaries are in fact receiving the
benefits planned. It also looks more broadly at the impact
(positive or negative) the project may be having on other people,
on the environment, on the regional or national economy.
The following notes look at the role RRA may have in PBME of
agricultural development projects.
Weaknesses of "Conventional" PBME
The typical M&E programme has a number of shortcomings:
- use of expensive random sample surveys;
- use of randomly selected crop-cutting samples;
- long delays in presentation of results and reports;
- a need to drastically reduce the data collection programme
after two or three years, when it is realized that the
original programme was over-ambitious;
- detachment of the M&E unit from the day to day activities and
preoccupations of project management and other line staff;
- a tendency for routinisation and for the M&E unit to become
the project report writing unit;
- concentration of PBME activities during the donor funded
period of project activities;
- low priority given to PMBE by the host government once donor
funded activities have ceased.
The Potential Contributions from RRA
A. Cost Effective Data Collection
For certain kinds of data that have "conventionally" been
collected by random sample survey methods RRA techniques
offer a cheaper, faster method of data collection, that has
been shown to provide data of acceptable quality for PBME.
These data areas include:
crop and variety preferences;
cropping constraints, and problem identification in
S crop and input prices.
These forms of data tend either to be qualitative but with
relatively small expected standard deviations.
B. Cropping Parameters
Three key parameters usually need to be measured if an
agricultural development project's impact is to be monitored
changes in cropping patterns;
changes in cropping intensity;
changes in crop yields.
Various RRA techniques have been tried to measure changes
in the three parameters:
ground transects (village walks; along irrigation bunds);
aerial survey or photography, including use of video
interviews with key informants;
the use of secondary indicators;
Where the extent of change in cropping patterns and intensity
is substantial, secondary indicators and group interviews
have proved effective. However the effectiveness of all RRA
techniques to contribute in this area is limited by three
Project boundaries often do not coincide with the
boundaries used to assemble secondary data, or perceived
The scale of annual changes anticipated is often small.
For example a 50% increase in yield over ten years is
only a 4.1% per year increase. As cropping data from a
sample of farms tends to be subject to substantial
variation, only data from a large random sample is ever
likely to provide data to a degree of accuracy that could
reliably trace such small trends.
Short term factors tend in any case to disguise trends of
change of this magnitude. Short term relative price
variations and climatic fluctuations can cause blips in
cropping pattern and yield trends which are far larger
than the "expected" 5% or 10% annual average increase.
C. The Use of Diagnostic Surveys
Diagnostic Surveys have been shown to make a substantial
contribution to improving M&E programmes, both at the outset
of project activities (to better define existing conditions
and needs) and during implementation to permit enquiries into
problems identified by project management, by project
technical staff or by the M&E unit itself.
The advantages of Diagnostic Surveys include:
allowing the scope of periodic surveys to be reduced,
thus accelerating the process of data processing and
adding flexibility to the PBME programme;
drawing the M&E staff closer to other project staff;
improving the motivation of M&E staff.
Hunting Technical Services Limited
Using RRA to Evaluate NGO Projects
The early 1980s saw a rapid increase in the level of support
provided by the EEC to NGOs under its programme of Small
Development Projects. This expansion took place in a largely ad
hoc fashion, without any real overall sense of direction or
priorities. In 1984, I was asked to work with a small team of
consultants which would review what had been achieved, and
suggest cleared guidelines for future action. Our first major
task was to design and test an evaluation system which would
reflect the distinctive objectives of NGOs. This was then to be
used in some 30 individual investigations, from which more
general conclusions could be derived.
The system we devised was strongly influenced by ideas about RRA
current at that time. If the exercise were to be repeated today,
it would almost certainly be modified to take account of more
recent developments. But although somewhat dated, the experience
may still be instructive, since accounts of the use of RRA for
evaluation, as opposed to identification, remain comparatively
We took the EEC's existing "general evaluation criteria" as our
point of departure, developing these in three particular
directions to reflect what we took to be central NGO concerns.
In the first instance, an attempt was made to go beyond
conventional pre-occupations with inputs, outputs and immediate
effects to look, in more detail than usual, at impacts. In the
case of a project designed to increase production, for example,
analysis would not be confined to identifying increased income,
but would aim to explore how this was allocated between different
types of expenditure, and at the consequences following from this
for those providing the goods and services consumed; as well as
for the future prospects and well being of the household itself.
Secondly, provision would be made to look at who was capturing
any benefits arising, with particular emphasis being placed on
the implications for women and the poorer inhabitants of areas in
which activities were located.
Thirdly, the analysis would extend beyond direct material
outcomes to explore institutional developments. A number of
dimensions were to be considered here, including:
S the viability of the structures created for the
implementation of the project itself;
S the consequences of participation for the weakening or
strengthening of the existing external relationships of those
households taking part;
the effect on relationships within the household.
To illustrate what this entailed, I shall use the example of the
preliminary study of an irrigation tank renovation project in Sri
Lanka, for which I was responsible. This was administered by the
Sri Lankan National Freedom from Hunger Campaign Board (FFHC),
and was jointly funded by the German freedom from Hunger Campaign
and the EEC. I was give 16 days for data collection and
analysis, followed by a further 14 days for report writing. I
was assisted in the field by a Sri Lankan anthropologist, who
worked half time, and a member of the FFHC staff who was
available throughout the initial period.
In outline, the material required was divided into three broad
- the context or environment;
- the project system;
- the impact.
Presented in the broad sequence in which they were used, the
methods employed for data collection included:
- analysis of secondary sources (including project records);
- interviews with project staff;
- direct observation and mapping;
- key informant interviews with local leaders and officials;
- separate group interviews with male and female participants;
- household case studies with male and female interviewees.
The content of individual categories and the ways in which they
were explored may now be considered in more detail.
(i) The Project Context
The particular objectives guiding the exercise led me to
devote a greater proportion of the total time available to
the investigation of context than is customary in project
evaluation. This was an essential pre-requisite for the
subsequent identification of the changes arising in
"external" social and economic relationships. In view of
the increasing number of factors, deriving from beyond the
project itself, which come into play as one moves further
down the chain of impacts, this was also important as an
aid in isolating project impacts from other changes taking
place during the same period of time.
The context was treated as a matrix of possible influences
(Table 1). Four areas of investigation were identified
along one axis, and four different levels, at which each
could be explored, were set out along the other. Entries
in individual boxes in the figure provide examples of the
types of data collected. The amount of time devoted to
different aspects of the context increased moving from the
left of the matrix to the right.
The national level was reviewed almost entirely from
secondary sources, and was relevant primarily as a means
of assessing potential replicability. The physical
environment of the project area was explored mainly through
a series of walks, which were undertaken during the first
two days in the field. The structure and evolving nature
of production activities, and of economic, political and
social relations at the area level, were explored
initially through key informant interviews with leading
male and female residents. Findings were then checked in
subsequent semi-structured group interviews. Some groups
only involved women, and these were used as the primary
source of information about intra-household relations.
(ii) The Project System
Although the context precedes the project system in the
formal presentation of the methodology, logistics dictated
that there should be a considerable degree of overlap in
the way in which they were investigated in practice.
Enquiries started here with initial interviews with
project staff, combined with inspection of project records.
These were used to construct a preliminary picture of
overall objectives; to capture basic input/output
relationships and to understand the strategy for building
the institutions which would be required to implement the
project and sustain activities in the post-project period.
Preliminary impressions were then fleshed out in the group
interviews, where it was possible to look in greater depth
at the sequence of events in individual instances, starting
with group formation, and working through the tank
construction phase, to the provisions made subsequently for
the organisation of irrigated agriculture.
Table 1 The Project Context
(iii) The Project Impact
The group accounts of the functioning of the project
system, and especially of the manner in which the problems
encountered had been overcome, led automatically to an
appreciation of key institutional impacts. The women's
groups provided additional insights into transitory and
enduring shifts arising in the division of labour within
households. Finally, the interviews also served as a
medium for identifying the general types of impacts arising
for individual households, which were then taken as
hypotheses to inform the detailed case study.interviews
which completed the investigation.
Six households were used for this purpose, selected on the
basis of quality of housing, to represent the upper, middle
and lower points in the economic spectrum from which
participants were drawn. In a series of interviews taking
some three to four hours in all, questions were asked
about: land holdings; production activities and relations;
exchange relations (i.e. trade, loans and debts); patterns
of consumption; patterns of expenditure; patterns of
investment and health. In each instance, an attempt was
made to establish the situation prior to the start of the
project, to compare that with the present, and to account
for any differences which might have arisen.
The comparative aspects were dealt with through closed
questions, whilst those exploring causes were open-ended.
Where possible, and where it appeared likely that
understandings and interpretations would diverge along
gender lines, both male and female informants were
interviewed, although in some cases the absence of the
household head made this impossible.
In the course of exploring what, for the most part, may
appear to be a rather narrowly economic range of issues,
the format in fact yielded a considerable amount of data on
institutional change. Increased food production was shown,
for example, to reduce purchases and lessen dependence upon
traders. Similarly, increased opportunities for productive
work on participants' own land reduced the need to seek
work on large farms in the neighbourhood, and improved
their bargaining position vis a vis their employers.
Increased incomes created opportunities for ceremonial
expenditures, which, in turn, contributed to the creation
of a sense of community where little had previously
existed, and so forth.
What has been described is an attempt to deal with a relatively
complicated set of questions in a comparatively short period of
time. Wherever possible, different methods and different types
of informant were used to counteract the dangers of bias and
false inference. To this extent, the approach corresponds to RRA
as it is currently practiced. But readers will have no
difficulty in recognizing points at which it could be
Diagrams could have been used at an early stage as a means of
representing key aspects of the project environment, and of
confirming or refuting hypothesised relationships with
informants. Wealth ranking would have sharpened our perception
of economic differences and probably lead to a more satisfactory
selection of case study households. A final meeting to report
back on major findings would almost certainly have uncovered
errors and revealed significant relationships and changes, which
had not hitherto been apparent.
It is difficult to see how these modifications could have been
made without some net addition to the time required to conduct
the exercise, although this would not necessarily have been very
large. This, however, could probably have been justified in the
light of the improved quality in the data obtained especially
in view of the fact that this was a pilot exercise for a much
larger intended programme.
One other limitation should also be noted. Although the approach
was applied to reasonably good effect in the case described, it
was hardly used at all in the 30 follow up studies which
constituted the major data collection part of the overall
programme of work. A number of factors which have no bearing on
the present discussion were partially responsible for this
omission, but even when these are allowed for, it became
abundantly clear, in retrospect, that a manual, by itself, cannot
even come close to providing a sufficient basis for the
introduction of a new system of evaluation.
Significant advances can only be achieved where the basic text is
brought to life through various forms of "hands on" experience.
At the very least, this requires the kinds of workshops with
which RRA is now increasingly associated, but which have yet to
gain widespread acceptance in the project dominated procedures of
the larger agencies. Ideally, it would go beyond this to
encompass a process of field testing, followed by modification in
the light of experiences. This point needs to be made even more
strongly in relation to the more innovative self- or
participatory forms of evaluation in which NGOs are now showing
an increasing interest.
Carr, M., de Crombrugghe, G., and Howes, M. 1984. Assessing
Rural Development Projects: an approach to evaluation as if
people mattered. EEC
Howes, M. 1984. An Evaluation of the Sri Lankan Freedom from
Hunger Campaign Board Small Tank Renovation and Rural Development
de Crombrugghe, G., Howes, M. and Nieuwkerk, M. 1985. An
Evaluation of EEC Small Development Projects. EEC
Institute of Development Studies
University of Sussex
NOTES ON THE DISCUSSIONS HELD BY THE 35 PARTICIPANTS AT THE
JOINT IDS/IIED RRA REVIEW WORKSHOP, 19-20 JUNE 1989.
A brief disclaimer: these notes on the discussions at the
workshop have been compiled by Jules Pretty, Robin Mearns,
Jennifer McCracken and Ian Scoones. They do not necessarily
represent consensus opinions, and some of the points are
contradictory. We have attempted to be as faithful to the
discussions as possible, but of course take full responsibility
for any mistakes or misrepresentations.
Notes on Objectives
List of topics arising from discussion of objectives for the
review workshop day, plus those arising during presentations and
- philosophy underlying RRA
- institutionalising RRA
- the project cycle
- links with other research methodologies
- quality control and failures
- new applications
- the dangers of RRA
Notes of aerial photo session
* Photos are useful for eliciting information.
* Are there any other uses for land photographs? In Nigeria
pictures of each species of 21 grasshoppers were presented to
farmers, and everywhere people could identify easily those
that caused problems. In one week more information of a
higher quality was gathered than in 6 months of conventional
laboratory based research.
* In another example time series photos were useful in Nigeria.
A grasshopper pest of Cassava was known to move from a weed
Eupatorium to the crop. Photos were used to demonstrate to
farmers what was the difference between fields with and
without the Eupatorium.
Notes on groups and group interviews
* A tip for involving all the members of a group being
interviewed ask everyone in turn the same quickfire
question, such as what was the price for ... last year? This
brings everyone into the interview.
* Another kind of 'fix': invite leaders/outspoken members of a
group to sit apart from the main group to hold own discussion
i.e. extract and promote them (perhaps sit on chairs, not on
ground). Kenya Woodfuel Development Programme in Kisii use
exactly this technique in their community meetings.
* But the fact that one or two people dominate may be very
culture specific. Often a group will predetermine who will
speak, and often they really do speak for everyone present.
It may be helpful; to let people/respondents know what issues
are to be discussed some days ahead of time, since husbands
and wives frequently discuss matters in private, for which
the husband is often the public spokesperson. This may give
the false impression to outsiders that the men dominate
discussion to the extent that women's views are not
represented at all.
* The nature and structure of any questions could make all the
difference to the interview.
* Need to identify interest groups as well as focused group.
Projects are not 'neutral' or inheritantly 'successful' in
their technical make-up; 'success' may simply mean a project
secured the support of powerful interest groups. The
technique of focused group interviewing does appear to have
great potential for getting a handle on this issue. There is
no underlying 'truth' to be revealed.
* Just because people are being polite does not mean they are
telling the truth. Both sides interviewers) and
interviewee(s) are trying to establish what it is that the
other side wants/expects out of the process. Cannot get rid
of biases, but can be aware of this problem.
* But never promise anything. Danger of raising expectations.
* A good interview involves an exchange of information.
* Ranking in itself is not enough need always to ask why and
derive criteria. It is this process that is so important.
Don't forget the value of just asking "But why?"
* Ranking may bias what you are looking for, may not allow you
to discover unusual things.
* Statistical correlation techniques may help with some
complex DMR procedures, although of course, a poor
correlation does not mean the answers are in any way 'wrong'
Discussion on validity of statistical techniques.
* Routine of carrying out these procedures can be easily learnt
by outsiders and may in fact bring about professional
reversals, through the nature of the technique itself it
forces one to appreciate local valuation etc. and may force
one to learn from local people.
* Can the technique be easily learnt? Complexity of some
variants raises issues of training needs e.g wealth ranking.
* What is the purpose of RRA in general? It is simply to
identify the right questions to ask, or is it more than that?
Discussion on health
* Was RRA successful in converting the medical professionals
from their desire to tell people what to do, to just
listening? Not entirely. The final proposal did not reflect
* Problems may arise if the community realises the team have
come up with something completely different to the community.
In fact if the same thing happened in a rural area the
final outcome failing to take the RA and views of local
people into account there would be great resentment from
the local community, jeopardising future relations with
Discussion on M & E
* RRA is excellent for determining whether or not farmers are
prepared to adopt/will benefit from project outputs.
* Integrating farmers themselves into M&E process numerous
examples show their analyses may be more accurate. Walking
transects with farmers, at the very least.
* There are difficulties in evaluating small expected changes
around a very variable mean. Farmers might be able to sort
through year-to-year variation, perhaps through the use of
* Farmers' measures of yields are they better or worse than
those of experts?
* Can project benefits be measured anyway? Perhaps not if the
expected levels are to be small.
* Also people's attitudes may have changed, which may not
appear for some years. Also these may not necessarily be the
things that the project will be monitoring.
* Emphasise the value of quick diagnostic instant survey to
monitor what is happening. Diagnostic RRAs are most valuable
at the time of project appraisal. Too time consuming if use
baseline surveys. RRA has the advantage of being potentially
part of the process they can be repeated.
* What about M&E of small-scale NGO projects or within
government. Institutions that cannot afford near-full time
M&E staff? Perhaps a useful approach here is to hold regular
workshops with farmers/extensin agents. What turns this into
M&E is the close documentation of proceedings (e.g. the CARE
International/FAO Agroforestry Monitoring & Evaluation
Methodology Project, funded by Ford Foundation, is moving
more in the direction of simpler qualitative M&E techniques
for these kinds of projects).
Notes on Participation
* Have any of these participation projects been evaluated?
VRMPs will be baseline document. Will be able to use this as
a measure of who has/has not done what they said they would.
* Do you attempt to evaluate the process or the physical
outcome of the process as an indicator of the effectiveness
of the approach? The timescale is vital here: the process
may not yield highly visible outcomes for a long time. The
important thing is to change attitudes regarding local
people's capacity to act for themselves.
* Participatory approaches in health have been shown to improve
health in some communities, but probably only where NGOs play
a role. These also tend to be very site specific, making it
difficult to scale up to a wider area.
* Are there any preconditions within the community for good
issues of leadership
issues of appropriate institutions
* Too stong a leadership within a village not necessarily good.
Factionalism might also be destructive, though RRA type
process involving people in the research may actually break
down some of this factionalism.
* Indigenous rapid appraisal: local people are way ahead of us
in terms of monitoring their own environment
Farmer and soil types story from Nigeria: an old man who
identified different soil types by the respective changes
in his footprints after rain.
-farmers have also dug soil profiles to monitor water
changes over the seasons.
* Who is participating in whose development? RRA and
consultants vs locals. Great potential for local use.
* Although participatory development implicitly involves a
notion of empowerment of people in their own development some
distinction should be made between empowerment and
* Importance of historical context: case of toxic waste dump
in Appalachians, USA, where an old man with intimate
knowledge of local herbs, fish and woodlands identified its
existence and negative effects long before the
professionals/authorities. Other local people thought him
mad and dismissed his fears since the authorities had said
nothing about it. Only when it was pronounced an official
problem was the old man's view vindicated. Community
development project responded by teaching people to read
medical dictionaries so that they could critically question
the health professionals' statements on levels of risk to
their health. This worried the professionals for a while
Also diagramming of oral knowledge was used.
Notes on Ideology
* What drives RRA? Is it just a more effective research tool,
or does it have an ideological framework?
There are perhaps three levels in RRA:
tools that are value neutral and can be used by anyone
clusters of cultures of RRA, in which the range and
diversity is a thoroughly good thing
overarching philosophy that resonates with changes in
ideas in development.
* What is 'rapid'? Tools and techniques can be conducted
quickly, but it does not necessarily speed up development.
It does imply a move to a more process orientated approach.
* The name is a problem: RRA is a lable, but it is too late to
* The selling of RRA: is RRA sold or presented in
contradistinction to formal surveys, anthropological
participant observation or other methods? Why is RRA
different to rural development tourism?
RRA is complementary to long-term, anthropological work and
conventional surveys, but it is a more structured and
rigorous alternative than what short term consultants do at
present and has a greater potential for using local
professional and farmer expertise.
* Indeed should it be sold at all? Participants drew attention
to the paradox between promoting RRA over the years in order
for it to achieve a measure of respectability, yet not over
promoting it such that it becomes mystified.
* RRA is always evolving. There is no single model. This is
why it is currently still impossible for anyone to produce a
single training manual that encompasses all of RRA.
* The actual differences between RRA and other conventional
methods: these may not appear too great a number of
participants quoted from practitioners and consultants heard
to say that they had been conducting RRAs "for years". Yet
there are key differences in:
* RRA must continually be demystified.
* RRA legitimises curiosity.
* What is different about RRA?
everyone is involved in the process, or a single person is
involved in all parts of the process
those who are collecting data have to take rural people
* The repetoire of methods has enormously expanded over recent
years, creating a growing pool of common knowledge. These
methods allow sharing and cross-fertilisation of experiences.
* Two roles distinguished for RRA?
acquiring good information, and
* Territory and ownership issues. It is critical that there
are no ownership issues over techniques and tools. This is
central to the RRA philosophy. This is also the reason why
RRA Notes is produced.
* Some discussion over development that deals with 'problems'.
Maybe this kind of approach creates blinkers for experts
trying to understand what is happening on farms. We should
not let the negative connotations of problem identification
stop us looking for more than a problem. For example looking
for opportunities. But even this implies coming from the
outside with changes. We should first understand some of the
tremendous achievements of local people in managing their own
Notes on the Dangers of RRA
* One participant quoted from a recent donor agency document:
"Rapid appraisal should be used in preference wherever
possible by all consultants for the agency".
Another quoted from a recent project proposal made to an NGO,
namely that the contractors would conduct a standardised
survey "in an RRA format".
More thought should be spent on the drawing up of Terms of
* There are limitations in the current project cycles. A good
appraisal still needs its findings/best bets to be seen
through. The question is, do current institutional
structures inhibit good follow-up? RRA may just be raising
expectations beyond what can reasonably be achieved.
* We do not know enough about cultural specificity and RRA
tools and techniques. Dangers relate to techniques that
appear standard but have not been tested. Also a danger of
missing the opportunities of learning from culture specific
techniques, eg the Ayo board.
* Take warning from FSR. This got overdone, ultimately at the
expense of quality. There is a severe danger that as soon as
something is adopted as an orthodoxy it will be killed. To
learn from this experience RRA should not be oversold, and
should be well insititutionalised. How can enthusiasm for
RRA be moderated to prevent the fall from grace? In
particular, there was a large concern that RRA could very
quickly be swamped by large amounts of money.
* A method of moderating RRA emphasise ifs and buts emphasise
what RRA cannot do.
* It is incumbent upon all of us as practitioners, especially
consultants, to use all RRA techniques and tools well.
Always document methodology very closely.
* On the question of quality control: this may be more
straightforward when using conventional surveys. It is clear
what can or should be measured, and the questionnaire is
probably present in an appendix of the report. So how do we
evaluate whether an RRA has been done well?
* Self-criticism, unfortunately, is not conventionally rewarded
as it should be. In fact, it adds greatly to the credibility
of the piece. An indicator that an RRA has been done well:
search for self-criticism on behalf of the authorss. If an
author exposes themself through self-criticism then this adds
strength to the report. However this is not easy just such
self-criticism in a report has been used to attack the
overall approach and people involved.
* There have not been enough studies debunking bad standardised
surveys, especially drawing attention to non-sampling error.
Such studies could help to ensure that RRA is not seen as an
alternative, but is genuinely attempting to further
Notes on Training in RRA
* Who trains and who is trained? All different levels can be
trained. General consensus: local trainees are best, helps
them gain knowledge and power.
* Problems with outside trainers: insufficient local
knowledge, lack of appropriate language. Need of development
and support of Southern organizations. NGO sector should be
* Many NGOs already use RRA-type approaches in project design.
Can learn a great deal from them. But we should be aware of
excessive faith here too.
* Perhaps all training manuals should carry a health warning.
* RRA is not simple to learn, not easy to use.
* There is a need to follow up training, perhaps over a period
of years, to ensure that trainees are able to conduct RRAs
well. There is a trade off between quality of training
involving good follow up and quantity involving trying to
train too many people: the consensus is for quality.
* How successful have trianing programmes really been? Most
felt it was still too soon to say.
* One problem is that there may be a strong personality factor
in the likelihood of successfully training people in RRA.
For example, perhaps only some people fall easily into the
open endedness of RRA.
* Training in the future should look closely at multi-media
approaches, such as use of photographs, videos.
* Good RRA training is manpower intensive, and should where
possible make use of both workshops and field work.
* A further problem is that we do not as yet know whether
trainees actually go away with adequate knowledge of RRA.
Has the training course initiated a change in their
behaviour? Any follow-up needs to be aimed at institutions
rather than individuals.
* RRA is rapidly evolving, and thus it is very difficult to
conceive of producing training manuals that are capable of
dealing with the whole topic in a sufficiently open-ended
fashion. Training manuals that are both flexible and NOT
* Two central issues networks and cautions.
* A model suggested might be centred upon participation,
empowerment, diversity and democracy.
* Avoid tendency to standardise, centralise, and in the
agricultural field there is a reasonable chance that these
approaches can achieve the required increases in productivity
and sustainability for those areas of the new paradigm of the
* Take the point of diversity: a serious limiting factor might
still be information flows, particularly between different
languages. The workshop strongly recommended that RRA Notes
should be produced in French, so as to give the opportunity
to many francophone researchers the advantage of the
techniques, as well as contribute from their tradition. It
was resolved to search for ways of increasing translation of
* Keep publications looking non-polished encourages the
innovative/inventive aspect which is central to the success
* Salute for IIED's RRA Notes great source of information on
diversity of applications and disciplines.
* No principal author yet in RRA Notes has been from the
South those in the South ought to be encouraged and
supported to write up their work.
* The conditions for achieving good RRAs include some or all of
avoid the production of overglossy manuals and documents
encourage a learning approach in all projects
help people create their own way of doing things much
good RRA is inventive and innovative.
* Need to get away from labels RRA especially since this
obscures the real principles involved. Labels are like
trademarks too, and let us not get into the whole business of
proprietary attitudes over ideas and approaches.
* Those who are collecting data have to take rural people
seriously putting the farmer first in the process.
* As put by a participant from a developing country at the end
of the workshop: the bottom line is that developing country
people have to live with the development.
RRA Review Workshop
IDS, Sussex & Univ of Ilorin,
Emerson College, Brighton
Aga Khan Foundation, London
OFI and Nairobi
University of Hohenheim, FRG
DSU, University of Stockholm
Univ of Tennessee, Knoxville
Huntings Technical Services
London School Hygiene and
DSU, University of Stockholm