Ethiopia

R. PERCY, LECTURER,
Agricultural Extension and Rural Development Department,
University of Reading, and FAO Extension Education Adviser
(1994-1996) to teh Project

THIS CASE STUDY IS ABOUT A SMALL pilot project conducted in just a few areas of Ethiopia. Its aim was to initiate client-oriented extension planning based on the use of participatory approaches and gender analysis. The project - which took place from December 1994 to December 1996 at a cost of just under US$600 000 - was called Improving Client-Oriented Extension Training in Ethiopia (GCP/ETH/051/NET).

The objectives of the project were twofold: to improve the capability of the Agricultural Extension Department of the Ministry of Agriculture (MOA) to design and implement client-oriented extension training focusing on gender issues and using participatory approaches; and to develop and integrate a gender-sensitive analytical framework into the planning and implementation of agricultural programmes.

The main focus of the project was to develop a suitable extension approach based on an "action-reflection-action" learning cycle. This involved training a multidisciplinary team in gender analysis and participatory approaches, testing the approach and, subsequently, reviewing and improving it. Project planning was participatory - national staff worked with regional and zonal coordinators from the pilot areas. The first phase of the project was considered successful by the external evaluation team, and MOA has put together a proposal for a second phase.

The pilot project was very much a learning experience for all concerned, and this learning is expected to continue. The outcomes of the pilot project are therefore not an end-point in themselves, but rather a stage in the ongoing action-reflection-action learning cycle. It is hoped that the lessons learned will prove valuable to others working towards making agricultural extension more gender-sensitive and extension planning both more participatory and more client-oriented.

Background on Ethiopia

OVER 80 PERCENT OF ETHIOPIA'S 57 MILLION people live in rural areas, where most of them are engaged in subsistence farming or pastoralism. Altitudes in Ethiopia vary greatly. The lowlands are characterized by dry, sometimes drought-stricken, areas occupied mainly by nomadic or semi-nomadic pastoralists. In the mid-altitude areas such crops as maize, cotton, sorghum and legumes grow. The high-altitude areas are most suited to wheat, barley and coffee cultivation. The staple cereal crop in Ethiopia - teff - grows in mid- to high-altitude areas as does chat which is grown for commercial as well as domestic use. Ethiopia has the highest cattle population in Africa, with large herds providing status to their owners.

Since the 1970s, Ethiopia has been periodically struck by drought, and many areas consistently suffer from erratic and unpredictable rainfall. Pressure on the land is very high, with an average landholding per household in mid- to high-altitude areas of only about 0.2 to 0.6 hectares. Households are typically large with an average of seven members. The use of family planning is not yet widespread in the country, and in rural areas, especially in Muslim communities, early marriage is very common. Girls are often married or committed to marriage at between eight and 12 years of age. Health facilities are limited and sparse and many people die of preventable diseases such as malaria, cholera and typhoid as well as malnutrition.

Orthodox Christianity and Muhammadanism are the most widespread religions in Ethiopia, although there are many converts to modern Protestantism. Religious tolerance is generally high with Muslims and Christians living in harmony within villages. Culture and tradition tend to override religious differences.

The country has suffered severe deforestation with forest cover dropping from an estimated 30 percent at the beginning of the century to under 4 percent today. This, combined with overgrazing in places and intense pressure on small landholdings, has led to severe erosion in many parts of the country.

Until the 1970s, Ethiopia had an imperial regime. This was overthrown in 1974, when Haile Selassie I was overthrown and a socialist state was established by the Workers' Party of Ethiopia. In 1987 the country was renamed the People's Democratic Republic of Ethiopia. In 1991, 17 years of socialist rule came to an end when Mengistu Haile Mariam, president and chief of the armed forces, was overthrown by the Ethiopian People's Revolutionary Front (EPRDF). A new constitution was promulgated in December 1994.

The country has recently undergone a process of regionalization as part of the decentralization process. There are now ten regions, mostly based on ethnic divisions. Regionalization has led to changes in the roles of MOA at central and regional levels. At the central level, the Ministry's work is now focused on national policy issues, while coordinating and facilitating activities are carried out at the regional level. The regions have much more autonomy than previously, as have the zones within the regions.

Ethiopia has long been a major recipient of food and, more recently, development aid. Although many aid agencies now take a longer-term, more sustainable approach to aid, wherever possible, an unfortunate negative outcome of the food aid that Ethiopia has received (which was vital at the time) is the strongly dependent attitude of rural dwellers towards the government and aid agencies. Rural people usually expect the government and agencies to solve their problems and do little to tackle them themselves.

Although gender roles vary according to ethnicity, income, status, location and other factors, there are some overall distinctions between male and female roles. For example, women are responsible for nearly all reproductive tasks such as fetching fuelwood and water, cooking, washing, cleaning and child care, although there are instances where men help (for instance, after childbirth or if fuelwood sources are very distant). Men are the heads of households and are, therefore, the main decision-makers in the household, although consultation sometimes takes place within the household on certain issues. Women have longer working hours than men as, in addition to carrying much of the burden of reproductive work, they also engage in productive work. They are frequently responsible, along with children, for the care of small livestock and the production and marketing of butter, cheese and vegetables. They may also engage in non-farm income activities such as petty trading, beer brewing and leather working. However, certain agricultural activities, such as ploughing and threshing, are only done by men.

Although there is some male (and female) migration from rural areas this is not such a major factor as it is in many other sub-Saharan African countries, where the proportion of female-headed households is significant and increasing every year.

Recently, a National Policy on Ethiopian Women has been formulated with the aim of addressing women's strategic gender needs through improving their access to resources and their decision-making power. Rural women's affairs departments have been set up in the relevant line ministries as well as in the Prime Minister's office.

Various approaches to agricultural extension have been taken over the decades. The integrated approach of the 1960s and 1970s was followed by the adoption of the "training and visit" system which is now institutionalized into the structure of the bureaux of agriculture in most regions. The new system for operating at both central and regional levels is based on the package approach and is called the "participatory demonstration and training extension system" (PADETES). This system combines technology transfer and human resource development. One of its objectives is "to empower farmers to participate actively in the development process".

PADETES operates in three different ecological areas: the moisture-reliable, the moisture-stressed and the nomadic pastoralist ecosystems. In the first two of these, PADETES involves the introduction of better farming practices such as the use of improved varieties, fertilizers, pesticides, herbicides and efficient planting methods. Good cultural practices such as intercropping, planting of legumes, mulching and other soil conservation measures are being advocated. A similar package of recommended practices has been drawn up for the nomadic (lowland) areas.

PADETES has been adopted during a period of decentralization and regionalization, in which the Government of Ethiopia is promoting more local-level planning. As part of its overall priority to achieve food security, the government is pursuing a strategy of agricultural development-led industrialization (ADLI) which ensures that priority is given to both small-scale and commercial agricultural development. Such issues as distribution of inputs, storage and the processing and marketing of agricultural produce are all being addressed. In adopting PADETES, the government is accepting, at a policy level, the concept of participation in development planning and implementation.

Project design and objectives

THE IMPROVING CLIENT-ORIENTED EXTENSION TRAINING PROJECT came about, in part, as a result of the finding that Ethiopian women, despite their significant involvement in agricultural production, have very limited access to relevant extension advice. Ethiopia's agricultural extension service is male-dominated and oriented predominantly towards advising and working with male heads of households. This is in spite of the fact that women (whether they are wives, female heads of households or labourers) contribute as much as 48 percent to the country's agricultural production.¹

In spite of the thorough analysis of, and ongoing change within, the country's extension approaches, the need for farming women to have access to extension advice has not been properly considered. In most regions, the bureaux of agriculture employ (mostly female) home agents who advise rural women in the areas of health, nutrition, sanitation, food processing, vegetable growing and the management of small livestock. However, women do not receive advice on field crop production, even when they are significantly involved in certain stages of the cropping cycle. Women's contribution in this area has, on the whole, been disregarded by extension staff and, even when recognized, it has commonly been assumed that any advice given to the male head of household would reach female household members.

Meanwhile, as part of its efforts to achieve food security, the government has realized the importance of granting access to relevant and useful advice to all those involved in agricultural production. The Government of Ethiopia therefore reached an agreement with FAO to carry out a pilot project, which was funded by the Government of the Netherlands and implemented by the Extension Division of MOA from December 1994 to December 1996 in three regions of Ethiopia.

The original project document recognized that rationalizing women's position in agriculture was crucial from environmental, economic, social and population perspectives. A better understanding of men's and women's roles was seen as essential, not only for improving agricultural production but also with regard to resolving environmental problems and population growth. Women's and men's respective contributions to, and participation in, planning and decision-making activities were seen as important factors in improving their quality of life.

The policy environment in Ethiopia was conducive to a project that focused on a client-oriented approach with an emphasis on gender, owing both to the government's emphasis on decentralization and local-level planning, and to the establishment of a National Policy on Ethiopian Women. It was recognized that gender issues had to be integrated into rural planning and, therefore, that staff had to be trained on gender-sensitive participatory approaches in agricultural extension.

The overall aim of the project was to improve the quality of life of the rural population, through the introduction of participatory, client-oriented extension services. Within this overall aim the project had two objectives:

• to improve the capability of the Agricultural Extension Department, at both central and regional levels, to design and implement client-oriented extension training which focuses on gender issues using participatory approaches, and
• to develop a gender-sensitive analytical framework and integrate it into the planning and implementation of MOA's central- and regional-level agricultural programmes.

The project did not stop at capacity building (first objective), but also covered application of the participatory, gender-sensitive approach to the planning of extension programmes (second objective).

As this was a pilot project, staff training and implementation of the approach took place in a limited area. Only three regions were involved (Amhara, Oromia and Southern regions). Within each of these only one zone was selected (North Wollo in Amhara, West Harerghe in Oromia region and North Omo in the Southern region) and within each of these zones, project activities took place in just four woredas (local areas).

The implementation strategy involved training of trainers (TOT), involving MOA staff from regional and zonal levels, in client-oriented participatory extension methodology. This was followed by implementation of the approach, review and analysis of implementation, and improvement of the approach and related training programme. Networking was an important component of the strategy, not only with agricultural research stations and the Rural Women's Affairs Department of MOA but also with other line departments (health, education, natural resources), donor agencies and woreda councils. A separate component of the project consisted of TOT in design and production of such extension materials as pamphlets, posters, flip charts, slide sets and videos.

The regionalization process was taking place during the project lifetime, thus each region could decide how to work with the project. Each region nominated two extension experts, one at the regional level and one from the pilot zone, as project coordinators. Project planning was then done in a participatory manner involving a national programme coordinator, an (international) extension education adviser and the six coordinators working as a team.

The table on the following page lists the project's expected outputs.

A project framework was prepared early on and used as a point of reference during implementation. In early 1996, the matrix was revised, as part of the project's normal monitoring and planning process, to ensure extension staff had sufficient time to complete duties aside from those related to the project.

Since the completion of the two-year pilot project, a proposal for a second, three-year phase, has been prepared for donor consideration. In this phase, the approach, which is to be linked with the promotion of food security, will be expanded into new woredas, zones and regions. This second phase has rural youth, as well as women, as specific client groups. The proposal was prepared by the Extension Division of MOA on the basis of the positive assessment of the project by the external evaluation team. Staff within the regions and zones involved in the project had expressed an interest in continuing the project, as had FAO and the donor.

Project implementation

IMPLEMENTATION OF THE PROJECT involved a series of stages indicated in the figure on the following page. These implementation stages built on each other through a process of reflection or learning from experience. The points at which this took place are indicated by an "R" (for reflection). The process was accompanied by various events at different levels (national, regional, zonal) designed to raise the awareness of policy-makers and facilitate networking.

Initiation

The project started with two activities: the convening of a national workshop (discussed later); and familiarization of both project staff and project clients (staff and farmers) in order to prepare the ground for the case studies. Project staff from the national level travelled to the regions and pilot zones and, with the coordinators, briefed staff on the project's aims and objectives, assessed training needs and helped to select the case study woredas. The process was repeated within the selected woredas to choose village development centres.

For the development centres, criteria for selection included the need for the village to be accessible in all seasons and fairly typical of villages in the woreda, and to have a strong small-scale farmers' association. Development centres and woredas classified as being "self-sufficient" or "food-surplus" were selected over those classed as "food-deficit."²

Project implementation stages

Project activation

Implementation involved three training-village implementation cycles, each building on the experience of the preceding one through reflection on the lessons learned. The first cycle involved research to prepare training case studies and the training of teams in how to carry them out. The research provided an ideal opportunity to learn from experience and to start to develop a client-oriented extension approach tailored to the Ethiopian context. This experience formed the basis of the training materials used in the first TOT. Once these new trainers had trained grassroots staff to implement the approach in a number of villages, a second TOT workshop provided the opportunity for reflection on, and subsequent improvement of, the client-oriented approach developed. Grassroots staff were given further training in the light of the improved approach and implemented it in other villages.

The action-reflection cycle also occurred within each village study, including the case studies. Every day, new information gained through work in the village was analysed both by the team and, through verification, by the villagers. This analysis guided the direction of the study on the following days.

As can be seen from figure on the facing page, a sequence of training occurred. The training objectives set were in-line with those of the project and the training itself blended theory and practice concerning participatory approaches and gender analysis with experience gained during implementation to produce the client-oriented methodology. Experienced national trainers were recruited for the two TOTs. Much of the training was given in the Amharic language and most of the printed material was also in Amharic.³ Active learning techniques were employed, including group work and role play, and a participatory evaluation technique was used throughout the training.

In preparation for the first TOT, a guide entitled How to make your extension programme client-oriented was prepared and published in Amharic. The guide was designed for use by woreda staff and development agents and as a supplement to the training. In the guide, gender, the gender analytical framework and rapid or participatory rural appraisal (R/PRA) were covered, as were the steps to be followed in using the gender analytical framework to design a client-oriented extension plan. The guide drew on the case study experiences with all examples (PRA tools and the gender analytical framework) stemming from the case studies.

A 25-minute video was also prepared on client-oriented extension planning. The video was made during the first two case studies. Additional material was added from an implementation site in the third pilot area. The steps in designing a client-oriented extension plan were indicated in the video which introduced the need for extension to reach women farmers, the use of various R/PRA tools and gender analysis, and communities' validation and ranking of development constraints. The video was used in TOT and awareness-raising workshops.

The three major areas of training were:

• participatory approaches, tools and techniques;
• gender, gender roles and needs, and gender analysis; and
• extension programme planning.

The level of learning by experience that took place during training increased as the project progressed. Thus, by the second TOT, one-third of the time was spent on reviewing how to test the client-oriented extension approach and improve it. To this end, woreda staff also participated in this training (unlike the first TOT where only zonal and regional staff participated). Furthermore, during this second TOT, three staff who had received training overseas provided five days of training. The last part of the training sequence - the second training of woreda staff - was not included in the project document or matrix, but was regarded as vital by all involved in the second TOT. Funds were reallocated to enable training sessions to be run in each of the three pilot zones during which reports from the first round of testing of the approach were used for group work.

As the case studies were primarily a learning experience for all concerned (the teams were new in each location), they were each conducted slightly differently so as to try to identify the best and most practical approach. Client groups therefore varied from case to case. In two of the cases, the team commuted to the village, but in the third (in North Omo) they resided in the village. Each case study involved spending ten to 14 days in the village, including time spent in documentation and analysis. Unfortunately, owing to the tight schedule of project activities, all three case studies took place in the rainy season when villagers were busy, so the team had to be prepared to work with villagers when they were available and it was convenient to them.

Implementation took place in the first 14 villages as soon as possible after training. Implementation style varied among regions. In West Harerghe, staff devised a very ordered, systematic and concise methodology. They also stayed in town and travelled to the site daily. Using only three focus groups, they completed implementation - including analysis - within one week in each village. In most instances, implementation took longer in North Omo and North Wollo, especially when teams chose to work with many different focus groups. However, despite working with more focus groups, implementation was completed on average within eight to ten days.

After the second TOT, implementation took place in a further four villages in each of West Harerghe and North Wollo and in two villages in North Omo along the lines agreed upon during the training. Findings were then incorporated into the relevant extension programmes.

Raising the awareness of policy-makers

At the beginning of the project a national inception workshop was held. The purposes of the workshop were to raise awareness about the project and, more importantly, to learn from participants about their activities with regard to the use of participatory approaches and gender analysis and of producing extension materials. Staff from relevant line departments, agricultural research stations and aid agencies working in the pilot zones participated in the workshop.

Useful contacts made were maintained throughout the length of the project. This meant that training and implementation activities complemented and reinforced the activities of the aid agencies in the localities, and the project also benefited from the experiences of the agencies in the pilot areas. Regional workshops, held at a later stage, basically served the same purposes but at the regional level.

Participants at the second TOT stressed the importance of involving local decision-makers because, during implementation, clients raised all kinds of constraints - not just ones related to agriculture - and a mechanism was needed whereby these constraints could be brought to the attention of the relevant line departments and aid agencies. Consequently, funds were made available for zonal workshops to which local policy- and decision-makers were invited.

During the workshops, which took place in two of the three zones, the project's approach and achievements were outlined and findings made in the implementation villages were presented. Constraints faced in the villages that could be solved by line departments or aid agencies, were stressed. The workshops laid the ground for greater cooperation in the future between MOA and the woreda and zonal councils, allowing for more of the constraints of rural clients to be addressed by, or through, the councils.

Lessons learned

Entry point

The project involved training of regional and zonal staff as trainers who then trained woreda staff and development agents at the grassroots. The entry point was, therefore, the development agent's initial contact with village leaders, followed by the introductory meeting between community members and the evaluation team that was held prior to conducting the gender analysis. Training emphasized the participation of different client groups in the study, to allow for subsequent bottom-up planning.

This entry point had two implications. The first of these concerned the objective of the study or implementation. During the three case studies and the subsequent first round of implementation, which followed on from the first TOT, the approach veered towards becoming extractive rather than empowering, i.e. it could be considered RRA rather than PRA. Although PRA tools were used to learn from villagers, the primary aim was to improve extension services offered, not help or empower villagers to plan their own development, with or without project or other assistance. Analysis of findings and design of the gender analytical framework, including identification of successful extension interventions, was done by the team alone, and only later discussed and verified with the community.

The second implication concerned the nature of the study team. Although the team was multidisciplinary to a degree, members were mostly, if not all, MOA staff. This reflected the facts that the project was located within MOA and that the purpose of the study was to improve the orientation of the extension services. However, the lives of rural dwellers cannot be neatly compartmentalized into agriculture, health, education, natural resources, etc. All these sectors are interlinked within the livelihood system of rural dwellers. The use of a participatory approach allows villagers to express themselves and encourages them to relate and analyse all their constraints. These constraints may be outside the capacity of MOA to solve, such as lack of water, clinic and access roads. Yet, by listening to the villagers and discussing their constraints with them, the team was in danger of implying that all problems could be solved.

The second TOT, in which implementation was reviewed by all the staff concerned, provided an opportunity to consider these implications and make important changes to the approach and methodology.

First, it was agreed that the gender analytical framework should not be used to identify "implications for successful extension interventions and improvement" but to identify "opportunities for development". These latter were subdivided into four groups according to who would have responsibility for implementing the recommended solution:

• community (the particular client group could be specified if relevant);
• development agent or woreda office;
• zonal MOA office;
• other (line ministries, aid agencies, etc.).

The second change made was that the level of participation of the client group was increased. It was the client group that now participated in the analysis of development constraints, described above, rather than the team conducting the analysis on its own. One outcome of this was the identification of activities that villagers could carry out themselves, so the participatory approach was empowering. Ways in which clients could participate more in PRA processes and actions were identified, and included involving literate villagers in copying maps, calendars, etc. on to flip-charts, rather than having team members do this, and making sure that everything was validated with the concerned client groups.

The third change involved the training of all staff in the planning of extension programmes. This enabled staff to incorporate development opportunities that development agents and woreda staff could act on into the woreda extension programmes, plans and calendars of work. Villagers could now benefit from the time and effort they had put into the study through improved extension services at a fairly early stage. However, problems with this process came to light during the institutionalization stage which is discussed below.

The fourth important change was to increase the involvement of policy- and decision-makers, i.e. heads of other line departments, members of the woreda council and representatives of aid agencies operating in the locality. The value of awareness-raising workshops was stressed, both prior to the start of any study and afterwards. Other ways of involving decision-makers included inviting them to open staff training sessions or attend introductory or final community meetings. Important stakeholders could also be invited to assist in the development constraints analysis, but only if their presence did not intimidate the villagers.

Tools and methods

The main tools used to facilitate the participation of rural women and men in agricultural planning and to learn about and document gender issues in agriculture were the project matrix; participatory approaches, tools and techniques; and gender analysis. At a later stage it was realized that staff also needed skills in extension programme planning which, therefore, could be considered a fourth tool.

Implementation followed an action-reflection-action approach. First, the development agent, and then the team, met with the community and explained their purpose and what the study would involve. It was stressed that the outcome of the case study was to be the provision of a more client-oriented extension service and not a funded project. Any questions community members had were clarified, and the procedure for identification of client groups was agreed.

The next stage involved getting an overview of the community and its constraints and opportunities through the use of such R/PRA tools as transect walks, historical profiles, resource and social mapping, Venn diagrams, observation, use of secondary sources of information and triangulation. The team split up into subgroups for this, then drew up and presented the findings they made each day to other team members. The whole team then analysed the findings and prepared checklists for the next day's work, which included any issues that needed further clarification. The overview usually took two to three days to complete.

Next, subgroups of the team worked with various client groups to learn about their activities as well as their access to, and control over, resources. There tended to be a minimum of three client groups: men, married women and female heads of households. In one case study, wealth ranking was used as a R/PRA tool and high-, middle- and low-income client groups of men and women were identified. Owing to the cultural context in Ethiopia, in which most rural women are not free to speak out in front of their husbands, all work with focus groups was with single-sex groups. Again, any data gathered during the day were presented and analysed. The R/PRA tools used were daily routines, seasonal calendars and proportionate piling.

The next stage involved using the information gained to conduct gender analysis using the Harvard Analytical Framework. This was adapted to indicate gender-sensitive implications for successful extension interventions.

Finally development constraints, identified by the team through their interactions with the community and their observations while in the village, were shared at a community meeting. Villagers validated and ranked the constraints. In some instances this validation and ranking was done by each focus group prior to meeting with the whole community so that women's views could be heard. The appropriate extension interventions and specific intended client groups were discussed with the community during the meeting as, when time permitted, were the gender considerations leading to the suggested interventions.

As for the appropriateness of the various tools, the project matrix served as a useful and flexible planning and evaluation tool for participatory planning of project activities, while R/PRA proved to be an appropriate means of learning about activities, access to and control over resources, and development constraints for the following reasons:

• The involvement of a MOA team in the learning process ensured that staff were motivated and that a comprehensive picture was gained.
• The use of R/PRA enabled all client groups to express themselves equally well, without inequalities according to gender, literacy, level of education, etc. Use of R/PRA tools with all-female client groups overcame both women's hesitancy to speak out and their high (commonly above 95 percent) level of illiteracy. As most women were hesitant to speak during the common community meetings held at the beginning and end of each study it was important to validate and rank constraints and possible solutions with each client/focus group as well as in the final meeting so that the women's viewpoint could be heard.
• Comparative rather than quantitative data are useful for gender analysis, and R/PRA is well-suited to eliciting appropriate data, even on such sensitive topics as control over resources and relative decision-making powers.
• R/PRA provides a forum for group reflection and analysis which can be useful for the team and empowering to the clients involved. Analysis of learning points by the team as a group led to deeper insights than working as individuals would have achieved.

Gender analysis was the most appropriate means of learning about the activities and constraints of men and women in the rural communities. Moreover, it allowed for an analysis of the differences in activities and constraints according to the income, age and marital status of male and female client groups. Thus, it was a useful framework for analysing the situation and needs of different client groups. The FAO adaptation of the Harvard Analytical Framework was used.

During the training in gender, the first reaction of many male trainees was amusement. However, once they had understood the differences between reproductive, productive and community management gender roles and between strategic and practical gender needs, they applied this understanding to their analyses of daily routines and began to appreciate women's triple gender role and the demands it makes on them. In conducting gender analysis, male trainees also began to see how women's lack of decision-making power and relatively limited access to resources meant that women face different constraints and, therefore, have different needs.

In terms of practice, both gender and participatory approaches and tools were new to most team members and training was not comprehensive. Yet, although no one was an expert in either area, in each of the implementing villages useful insights were gained for the planning of more client-oriented extension.

Through the action-reflection cycle of implementation followed by training, staff were able to define recommendations on how best to conduct the study. These recommendations included the needs to:

• conduct the study during a slack period for the farmers;
• limit the number of focus groups, although the minimum should be three;
• stay in the village during implementation, if possible;
• have a team of 11 to 15 members of which at least one-third should be women;
• collect only necessary and usable information;
• involve staff from local agricultural research stations from the start, if possible;
• ensure that the community fully understand why the study is taking place and what they can expect from it.

The guide and video were produced as additional tools for grassroots staff rather than for communities. However, as is already clear, being a pilot project, this was a learning process for all concerned, tools being selected, rejected, adopted or adapted as time went by. There was, therefore, no single "correct" way of carrying out the study, the methods being dynamic or fluid.

One problem with the use of R/PRA is that it provides information that is very context-specific. Constraints, gender implications and solutions may vary from village to village, but it is not feasible for staff to conduct studies in every village. To solve this problem, each woreda can be zoned according to socio-economic and agro-ecological similarities and studies carried out in a few villages in each zone. A comprehensive woreda extension programme could then be extrapolated with sufficient accuracy.

Extension programme planning was later recognized to be essential because it provides staff with the skills to incorporate findings from the client-oriented approach into the development centre and woreda extension programmes, plans and calendars of work.

Capacity building

This was one of the two main project objectives. Capacity building was based on a sequential system of TOT, training of field staff and implementation of the client-oriented approach at village level. Training was supplemented and reinforced by the video and the guide.

Capacity building was, therefore, essentially vertical, involving development agents and woreda, zonal and regional staff of MOA. A few research staff and staff of non-governmental organizations (NGOs) working in the pilot areas were also trained. Overall, the project contributed significantly to the raising of awareness of staff at all levels in the woredas, zones and regions where it was operating, concerning the use of participatory approaches and tools and gender analysis in client-oriented extension planning.

Some horizontal capacity building took place through the exchange of implementation coordinators from one region to another. Capacity building among the clients concerned in implementing villages was insignificant, although some of the client groups may have noticed the benefits of being recognized, respected and listened to.

Although researchers from national agricultural research centres attended the TOT and national and regional awareness-raising workshops, it was difficult to involve researchers in implementation, as intended, because they were busy with other duties. A much stronger link between researchers and extension would have been preferable. For instance, during implementation, it would have been very useful to have included a researcher from a local research station in the team that observed the effects of particular crop pests, crop and livestock diseases, weeds and problems related to varieties not suited to erratic and low rainfall. Such researchers may be able to suggest solutions, solve the problem at the research station or, more appropriately, embark on participatory research with the farmers in question.

The main lessons learned concerning capacity building can be summarized as follows:

• The step-by-step cycle of TOT, training of grassroots staff, implementation, further training and review followed by implementation was critical. A single training session followed by implementation, without any follow-up, is unlikely to have been nearly as effective.
• The pilot project was very much a learning process which relied on staff being able to analyse implementation, see where mistakes were being made and improve the approach. This would not have been achieved to the same extent if woreda staff, with their detailed knowledge of the villages in which implementation took place, had not been invited to the second TOT.
• Participatory project planning was vital to the whole capacity-building process. Training objectives, content, location, length, etc. were all decided by zonal and regional coordinators, the project adviser and the national programme coordinator as a team. Implementation locations, timing and teams were coordinated by zonal coordinators. Funding and reporting arrangements could be adapted to take account of regional preferences and needs. The project planning style was, thus, participatory and bottom-up which allowed for the project to be flexible and sensitive to other demands on extension staff in the field. For instance, the project matrix was modified in the second year to provide time for extension staff to complete other duties.
• Capacity building would have been greatly enhanced if there had been stronger links among extension, research and farmers. This would have led to more specific and appropriate solutions to many of the agricultural constraints being faced.

Gender information

Each of the 27 village studies contains gender information about its respective village. The following outlines some, but certainly not all, of the information gained about men's and women's activities, access to and control over resources, and constraints.

All three case studies revealed that common development constraints tended to have stronger consequences on women or on men. Thus, the constraint of erratic rainfall affected men in particular because it meant that they had to plough several times. Shortage of farmland often forced women to pursue non-agricultural income-generating activities, such as petty trading and beer brewing. (An exception to this was seen in a later implementation site in North Omo, where men have taken up weaving, leaving women primarily responsible for farming. Among the reasons for this unusual situation is the fact that there are no oxen in this particular area so land cultivation is done with hoes, by women.) Deforestation had a greater direct impact on women and girls who were always responsible for collecting fuelwood and cooking. In all cases, inputs, credit and extension services were far less readily available to women than to men, with many farming wives having no access at all to these resources. However, all three cases revealed that women do have a role in agricultural production, through their activities in livestock management, horticulture, processing of animal by-products, transport and storage of field crops, and marketing.

The case studies also revealed clearly how burdened women are by their triple gender role. In peak agricultural seasons women were found to be working double the hours of men and, in off-peak seasons, more than double. This was owing to the many hours they needed every day to fulfil their reproductive roles of fetching water and fuelwood, cooking, cleaning, child care, etc. in addition to their productive activities. The differences in women's and men's workloads are clearly depicted in the table below, taken from the case study in Damotgale village, North Omo zone in the Southern region.

The use of participatory approaches and gender analysis is, however, context-specific. As a consequence, differences in gender roles among the case study areas were revealed through the gender analysis. For example, the highest involvement of women in agriculture was seen in the North Omo village and the lowest in the West Harerghe village. This may have been the result of local culture and the long distances women had to travel for fuelwood and water in the West Harerghe case, which left them with no time to help with field crop cultivation. However, women in West Harerghe still had responsibility for small livestock, transporting crops from the fields and post-harvest processing and storage.

In the North Wollo case, the use of wealth ranking, followed by discussion of activities with high-, middle- and low-income men and women, revealed another interesting difference; the women most active in farming were those in the middle-income group. Low-income women tended either not to have land or to have let it out, whereas high-income women had no need to assist their husbands with field cropping, as their families could afford to employ labourers when needed.

A different picture emerged, however, for female-headed households in the three areas. As has already been noted, ploughing is very much a male role, and this leaves widows or divorcees in a difficult situation. The women in West Harerghe get around this by availing themselves of a local tradition of group working (debo) to get their ploughing done so that they could keep control of their land. In North Wollo, however, women in a similar position had to relinquish control of their land, renting it out to wealthier farmers with oxen and labour. Although these women did gain half of the crop produce of their land in payment, they did not get any of the other benefits (such as crop residues to use as fuel or fodder).

Among the case study areas, differences were also observed in terms of women's access to credit. In North Wollo there was no source of credit for farming wives (female-headed households could, at least in theory, get credit from the service cooperative), but funds for women's income-generation activities were available through International Fund for Agricultural Development (IFAD) projects in the other two case study areas.

Although not all the information was new to staff, analysing it in this gender-sensitive way gave them a new perspective on the position of women in rural areas.

Whereas, previously, activities had taken the position and roles of women for granted, they now took into account the inequitable daily work burden that women carry. They also began to appreciate the extent to which women are involved in productive work. Until they took part in the client-oriented project, staff did not sufficiently recognize the value of some of the productive work that women engaged in. The growing and marketing of vegetables, care of small livestock, processing and selling of butter, etc. were not taken seriously as productive activities. Neither did staff acknowledge the many and varied contributions women make to field crop production.

Staff found the daily routines and seasonal calendars very useful in revealing when men and women have less work to do and, therefore, more time for extension campaigns or training sessions. In the case of men, this was most often in the off-peak agricultural season. Women's work burden was slightly lighter in the off-peak season, but their free time was still very limited. The use of daily routines, seasonal calendars and other R/PRA tools revealed the difficulties women would face in leaving the home for training, and indicated to staff that they would have to find alternative, culturally acceptable ways of reaching women within their daily schedule. For instance, the daily routines drawn up by female client groups revealed certain times of day when women are together in groups and when it would be possible, and culturally acceptable, for male extension agents to approach them.

Gender analysis also revealed new information to staff regarding access to and control over resources, and helped them to overcome preconceived, and inaccurate, ideas. It was taken for granted that men are the heads of households and, therefore, the final decision-makers. Gender analysis revealed the implications of this for rural women, who have different constraints as a consequence. One implication concerned whether or not women would respond to a client-oriented gender-sensitive extension system when their resulting increased efforts, for instance in helping their husbands in field crop production, may not bring any advantage to them or their children. (This would be the case when men have decision-making power as to how the crop proceeds are used.)

With their new insights into gender roles and the needs of women and men, extension staff were able to plan extension programmes in a more gender-sensitive way. They identified the areas in which certain client groups, especially women, needed extension advice, and recognized women's cultural, time and mobility constraints, as well as their lack of access to extension, inputs, credit and other resources and their lack of decision-making power.

As the process of learning from the community was participatory, it led to a degree of gender awareness in some communities. For instance, although men stated that they had seen how women got up earlier and went to bed later than they did, they did not always accept that women were actually working longer hours than men.

In one location, men noted the shortage of water as their village's most serious development constraint and yet went on to comment that they regularly beat their wives for taking so long to fetch water, in spite of knowing that the water source was very distant. During a community meeting at a neighbouring village, on the other hand, men accepted that women spend a long time fetching water and that there could be some ways in which they could help. Although carrying water on their backs would not be culturally acceptable for men, transporting it in jerry cans by foot, bicycle or donkey would be.

However, the topic has to be approached with care. In the North Omo case study village, monogamous men objected to team members meeting with their wives as a client group, implying that it would stir the women up and make them dissatisfied. Thus, men disapproved of the validation of gender considerations with focus groups, stating "you should not mention anything concerning the female condition". Indeed, in this same village, attempts to draw up daily routines in mixed groups failed. Normally, women and men were interviewed separately, but in this case the team decided to try to consult men and women together and then to draw up separate male and female routines. This attempt to raise gender awareness within the community failed because the women would not speak out. Nevertheless, when interviewed separately, women were very vocal about their workloads and constraints, possibly appreciating the opportunity to express themselves and be heard on these topics.

Linkages

As can be seen from the figure on the facing page, the project promoted two-way communication between rural clients and development centre and woreda staff. It also promoted links among rural clients and woreda councils, aid agencies and other line departments, through awareness-raising workshops.

The figure on the following page also shows that links were established and maintained with councils at the various levels as well as with the Rural Women's Affairs Department of MOA and other relevant line departments. The links were strongest at woreda and zonal levels, weakening somewhat at the regional level. At the national level there was no longer a link with other line departments.

Linkages between rural clients and extension division-based project activities at all levels

Links with aid agencies doing similar work in the pilot areas were strong at all levels. These links were established through the national and regional awareness raising-workshops and a familiarization tour. They were maintained by inviting representatives of the aid agencies to attend TOT sessions, training of teams and implementation. The links served two purposes: they ensured that ongoing capacity-building work was reinforced or complemented, rather than contradicted, by project activities; and it was hoped that the aid agencies would solve some of the constraints revealed during the client-oriented studies, although this was complicated by the fact that each project or programme has its own objectives, plan of action and planned outputs.

The project had links with research stations at all levels, but was not successful in promoting a stronger link between research and farmers. However, this was recognized as a primary aim for future phases of the project.

The following is a summary of the main lessons learned concerning linkages:

• The linkages between rural clients and research and extension are very important to ensure that specific agricultural constraints can be addressed as quickly as possible.
• Linkages with such decision-makers as council members and heads of other line departments are very important, especially at woreda and zonal levels. If council members appreciate the need to address the constraints raised by villagers during the client-oriented approach, they are more likely to pressure the relevant line departments of local NGOs to assist in solving them.
• Linkages with aid agencies working in areas in which the client-oriented approach is taken are vital for the sharing of information and experience, but cannot be expected to lead to financial or technical assistance from the agencies of the clients concerned.

Institutionalization

Table 3 summarizes the changes that have already taken place or should still take place if an enabling and sustainable environment for mainstreaming gender-responsive participatory approaches to agricultural development planning is to be mainstreamed.

Conclusions

THE EXPERIENCE IN ETHIOPIA demonstrates the benefits of training teams of MOA staff to use participatory approaches and tools to learn about activities and resources within different client groups in communities and to analyse this information using the gender analytical framework. From this it is possible to identify development opportunities that the development agent/MOA woreda office can take up in a gender-sensitive manner in extension programme planning. The process can be used to identify opportunities for all kinds of client groups (youth, elderly, landless, etc.) not just men and women per se.

Many lessons were also learned from this project about how to support a process of learning from experience. These, combined with the participatory way in which project planning took place, led to a number of outputs that had not been envisaged in the project document, including the training of staff in extension programme planning, and the workshops held towards the end of the project at the zonal level. It is likely that, by the end of the proposed second phase of the project, the client-oriented approach will have developed much further as a practical planning tool.

The following are suggestions for those wishing to support participatory approaches to gender-responsive agricultural development planning:

• Avoid the marginalization of any initiative by gender-blind decision-makers who may equate "gender" with "women" and attempt to slot the programme into a possibly underresourced Women's Division from where it would be difficult to influence the mainstream - mostly male - extension services.
• Ensure that decision-makers are fully aware of the programme's objectives and intended outcomes in relation to production and male, as well as female, rural clients. Consider including a component covering the gender-sensitization of policy- and decision-makers at all levels.
• Use an experiential learning methodology in the programme whereby trained staff implement the approach then analyse, review and improve it before implementing it again in an improved manner (i.e. an action-reflection-action cycle).
• A TOT system is the best way of building up capacity and ensuring the programme's sustainability.
• Consider including methods of raising gender awareness among participating rural clients. Only then will they make the most of gender-sensitive agricultural planning.
• Establish strong links with local and national agricultural research stations, especially their farming systems or farmer participatory research sections. This will be useful in solving some of the agricultural constraints being faced by clients.
• Try to avoid raising expectations among rural clients through involving them in sharing and analysing their development constraints. In addition, try to address their constraints as soon as possible after completing participatory studies so that they do not feel their time and effort has been wasted. Encourage a sense of ownership among rural clients so that they draw up a development plan which they themselves can implement to address some of their constraints.
• Remember that there are strong economic arguments for the use of gender analysis in agricultural development planning. In too many cases funds have been wasted on projects that have failed because of gender-blind planning. The use of gender analysis in planning ensures that the right clients are reached with the greatest effectiveness.

Acknowledgements

The author would like to acknowledge the support and cooperation of the Extension Department of MOA, Addis Ababa, as well as the enthusiasm and dedication of the project coordinators from regional and zonal levels and all staff and clients involved in this pilot project. Special thanks go to Ato Belay Ejigu, Head of the Extension Division of MOA, for going through and commenting on the draft case study.

The original version of the document was edited by Sally Sontheimer.

Endnotes

¹ These data come from a case study carried out by the Planning Department of the Ministry of Agriculture in 1992.

² Villages from food-surplus areas were selected because their inhabitants were expected to be more interested in working with the team and looking beyond their daily survival problems than were villagers from food-deficit areas.

³ At the start of the project, Amharic was the official national language. However, during the regionalization process, some regions moved towards re-establishing the use of their own language as the official language within that region. This was the case in Oromia region - hence training materials for grassroots staff in that region were translated to Oromiffa.