Jamry- March 1987
CIMMYT EASTERN AFRICA ECONOMICS PROGRAMME
INTERNATIONAL MAIZE AND WHEAT IMPROVEMENT CENTRE (CIMMYT)
P 0 Box 25171, Nairobi, Kenya, Telephone 592054/592206
The Involvement of Farmers and Extension Personnel
In Carrying Out On-Farm Research (OFR)
D. Sungusia & Larry Lev
On-farm research (OFR) has gained favor in recent years as a needed
supplement to the traditional on-station research (OSE) for two main reasons:
1. It permits the researcher to test new technologies over a variety of
2. It permits the researchers to incorporate farmers' circumstances (in the
broadest sense), management, and evaluation into all phases of
You will note that the inclusion of the second point distinguishes OFR from
multilocational testing which does involve the sampling of environments but do
not ordinarily incorporate a farming systems' perspective.
Successful OFR requires the combined efforts of a variety of participants
including adaptive and commodity researchers, extension agents, village and
national political leaders, and of course the farmers themselves. In this
paper we focus on our experiences in Tanzania in insuring the effective
participation of two of these groups, extension and farmers, in the OFR
process. Although there are no hard and fast rules for conducting OFR, the
FSR community has made substantial progress in defining workable research
methods. This paper should be seen as adding to these general guidelines.
The USAID funded FSR Project is located within the Tanzania Agricultural
Research Organization (TARO), a parastatal organization. The Project is three
years old and was designed as a pilot activity which operates in three districts
(Kilosa, Moshi and Dodoma) out of the eighty seven districts in the country.
The TARO/FSR teams, which are comprised of biological and social scientists,
operate from two zonal Research Stations (Ilonga and Lyamungu) and conduct
their research activities jointly with extension agents, commodity researchers
and farmers in the area. By the end of the USAID component of the project
on September 30th, all phases of the OFR process will have been completed in
each of the pilot districts and OFE results will have been generated for two
years (four seasons) in Kilosa, two years in Moshi and one year in Dodoma.
Presented at the Networkshop on On-Farm Research, Ethiopia.
August 3 10, 1986.
Coordinator, TARO/FSR Project, P.O. Box 9761, Dar as Salaam, Tanzania
3 Chief of Party, USAID/TARO/FSR Project, P.O. Box 9130, Dar es Salaam, Tanzania.
The Involvement of Extension in OFR
Whereas agricultural research activities are primarily carried out within
parastatals in Tanzania, agricultural extension is the direct responsibility
of the Ministry of Agriculture and Livestock Development (MALD). Operationally,
extension functions in a decentralized fashion as one moves from the national
level to the region, district, ward, and finally the village. Overall there
is one extension agent (Certificate/Diploma holder) to service approximately
1000 rural households. Quite obviously the individual agent faces a
The extension service has been often criticized in Tanzania for failing to
achieve its primary task which is to disseminate new technologies. It must
be recognized, however, that the extension agents can only carry out this
task if attractive new technologies are developed and made available to
them. It is not clear that this has been the case in the past.
From the beginning, the TARO/FSR staff recognized the critical role of
extension personnel in the successful implementation of all phases of the OFR
process (Sungusia and Acker, 1985). In the initial phase they should
participate in the zouation and selection of priority research areas. Next,
during the diagnostic stage in which priority problems/opportunities are
identified, the extension agents can serve as both resource persons as well
as interviewers. During the planning stage, the extension agents should
insure that the research program continues to focus on farmers' priority
rather than be diverted to issues of interest to researchers alone. The
knowledge which extension agents have of farmer circumstances should be
particularly helpful during the prescreening of potential solutions.
It is during the experimentation phase that the critical role of the extension
agents is highlighted. In contrast to researchers, the extension personnel
reside in villages. Thus whereas the researchers are continually plagued by
transport problems and uncertainty about current conditions in the village,
the extension agents do not suffer these problems. This is particularly
important in insuring the timely planting of trials with the beginning of the
rains. In specific terms the extension agents can assist with:
-- the selection of farmers and fields,
the explanation of field instructions,
-- the laying out and managing experiments,
monitoring and record keeping,
-- the evaluation of results,
the modification of the experiments for future years, and
-- follow up to ascertain actual farmer adoption of new technologies.
Throughout, the extension agents provide feedback from farmers to researchers
and vice-versa. In addition, they can also:
help in carrying out surveys and special studies,
supervise record keeping and climatic monitoring, and
- organize field days.
The problems which we have experienced in Integrating the extension service
into the OFR work can be divided into two basic categories: organizational
(linkage) problems and training problems. The organization problems stem
from the fact that under the existing institutional structure there is no
simple means of functionally integrating research and extension personnel into
a single team. Without clear directions from the top, it is difficult to
attract the attention of the field extension agents who are already saddled
with a full work load. A formal agreement must be worked between TARO and
MALD which will relieve field agents from some of their other duties and
provide them with adequate financing so that they can participate as full OFR
The training question is equally important. We must not forget that OFR
work is fundamentally different from the traditional work of extension agents.
Instead of disseminating a new technology, they are being asked to participate
in the development of new technologies. If the extension agents do not
understand this distinction, it is impossible to expect that the farmers will.
In conducting research, extension personnel must overcome a number of their
ingrained tendencies. They must realize first of all that many of the new
technologies will fail. They also must recognize that it is important that
the experimental work is conducted with representative farmers, not just the
most advanced farmers. Finally, they must realize that their job entails
developing a better understanding of how the current system operates and can
be improved rather than simply acting as a conduit for information. In order
to prepare and retrain extension agents to serve on OFR team, it will be
essential to arrange special seminars and field practical. To date this has
not been done, as the extension people have been attending common training
sessions with the researchers.
The involvement of Farmers in OFR
Since farmers are the ultimate clients of OFR, it is important that they be
involved throughout the process. This is the best way of insuring that they
will adopt the new technologies which are developed. In concrete terms this
implies that the priority problem areas are identified based upon their
priorities and that their participation in the testing and evaluation permits
them to contribute to the fine turning of the technologies. Lightfoot (1985)
notes that farmer involvement tends to be highest at the beginning and the
end of the OFR process but is often neglected during the actual experimentation
phase. This would appear to be a serious oversight.
Let us begin by looking at the role which farmers should play in the'
formulation of the research program. Our work in Kilosa District provides a
good example of the importance of involving the farmers in the priority
setting stage. Whereas station based researchers had spent thirty years
working on maize production during the main rains, the farmers just outside
the station fence concentrated virtually their entire production effort"
during the short rains. The OFR team was able to generate considerable
excitement by promoting a short season maize variety which was languishing on
the research shelves. Until OFR came along no one had bothered to ask
farmers about things like the "hungry season" and labor conflicts which so
greatly influence their cropping pattern. (Lev, 1985).
Our experience has in fact shown that farmers are avid experimenters in their
own right. Thus in our own program many of the experiments, which we have
planned and carried out, are, in fact, modifications of locally generated
technologies found during our initial and follow up farm visits. Some examples
of these are the work on maize/cotton and cowpea/cotton intercropping work in
Kilosa District. In both instances the farmers were far ahead of the
A slightly different example is the maize/bean paired row intercropping
trial which is being conducted in Moshi District. In that area, our on-farm
visits indicated that farmers were practicing two distinct types of maize bean
intercropping. In the first or standard pattern, single rows of beans were
intercropped between single rows of maize; the bean were regarded as a
secondary crop. On a smaller but significant percentage of the maize/bean
land, multiple rows of beans were intercropped bweteen widely spaced rows of
maize and the two crops were regarded as having roughly equal importance.
Building upon this second technology, the OFR team (in conjunction with the
maize and bean programs) planned and carried out an experiment in which
multiple rows of beans were intercropped between paired rows of maize. Although
it is too early to report any conclusive results, the initial indications are
It is also clear that farmers experiment with their own modifications to
introduce technologies. In Kilosa, when we introduced the new short maturity
maize variety, we found that even in the first year some farmers were
experimenting with alternative patterns of integrating the new seed. Among
the options observed (all of which may help to reduce the risks of crop
failure) were growing the new seed in adjoining plots with a full maturity
variety or in alternate rows or, finally grown mixed in the same hole.
The levels of farmer involvement can be defined as Researcher Managed Trials,
Researcher/Farmer Managed Trials and Farmer Verification Tests. Researcher
managed trials may be conducted with techniques similar to those used for
on-station experiments. In these trials, the researcher applies both
experimental and non experimental variables while the cooperating farmer
carries out the rest of the crop management according to prescribed specifications.
In the case of researcher/farmer managed trials, the experimental variables are
superimposed on farmers current practices and thus the non experimental
variables are managed by the farmer. Finally in verification tests the farmer
is largely left on his/her own.
The wide range of activities in which farmers are involved with OFR are'
-- as participant during identification of technology on design,
management and execution of trial;
-- key informant, contributing towards experimental treatment ideas and
setting non-experimental variables;
as an observer for researcher planned and researcher implemented
trials, i.e. exploratory type;
as a supplier of resources in form of land, labour, and risk bearing;
as an evaluator to provide ongoing assessment of technology
using their own criteria;
-- as an originator of new technologies.
In our program the OFR teams pick farmers to participate in OFE in
consultation with extension and the village chairperson. However final
decision on whether to select a given farmer rest with the researcher. In
order to ensure their cooperation the following check list is cleared with them:
the objectives of the trials should be clearly defined and explained;
-- participation should be voluntary;
the responsibilities should be clearly spelled out and agreed regarding
risk, timing, and management;
-- maintain a good relationship.
In most cases farmers have offered portions of their fields as experimental
sites leaving the researcher to decide on its suitability. In some cases
we have missed planting dates with researcher managed trials in Kilosa
because the site was inaccessible due to flooding or inavailability of fuel.
In Dodoma several attempts to plant the experiments were necessitated by the
delayed onset of the rains and then drought.
Working with farmers has its share of problems. Many times experiments do not
go as planned and data are highly variable or even completely lost. We have
had numerous occasions on which the farmers harvested prematurely. Clearly our
prior arrangements were faulty. One strategy which we are considering is to
enter into more formal written arrangements with farmers in which all
responsibilities for both sides are spelled out.
Even from failures thereare often hidden lessons to be learned. One good example
of this is a maize/cowpea intercropping trial which "failed" in Kilosa District.
On-station work at Ilonga ARI demonstrated that quite high returns to land could
be earned by intercropping cowpeas between maize rows. Although our survey did
not show this to be a widespread practice in the area, we proceeded to plan and
plant an intercropping trial with a number of farmers. The trial was quite
simple in design and entirely managed by the farmers. We were quite surprised
by how the fields progressed. Initially the cowpeas did quite well. Later,
however, the farmers neglected to carry out an essential second spraying and
to weed the cowpeas adequately. As a result they obtained extremely low yields.
Was the trial a failure? Well yes and no. The treatment failed but we learned
something about the place of cowpea in the system. Most farmers regard it as
a minor crop and are unwilling to provide the management care required. This
sort of understanding can not be gained in isolation from farmers. We must
never forget that it is ultimately the farmers who provide the final evaluation
of any new technologies. Therefore we must integrate them into the evaluation
process as early as possible.
The TARO/FSR project has had two years experience with OFR activities in three
Districts. Although many problems have arisen, much of the success which we
have achieved can be attributed directly to the efforts of the extension and
the cooperating farms. Both have provided manpower and expertise to the OFR
process. We are continuing to search for new ways to tap these vital
Lev, Larry. 1985. "The Kito Story". Tanzania Farming System Project
Publication No. 114.
Lightfoot, Clive. 1985. "Farmer Participation in On-Farm Trials". Paper
Presented at 1985 KSU FSR Symposium.
Sungusia, Don and David Acker. 1985. "A Study of the Role of Extension in
Farming Systems Research in Tanzania". Tanzania Farming Systems
Publication No. 112.