World Development Vol. 40, No. 2, pp.

414–427, 2012
Ó 2011 Elsevier Ltd. All rights reserved
0305-750X/$ - see front matter
www.elsevier.com/locate/worlddev
doi:10.1016/j.worlddev.2011.05.005

The Empowerment Route to Well-being: An Analysis of Farmer

Field Schools in East Africa
ESBERN FRIIS-HANSEN
Danish Institute for International Studies, Copenhagen, Denmark

and

DEBORAH DUVESKOG *
The Swedish University of Agricultural Sciences, Uppsala, Sweden
Summary. — The study explores empowerment and well-being related outcomes of Farmer Field Schools (FFS), an extension approach
that has gained popularity with agriculture development programs in many African countries. This is done by examining the empirical
relationships between FFS participation and increased well-being; as well as FFS participation and empowerment; and finally between
empowerment and enhanced well-being. Data analysis from the two thousand household questionnaires show a relationship between
these aspects, despite contextual differences in the three countries studied. It is thereby argued that there could be scope to talk about
an empowerment route to well-being. The paper further suggests that the most significant impact of FFS could be viewed in terms of
building the capacity of local people to make choices and make decisions that ultimately lead to increased uptake of agricultural inno-
vations, access to services, and market access as well as collective action. A major conclusion of the study is that agricultural develop-
ment programs should focus more on processes of empowering farmers as opposed to technical solutions that characterize most
programs, in order to create an appropriate mix of technological and social advancement for a development process that is sustainable
in nature.
Ó 2011 Elsevier Ltd. All rights reserved.

Key words — Farmer Field Schools, empowerment, poverty, well-being, East Africa, agricultural extension

1. INTRODUCTION In East Africa there are a number of national extension pro-

(a) Introduction to key issues
A majority of the rural population in East Africa continue
to live in poverty despite the increased efforts by governments
to achieve the millennium goals. Reducing poverty, therefore,
requires a thorough understanding of the factors that generate
poverty and an assessment of policies and interventions de-
signed to support rural poverty reduction. The World Develop-
ment Report 2008 (WDR, 2008) marked the culmination of a
long row of international reports that all point toward small-
holder-based agricultural growth as being the most effective
way of reducing poverty in Africa (IAASTD, 2009; IFPRI,
2006; NEPAD, 2007). However, how best to facilitate agricul-
tural growth among smallholders still remains widely debated.
A large proportion of small-holder farmers are not able to en-
gage with profitable markets for agricultural products, nor can
they effectively engage with local agricultural service provid-
ers. Scoones and Thompson (1994), Chambers (1993) and
Leeuwis (2004) argue that existing approaches to transfer of
technology do not fit the resource-poor farming context of
the South. To achieve agricultural and rural development,
new methods for extension and training are needed that make
better use of knowledge among farmers and provide for them
a stronger voice to demand advice, services, and negotiating
power (Christoplos, 2003; Haug, 1998).
East African countries are undergoing a progressive policy
change toward more demand-driven and market-oriented
agricultural services. This includes a policy shift from central-
ized extension systems, for example, Training and Visit, to
decentralized, demand-driven agricultural advisory systems.
414
grams that all recognize the important role of farmer empow-
erment, public/private partnerships and local participation. In
Uganda the National Agriculture Advisory and Development
Services (NAADS) program started in 2001 and is today pres-
ent throughout the country. In the case of Tanzania, the Agri-
culture Sector Development Program (ASPD) started in 2007
and is currently expanding its scope. In Kenya the National
Agricultural and Livestock Extension Program (NALEP) is
operating nationwide extension services. To be effective these
programs assume that farmers are organized and capable of
articulating informed demands to external service providers.
However, experience from NAADS indicates that, without a
very deliberate external support for farmer empowerment,
farmers are often subject to manipulation by external actors,
and as a consequence access to agricultural services remains
limited (Government of Uganda, 2005).
Farmer empowerment is thus generally seen as an important
element in developing demand-driven advisory services
(Barlett, 2005). The concept was first recognized by the Bank
in its World Development Report 2000/2001 (Bank 2000) as
one of the three pillars of poverty reduction. Empowerment
is an advanced form of participation that entails farmers mak-
ing their own decisions rather than adopting recommenda-
tions (Friis-Hansen, 2004) from others. Despite the lack of
robust data (Alsop & Heinsohn, 2005), empowerment is
increasingly seen among donors and development actors as a
major contributor to development outcomes (World Bank,
2002; Narayan, 2005).
* Final revision accepted: May 10, 2011
 THE EMPOWERMENT ROUTE TO WELL-BEING:
This study rests on the assumption that the need for individ-
ual and collective agency among smallholder farmers in East
Africa is on the increase. The current transition in small-scale
farming requires farmers that are innovative and able to adapt
to changing situations. Rivera and Alex (2004) conclude that,
before demand-driven extension systems can take root in prac-
tice, farmers must be empowered to develop their capacity to
articulate their demands and exert pressure on the system to
deliver what they want. The privatization of markets has cre-
ated opportunities for increasing agricultural production and
marketing. However, a high level of individual and collective
agency and coordinated action is required if small-scale farm-
ers are to engage successfully with commodity markets, for
example, by meeting quantity and quality standards and nego-
tiating effectively with traders (Leeuwis, 2004). Secondly, ac-
cess to local public services, such as education and health,
also increasingly requires the active participation of citizens.
Despite their recognized importance, efforts to assist the
empowerment of farmers and their organizations still seldom
form an integrated part of agricultural programs and projects.
Conceptual support for farmer empowerment comes from do-
nor agencies and only rarely from governments and their agri-
cultural ministries. Likewise the empowerment components of
joint government–multi-donor projects are most often funded
by external donors. Finally, the implementation of such com-
ponents is frequently sub-contracted to national or local
NGOs. A low priority is also given to human resource devel-
opment support in new agricultural development policies and
a lack of a “human” side of the poverty debate in, for example,
policy processes such as the NEPAD comprehensive plan for
agricultural development. Furthermore, major global invest-
ment programs such as the Alliance for a Green Revolution
in Africa (www.agra-alliance.org/) funded by the Gates Foun-
dation and the Rockefeller Foundation primarily emphasize
input and technology options rather than capacity-building
as ways of solving rural poverty challenges.
Following the collapse of T&V (Anderson, Feder, &
Ganguly, 2006; Gautam, 2000) there has been a search for
improved methodologies that respond better to farmers’ de-
mands and a shift toward more broad based, participatory,
and group focused approaches (Davis, 2006; Friis-Hansen,
2004). Farmer-to-farmer extension, group extension methods,
Participatory Rural Appraisal and Farmer Field Schools
(FFS) are some of the recent methodologies applied on larger
scale (Neuchatel Group, 2006; Qamar, 2006). Working with
farmer groups has been found far more effective than working
with individual farmers (Friis-Hansen, 2004; Heemskerk &
Bertus, 2004; Heemskerk et al., 2003; Leeuwis, 2004; Umali-
Deininger, 1997) and thereby most extension methods of to-
day are group based. Current thinking in extension also
emphasizes an innovation system where research-extension-
farmer experimentation is closely interlinked rather than seen
as a one-way flow of communication (Leeuwis, 2004). Client
orientation is encouraged and extension messages need to have
a broadened scope and cover rural livelihood in general as op-
posed to just agricultural techniques (Anderson et al., 2006;
Qamar, 2006). Market oriented agricultural advisory services
are also increasingly mainstreamed in extension (Neuchatel
Group, 2008). Within this context, FFS is gaining attention
among development actors in East Africa as a community-
based, demand-driven, non-formal education program that
appears to stimulate both empowerment and agricultural
growth. The FFS approach embraces most of the concepts
above and is fully in line with the current thinking in exten-
sion. However, a distinct difference between FFS and many
other modern approaches is the focus in FFS on experiential
415
learning tools and development of analytical and problem
solving capacity among farmers and the use of highly trained
facilitators rather than technical advisors to support farmer
group activities. Due to the focus in FFS on human develop-
ment this study found it appropriate to use FFS as the entry
point to explore further the linkages between farmer educa-
tion, empowerment, and the increased well-being of the rural
poor.
(b) Accumulated knowledge and gaps
A wide range of unpublished literature describes the suc-
cesses and impacts of the FFS approach, as well as the emerg-
ing challenges. Aspects commonly pointed out include both
increases in agricultural production and individual and collec-
tive agency. Published research indicates substantial impacts
of FFS in terms of increases in farm productivity, reducing
farmers’ use of pesticides, and improved farming knowledge
(Mwagi, Onyango, Mureithi, & Mungai, 2003; Praneetvatakul
& Waibel, 2003; Rola, Quizon, & Jamias, 2002). A number of
studies discuss the role of FFS as an extension model, though
with contradictory arguments. For example, Quizon, Feder
and Murgai (2001) challenge the fiscal sustainability of the ap-
proach when implemented on a large scale due to the high
costs per trained farmer. As a response, van den Berg and
Jiggins (2007) have argued that FFS should not be considered
as mainly an extension model but as a complementary educa-
tional instrument that provides intangible public goods that
cannot be measured only in agricultural terms.
Few studies have focused specifically on empowerment and
FFS, but wider developmental benefits are reported in terms
of poverty reduction and human and collective action
(Mancini, van Bruggen, & Jiggins, 2006; Van den Berg &
Jiggins, 2007; Züger, 2004). A recent study carried out by IF-
PRI in East Africa demonstrated significant impacts of FFS
on the lives, productivity, and incomes of especially women-
headed households and people of low literacy levels. While
the study refers to empowerment-related impacts, it lacks con-
crete measures of them (Davis et al., 2010). Friis-Hansen’s
(2008) study of FFS and NAADS groups in Soroti Uganda
shows that FFS served as a platform and catalyst there for
the success of demand-driven advisory services. This study
also points at poverty reduction among the studied groups,
but it does not attempt to explain the relationship of FFS to
potential empowerment and poverty reduction. Overall, estab-
lished research on the FFS approach covers bits and pieces of
the anticipated impact. However, this study is the first to aim
more systematically at establishing links between FFS,
empowerment, and increased well-being among participants.
(c) Study objectives
This study sets out to determine whether the effectiveness of
external investments in support of African agricultural devel-
opment can be enhanced by being combined with support
for farmer empowerment and rural institutional development.
This contradicts a main line of thinking in current interest and
investment support for African agriculture that continue to be
based on the hypothesis that there is a “transfer-of-technology
and liberalized market” pathway out of poverty. The entry
point of this study is a critical view of purely technical ap-
proaches that lack support for human resource development.
The hypothesis of this study is that a joint learning process
leads indirectly to enhanced well-being among resource-poor
communities through empowerment that translates into trans-
formation and action at both personal and collective levels,
416 WORLD DEVELOPMENT
ultimately impacting positively on the lives of the resource
poor. This article thus sets out to explore the validity of an
alternative or complementary route to enhanced well-being
among the poor, that is, the empowerment route to well-being
(see Fig. 1).
The article examines this route in greater detail by examining
relationships between (1) FFS participation and increased well-
being; (2) FFS participation and empowerment; and finally (3)
empowerment and enhanced well-being. The study also recog-
nizes that the extent to which this process takes place depends
on the external context, that is, on opportunity structures in
terms of institutional, social, or political conditions.
2. FARMER FIELD SCHOOL
(a) The Farmer Field School approach
The FFS approach was originally developed in Asia in the
1980s as a response to the failure of the commonly applied
Training and Visit (T&V) extension model to prevent disease
outbreaks. The hands-on practical learning in FFS emerged
as a means of facilitating critical decision-making skills among
farmers to deal with complex farming problems (Gallagher,
2003). FFS uses a learner-centered, problem-based approach
to teaching involving field observations, the relating of these
observations to the ecosystem, and applying previous experi-
ence through group discussions with new information being
available to make informed crop or livestock management
decisions (Duveskog, 2006). A group of farmers who meet reg-
ularly (usually weekly) in the field form the field school, while
plants or animals at the learning site form the main study
materials. The learning takes place under the guidance of a
trained facilitator, who helps promote active participation,
group dialog, and reflection. Critical reflection is promoted
through engagement in comparative experiments, the regular
agro-ecological system analysis (AESA) exercise and discov-
ery-based activities which further stimulate participants to
question preconceived beliefs and norms about farming. Apart
from the farming-related content, small-group activities and
discussion sessions address “special topics” relating to non-
agricultural issues (e.g., HIV and domestic violence). In addi-
tion, to enhance learning, song and dance is often used as a
complement to the problem-based learning and to take advan-
tage of local ways of knowing (Duveskog, 2006).
There are currently a multitude of FFS initiatives in more
than 27 countries in Africa (Braun et al., 2005) funded by var-
ious development agencies, and the approach is gaining in
popularity. The approach was first implemented in Kenya in
1995 and has since gained popularity in most East African
countries and been institutionalized within the national exten-
sion system in Tanzania.
Fig. 1. A schematic overview of the empowerment route to well-being.
(b) The FAO/IFAD FFS program in East Africa
The empirical framework for this article was the IFAD-
FAO FFS project in Kenya, Uganda, and Tanzania that oper-
ated in two phases; the first one between 1999 and 2002 and
the second phase between 2005 and 2008. It was one of the
first large-scale FFS programs in Africa, and the entry point
was the South-East Asian approach to FFS, which was
adapted to meet the varied requirements of East Africa. The
project was implemented jointly between the Food and Agri-
culture Organization (FAO) and the ministries of agriculture
in the three countries. The objectives of the first phase, which
involved approximately 25,000 farmers across the region, a
majority of whom were women farmers (70%) were to increase
the competence of the extension system, establishing network-
ing capacity for exchanging FFS experiences and contribute to
knowledge on the effectiveness of the approach (IFAD, 1998).
The second phase that led to the establishment of 330 new
FFS groups aimed at development and expansion of a low-
cost, sustainable, and broad-based model for farmer education
and empowerment. Technical entry point in the project was
originally Integrated Pest Management in staple crops, such
as vegetables, potatoes, and beans. Gradually as per partici-
pants’ demands the scope of the project broadened out widely
to address a very broad range of crop and livestock enterprises
(Okoth, Khisa, & Thomas, 2002). Many of the other FFS pro-
grams in the region, supported by NGOs, Governments or
other actors, relied on this IFAD program and its participants
as a source of expertise in the approach. The second phase in-
cluded development of a number of innovative aspects of FFS,
such as low-cost and self-financed FFS model, market ori-
ented services, and farmer association building (Davis et al.,
2010).
The reasons for selecting the IFAD program as the empiri-
cal framework for this study included:
This project spearheaded the development of demand-dri-
ven services in the region and applied a range of innovative
aspects, such as a broadened curriculum, demand-side
financing, market orientation, and participatory learning.
The project was the longest running FFS project in the
region, therefore, making it possible to evaluate effects
and impacts that take time to emerge, such as social
changes.
The project implementation strategy and implementation
modalities had been largely identical in the three countries,
thereby making it possible to compare FFS impact across
districts and countries.
The project had a strong commitment to learning and
allowed the necessary flexibility in implementation modali-
ties in order to ensure that it could evolve according to the
demands of participants.
 THE EMPOWERMENT ROUTE TO WELL-BEING:
3. METHODOLOGY
(a) Evaluation framework
The analysis of relationships between FFS participation and
increased well-being; between FFS participation and empow-
erment; and between empowerment and enhanced well-being
was in Tanzania and Kenya done by comparing surveys of
FFS members pre- and post-FFS (before and after the inter-
vention) but in relation to different individuals, that is, not
longitudinal analysis of panel data. Analysis in all three coun-
tries is also done of FFS graduates compared with a control
group (with and without).
FFS members join the groups on voluntary basis, and not
randomized in the community. Thus a methodological prob-
lem had to be overcome, that a simple comparison of means
of the outcome participating in FFS and control group may
be biased due to the issue of self-selection. There could be a
possibility that households that were better off before joining
FFS were more likely to participate than poorer households.
To address this bias a control for the selection of participants
of FFS was needed. The mainstream treatment literature as-
sumes selection on observables but assume no selection on
un-observables, also called un-confoundedness or conditional
independence assumption. In order to relax this assumption
the characteristics of the participants of FFS groups, as com-
pared to random samples of community members was estab-
lished. This was done in order to assist in attributing
differences among FFS members pre- and post-FFS to the
FFS learning process, as well as to clarity and possibly rule
out selection bias in the sampling procedure.
Figs. 2 and 3 show the proportion of non-FFS members and
FFS pre-members in Kenya and Tanzania belonging to the
various poverty categories commuted by the poverty index.
The data show the poverty status, based on the well-being cat-
egories, among the randomized community members as com-
pared to participants enrolled for FFS. In Tanzania, the
sample of FFS pre-members demonstrated a slightly higher
proportion of very poor and slightly lower proportion of
non-poor as compared to non-FFS members. In Kenya, on
the other hand, the sample of FFS pre-members included a
slightly higher proportion of non-poor and a lower proportion
of very poor individuals than the non-FFS group.
However, the differences between non-FFS and FFS pre-
members in the two countries were not statistically significant,
and it can, therefore, be assumed that there is no significant
difference in terms of poverty status among the average com-
munity members and individuals who enroll in FFS participa-
tion. This assumption allows us to compare FFS pre-members
with FFS graduates, even when the two groups do not repre-
60
50
41.7
40
29.2
30
20 18.6
10
0
non-poor
Non-FFS (n=120)
Fig. 2. Proportion (%) of non-FFS and FFS re-members among the various poverty categories, in Tanzania.
417
sent the same individuals (pre and post longitudinally), as is
the case for the findings reported in the subsequent sections.
Similar baseline data do not exist for Uganda, however, qual-
itative interviews indicate that there were no major differences
in well-being between those joining FFS in Uganda and those
who did not.
The data source for this study is a combination of face-to-
face questionnaire surveys comprising a total of 1,203 house-
holds carried out in Kenya, Uganda, and Tanzania during
2004–07. The samples in the different countries included
non-FFS members, FFS pre-members, and FFS members
(FFS/NAADS group members in Uganda).In Kenya and
Tanzania data in each country were collected in 2006 through
a stratified random sampled survey with about 280 individuals
signed up for FFS (but not yet having commenced participa-
tion in FFS, i.e., FFS pre-members) and 120 non-FFS house-
holds. A two-stage random sampling technique was applied to
select FFS pre-members with 20 FFS groups per country ran-
domly selected, divided proportionally per district. Thereafter
household members were selected based on lists of households
in the selected FFS groups. Non-FFS participants were ran-
domly sampled in neighboring villages (without FFS activities
ongoing) to the selected FFS groups, by means of village and
household name lists obtained from the local administration.
In addition a survey of about 300 graduated FFS members
in Kenya and Tanzania each was carried out in 2007 with ran-
domly sampled FFS participants from FFS groups started
back in the years 1999–2002. This dataset was used to compare
FFS pre- and post-situation.
In Uganda (Soroti district) a survey questionnaire was
implemented in 2007 with 403 respondents. Respondents
were randomly selected in the district irrespective of FFS
membership status. During data analysis, groups of FFS
graduates and non-members were then separated for compar-
isons. In Uganda the NAADs program provides the domi-
nant framework for collective activities among farmers, and
most FFS groups that started in 1998–2001 had turned into
NAADs groups. Therefore, the sample of FFS members in
practice included both FFS and NAADs participants, while
the non-member groups included neither FFS nor NAADs
members.
Interviews were conducted with the help of a formally struc-
tured questionnaire under the supervision of the researchers.
All surveys were carried out by trained enumerators, knowl-
edgeable in the local language. The survey was field-tested be-
fore being implemented in the countries. The questionnaire
included a range of aspects, such as poverty indicators, the
adoption of agricultural technologies, economic and institu-
tional issues, personal and collective agency, attitudes, percep-
tions of power, etc. The content of the surveys in Kenya and
50
31.4
29.2
poor very poor
FFS Pre-member (n=280)
418 WORLD DEVELOPMENT
60
51.3
50
40
35.8
30 28.6
20
10
0
non-poor
Non-FFS (n=121)
Fig. 3. Proportion (%) of non-FFS and FFS Pre-members among the various poverty categories, in Kenya.
Tanzania was largely identical, while there were a few varia-
tions in the Uganda survey format.
(b) Conceptual approach to study
(i) Measuring well-being
Poverty has traditionally been measured mainly in terms of
income and consumption dimensions, and poverty threshold
measurements such as living on less than $1 a day have been
widely used to define who is poor and who is not (http://
www.undp-povertycentre.org/). More recently a wider and
more holistic concept of poverty, based on a broader defini-
tion of resources, abilities, and vulnerability among house-
holds and individuals, has emerged under the term well-being
(Coudouel, Hentschel, & Quentin, 2002). There has been a
growing recognition that processes involved in alleviating
poverty are highly complex (Kristjanson, Place, Franzel, &
Thornton, 2002). When the poor themselves are asked what
poverty means to them, income is only one of a range of the
aspects they highlight (Chambers, 1987). Others include a
sense of insecurity and vulnerability; a lack of voice vis-à-vis
other members of their household, community or government;
and low levels of health, literacy, education, and access to as-
sets, many of which are influenced by the scope and quality of
service delivery (Farrington, Carney, Ashley, & Turton, 1999).
Recently poverty has been defined in terms of an absence of
the basic abilities to meet these physical needs, as well as to
achieve the goals of participating in the life of the community
and influencing decision-taking (Farrington et al., 1999).
The Voices of the Poor study (Narayan, 2000; Narayan
et al., 2000), conducted in 60 countries, showed that voiceless-
ness and powerlessness are pervasive among the poor, affect-
ing every aspect of their lives. Sen (1997) highlights
information as a crucial aspect of poverty, since it is only when
poor people know what monies and services are available that
they can hold program functionaries (public or private) to ac-
count. He also makes a strong argument that there are com-
pelling reasons to define poverty in terms of the lack of the
basic abilities to avoid hunger, malnutrition, and poor health,
as well as to be adequately clothed, to partake in the life of the
community, to feel safe and secure, and so forth.
(ii) Measuring empowerment
The broader understanding of what poverty is, according to
poor peoples’ own perceptions, links directly to concepts
about power and human empowerment. Giddens (1976)
speaks of “power as transformative capacity,” a reference to
the fact that every human being has the right and capacity
to act and get certain things accomplished, as well as to partic-
ipate effectively in shaping the social limits that define what is
52.8
20.2
11.3
poor very poor
FFS Pre-member (n=280)
possible (Hayward, 1989). This type of power is also called
human agency (Leeuwis, 2004) and is the field of power
around which this study revolves. The concept of agency stems
from the idea of the “human agent,” or, as Sen expresses it
(1999), “somebody who acts and brings about change, and
whose achievements can be judged in terms of her own values
and objectives.” If power means control, then empowerment is
the process of gaining control (Sen 1997). This means that
facilitating empowerment means supporting people in becom-
ing agents in their own development. There are many defini-
tions of empowerment in use, but the definition used in this
study to frame parameters of measurement is as follows: “A
process that increases the capabilities of smallholder farmers
and farmer groups to make choices and to influence collective
decisions towards desired actions and outcomes on the basis of
those choices” (Friis-Hansen, 2004).
Measurements of assets and institutions provide intermedi-
ary indicators of empowerment. According to Alsop and
Heinsohn (2005), direct measurements of degrees of empower-
ment can be made by assessing the existence of choice, the use
of choice, and the effectiveness of choice. To be sustainable,
the empowerment process must alter both people’s self-percep-
tions and their control over extrinsic resources, as well as pro-
viding greater autonomy and authority in decision-making
and assertiveness (Sen, 1997).
This study frames the measurement of empowerment based
on three main building blocks, which relate closely to recent
concepts of empowerment presented in the World Bank pub-
lication “Measuring Empowerment” by Narayan (2005).
These building blocks are the agency of the poor in terms of
assets and capabilities at both (1) the individual and (2) the
collective levels, and (3) opportunity structures in terms of
institutional climate and social and political structures.
Opportunity structures refer to the formal and informal con-
texts within which rural farmers operate (Alsop & Heinsohn,
2005) and influence strongly the development outcomes
achieved (Friis-Hansen, 2004). At the individual level, empow-
erment has been defined in terms of factors that give one great-
er control over one’s life. Factors include an individual’s
knowledge base, resources, rights, and assets. Reference is also
made to the sense of well-being in terms of status and self-es-
teem that are both facilitated by and give further support to
the capacity to control key aspects of one’s life (Friis-Hansen,
2004). Central to collective agency are processes that build col-
lective assets and improve the efficiency and fairness of the
organizational and institutional context which governs the
use of these assets (Friis-Hansen, 2004). Collective capabilities
and organizations are often critical in helping poor people
break through the constraints of powerlessness and voice-
less-ness (Narayan, 2005). Farmer organizations increase
 THE EMPOWERMENT ROUTE TO WELL-BEING:
people’s access to ideas and information and strengthen their
capacities for planning, decisionmaking, collective action
etc., as well as expanding their ties to other networks and re-
sources.
(c) Research sites
Data collection was undertaken in the IFAD-FAO FFS
project sites in Kenya (Kakamega, Busia and Bungoma dis-
tricts), Uganda (Soroti, Kaberamaido, and Busia districts),
and Tanzania (Bukoba, Muleba and Missenyi districts). These
areas were originally chosen as project sites based on the pre-
vailing crop and farming system, the need for extension sup-
port and testing of FFS under current government structures
as well as potential linkages with other IFAD programs. All
the sites have fairly high agricultural potential, with a rainfall
in the range of 1000–2000 mm/year and medium to high pop-
ulation pressure. Agriculture is the main economic activity in
all the sites, predominantly subsistence agriculture, but also
with increasing degrees of commercialization and production
for markets. In all sites poverty levels are high. A national
study of poverty found Western Province in Kenya to be
one of the poorest in the country, with an estimated 50% of
the population in poverty (Republic of Kenya, 1997). Over
the past year, Soroti district in Uganda has experienced a
boom in agricultural commercialization with a resulting de-
crease in poverty. The spread of commercial agricultural enter-
prises is closely associated with farmer institutional
development, including FFS, as well as the availability of
external funding through National Agricultural Advisory
and Development Services (NAADS) and Serere Research
Station (Friis-Hansen, 2008). The Kagera region where the
Tanzanian sites are located is geographically isolated, and its
proximity to three foreign countries has made it vulnerable
to foreign influences and in particular to influxes of refugees,
which has caused environmental degradation and conflicts.
In recent years Uganda and Tanzania have experienced a pro-
cess of decentralization, especially of extension services, while
in Kenya public community support has continued to be more
centrally managed.
(d) Survey methods
(i) Well-being ranking methodology
One part of the questionnaire related to capturing the overall
well-being of the households. This part builds on recent inno-
vative experience in East Africa with developing well-being
indicators identified by farmers. Based on a rigorous selection
procedure, people’s perceptions of well-being are extrapolated
into clusters, validated, and combined into a poverty index.
The methodology has four steps: in step one, expressions of
well-being are collected through qualitative interviews of
groups of people geographically representing a given area.
Small groups of informants are asked to rank all households
in their community into three groups using a card-sorting
method and thereafter asked to describe the well-being of each
group. Step two transforms these statements into a number of
thematic well-being indicators each comprising a set of gener-
ally agreed perceptions describing the situation of non-poor,
poor, and very poor for each of the indicators. Step three is
to translate these indicators into questions in a household
questionnaire. Step four is to develop a poverty index based
on the poverty indicators. A household’s poverty index is
computed as the mean of its scores for each of the well-being
indicators (ASPS, 2003; Friis-Hansen, 2005; Ravnborg et al.,
2004).
419
(ii) Variables for measurements of empowerment
Variables for measurements of empowerment were defined
through a combination of theoretically informed expressions
of empowerment and participatory development of indicators
with farmers. A stakeholder workshop was carried out in each
country in 2004–05. The workshops ran for three days and in-
cluded about twenty individuals: FFS member farmers, farm-
ers from FFS networks, producer organizations, field
extension staff, and project coordinators. During the work-
shop, participants, through interactive facilitation, developed
indicators for expected outcomes of FFS education based on
their local perspective. Indicators related both to short-term
aspects, such as access to services, agency, and organizational
skills, and longer term outcomes such as improved livelihoods
and well-being. Even though there was some variation in indi-
cators in the three countries, the outcomes of the events were
similar, supporting the idea of the possibility of standardizing
survey content across the three countries.
Empowerment-related variables were separated into two
groups: (1) self-perceptions and attitudes among farmers to-
ward their power and agency in life, such as the power to influ-
ence their lives and community, trust and gender relations;
and (2) actual physical expressions of agency in their daily
lives, such as productive assets, knowledge, access to services
and the ability to plan. The questions in the survey relating
to self-perceptions and attitudes were captured by a three-
point summative scale (Likert, 1932): 1 = agree, 2 = neither
agree nor disagree, 3 = disagree. An example of the attitudinal
statements used was, “I feel I can make this village a better
place to live in!.” Questions related to physical expressions
of agency in everyday life were mainly included in the ques-
tionnaire as binary items (yes/no questions) such as “Are
you a member of a savings or credit organization?”.
(iii) Key informant interviews
Key informant interviews were used to capture data related
to institutional and policy issues and opportunity structures.
Semi-structured or fully open-ended interviews were held with
selected key informants, such as local and national extension
managers, extension workers, government officials, and village
leaders. Interviews were recorded and transcribed. Secondary
data were also reviewed and analyzed, including policy docu-
ments, extension management guidelines, and procedures at
the local and national levels. The aim was to gain an under-
standing of how institutional issues influence service provision
to farmers and constraints and facilitating factors in respond-
ing to farmers’ demands by the government and other exten-
sion actors. Further, background materials about the bio-
physical, socio-economic, and cultural contexts in the various
study sites were analyzed so as to gain a better understanding
of the local situation.
(e) Statistical data analysis
Data analysis was carried out using SPSS software. Partici-
pants whose responses were incomplete were excluded auto-
matically by SPSS in the data analysis. A factor analysis was
carried out in order to aggregate empowerment factors, and
levels of significance were tested on both the emerged factors
and categorical variable items by cross tabulation and Pearson
Chi-square tests.
(i) Factor analysis
The eleven questionnaire items with a summative scale on
perceptions and attitudes were subjected to principal compo-
nent analysis (for each country dataset) to segment the
420 WORLD DEVELOPMENT

variables into fewer factors of self-perception of agency. Cor- 3. EMPIRICAL FINDINGS

relations between the input variables were first checked with
Spearman, and all variables were shown to correlate with at
least a few others. Spearman and Pearson correlations were
checked to confirm that the limited three-point ordinal dataset
was applicable for factor analysis. One variable was removed
from the dataset after its correlation proved weak from check-
ing communalities and total variance. Models for four, five,
six, and seven factors were computed (factor rotation method
Equimax with Kaiser Normalization). The five-factor solution
was chosen because it parsimoniously explained the most var-
iance (min 52% variance explained) with the fewest factors,
and gave the same or similar factors across different countries.
Correlations and factor-loading coefficients were used to
understand the nature and structure of the four factors. Final-
ly the factors were labeled and saved as individual variables
subjected to the cross-tabulation in the same way as the binary
dataset items (Table 1).
(ii) Cross tabulation and Pearson Chi-square tests
Decision-making is a complex process that is not easily re-
duced to simple questions of existence-, use of-, and effective-
ness of choice. The indicators of empowerment used in the
household questionnaire were extracted from in-depth qualita-
tive interviews analyzed through NVIVO software. The binary
data items and emerging factors were analyzed through cross-
tabulation to identify dependences between various variables,
such as between the empowerment variables and poverty level
categories. In comparing the categorical variables and testing
of significant differences between groups, the Pearson chi-
squared (v2) test was used.
(a) Links between FFS membership and poverty level
Poverty levels, based on the well-being ranking methodol-
ogy, among FFS pre-members and FFS graduates were com-
pared in order to evaluate to what extent FFS might have
contributed to changes in poverty levels among participants.
Based on qualitative interviews, the dropout rate of FFS
groups is assumed to be low and not influencing the analysis.
The comparison was done through cross-tabulation, compar-
ing the distribution of non-poor, poor, and very poor among
the two sample groups for each of the three countries (see
Fig. 4). In all countries the proportion of very poor was lower
among FFS graduates than among FFS pre-members. Simi-
larly the proportion of non-poor was higher among FFS grad-
uates. In all cases the differences between the two groups were
significant. It can, therefore, be assumed that FFS graduates in
the study were demonstrated to be less poor than FFS pre-
members. Since the members and non-members shown above
were not to be significantly different in terms of poverty before
the FFS interventions, this change between pre- and post-FFS
groups can be assumed to be related to their participation in
FFS. The picture varied slightly between the countries. Tanza-
nia showed on average (across the two sample groups) the
highest proportion of very poor, with Kenya showing the low-
est levels. In all three countries, the FFS pre-member group
showed that the largest poverty category was in the middle sec-
tion, that is, the poor. In Tanzania this remains the case also
among FFS graduates, while among FFS graduates in Uganda
and Kenya the non-poor form the largest poverty category.

Table 1. Explanation of input variables included in the various factors that emerged
Factors Survey input variables K T U
Household decision-making capacity 1. Decision on farming activities and practices x x
2. Decision on education and health x x
3. Decision on household expenditure x x
Power to influence community 4. You have control in making decisions that affect your everyday life x x x
5 You can make this village a better place to live x x x
6. Men and women participate equally in community decision-making x
7. Able to understand your own problems and find solutions
8. Men and women equally involved in leadership in this village
Gender equity and trust 9. Help friends or neighbors in unfortunate situation x
10. Men and women are equally involved in leadership in the village x x
11. Village/neighborhood people trusted x
12. Men and women participate equally in decision-making in your HH x
13. Conflicts between man and wife in the HH rare x x
14. Trust people in groups in matters related to lending money x x x
15. Have power to make important decisions x
16. Do groups have a bank account? x x
Individual agency 15. Have power to make important decisions x x
4. You have control in making decisions that affect your everyday life x
9. Help friends or neighbors in unfortunate situations x
10 Men and women are equally involved in leadership in the village
11. Village/neighborhood people trusted x
17. Part of the decision-making processes for community action regarding election x x
Trust in community and local authorities 18. Your trust in local government official is high x
19. Your trust in local politicians is high x
Source: Stratified random survey of 701 household in western Kenya 2007, 697 household in Kigoma Regions, Tanzania 2007, and 403 household in
Soroti District, Uganda, 2007.
Explanatory note: K, Kenya; T, Tanzania; and U, Uganda. x indicates if the variable was used as empowerment indicator.
Statistical note: Extraction Method: Principal Component Analysis; Rotation Method: Equamax with Kaiser Normalization; Total Variance Explained
60% in Kenya and 58% in Tanzania and Uganda; Variables saved using regression; Visual Binary Transformation of data with cut points based on equal
percentiles.
 THE EMPOWERMENT ROUTE TO WELL-BEING: 421

Kenya
60 57
52.8
50

40 38.7
35.8
30

20
11.3
10
4.3
0
FFS Pre-members (n=280) FFS Graduates (n=300)

non-poor poor very poor

Tanzania
70

60 57.3
50
50

40
31.4
30 25.3
20 18.6 17.3

10

0
FFS Pre-member (n=280) FFS Graduates (n=324)

non-poor poor very poor

Uganda
60
54
50 49

40 38
36

30

20
13
10
10

0
Non-FFS members (n=x) FFS Graduates (n=x)

non-poor poor very poor

Fig. 4. FFS pre and post comparison: % of sample within various poverty categories.

Table 2 shows the percentage distribution within each of
the 13 well-being indicates members of FFS and the control
group for each of the three countries. The shaded areas indi-
cate that there is a correlation (at <0.05 level) between
belonging to an FFS group and the level of well-being.
The table shows considerable differences between the three
countries in the patterns of well-being, and thus the pro-
cesses that bring them out of poverty. This is probably a
reflection of the differences in socio-economic context. All
three countries indicate significant change for poverty indica-
tors that can change without the need for capital and/or a
long period of time, for example, stop working as casual la-
borer, hire of labor, quality of diet, household food security,
family health, and standard of family clothing. Even more
impressive is the significant change in asset based poverty
indicators, including housing standards, children’s education
level, and ownership of livestock. In addition for Kenya, Ta-
ble 2 shows a significant difference in access to non-agricul-
tural income.
(b) Links between FFS and empowerment
In order to explore potential links between FFS and empow-
erment the various defined expressions of empowerment were
correlated with FFS membership through cross tabulation,
comparing FFS members with the control group (non-FFS
members and FFS pre-members). Levels of significant differ-
ence between the two groups were then observed. Both the fac-
tors derived for empowerment from the Likert scale items in the
questionnaire and the categorical variables are reported below.
As shown in Table 3 the empowerment factors showed sig-
nificant differences between FFS graduates and the control
group in terms of higher levels among FFS graduates for the
factors, such as gender and trust, critical thinking, and house-
hold decision making capacity. However, power and influence
beyond the individual or household such on community level
did not show significant differences between the two groups.
Trust in community institutions and local authority was only
tested in Uganda where significant difference was observed.
422 WORLD DEVELOPMENT

Table 2. FFS by poverty indicators

Well-being indicators Kenya Tanzania Uganda
Control (%) FFS (%) Control (%) FFS (%) Control (%) FFS (%)
Land ownership
High 10 9 6 5 37 39
Medium 79 83 76 84 56 56
Low 11 9 18 11 7 5
Non-agricultural income
High 37 37 32 26 19 26
Medium 29 45 31 39 57 43
Low 34 18 37 35 24 31
Work as casual labor
High 59 61 62 67 37 55
Medium 20 23 19 22 32 32
Low 21 16 19 11 30 13
Livestock ownership
High 69 74 23 23 50 62
Medium 29 24 63 67 37 33
Low 2 2 14 9 14 4
Hire casual labor
High 46 57 37 42 38 67
Low 54 43 63 58 62 33
Household food security
High 22 29 29 23 49 70
Medium 39 42 16 45 25 15
Low 39 29 55 32 27 14
Quality of diet
High 22 27 8 14 5 13
Medium 69 73 63 82 84 78
Low 9 9 3 11 9
Housing standard
High 11 22 28 43 5 3
Medium 64 78 71 57 89 93
Low 25 2 5 3
Family health
High 82 87 94 84 87
Low 18 13 100 6 17 13
Level of children’s education
High 36 53 19 32 26 39
Medium 55 45 65 67 59 46
Low 9 2 16 2 15 15
Standard of family clothing
High 31 45 28 29 31 39
Medium 46 38 72 71 65 59
Low 23 17 4 2
Marital status
High 86 87 86 82 84 87
Low 14 13 14 18 17 13
Age of household head
High 74 68 80 75 93 91
Low 26 32 20 25 7 9
Note: in Tanzania and Kenya: control, baseline and FFS, graduate FFS members; in Uganda control, non NAADS/FFS members and FFS, NAADS/
FFS members. The choice to use the mean instead of assigning different values to the thirteen well-being indicators is contextual and not theoretically
based. It would be very difficult to carry out a participatory prioritization process of the indicators because these are not known to the informants, as they
are based on a consolidation process of statements about well-being. Figures are rounded up or down to nearest whole figure.

The data presented in Table 4 represent aspects that through Aspects related to innovation uptake express increase in
the qualitative field work were identified as expressions of emancipation and motivation. Since FFS provides farmers
empowerment. opportunity to test practices and technologies in a safe
 THE EMPOWERMENT ROUTE TO WELL-BEING: 423

Table 3. Empowerment factors by FFS membership, level of significance strongly significant levels on most. Access to services such as
difference bank account, savings/credit means, and receipt of advice indi-
Empowerment factors Kenya Tanzania Uganda cate negotiation skills and openness as well as determination
*** and drive. For example in Kenya 77% of FFS graduates had
Household decision-making capacity 0.000 0.000*** NA
bank accounts as compared to 45% among the control group.
Power to influence community 0.128 0.545 0.046
Gender and trust 0.000*** 0.000*** 0.000***
A move from subsistence farming where production is
Critical thinking 0.001*** 0.003*** 0.001***
mainly carried out for the sake of survival to a more market
Trust in community institutions NA NA 0.005**
driven commercial farming was identified as an indicator of
and local authorities motivation, self-confidence, and determination. In Kenya
FFS graduates showed a significant higher level of commer-
Significance level between FFS graduates and the control group derived cialization than the control group on all aspects tested. How-
from Pearson Chi-square test when testing the difference. Highlighted
ever, this picture was not confirmed in Tanzania and Uganda
values represent significance at min of 10% level.
**
p < .005 where differences were observed but not at statistically signif-
***
p < .001 icant levels.
An important aspect of personal empowerment is the level
of involvement in collective action and societal involvement.
group-based environment, uptake can be seen as an expression This was in the study examined though involvement in collec-
of collective decision making processes. FFS graduates tive marketing of produce, tenure of leadership positions, and
showed higher levels of innovation uptake than the control participation in voting. In Kenya FFS graduates showed sig-
group in most countries and aspects tested, but only at statis- nificant higher levels than the control group on all tested
tically significant levels in a few of the items, such as uptake of items. In Tanzania and Uganda some differences were ob-
improved crop varieties in Kenya and Uganda, vaccination of served but not at significant levels.
livestock in Kenya and soil fertility measures in Uganda. For
example in Uganda 46% of FFS graduates had started using (c) Links between poverty and empowerment
improved crop varieties as compared to 17% of farmers in
the control group. In one case a contradictory result stands The potential links between poverty status and empower-
out in Tanzania where the control group showed significant ment were explored through cross-tabulation between the
higher levels of vaccination of livestock the than FFS gradu- two aspects, comparing the differences between the poverty
ates. The reasons for this are not clear but might be related categories of “better-off” and “poor” in terms of the range
to a parallel IFAD agricultural development program that of empowerment factors and variable items. Levels of signifi-
operated during the same time period as FFS targeting non- cant difference between the two groups were then observed
FFS members. and are indicated in Table 5. The empowerment factors did
In relation to access to services FFS graduates showed high- not demonstrate frequent significant correlation with poverty
er levels than the control group on all items tested with levels except in terms of household decision-making capacity

Table 4. Differences between FFS graduates and the control group in terms of empowerment aspects
Kenya Tanzania Uganda
FFS (Control) FFS (Control) FFS/NAADS (Control)
Innovation uptake
Uptake of improved crop varietiesa 86% (71%)*** 43% (38) 46% (17%)***
Vaccination of livestockb 72% (44%)*** 2% (19%)d,*** 25% (9%)
Improvement of soil fertilityc 86% (77%) 66% (61%) 55% (23%)***
Access to services
Obtained agricultural advice or assistance in the last two years 84% (38%)*** 88% (61)*** 91% (41%)***
Obtained advice from other farmers 49% (20%)*** 53% (16)*** 60% (64%)
Membership in savings/credit org.c 77% (45%)*** 40% (27%) 43% (19%)***
Have a bank account 31% (14%)*** 13% (8.3) 50% (15%)***
Engaging with markets
Sold farm produce in past two seasons 89% (78%)** 85% (89%) 81% (74%)
Occasional or regular sale of produce 91% (71%)*** 57% (53%)
Store produce to fetch higher price 56% (44%)* 36% (23%)
Add value/process products 20% (5%)*** 39% (15,9) 1% (0%)
Collective action/social relations
Involved in collective marketing 14% (5)** 30% (28%) 14% (12%)
Hold leadership position 63% (40%)*** 58% (48%) 57% (43%)
Voted in the last local election 96% (89%)** 90% (97%)
a
xx for Tanzania, yy for Kenya and citrus for Uganda.
b
Vaccination of livestock in Tanzania and Kenya, improved poultry production in Uganda.
c
Uganda data from 2004 survey.
d
Significant in other direction, that is, control group higher than FFS.
*
p < 0.01.
**
p < .005.
***
p < .001
424 WORLD DEVELOPMENT

Table 5. Empowerment factors by poverty category Tanzania 53% of very poor farmers had obtained agricultural
Empowerment factors Well-being level advice in the last two years compared to 83% of the non-poor.
Commercialization of agriculture and collective action items
Kenya Tanzania Uganda also appeared to be linked to poverty levels. For example, in
**
Household decision-making capacity 0.001 0.555 NA Kenya 70% of non-poor farmers hold some kind of leadership
Power to influence community 0.197 0.217 0.598 position, while among the very poor the proportion was only
Gender and trust 0.457 0.546 0.007* 35%.
Critical thinking 0.530 0.000*** 0.051
Trust in community institutions NA NA 0.004**
and local authorities 4. DISCUSSION ABOUT ATTRIBUTION AND SOCIO-
Significance level between FFS graduates and the control group derived ECONOMIC CONTEXT
from Pearson Chi-square test when testing the difference. Highlighted
values represent significance at min of 10% level. The changes described in this study in terms of empower-
*
p < 0.01. ment and well-being among farmers appear to be attributed
**
p < .005. to the FFS learning process. The relationship between FFS
***
p < .001.
membership on the one hand and well-being and empower-
ment on the other includes comparisons between members
of FFS and farmers in a control group that is assumed to have
in Kenya, critical thinking in Tanzania, and trust in commu- characteristics similar to those of FFS members before they
nity institutions and local authority in Uganda. entered FFS, for example, a with-or-without assumption.
When cross-tabulating poverty levels with the empowerment However, this assumption can be challenged.
items, more frequent relationships appeared (see Table 6). The Firstly, FFS members could have been better off already be-
links between the uptake of innovations and poverty levels ap- fore they joined FFS. The study rejects this by showing that in
peared to be significant in all countries, with the non-poor the base line survey there were no significant socio-economic
showing a higher frequency of innovation uptake. For exam- differences between farmers who joined FFS and farmers in
ple, in Kenya only 32% of the very poor vaccinate their live- the control group (in the cases of Tanzania and Kenya). Sec-
stock as compared to 72% among the non-poor. The same ondly, as farmers actively chose to become members of FFS
situation was demonstrated in terms of access to services, groups, it cannot be ruled out that the two groups may have
which appeared to be linked to poverty levels, with the stron- had different physiological characteristics, for example, in
gest correlations in Kenya and Tanzania. For example, in terms of motivation and interest in learning, which may have

Table 6. Difference between non-poor and very poor in terms of empowerment variablesd,e
“Non-poor” (vs. very poor)
Kenya Tanzania Uganda
Innovation uptake
Uptake of improved crop varietiesa 89% (48%)*** 50% (31%)** 46% (14%)***
Vaccination of livestockb 72% (3%)*** 20% (5%)*** 26% (6%)*
Improved goats 6% (4%)
Improvement of soil fertilityc 93% (64%)*** 83% (54%)*** 32% (11%)**
Access to services
Obtained agricultural advice or assistance in the last two years 84% (58%)*** 84% (53%)*** 88% (62%)***
Obtained advice from other farmers 52% (35%) 46% (27%)*** 60% (53%)
Applied for credit 41% (32%)*** 34% (23%)**
Membership in savings/credit orgc 81% (60%)*** 41% (40%) 27% (13%)
Have a bank account 42% (12%)*** 33% (5%)*** 49% (24%)
Engaging with markets
Sold farm produce in past two seasons 89% (80)* 90% (89%) 88% (57%)***
Occasional or regular sale of produce 90% (76%)** 55% (59%)
Store produce to fetch higher price 63% (28%)*** 44% (21%)***
Add value/process products 19% (6%)** 43%% (20%)*** 1% (0%)
Collective action/social relations
Involved in collective marketing 16% (9%) 45% (9%)*** 17% (2%)*
Hold leadership position 70% (35%)*** 63% (38%)*** 58% (28%)***
Voted in the last local election 94% (90%) 95% (97%)
a
Improved banana for Tanzania, improved maize for Kenya and citrus for Uganda.
b
Vaccination of livestock in Tanzania and Kenya, improved poultry production in Uganda.
c
Uganda data from 2004 survey.
d
The whole dataset was used for this analysis, that is, both non- and FFS members.
e
The values in the table represent the % among “better-off” versus among poorest (value in brackets) for each indicator and the significance level of
correlation between the two variables.
*
p < 0.01.
**
p < .005.
***
p < .001.
 THE EMPOWERMENT ROUTE TO WELL-BEING:
affected the outcomes. The study does not deal with issues of
mindset, critical thinking, etc. from a psychological perspec-
tive and thus may not capture individual differences in such as-
pects. To minimize this risk, the control group in the 2007
survey consisted of farmers who had recently joined FFS
but had not yet been influenced by it (in the cases of Tanzania
and Kenya). Thirdly, the linear relationship in which learning
contributes to empowerment that produces well-being im-
pacts, for example, an all-other-equal assumption, can be chal-
lenged. In practice it probably is more of a parallel change
process. The socio-economic context in the three study areas
is complex, and there are likely to be many varying factors
contributing to people moving in or out of poverty. It can
be argued, however, that the socio-economic context is the
same for FFS members and non-members. The key difference
is that FFS members seem to be better at taking advantage of
opportunities and negotiating obstacles.
In addition to these challenges, one limitation of the study is
that it does not address the dynamics of poverty. What we
measure is the net result of people coming out of poverty
minus those who fall into poverty. It is likely that some farm-
ers fall into poverty, both among FFS members and within the
control group. The reasons for this are not addressed in this
study, nor do our data indicate whether any individuals moved
directly from being very poor to non-poor or whether they
gradually moved from very poor to poor and thereafter to
non-poor.
5. CONCLUSIONS: AN EMPOWERMENT ROUTE TO
WELL-BEING
The study confirms the hypothesis that group-based learn-
ing in FFS can lead to empowerment and act as a pathway
toward increased well-being. The fact that the data from
the three countries all point toward the same trend, despite
contextual differences in the countries studied, strengthens
this finding, and justifies making generalized conclusions
about a possible empowerment route to well-being. The
hypothesis was tested by examining three sets of relation-
ships: FFS and empowerment, FFS and enhanced well-being,
and empowerment and well-being, as well as by using two
types of data: farmer’s perceptions, and actual expressions
of empowerment.
The link between FFS participation and empowerment in
terms of both perceptions and expressions of power in every-
day life was very apparent in Kenya and to certain extent in
Uganda and Tanzania as well. All countries showed linkages
between innovation uptake and increased access to services
and FFS membership. Kenya, however, was the only country
that also showed significant differences between the two
groups in the aspects of engagement with markets, collective
action, and social relations. At the individual level, FFS
showed significant impacts across the countries on changes
in gender, trust, critical thinking, and household decision-
making capacity, which is in line with a more qualitative study
of FFS in Kenya (Duveskog, Friis-Hansen, & Taylor, in
press). This study also confirms that high quality of the train-
ing that fosters the process of empowerment is crucial. Loss of
quality when scaling up processes of empowerment is well rec-
ognized in the critical participation literature (Cooke &
Kothari, 2001; Parkinson, 2009). However, it should be noted
that power and influence beyond the individual or household
domain on the community level did not demonstrate a strong
relationship with FFS, apart from an increase in leadership
positions among FFS graduates, particularly in Kenya.
425
This study demonstrates a relationship between the FFS
learning process and poverty levels. In all three countries,
FFS graduates proved less poor than their fellow community
members. It was also shown that typical FFS members were
not significantly different from the average community mem-
ber in terms of the well-being indicators studied here. This
strengthens the conclusion of dependence between FFS partic-
ipation and increased well-being.
The links between empowerment and poverty are less clear
than the relationships described above. Possibly this indicates
that the relationship is a complex one with many additional
factors influencing the dependency between the two aspects
looked at in this study. However, the study results still pro-
duce some interesting findings on this issue. With some excep-
tions, the factors of perceptions of and attitudes to power do
not show a significant link to poverty level: that is, the less
poor do not perceive themselves to have more power in their
household or their community than the very poor. However,
when looking at actual expressions of empowerment in terms
of innovation uptake, access to services, engagement with
markets, and involvement in collective action, etc., a clear link
between poverty level and these variables was observed in all
three countries. The study is not able to determine the reasons
for this contradiction. However, possible explanations could
be that the very poor, if or when excluded from decision-mak-
ing processes and social engagement, do not perceive power as
a problem, thus giving a higher score in the study than the real
situation. On the other hand, when starting to engage in and
push for a change in the household or community, the non-
poor might encounter more challenges and obstacles in terms
of power relations and, therefore, rate their level of power low-
er than the real situation. The reason could also simply be that
feelings of power, including the aspects of self-confidence and
trust, etc., are not related to poverty levels, but are individual
attributes that cross-cut well-being categories. It is also impor-
tant to remember that the poverty-rating syntax in this study is
based mainly on indicators of assets and standards of living
and does not include more physiologically related items such
as feelings of power in everyday life.
Though the direct link between empowerment indicators
and well-being appearing weak in the study, the fact that the
links between FFS participation and both empowerment and
well-being indicators are strong, the authors still feel confident
to support the concept of an empowerment route to well-
being. Further research, that take into account a greater num-
ber of external factors in the institutional, social, and political
context is needed to un-pack the complexity and inter-con-
nectedness between empowerment and well-being.
6. IMPLICATIONS FOR DEVELOPMENT POLICY AND
PRACTICE
Three key sets of policy implications can be drawn from this
study. These are concerned with how processes of farmer
empowerment can be stimulated, with implications for
empowerment in the poverty debate and for how investment
in a technological route to well-being could benefit from in-
creased attention to the building of human resources.
The inter-connectedness between an empowering learning
process and enhanced well-being that this study demonstrates
indicates that enhancing human resources among poor farm-
ers is a crucial element in allowing them to access services
and to benefit from development investments. The study thus
indirectly questions the current widespread faith in mainly
technical solutions to poverty problems, as is for example
426 WORLD DEVELOPMENT
the case in current attempts to re-launch a green revolution
promoted by the Gates Foundation, AGRA and others, and
it calls for increased attention for empowerment of the poor.
In this light there is a need to find a balance between technical
and social innovations and recognition for the complex inter-
connectedness between the two.
Lessons can be drawn from this study for how to support
informal education or community learning for empowerment
outcomes. While the concept of empowerment has generated
considerable policy interest and is an included component in
many development programs, it has proved difficult to achieve
in practice. While external support for community empower-
ment is not easy, experiences with the IFAD/FAO supported
FFS program in East Africa show that it is possible. Lessons
from FFS indicate that stimulating empowerment requires a
comprehensive approach combined with high-quality facilita-
tion. Many empowerment initiatives place insufficient empha-
sis on both of these elements, for example, on-off visits by
NGOs to sensitize about a given project, and as a result they
have had few lasting effects in terms of empowerment. Lastly
local and national political elites seldom support empower-
ment efforts and may undermine the efforts of non-govern-
mental groups if they feel that their power base is
threatened. Further, the study contributes to the methodolog-
ical field of the measurement of empowerment in terms of its
attempt to produce definitions through a combination of qual-
itative and quantitative process indicators and expressions of
empowerment in the rural small-holder farming context. This
REFERENCES
Alsop, R., & Heinsohn, N. (2005). Measuring empowerment in practice.
Structuring analysis and framing indicators. Washington, DC: The
World Bank.
Anderson, J. R., Feder, G., & Ganguly, S. (2006). The rise and fall of
training and visit extension: An Asian mini-drama with an African
Epilogue. World Bank Policy Research Working Paper 3928.
ASPS. (2003). Gender and poverty impact monitoring for the Agricultural
Sector Programme Support (ASPS); Uganda: ASPS Component
Impact Processes. Copenhagen and Kampala., Department of Agri-
cultural Economics, Makerere University (DAEMUK) and Centre for
Development Research (CDR).
Barlett, A. (2005). No more adoption rates. In N. Lilja, J. Dixon, J.
Manners, J. Hellin, & H. S. Feldstein (Eds.), New avenues in impact
assessment. Mexico City: CIMMYT.
Braun, A., Jiggins, J., et al. (2005). A global survey and review of
Farmer Field School experiences. International Livestock Research
Institute.
Brembeck, C. S., & Thompson, T. J. (Eds.) (1973). New strategies for
educational development; the cross-cultural search for nonformal alter-
natives. Mass: Lexington.
Chambers, R. (1987). Sustainable livelihoods, environment and develop-
ment: putting poor rural people first. IDS Discussion Papers 240.
Chambers, R. (1993). Challenging the professions; Frontiers for Rural
Development. London: ITP.
Christoplos, I. (2003). Common framework for supporting pro-poor
extension. Lindau: LBL, Swiss Centre for Agricultural Extension.
Cooke, B., & Kothari, U. (2001). Participation: The new tyranny. London:
Zed Books.
Coudouel, A., Hentschel, J. S., & Quentin T. W. (2002). Poverty
measurement and analysis, in World Bank (2002). A sourcebook for
poverty reduction strategies. Volume 1: Core techniques and cross-
cutting issues. Washington, DC: World Bank.
Davis, K. (2006). Farmer Field Schools: A boom or bust for extension in
Africa. Journal of Agricultural Education and Extension, 13(1).
Davis, K., Nkonya, E., Kato, E., Mekonnen, D.A., Odendo, M., & Miiro,
R. (2010). Impact of farmer field schools on agricultural productivity
and poverty in East Africa. International Food Policy Research
Institute.
is an important contribution given the global lack of practical
tools and processes for measuring the social impact of capac-
ity-building efforts.
The fact that FFS is a non-formal education process does
not imply that what students learn is of less importance than
formal education, such as primary and secondary school. With
regard to empowerment, the non-formal setting gives FFS an
advantage over formal education because of its propensity for
immediate action, providing learning opportunities that have
direct application, and it is often close in proximity and acces-
sibility for those that need it (Brembeck & Thompson, 1973).
The study finally offers policy implications for the effective-
ness of support for demand-driven services. Demand-driven
agricultural advisory programs, such as NAADS in Uganda,
ASDP in Tanzania and NALEP in Kenya, as well as many
other participatory rural development programs, require farm-
ers who are able to articulate informed demands if they are to
benefit fully from services. While most programs include smal-
ler components that support sensitization and institutional
support, support to farmer empowerment in the sense of the
production of knowledge for a framework of action, as is
the case in FFS, is seldom given adequate attention by donor
agencies or national governments. The results of this study
indicate that there is a link between empowerment and poverty
alleviation and that the cost-effectiveness of agricultural pro-
grams could, therefore, probably be enhanced with a stronger
focus on investment in human resources, thus informal educa-
tion that builds human and collective capacities.
Duveskog, (2006). Agricultural education as entry point for change and
transformation among resource-poor: An analysis of Farmer Field
Schools in East Africa. A procedural paper. The Swedish University of
Agricultural Sciences Department of Urban and Rural Development.
Duveskog, D., Friis-Hansen, E. & Taylor, E. (in press). Impact of farmer
field schools on agricultural productivity and poverty in East Africa,
Journal of Development Studies.
Farrington, J., Carney, D., Ashley, C., & Turton, C. (1999). Sustainable
livelihoods in practice: Early applications of concepts in rural areas.
ODI Perspectives, (42).
Friis-Hansen, E. (Ed.), (2004). Farmer empowerment: experiences, lessons
learned and ways forward, Volume 1: Technical papers, Danida.
Friis-Hansen, E. (2005). Agricultural development among poor farmers in
Soroti district, Uganda: impact assessment of agricultural technology,
farmer empowerment and changes in opportunity structures. The Nordic
Africa Days. Uppsala: Nordic Africa Institute.
Friis-Hansen, E. (2008). Impact assessment of farmer institutional
development and agricultural change: Soroti district Uganda. Devel-
opment in Practice, 18(4–5), 506–523.
Gallagher, K. (2003). Fundamental elements of a Farmer Field School.
LEISA Magazine 2003(3).
Gautam, M. (2000). Agricultural extension the Kenyan experience: An
impact evaluation. Washington, DC: The World Bank.
Giddens, A. (1976). New rules of sociological method: A positive critique of
interpretative sociologies. London: Routledge.
Government of Uganda (2005). Proceedings of the mid-term review of
National Agricultural Advisory Services. Kampala: Government of
Uganda.
Haug, R. (1998). Agricultural extension – the power of knowledge in
processes of change: Lessons learned and future challenges for Africa.
NORAGRIC: Agricultural University of Norway.
Hayward, C. R. (1989). De-facing power. Polity, 31(1), 1–22 (IFPRI 2006).
Heemskerk, W., & Bertus, W. (2004). Building social capital for agricul-
tural innovation. Amsterdam: KIT Publishers.
Heemskerk, W., Lema, N., Guindo, D., Schouten, C., Semgalawe, Z.,
Verkuijl, H., et al. (2003). A guide to demand-driven agricultural exten-
sion. The client-oriented research management approach: Rural service
delivery for agricultural development. Amsterdam: KIT Publishers.
 THE EMPOWERMENT ROUTE TO WELL-BEING:
IAASTD. (2009). International assessment of agricultural knowledge,
Science and Technology for Development Sub-SaharanAfrica (SSA)
Report. Washington, DC: Island Press.
IFAD (International Fund for Agricultural Development). (1998). Report
and recommendation of the president to the executive board on a
proposed technical assistance grant to the Food and Agricultural
Organization of the United Nations for the East African pilot project
for farmer field schools in Kenya, the United Republic of Tanzania, and
Uganda. Rome: IFAD.
Kristjanson, P., Place, F., Franzel, S., & Thornton, P. K. (2002). Assessing
research impact on poverty: The importance of farmers’ perspectives.
Agricultural Systems, 72, 72–92.
Leeuwis, C. (2004). Communication for Rural Innovation: Rethinking
agricultural extension. Blackwell.
Likert, R. (1932). A technique for the measurement of attitudes. Archives
of Psychology, 140, 1–55.
Mancini, F., van Bruggen, A. H. C., & Jiggins, J. L. S. (2006). Evaluating
the integrated pest management (IPM) farmer field school outcomes
using the sustainable livelihoods approach in India. In F. Mancini
(Ed.), Impact of integrated pest management farmer field schools on
health, farming systems, the environment, and livelihoods of cotton
growers in Southern India (pp. 69–84). Published doctoral dissertation,
Wageningen University, The Netherlands.
Mwagi, G. O., Onyango, C. A., Mureithi, J. G., & Mungai, P. C. (2003).
Effectiveness of FFS approach on technology adoption & empower-
ment of farmers: A case of farmer groups in Kisii District, Kenya.
Proceedings of the 21st Annual Conference of the Soil Science Society of
East Africa, Eldoret, Kenya.
Narayan, D. (2000). Voices of the poor, can anybody hear us?. New York:
Oxford University Press.
Narayan, D. (2005). Measuring empowerment, cross-disciplinary perspec-
tives. Washington, DC: The World Bank.
Narayan, D., Chambers, R., et al. (2000). Voices of the poor, crying out for
change.. Washington, DC: The World Bank.
NEPAD (2007). Comprehensive African Agricultural Development Pro-
gram (CAADP). Addis: OAU.
Neuchatel Group. (2006). Demand Driven Agricultural Advisory Services
Lindau, AGRIDEA.
Neuchatel Group. (2008). Common framework on market-oriented
Agricultural Advisory Services, Lindau, AGRIDEA.
Okoth, J. R., Khisa, G., & Thomas, J. (2002). Towards a holistic Farmer
Field School approach for East Africa. LEISA Magazine, 18(3).
Parkinson, S. (2009). When farmers don’t want ownership: Reflections on
demand-driven extension in Sub-Saharan Africa. The Journal of
Agricultural Education and Extension, 15(4), p417–429.
Available online at www.sciencedirect.com
427
Praneetvatakul, S., & Waibel, H. (2003). A socio-economic analysis of
Farmer Field Schools (FFS) implemented by the national program on
integrated pest management of Thailand. In Proceedings CIMMYT
Impact Assessment Conference, Costa Rica (4–7 February, 2002).
Qamar, M. K. (2006). Agricultural extension at the turn of the millennium:
Trends and challenges. <www.fao.org/docrep/003/X7925M>.
Quizon, J., Feder, G., & Murgai, R. (2001). Fiscal sustainability of
agricultural extension: The case of the Farmer Field School approach.
Journal of International Agricultural and Extension Education, 8(1),
13–23.
Ravnborg, H. M., Boesen, J., Sørensen, A., Akello, Z., Bashaasha, B.,
Kasozi, S., et al. (2004). Gendered District Poverty Profiles and
Poverty Monitoring: Kabarole, Masaka, Pallisa, Rakai and Tor-
oro Districts, Uganda. DIIS Working Paper 2004: 1. Copenhagen:
DIIS.
Republic of Kenya (1997). Welfare Monitoring Survey III. Nairobi:
Government Printers.
Rivera, W. M., & Alex, G. (2004). Volume 3: Demand-driven approaches to
agriculture extension. Washington, DC: The World Bank.
Rola, A. C., Quizon, J. B., & Jamias, S. B. (2002). Do farmer field school
graduates retain what they learn? An investigation in Iloilo, Philip-
pines. Journal of International Agricultural and Extension Education,
9(1), 65–76.
Scoones, I., & Thompson, J. (1994). Knowledge, power and agriculture –
Towards a theoretical understanding. In I. Scoones, & J. Thompson
(Eds.), Beyond Farmer First. Rural people’s knowledge, agricultural
research and extension practice. ITDG Publishing.
Sen, G. (1997). Empowerment as an approach to poverty. Bangalore: Indian
Institute of Management.
Sen, A. (1999). Development as Freedom. Oxford: Oxford University
Press.
Umali-Deininger, D. (1997). Public and private agricultural extension:
Partners or rivals. World Bank Research Observer, 12(2), 203–224.
van den Berg, H., & Jiggins, J. (2007). Investing in farmers: The impacts of
Farmer Field Schools in relation to integrated pest management.
World Development, 35, 663–686.
World Bank (2002). Empowerment Source Book. Washington, DC: World
Bank.
World Bank (2008). World Development Report. Washington, DC: World
Bank.
<www.agra-alliance.org/>.
<www.undp-povertycentre.org/>.
Züger, R. (2004). Impact assessment of farmer field schools in Cajamarca,
Peru: An economic evaluation. Social Sciences Working Paper No.
2004-1. International Potato Centre, Lima.