Citation
Women in development

Material Information

Title:
Women in development A.I.D.'s experience, 1973-1985
Series Title:
A.I.D. program evaluation report
Creator:
Goddard, Paula O
Carloni, Alice Stewart
United States -- Agency for International Development
Place of Publication:
Washington D.C
Publisher:
U.S. Agency for International Development
Publication Date:
Language:
English
Physical Description:
v. : ; 28 cm.

Subjects

Subjects / Keywords:
Agricultural development projects -- Women ( lcsh )
Economic development projects -- Women ( lcsh )
Women in development -- Developoing countries ( lcsh )
Women -- Economic conditions -- Developing countries ( lcsh )
Genre:
federal government publication ( marcgt )
bibliography ( marcgt )

Notes

Bibliography:
Includes bibliographical references.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
21026095 ( OCLC )

UFDC Membership

Aggregations:
Women in Development
University of Florida

Full Text







A.I.D. Program Evaluation Report No. 18


Women in Development:
A.I.D's Experience, 1973-1985
Vol. 1. Synthesis Paper


April 1987


Agency for International Development (A.I.D.)


Washington, D.C. 20523


PN-AAL-o87


I



















WOMEN IN DEVELOPMENT: A.I.D.'s EXPERIENCE, 1973-1985
VOL. I. SYNTHESIS PAPER




A.I.D. PROGRAM EVALUATION REPORT NO. 18








by

Alice Stewart Carloni


U.S. Agency for International Development

April 1987




The views and interpretations expressed in this report are those
'of the author and should not be attributed to the Agency for
International Development.








TABLE OF CONTENTS


Page

Foreword ................................... ... .......... vii

Preface .............................. .................... viii

Acknowledgments ........................................... xii

Summary ................................................... xiii

1. Introduction ....... ...... ............................... 1

1.1 Background ..................................... 1
1.2 Methodology ..................................... 2
1.3 Terminology ..................................... 4
1.3.1 Women Versus Gender ...................... 4
1.3.2 Participants Versus Beneficiaries ........ 4
1.3.3 Achievement of Project Objectives Versus
Achievement of Goals ..................... 5
1.4 Conceptual Framework ............................. 6
1.4.1 The Baseline Situation ................... 6
1.4.2 Gender Analysis .......................... 8
1.4.3 Project Adaptation ..................... 8
1.4.4 The Gender of Project Participants ....... 9
1.4.5 Achievement of Project Purposes .......... 10
1.4.6 Impact on Women and Achievement of
Socioeconomic Goals ...................... 10

2. The Importance of Gender for Achievement of Project
Purposes .............................................. 11

2.1 Agriculture ..................................... 11
2.1.1 Farm-Level Projects ...................... 11
2.1.2 Mainstream Institutional-Development
Projects ................................. 17
2.2 Income Generation and Employment ................ 18
2.2.1 Job Training ........................... 18
2.2.2 Credit Projects ......................... 19
2.2.3 Income-Generating Projects ............... 20
2.3 Education and Training .......................... 21
2.3.1 Basic Education ......................... 21
2.3.2 Technical Education and Training ......... 22
2.4 Energy .......................................... 22
2.4.1 Cookstove Components ..................... 23
2.4.2 Fuelwood Components ...................... 23
2.4.3 Biogas ................................... 23
2.4.4 Reforestation ............................ 24








TABLE OF CONTENTS (cont.)


Page

2.5 Water Supply and Sanitation ..................... 24
2.5.1 Design of Water Supply Systems ........... 25
2.5.2 Construction and Maintenance of Water
Supply Systems ........................... 25
2.6 Regional Variations............................ 26
2.7 Conclusions ................................... 27

3. The Importance of Gender for Achievement of
Socioeconomic Development Goals ....................... 28

3.1 Increasing Agricultural Production .............. 29
3.1.1 Gender Access to Resources as a Factor
in Production ........................... 29
3.1.2 Gender-Linked Labor Constraints as a
Factor in Agricultural Production ........ 30
3.1.3 Gender Differences in the Control of
Produce as a Factor in Agricultural
Diversification .......................... 32
3.1.4 Gender Differences in Incentives as a
Factor in Production ..................... 32
3.1.5 Gender as a Factor in the Development
and Transfer of Technology to the
Rural Poor ............................... 33
3.2 Developing and Conserving Natural Resources ..... 34
3.3 Generating Employment ......................... 35
3.4 Raising Incomes of Low-Income People ............ 36
3.5 Reducing Malnutrition ........................... 37
3.6 Conclusions ..................................... 38

4. Practical Suggestions for Project Design and
Implementation ....................................... 40

4.1 Gender Analysis: Timing in the Project Cycle ... 40
4.2 Data Requirements and Costs ..................... 42
4.3 The Process of Gender Analysis: Ten Steps ...... 42
4.4 From Gender Analysis to Adaptations in Project
Design and Implementation ........................ 44
4.4.1 Women-Only, Women's Component, and
Integrated Approaches ................... 44
4.4.2 Targeting Resources to Women ............. 46
4.5 Project Adaptation .............................. 47
4.5.1 Change in the Focus of Project
Activities ............................. .. 47
4.5.2 Change in the Number of Women in the
Pool of Eligible Participants ............ 48
4.5.3 Adaptation of Credit Components .......... 49
4.5.4 Outreach of Existing Delivery Systems .... 49








TABLE OF CONTENTS (cont.)


Page

4.5.5 Location of Project Activities and
Services ................................. 50
4.5.6 Timing and Duration of Activities ........ 50
4.5.7 Facilities for Sleeping and Child Care ... 50
4.5.8 Choice of Language and Communication
Network .............................. 51
4.6 Conclusions ................................... ... 51

Appendixes

A. Findings
B. Methodology
C. List of Projects
D. Desk Reviews and Field Studies


References











-vii-


FOREWORD


A.I.D. was among the first donor agencies to recognize the
central role of women in economic and social development.
A.I.D.'s legislation and policy guidance on women in development
have served as models to others. The Agency has supported a
wealth of research and information sharing on this topic over
the years, and has acquired a significant amount of project
experience. It is this experience that is presented and
analyzed here.

The findings and conclusions of the study will be of
interest not only to scholars and practitioners in the field of
women in development, but also to the wider development
community. The study provides new insights on the question of
project effectiveness and impact. It suggests that attention to
gender is a necessary element in meeting project objectives and
in achieving overall development goals. With the publication of
this study, A.I.D.'s Center for Development Information and
Evaluation is pleased to be able to carry on A.I.D.'s tradition
of leadership in this important subject.


W. Haven North
Associate Assistant Administrator
Center for Development Information
and Evaluation
April 1987








-viii-


PREFACE


This report synthesizes information from 102 Agency for
International Development (A.I.D.) projects covering a 12-year
time span. It is not an evaluation of A.I.D.'s women-in-
development projects, nor is it a study of the Office of Women
in Development or its program. The report documents how A.I.D.
has interpreted its women in development mandate over the years
and how A.I.D. policy in this area is being implemented.
A,I.D.'s many years of leadership and experience with women in
development provide a basis for a much better understanding of
the role gender plays in the achievement of development goals.

The Center for Development Information and Evaluation
(CDIE) began studying A.I.D.'s women in development experience
during the summer of 1984 in anticipation of the 1985 Nairobi
World Conference To Review and Appraise the Achievements of the
U.N. Decade for Women. CDIE planned a year-long study that
would include a desk review of project documents in five sectors
and fieldwork in selected countries. CDIE formed a working
group of expert consultants and A.I.D. staff who were known for
their expertise in the women in development subject to advise on
the design of the study and to produce the reports.

But halfway through the year it became apparent that the
work could not be completed in the original time frame.
Significant differences emerged among the members of the working
group on the meaning of even common terms used in the existing
literature on women in development. The definitional and
conceptual problems were basic. What is "women in development,"
and to what does the term refer? Does it imply a program or a
project for women only? A goal of the development process? Or
a means to achieving development itself? Is the issue women or
gender? The Percy Amendment to the Foreign Assistance Act
directs A.I.D. to "integrate women into their national
economies"; however, what does "integrate" mean in a project
context? And once integration is achieved, does the term "women
in development" still apply? A.I.D.'s 1982 Women in Development
Policy Paper stresses the importance of women's contribution to
the success of projects. But the policy also emphasizes
equity--bringing the benefits of development to women--as an end
in itself. Which should be the focus of the study?

These problems slowed the preparation of the final
synthesis paper considerably. The 1-year schedule stretched to
nearly 3-years as the working group went through a painstaking








-ix-


process of defining and redefining terms. The group met
periodically to discuss the issues and to refine the evolving
study approach through many drafts of the final synthesis
paper. Defining and measuring women's participation in
development projects, for example, was a particularly
troublesome issue. What exactly is participation? How should
participants be counted? Are all household members living in
the area of the project counted equally as "participants"? Or
are participants only those who provide labor in a project? If
so, are unpaid laborers also considered "participants"? How
then do we differentiate between a person contributing labor to
a project and one receiving project resources, such as
training? What is the difference between "participating" in a
project and "benefiting" from a project, or are they one and
the same? A.I.D. documents usually interchange the two.

In the end, a consensus emerged on the best way to present
these key terms in the synthesis. The definitions are given in
Section 1, Introduction. However, the evaluation process
revealed a lesson with relevance beyond this study: ambiguity
and confusion persist in A.I.D.'s use of terms in project
design and evaluation, in particular concerning the "people
side" of development. Design and evaluation documents need to
be much clearer in defining what is meant by participation in
development activities, in describing beneficiary groups, and
in articulating the linkages between project participants,
project activities, project outcomes and benefits, and
development goals.

The study was also difficult conceptually. A.I.D.'s
women-in-development experience could not be analyzed neatly,
for example, in a particular sector. The Policy Paper
establishes women in development as a cross-cutting issue.
"The responsibility for implementing A.I.D.'s Women in
Development Policy rests with all of A.I.D.'s offices and
programs at all levels of decision making. Implementation of
this policy must be understood to be an important qualitative
aspect of A.I.D.'s overall program, one which is crucial to the
achievement of the Agency's goals. It is not a concern which
can adequately be addressed in any one sector alone or by any
single office or officer." (A.I.D. 1982, 9).

To be consistent with this policy, we selected the A.I.D.
activities for the study that were representative of projects
in A.I.D.'s total portfolio. We did not limit the study to
women-only activities nor to any one type of project. But the
analytical approach required to evaluate this broad perspective
on A.I.D.'s women-in-development experience was slow to
develop, even in the hands of experts.









We knew from the beginning that asking the question "does
the A.I.D. program benefit women?" was insufficient. We needed
to probe further to examine the implications of benefits or
lack of benefits to women for overall development goals. The
Policy Paper states that the economic participation of women is
essential to balanced economic growth--that if women benefit
from A.I.D.'s assistance, the whole process of development is
advanced. But how? And what is the cost to the development
process if women are disadvantaged or ignored by development
programs?

To answer these questions we developed the conceptual
framework found in Section 1, which guided the study through
the final synthesis stage. When evaluation reports concluded
that women's participation in a project was low, the group
pushed the analysis further and asked, in effect, so what? Did
the absence or presence of women in a project make any
difference in the project's outcome? What relationship, if
any, could be seen between explicit attention to gender roles
at the design or implementation stage and the successful
achievement of project objectives? In other words, how does
the gender of project participants and beneficiaries affect the
project and its goal? This approach offered the best hope of
linking our findings to the thrust of A.I.D.'s two-pronged
policy, focusing on both equity and efficiency.

The study results are organized in the synthesis paper to
correspond to A.I.D.'s logical framework analysis method for
project design, with stated goals at two levels: immediate
project purposes and longer term development goals. Section 2
describes the role gender plays in the achievement of projects'
immediate objectives, while Section 3 discusses the impact of
gender on longer term development outcomes. The findings from
projects with an institutional development focus were separated
from those that provided direct service delivery because these
kinds of projects were so different analytically. The results
of our analysis revealed that, in fact, gender variables do
affect projects with an institution-building goal differently
than projects that involve direct contact with farmers or
grassroots organizations.

The analysis is also divided among the five sectors
highlighted in A.I.D.'s Women in Development Policy Paper
(agriculture, education, water and sanitation, employment and
income generation, and energy/natural resources). Women-only
projects, women-only components of larger projects, and
mainstream (i.e., integrated) projects are compared in Section
4, and regional differences are briefly described.







-xi-


The results of the study corroborate the assertions in the
Policy Paper that gender variables influence the success of
development projects and the quality of the entire development
effort. These findings suggest that a better understanding of
gender roles in developing country societies is a key to
understanding the development process itself.

A.I.D. has long known that when projects are appropriately
planned and adapted to reflect local conditions, the projects
are more likely to achieve their objectives. A.I.D. procedures
for project design and implementation have been refined over
the years to promote this better understanding, through the
introduction of social soundness analysis, social and
institutional profiles, and other analytic techniques to the
project development process. Yet this study and other recent
studies (Robert Cassen and Associates 1986, Devres 1987)
concluded that development activities continue to reflect poor
understanding of the local conditions in which project
activities must operate.

This study suggests that understanding gender variables in
the context of a project is key to understanding human
variables in development. Understanding the ways in which poor
women and men interact, divide responsibilities, allocate risk
and resources, share burdens, organize their labor, and plan
for the future is essential to planning effective development
programs. With the publication of this two-volume report,1
CDIE hopes to enhance this understanding among A.I.D. staff and
other development practioners. It is our hope that future
development activities will then more accurately reflect the
role that gender plays in the populations that A.I.D.
assistance is designed to serve and that the benefits of
development will more likely be assured for all.


Paula Goddard
Deputy Associate Assistant Administrator
Center for Development Information
and Evaluation
April 1987






-A companion volume to this report will be published by CDIE
as "Women in Development: A.I.D.'s Experience, 1973-1985,"
Vol II, "Ten Field Studies."








-xii-


ACKNOWLEDGMENTS


This study was a group enterprise. Rae Blumberg provided
the original conceptual framework and the desk review
questionnaire. The synthesis paper was written by Alice
Stewart Carloni with editorial assistance from Chris Krueger
and Jane Knowles. The paper is drawn from 6 desk reviews and
10 field studies whose authors are listed in Appendix D. A
working group comprising many of these people met periodically
to advise and guide the study.

CDIE coordination of the study was provided by Ana Maria
Long and myself. Several other A.I.D. staff contributed their
comments and participated in the workshops including Kay
Davies, Margaret Sarles, Bill Miner, and Jack Francis.

Meta de Coquereaumont and Pamela McDade of Professional
Management Associates prepared the paper for publication,
including word processing, editing, proof-reading, and layout.

CDIE gratefully acknowledges these important contributions
and the good spirit in which all who contributed gave of their
time and effort.


Paula Goaaard







-xiii-


SUMMARY


Introduction


The paper synthesizes an evaluation of more than a decade
of Agency for International Development (A.I.D.) experience with
women in development. Evaluation objectives were to

Examine the relationship between gender .variables and
the achievement of project purposes and goals in the
five sectors highlighted in the 1982 A.I.D. Policy
Paper Women in Development

Draw lessons for improved policy implementation at the
project level

In 1984, A.I.D.'s Center for Development Information and
Evaluation (CDIE) drew up a list of 416 projects from A.I.D.'s
data bank, using such key words as "women," "female," and
"gender," and randomly selected 98 projects for analysis by
sector in a set of desk reviews. Of the 98 projects analyzed,
82 percent are mainstream and 18 percent are women's projects or
have a special women's component. In 1985, CDIE chose 10
projects for detailed field studies--6 from the cases in the
desk reviews and 4 others at the suggestion of A.I.D. staff.
The synthesis paper sums up the findings of these analyses and
the conclusions of two seminars with the study consultants and
A.I.D. staff.

The basic hypothesis underlying the evaluation is that
attention to the differences between the roles, responsibilities,
and opportunities of women and men increases the probability
that projects will involve and benefit women, achieve project
purposes, and contribute to the achievement of long-range
socioeconomic goals. In the light of this hypothesis, a
seven-component conceptual framework was developed, which linked
the following elements:

The baseline situation within which every project
operates

Analysis of the differing responsibilities and
opportunities of women and men

Adaptation of projects to deal with these differences


-- Gender of project participants







-xiv-


Relationship of these four factors to

achievement of project purposes
impact on women
achievement of socioeconomic goals


Findings


The major finding of the evaluation is that

mainstream projects that ensure women's
participation in proportion to their roles and
responsibilities within the project's baseline
situation are more likely to achieve their
immediate purposes and their broader
socioeconomic goals than are projects that do
not.

This finding is examined in detail at both the project-
purpose and the project-goal levels.


Achievement of Project Purposes


In the agriculture sector, projects that delivered
resources directly to women in accordance with their role in the
local farming system were much more likely to succeed in
achieving their purposes than were projects in which women did
not receive resources. The degree of match between the gender
division of labor in the baseline situation and the gender of
project participants is a key factor for successful efforts to
raise productivity and to diversify small farm production.

In a project in which gender roles in agriculture
were ignored, crops were planted incorrectly and did
not grow, machinery provided by the project was not
used, and crops were inadequately fertilized. In a
project in which design was adapted to fit gender
roles, significant amounts of labor were mobilized
for timely soil and water conservation efforts.

But the process of adapting projects to the gender division
of labor in the baseline situation is by no means automatic. In
case after case, the evaluation showed that project planners
should never assume that female farmers will be automatically
included in training or extension activities simply because
there are no formal barriers to their participation.







-XV-


Even where a mainstream project focused on crops or
activities that were primarily women's responsibility,
women's participation in these projects was low unless
delivery systems explicitly earmarked resources or services
to women.

In the employment sector, projects that adapted mainstream
training and credit programs to enable them to reach women were
more effective in expanding women's employment than women's
projects that worked outside mainstream institutions.

Income-generating projects for women very rarely
achieved their objectives and hence were of little
economic benefit to the participants. There proved to
be a wide gap between identifying women's needs and
designing viable projects.

Job training projects stopped short of creating
employment because trainees lacked capital to set
themselves up in business. The exception was a project
that adapted recruitment procedures of a mainstream
industrial job training program to enable women to
acquire nontraditional skills and then helped place
them in jobs in the industrial sector.

Credit projects that adapted eligibility criteria to
enable informal sector entrepreneurs with few assets to
receive loans had high levels of female participation
and were also the most successful in increasing
employment and raising beneficiary incomes. Female
entrepreneurs were better at loan repayment and
creating jobs than were men.

In the energy and water/sanitation sectors, the degree of
match between the gender of project participants and the
division of labor in the baseline situation is highly correlated
with achievement of project purposes and goals. Most projects
focused too narrowly on women's domestic roles, viewing women
mainly as passive beneficiaries of improvements in water and
fuel supply and overlooking their potential contributions to
complementary activities such as the construction and
maintenance of water systems, afforestation, and watershed
protection. Designs based on baseline studies of women's actual
patterns of water/fuel use and their preferences and constraints
were more successful than designs based strictly on technical
criteria.







-xvi-


Cookstoves saved less fuel than anticipated, partly
because of faulty design and partly because of their
lack of versatility. Instead of abandoning their
traditional hearths, people continued to use them for
some part of their cooking because of their greater
versatility.

Biogas plant milling fees were too costly, and women
needed the manure for other purposes; as a result, the
plant was little used.

An afforestation component was unable to meet its
target because the timing of the tree planting campaign
interferred with women's responsibilities for planting
crops on their own farms.

As in the agriculture sector, energy and water/sanitation
projects that analyzed and adapted to the reality of women's
roles in the baseline situation were more successful in
achieving their purposes than were those that did neither.

In the education sector, the correlation between the level
of women's education and the goal of achieving improvements in
health, hygiene, nutrition, and family planning is clear.
Education is also correlated with higher productivity and
employment. However, the correlation between the level of
female participation and achievement of immediate project
purposes is difficult to document. Most of the evaluation
reports used for the study overlooked gender disparities in
enrollment when assessing educational institution-building
projects.


Achievement of Project Goals


The lack of a clear causal relationship between the
achievement of project purposes and long-term socioeconomic
goals continues to be a sensitive point for development
assistance. The relationship between them is neither linear nor
automatic. The findings from this evaluation show that gender
variables intervene at every step in the chain in critical
ways. The paper reviews the relationship of gender variables to
a set of socioeconomic goals and concludes that

Understanding gender factors in agricultural
production is crucial to the successful transfer of
technology into agricultural systems. These factors
include differential access to and control over







-xvii-


resources, gender-linked labor constraints, control
of income from sale of crops, and differing stakes in
and incentives for increasing output.

Understanding the various sources of income and the
control over the uses of income is crucial to raising
the levels of living of the poor. Understanding
gender roles in consumption is crucial to the
achievement of improved nutrition and family
well-being.


Conclusion


The paper provides 10 practical suggestions for project
design, implementation, and evaluation. They include adequate
analysis of gender differences in the baseline situation,
appropriate adaptation of project delivery systems in the light
of those differences, and tracking the results of such
adaptations. Finally, the paper reviews the relative success of
women-only projects, projects with women's components, and
mainstream projects that attempt to "integrate" women into their
activities. It concludes that

Integrated projects require gender-sensitive designs in
order to be successful in achieving project purposes
and in benefiting women. Of the three project types,
gender-sensitive, "integrated" (or mainstream) projects
are the most effective in promoting and utilizing
women's contributions to socioeconomic development.
Mainstream projects can often be made gender sensitive
by adjustments in the focus of project activities, in
their location and timing, and in support services.
The synthesis paper identifies a number of such
adaptations.

-- Women-only projects tend to be very small in scope and
disproportionately costly in staff time. The chances
of success can be enhanced by locating them in major
(i.e., mainstream) institutions rather than in women's
bureaus or local voluntary organizations. Women-only
projects are best suited to delivering training rather
than to increasing production or generating income.

Women's components in larger projects can be an
effective way of benefiting women if the components are
well integrated into the whole range of project
activities. Alternatively, if they focus only on
women's domestic roles, they can lead to tokenism and
distract attention from significant gender issues in
the overall project.







-xviii-

Evaluation findings show that project investments based on
gender analysis can have higher returns and provide benefits to
women. The returns could be much greater if institutional
barriers to women's participation and benefits were better
understood and projects were adapted to overcome the barriers.
Women-only projects and women's components of projects can be
useful in specific contexts. Gender-sensitive adaptation of
mainstream projects will most effectively include women in the
development process and also provide a higher return to project
investments.







1. INTRODUCTION


1.1 Background


Women in development has been a subject of particular
attention for the Agency for International Development (A.I.D.)
since 1973, when the Percy Amendment to the Foreign Assistance
Act required that U.S. bilateral assistance programs

be administered so as to give particular attention to
those programs, projects and activities which tend to
integrate women into the national economies of foreign
countries, thus improving their status and assisting
the total development effort.

Publication of the Women in Development Policy Paper in
1982 further refined the Agency's approach. The Policy Paper
emphasized that

For A.I.D. to undertake an effective strategy that
promotes balanced economic development, a focus on
the economic participation of women in development is
essential.... Research from the last decade portrays
a fairly consistent pattern of findings that in most
developing countries, females differ from men in
their:

-- Access to and control over productive resources

Stakes in development outcomes

Responses to incentives introduced to encourage
development

The Women in Development Policy Paper illustrates the
importance of women's economic roles in five sectors:
agriculture, employment and income generation, human resource
and institutional development, energy and natural resource
conservation, and water and health. It concluded that

The key issue underlying the women in development
concept is ultimately an economic one: misunder-
standing of gender differences, leading to inadequate
planning and designing of projects, results in
diminished returns on investment. Gender, therefore,
is a critical category of analysis in A.I.D.'s work,
one which has not received sufficient attention to
date.







-2-


Beginning in 1984, in preparation for the Nairobi World
Conference to Review and Appraise the Achievements of the U.N.
Decade for Women, the Center for Development Information and
Evaluation (CDIE) in A.I.D.'s Bureau for Program and Policy
Coordination undertook a systematic evaluation of more than a
decade of A.I.D. experience with women in development. The
objectives of the evaluation study were to

Assess A.I.D. experience with women and
socioeconomic development in the five sectors
highlighted by the Policy Paper

Examine the relationship between gender variables
and the achievement of project purposes and goals

Draw lessons for improved policy implementation
at the project level

The study generated a wide range of materials, including 6
desk reviews and 10 field-based case study reports. The desk
reviews are among CDIE's Working Papers; the case studies are to
be published together as Volume II of this report. The entire
set of documents is listed in Appendix D. The purpose of this
synthesis paper is to present the conceptual and practical
findings of the entire effort.

The paper is organized in four major sections and three
substantive appendixes. The introduction presents a basic
discussion of the methodology, terminology, and conceptual
framework. Sections 2 and 3 examine the key relationships with
which the study was concerned: the importance of gender issues
for the achievement of project purposes and of long-term
socioeconomic goals. Section 4 returns to the conceptual
framework and spells out more clearly the intervening
linkages--especially project adaptation--that must occur if
A.I.D. projects are to effectively promote and utilize women's
contributions to socioeconomic development. Appendix A presents
findings from the computer-based analysis of data from 98
projects. Appendix B is a detailed presentation of the
methodology employed in the sample selection and the analysis.
Appendix C is a list of the projects examined in this study.


1.2 Methodology


Using the Women in Development Policy Paper as a guide,
five sectors were chosen for analysis: agriculture,
employment/income generation, education, energy/natural resource
conservation, and water supply/sanitation. A list of 416
projects was drawn from A.I.D.'s document data bank, using such








-3-


key words as "women," "females," and "gender." It was necessary
to use these terms as identifiers to ensure that the project
documentation for the sample would provide information relevant
to the purpose of the study, that is, understanding the
relationship between gender variables and project outcomes.
Next, 98 projects were randomly selected from the list of 416.
It is important to highlight that 82 percent of the projects in
the sample were mainstream projects and only 18 percent were
women's projects or had a special women's component. Thus,
findings based on the analyses of the project sample can be
regarded as relevant for most A.I.D. projects. (See Appendix B
for a more detailed discussion of the sample and methodology.)

During the first phase of the study, six desk reviews--two
for agriculture and one for each of the other sectors
chosen--were made by experts familiar with women-in-development
issues in each sector. A questionnaire was used to elicit and
codify information contained in project documents for analysis
across sectors. Compilation of the the results of the
questionnaire revealed gaps in the documentation. In some
cases, the reviewer's first-hand knowledge of the project
allowed her to fill those gaps; in others, documentation gaps
were filled by consulting A.I.D. project officers. Nonetheless,
for some projects, incomplete documentation prevented full
analysis. For this reason, the findings from the computerized
analysis reported in Appendix A refer only to those projects on
which there was sufficient documentation to permit analysis.

As the desk review phase drew to a close, a second phase
was designed to carry out field studies of 10 projects: 6 of
these projects had been included in the desk review; 4 other
projects recommended by A.I.D. officers as likely to provide
insights relevant to the study were added. The full list of
projects covered in both phases is in Appendix C. Each time a
project is referred to in the text, an identification number
appears in parentheses; by referring to the list in Appendix C,
the reader can identify the country and the project title.

This synthesis paper pulls together findings from both
phases of the study and from discussions by study participants
and A.I.D. staff who joined in two seminars as this synthesis
was being developed. While a certain degree of conceptual
clarification and confidence has been reached to date, it is our
hope that readers will examine the findings against their own
knowledge and experience and contribute to ongoing efforts to
understand what constitutes effective project implementation and
socioeconomic development, and the role gender analysis and
project adaptation play in achieving them.









1.3 Terminology


Before discussing the conceptual framework, it is necessary
to comment on some key terms.


1.3.1 Women Versus Gender


The initial impetus for a focus on the roles and functions
of women in developing countries was a concern with equity--an
attempt to ensure that development projects and processes
provided benefits to women as well as to men. Over time,
however, it became clear that differences between the roles,
responsibilities, and opportunities of men and women have
implications that go beyond equity; they also affect projects'
ability to achieve their immediate purposes and long-range
development goals. The introduction of the more relational term
"gender" to complement emphasis on "women" is one important
indication of a growing concern with the broader implications of
differences between men's and women's economic roles.

A focus on "women" in isolation can obscure differences
among women stemming from age, socioeconomic status, and stage
in the household cycle. "Gender" is a broader analytic concept,
which not only encompasses concern with women but also
highlights women's roles and responsibilities in relation to
those of men. Gender, like age and socioeconomic status, is an
aspect of social organization that both reflects and is
circumscribed by the surrounding culture. A fuller treatment of
the concept of gender and its application to the process of
project design and implementation is one of the themes of this
paper.


1.3.2 Participants Versus Beneficiaries


In recent years, development theorists have come to
differentiate between direct versus indirect and intended versus
unintended beneficiaries of projects, but there has been no
comparable rigor in the distinction between project participants
and beneficiaries. Regardless of whether they play any active
role in a project, residents in affected areas are commonly
referred to as "participants"; regardless of whether or not they
actually benefit, participants in project activities are
commonly referred to as "beneficiaries."

This study and others have indicated the clear need for
more precise terminology. The simple equation of presence in
the project area with "participation" and the equation of








-5-


participation with' "benefit from" a project is particularly
troublesome in the case of women. For example, it can never be
assumed that if one family member participates in a project, the
whole family participates. Nor can it be assumed that if one
family member benefits, there is an automatic "trickle over"
benefit to other family members. Precision about the gender,
age, and socioeconomic status of project participants and actual
beneficiaries is important. Clarification of terminology is a
prerequisite for shedding light on how the distribution of
benefits within households affect women's and men's differential
incentives to undertake project activities and how these affect
project outcomes.


1.3.3 Achievement of Project Purposes Versus Achievement
of Goals


A.I.D. terminology distinguishes between a project's
immediate "purposes" and the long-range development "goals" the
project was intended to further. Even though, strictly
speaking, project designers and implementers can only aim at
achieving a project's immediate purposes, the nature of these
purposes and the strategies for achieving them must be judged by
their effectiveness in contributing to overall development
goals.

Increasingly it is recognized that development assistance
encounters serious difficulty precisely in this regard: many
projects that achieve their immediate purposes fail to have a
sustainable impact on the lives of the people they were intended
to help. This is usually explained by exogenous factors (events
beyond the project's control, such as bad weather, shifts in
government policy, or changes in the world market). It is less
often recognized that such factors as mistaken assumptions about
how a project's immediate objectives actually contribute to
well-being can also be responsible.

Mistaken assumptions about the roles and responsibilities
of men and women are a factor that deserves much more
attention. The literature on women in development shows that
even in cases where immediate purposes are achieved, projects'
contribution to overall development.can be minimized by failure
to take gender roles into account. In the worst case, achieving
project objectives while overlooking intrahousehold dynamics can
be counterproductive. There are cultural contexts in which
project designers can systematically harm family welfare if they
do not know how responsibilities are divided among family
members. For example, a project may achieve its objective of
income generation but fail to achieve its goal of alleviating
hunger if, because of family expenditure patterns, the income







-6-


generated by the project is not used for the purchase of food.
In later sections, this paper further examines the role gender
plays in explaining how projects can increase production and
raise income without actually improving well-being.


1.4 Conceptual Framework


From the beginning, the basic hypothesis underlying the
evaluation was that

Attention to gender differences increases the
probability that projects will reach women, benefit
them, achieve project purposes, and contribute to the
achievement of long-range socioeconomic goals

Overlooking gender differences decreases the
probability that projects will reach women and benefit
them; this in turn reduces the probability of achieving
immediate project purposes as well as broader
socioeconomic goals

In the light of this hypothesis, a seven-component
conceptual framework was developed, which linked the following
elements:

The baseline situation
Gender analysis
-- Adaptation of projects to deal with gender differences
-- Gender of project participants
-- Achievement of project purposes
Impact on women
Achievement of socioeconomic goals

The relationship among these factors is illustrated
schematically in Figure 1.


1.4.1 The Baseline Situation


Every project operates within a specific geographic,
economic, social, and cultural situation as well as a broader
policy environment. The gender-linked division of labor and of
other roles and responsibilities varies widely between countries
and among different social, economic, and cultural groups within
countries. It was hypothesized that recognition of such gender
differences would have a larger impact on projects concerned
with activities in which women's role is major than on those in
which their role is minimal.





-7-



>4 z z \ z
So a
E H 0 H 0
MZ < >
(H 0 m H u
O (i 1nmm o


uo aU8 ou ot uo
HH U O I O MU


MICm W4 00 3 WZ0o


I > I O







6E- 5
0 0

0z 0 p
UzB E UZOEOE

0 S O 0 0





I .
Ai 1 -----~1 ----:14- -- -


on o H









$4
0 rH


E-4 W-l W M
SU I u W u

0
O

a a

0 z


mo o



H tH
MW K^Wl< c

S 0 > mu w
3 m!OS U 0 C4
1H EZ H 0 A rE
M 0 P4m E 0 o P4






0E-4 z ? a

z H WA

W ~E1 ll W W W % E-1 M





^ m







tC0 U to i to










1.4.2 Gender Analysis


For the purpose of this study, gender analysis is defined
as the analysis of the intersection between male and female
roles/responsibilities and project goals, strategies, and
outcomes at any stage of the project cycle. During the desk
review phase of this study, project documents were examined for
the quality of gender analysis.

The following items were identified as essential for an
understanding of gender factors:

The division of labor
Access to and control over productive resources
Stakes and incentives in project activities
Contribution to household income
Degree of income pooling within the household
-- Responsibilities for different types of expenditure

Each project was coded first to assess how much was said
about women in project design, implementation, and evaluation
documents; second, to determine whether or not women's
productive as well as their reproductive roles were analyzed;
and third, to evaluate the linkage of this analysis to project
activities. For example, agriculture projects that analyzed the
gender division of labor and pointed out how integration of
women could enhance the project were rated "high" on gender
analysis; projects that mentioned women as potential
beneficiaries but failed to link their role to project outcomes
were coded as "low" on gender analysis.

The basic hypothesis was that a detailed gender analysis
would increase the probability that women would participate in
and benefit from the project, whereas superficial gender
analysis ("boilerplate") would not affect women's participation
or benefit very much. It was further assumed that even the best
gender analysis would not automatically increase women's
participation or benefit if constraints affecting outreach to
men and women of different ages and socioeconomic status were
not recognized and explicitly addressed.


1.4.3 Project Adaptation


This component of the conceptual framework is linked
closely with gender analysis. In fact, findings from this study
suggest that it can be so closely linked as to be overlooked.
When working hypotheses were formulated for the desk study, it
was assumed that if gender roles in the baseline situation had








-9-


been recognized and analyzed, project documentation would show
corresponding adaptations in project design and implementation.
Initially, only three possible adaptations were considered: a
"women's project," a "women's component" in a larger project, or
"earmarking" resources such as credit, equipment, personnel
slots, or training for womenI in a mainstream project. Later,
the concept of project adaptation was expanded to cover a broad
spectrum of possible interventions, including gender-responsive
changes in project components, institutions, delivery systems,
technical packages, and feedback mechanisms. It was
hypothesized that without some type of adaptation, mere analysis
of gender differences would have little effect on participation
or the distribution of benefits.


1.4.4 The Gender of Project Participants


One of the study's main objectives was to examine the
relationship between the gender of project participants at the
grassroots (or "beneficiary") level and project outcomes. The
interest grew out of insight into the gender-linked division of
labor in smallholder agriculture, informal sector employment,
domestic water, and fuel-related activities.

Women's level of participation was defined in terms of the
following:

-- The share of project inputs going to women (in relation
to men)

The proportion of women among persons taking part in
project-organized activities (such as education,
training, and demonstration)

The proportion of women among persons employed by the
project

Women's representation in organizations (such as
water-user groups, cooperatives, or farmers
associations) and their voice in project decision-making

-- Women's labor in major project activities (such as crop
production, soil conservation, or afforestation)

As the study progressed, it became increasingly clear that
women's absolute level of participation in project activities
was less relevant for project outcomes than the degree of match


IFull definitions are in Section 4.4.







-10-


between the gender of project participants, the baseline
division of labor and responsibilities, and the activities
targeted by the project.

Participation was judged to be "high" when the relative
proportion of women participating in project activities
reflected the gender division of labor/responsibilities in the
major activities targeted by the project. A project was judged
to have low levels of women participation when the proportion of
women did not reflect the division of labor in the baseline
situation (e.g., farm inputs are delivered only to male farmers
for activities that are women's responsibility in the baseline
situation). Women's participation in activities peripheral to
the project's main thrust (such as cookstoves in a watershed
management project) was expected to influence overall project
outcomes only slightly. Participation of female professionals
was expected to affect project success only when it improved
contact with village women.


1.4.5 Achievement of Project Purposes


A major concern of the study was to examine the importance
of gender variables for the achievement of immediate project
objectives in the five sectors covered by the Women in
Development Policy Paper. The hypothesis was that where women's
roles are extensive in activities targeted by a project (e.g.,
farming, nonfarm production, domestic fuel and water supply,
family health and nutrition), high levels of female
participation increase the probability of achieving project
purposes; where women's roles in targeted activities are
extensive but women's level of project participation is low,
there is a lower probability of achieving the same purposes.
The key variable is the degree of match between the
characteristics of project participants and the division of
labor/responsibilities in activities affected by the project.


1.4.6 Impact on Women and Achievement of Socioeconomic Goals


One of the novel and intriguing aspects of the conceptual
framework is the role played by gender variables in the linkage
between achievement of project purposes and broader socio-
economic goals. All project purposes, if achieved, are meant to
have a positive impact on larger societal aims. These larger
goals--increased income, improved health, better nutrition,
better standards of living, or reduction of poverty--are
supposed to flow from the achievement of individual projects'
purposes (although exogenous factors in the environment








-11-


surrounding projects also affect this linkage). The conceptual
framework suggests that the final impact of projects flows
through gender variables, which in turn influence whether
intended goals are reached.

An aim of the evaluation is to identify precisely where and
when gender variables matter for the achievement of these
broader development goals. It was hypothesized that projects
that achieve their immediate purposes and at the same time have
a positive impact on women are more likely to achieve their
goals than projects that achieve their purposes but have a
negative impact on women. Projects that fail to achieve their
purposes and have a negative impact on women are even less
likely to achieve their goals than those that have a positive
impact on women in spite of failure to achieve other purposes.


2. THE IMPORTANCE OF GENDER FOR
ACHIEVEMENT OF PROJECT PURPOSES


A.I.D. terminology distinguishes between a project's
immediate purpose and the long-term goals a project is meant to
further. This section of the paper examines the relationship
between gender variables and achievement of short-term project
objectives (or purposes) in each of the five sectors covered in
the Women in Development Policy Paper. It draws on the findings
of the 10 case studies of field projects in addition to the
projects covered by the desk review.


2.1 Agriculture


The greatest share of A.I.D. resources is devoted to
agricultural development. As a result, 40 of the projects
analyzed in the desk review were in agriculture; 7 of the 10
case studies of field projects were also concerned, wholly or
partly, with agriculture. Four of these case studies were
included in the desk review; the other three were suggested by
A.I.D. Missions as projects likely to provide insights relevant
to the scope of the study. In all of the projects examined,
women were responsible for some of the crops or activities
selected for project intervention.


2.1.1 Farm-Level Projects


According to the Women in Development Policy Paper, the
importance of gender variables for the success of projects aimed
at raising the productivity of small farmers derives from the
following factors:







-12-


Differences between men and women in their access to
and control over productive resources (such as land,
labor, capital, and expertise)

SDifferences in labor allocation (between different
crops, as well as between crop and livestock
production, off-farm work and domestic tasks)

Differences in stakes and incentives (depending on the
control of crops and which family member benefits from
a given activity)

It was expected that where women played a major role in
activities targeted by the project, it would be possible to show
a clear correlation between women's direct participation in
project activities and the achievement of project objectives.
Computerized analysis of the desk review sample strongly
supports the assertions of the Policy Paper. Table 1 shows that
when women's participation was high (i.e., substantial numbers
of women received training, credit, and extension), projects
were much more likely to achieve their objectives than when
participation was low.


Table 1. Relationship Between Level of Female
Participation and Achievement of Project Objectives:
Mainstream Direct-Servicea Projects in Agriculture
(n = 20 projects with information)


Level of Achievement of Objectives
Participation Low (n=6) Medium (n=6) High (n=8)


Low (n=ll) 5 4 2


Medium (n=5) 1 2 2


High (n=4) 0 0 4


aA distinction was made between mainstream projects having
direct contact with people at the grassroots level referredd to
as "direct service" projects) and those having no grassroots
contact (referred to as "institutional development" projects).

Another important finding is that token participation of a
few women is not correlated with achievement of objectives. What
matters is not whether the project was "equitable" in the sense







-13-


of participation by a few women, but whether it reached the right
people from the standpoint of gender roles in the farming system.
The degree of match between the gender division of labor and the
gender of project participants is the key factor for efforts to
raise productivity and to diversify small farm production.

Gender analysis by itself, however, failed to show the
expected relationship to achievement of project objectives (see
Table 2). Low levels of analysis were often associated with
failure to achieve objectives, but better analysis of gender
differences was no guarantee that objectives would be achieved.


Table 2. Relationship Between Gender Analysis and Achievement
of Project Objectives: Mainstream Direct-Service
Projects in Agriculture
(n = 21 projects with information)


Level of Gender Achievement of Objectives
Analysis Low (n=10) Medium (n=6) High (n=5)


Low (n=ll) 6 4 1


Medium (n=5) 2 1 2


High (n=5) 2 1 2



To understand why good gender analysis was not correlated
with achievement of objectives, we looked at the relationship
between gender analysis and women's participation. Low levels
of gender analysis showed some correlation with low participa-
tion, but high levels of gender analysis were not correlated
with higher participation. This led us to look for an
intervening variable that could explain why some projects with
the best gender analysis had low levels of female participation.

Table 3 shows the correlation between gender analysis and
project adaptation. Initially, we expected that adaptation
would flow automatically from analysis of gender differences.
Table 3 shows that this assumption was overly optimistic:
although the expected relationship was found between failure to
analyze gender roles and lack of adaptation, the projects that
did the best job of gender analysis nevertheless rarely made
appropriate adaptations to overcome barriers to women's
participation or to increase the benefit to women.







-14-


Table 3. Relationship Between Gender Analysis and Project
Adaptation: Mainstream Direct-Service Projects in Agriculture
(n = 22 projects with information)



Level of Gender Project Adaptation
Analysis Low (n=16) Medium (n=5) High (n=1)


Low (n=ll) 10 1 0


Medium (n=5) 3 1 1


High (n=6) 3 3 0



Further analysis revealed a high correlation between
project adaptation and levels of female participation (see Table
4). Not only was women's participation predictably lower when
projects failed to adapt training and extension methods and
messages, but--more important--female farmers' participation was
substantially higher when conscious efforts (i.e., adaptations)
were made to remove barriers to their participation.


Table 4. Relationship Between Project Adaptation and
Women's Participation: Mainstream Direct-Service
Projects in Agriculture
(n = 18 projects with information)



Level of Level of Women's Participation
Adaptation Low (n=ll) Medium (n=6) High (n=l)


Low (n=12) 9 3 0


Medium (n=5) 2 3 0


High (n=l) 0 0 1








-15-


Given the fact that projects with higher levels of direct
involvement of female farmers are more likely to achieve their
objectives (Table 1), project adaptation should be a high
priority for future action.

These findings are corroborated by Cloud in her desk review
of agriculture projects. She found that projects were much more
likely to achieve their objectives when the flow of resources to
women matched the division of labor. When women's role in
project-related activities was major but resources did not reach
them, objectives were not achieved. Likewise, women were much
more likely to receive resources in projects that explicitly
targeted them than in projects that did not.

In case after case, the evidence showed that project
planners should never assume that female farmers will be
automatically included in training or extension activities
simply because there are no formal barriers to their
participation. Even in cases where the project focused on a
women's crop or activity, project resources bypassed them when
they were not explicitly earmarked for women.

For example, in Burkina Faso (previously Upper Volta),
where women are responsible for most of the small-scale sheep,
goat, and poultry production, a village livestock project (84)
initially directed resources for small animal production to
men. As a result, the people responsible for production--
women--got no assistance. Toward the end of the project, a
consultant who was brought in to find out why the project was
not working recommended that future activities be directed to
women. From then on, efforts were made to adapt the project to
women, but little could be accomplished before the project
terminated. The only thing village women actually received from
the project were six roosters intended for cross-breeding, which
died after project assistance was phased out.

In Cameroon (78), the connection between gender and the
failure of a seed multiplication project to achieve its
objectives is less clear. It is known that the project failed
to achieve its objectives because the quality of groundnut seed
produced by small-scale contract growers was poor. It is also
known that the project delivered resources to men, who had
little or no previous experience growing groundnuts since this
had been women's work. Although it is impossible to draw a
direct connection from the data available, a strong associative
relationship exists.

Of the 20 cases, only one appears to contradict the
hypothesis. This was a small farmer credit project in Egypt
(76) that was highly successful in achieving its objectives
despite a lack of match between the gender of loan beneficiaries








-16-


and the division of labor in the baseline situation. What is
intriguing about the case is that project planners originally
anticipated that the majority of loans would go for field crops
(men's domain) rather than livestock (women's domain). What
actually happened is that the demand for livestock loans greatly
exceeded the demand for crop loans. In part, the lack of
interest in field crops reflected low prices, but widespread
male migration may also have been a disincentive for investment
in field crops. In the absence of conscious efforts to ensure
that the persons responsible for livestock--women--could qualify
for loans, only 8 percent of the loans were given in a woman's
name. In the other cases, it is possible that men took out
loans on behalf of their wife. Apparently, then, women received
credit for their livestock activities despite their lack of
formal access to loans. In this case we do not know, however,
whether the women controlled the earnings from the sale of the
livestock products.

In Fortmann's desk review, there were no cases with enough
information to conclude that a project had failed to achieve its
objectives because of lack of gender adaptation. At most,
failure could be attributed to a lack of understanding of the
target group and the local farming system in general, and the
tendency to overlook the role of women could be seen as
symptomatic of this larger problem (02, 03, 07).

The costs of overlooking gender are illustrated in the case
study of the Northeast Rainfed Agricultural Development project
in Thailand (94, see Blanc-Szanton, Viveros-Long, and
Suphanchainat). Project management assumed that men were the
principal farmers and trained them to carry out crop trials. In
reality, many men had outside income sources and were frequently
away from the farm. Because wives of "specialist" farmers
received no training, crops were planted incorrectly and did not
grow, the power tillers provided by the project could not be
used, and a nitrogen-fixing crop intended to fertilize rice did
not get planted. Even when the husband was present, advice on
crop production was incorrectly transmitted from husband to wife.

Women were never consulted about their interest in the
project. In some instances, when the wives found out how much
additional work the trials would entail, they pressured their
husband to drop out. Some trials fell a year behind schedule.
The loss of time could affect the project's ability to transfer
technologies to farmers during the life of the project.

In contrast, the benefits of adaptation to gender were
illustrated by the case study of the Arid and Semi-Arid Lands
project in Kenya (001, see Carloni and Horenstein). The
project's strategy for improving production and preserving the
agricultural resource base is to popularize bench terracing and







-17-


water conservation in the semi-arid highlands while it carries
out agronomic research to develop technical solutions suitable
for the arid lowlands. The social soundness analysis for the
Project Paper pointed out that women are the principal farmers
and that, because of male migration, women's self-help groups
would also be the main source of labor for project works such as
construction of terraces and water catchments. It warned that
if women were expected to supply free labor for soil and water
conservation during the peak agricultural seasons, targets would
not be met, and recommended that the project either pay for the
labor or suspend the works during the peak season.

The original project design ignored the recommendation.
Targets were set on the assumption that works would be carried
on throughout the entire year. Ultimately, project management
recognized that the original targets were not feasible and
suspended work during the peak season so that women could finish
their ploughing and planting. Recognition of women's economic
responsibilities and time constraints has been a critical factor
in securing their unpaid labor for soil and water conservation
works. Achievement of project objectives is a direct result of
women's high level of participation. The high level of
participation is a direct result of project adaptation, which in
turn was facilitated by gender analysis.

Project adaptation had an impact on the achievement of
targets for construction of terraces and water catchments. It
also had quantifiable economic benefits: the Kenyan Government
assessed the value of women's unpaid labor contribution to the
project at US$1.8 million, a direct savings to the project.


2.1.2 Mainstream Institutional-Development Projects


One of the main differences between institutional-
development projects and direct-service projects is that the
participants in the former are not farmers; they are government
planners, agricultural research personnel, ministry of
agriculture staff such as veterinarians and extension workers,
or teachers and students in agricultural colleges and training
institutes. The degree of match between the gender of project
participants and the gender division of labor at farm level is
less likely to be important for project efficiency. Under these
circumstances, women's level of participation is mainly a matter
of equity.

The desk review found little relationship between the
gender of participants and achievement of objectives in
agricultural planning, research, education, and extension
projects unless some of their activities reached the farm
level. In two agricultural sector planning projects (04, 95),







-18-


women's participation was high because women constituted a
relatively high percentage of the staff in the planning units of
the ministries of agriculture. The presence of women on the
planning staff had no relevance for the achievement of immediate
objectives. Nor did it appear to increase awareness in ministry
plans of the gender division of labor at farm level.

The case study of the Caribbean Agricultural Extension
project (09, Schminck and Goddard) illustrates how difficult it
can be in an institutional-development project to show whether
gender makes a difference for the achievement of project
objectives. After 7 years of planning and implementation, the
project is achieving its stated purpose of improving national
extension systems, but it is still too early to see the effects
of these improvements on actual farm practices. Interviews with
extension agents suggest that gender-linked farming practices
can and should be addressed in extension work and that certain
goals, such as increased vegetable production, will be affected
if such adaptations are not made. Yet, the institution-building
focus of the project so far allows for little, if any,
assessment of direct impact of improved extension at the farm
level.


2.2 Income Generation and Employment


A.I.D. gives high priority to expanding private sector
employment as a strategy for raising the incomes and improving
the quality of life of low-income people. In urban areas of
developing countries, a large share of the small-scale
enterprises in the informal sector are managed by female
entrepreneurs. In newly industrialized countries, the majority
of workers in labor-intensive, export-oriented industries are
women. Eleven projects in the desk review sample and two more
from the field studies aim at generating employment through
vocational education, credit, or promotion of community-based
income-generating activities (see the desk review by Lycette and
Self, the case study by Lycette [49], and Blumberg's Dominican
Republic case study [003]). Given the small sample and the
heterogeneity of project types, quantitative analysis of
relationships between variables in the conceptual framework was
not attempted. Instead, projects were examined according to the
type of intervention.


2.2.1 Job Training


There were three job training projects in the sample. Only
one project was successful in expanding employment for women.
This was achieved by adapting procedures for recruitment,







-19-


training, and counseling of women entering the Ministry of
Labor's industrial and commercial job training program in
Morocco (49, see Lycette's case study). Two projects were
unsuccessful in expanding women's employment. In Ecuador,
women's level of participation in job training activities of the
Tarqui Cooperative and Community Development project (27) was
low because the type of training offered to women was of little
value in getting a job. In Senegal, where women were expected
to participate in sewing and tailoring courses as part of a
youth job development project (38), no effort was made to adapt
recruitment procedures to ensure that women would actually
participate. As a result, women's level of participation was
lower than anticipated. In the sector paper, Lycette concludes
that it is much easier to train women successfully than it is to
generate employment for them. She suggests that projects that
provide credit so that trainees can set themselves up in or
expand existing businesses are more likely to achieve the
objective of employment creation than projects that only provide
training.


2.2.2 Credit Projects


Women's participation in credit projects was linked to the
number of female entrepreneurs operating in a given subsector,
the up-front cost of borrowing, collateral requirements, and the
minimum size of loans. For example, when loans were delivered
to enterprises with relatively high levels of assets or the
minimum size of loans was high, female entrepreneurs were
usually excluded. In Burkina Faso, where 40 percent of the food
processing loans from Village Development Funds were expected to
go to women's microenterprises, the minimum size of loans was so
high that microentrepreneurs were excluded. In contrast, the
Entente Enterprises project (16) in five Francophone West
African countries made no particular effort to reach women, but
because the terms of lending did not exclude microenterprises,
25 percent of all loans went to women. Projects that adapted
delivery systems and had high levels of female participation
(16, 26) were more successful than those where lack of
adaptation excluded women (28).

In the Dominican Republic (003), a project to provide
short-term working capital to poor street venders and
microentrepreneurs was unusually successful in removing the
barriers that usually prevent the poor from getting credit.
Specifically, red tape was minimal, loans were disbursed in a
matter of days, and almost none of the loans required
collateral. Although the project made no attempt to cater to
women or to seek them out, the adaptations made by the project
to remove obstacles to participation by the poor in general








-20-


resulted in substantial proportions of women becoming loan
beneficiaries. In total, 43 percent of the impoverished street
vendors (organized into "solidarity groups") and 17 percent of
the somewhat better-off microentrepreneurs receiving loans were
women. Moreover, the women proved to be as good or better than
men in their loan performance. Among the microentrepreneurs,
women created more jobs than their male counterparts (an average
of 1.5 new positions versus 1.3 for the men). Since employment
generation was a principal objective of the project, the level
of women's participation can be clearly linked to the
achievement of project objectives.


2.2.3 Income-Generating Projects


The sample included several projects aimed at promoting
income-generating activities for village women. Most were
designed as women-only projects. Only one of the income-
generating projects examined by the study was successful in
achieving its objectives; this was the Latin America Regional
Appropriate Technology for Women project (21, see the case study
by Flora). Under that project, Flora found that subprojects in
Bolivia were more successful than subprojects in Ecuador because
they carefully analyzed women's existing income sources, their
incentives to undertake new activities, and their time
constraints.

On the other hand, projects in Burkina Faso (85, 86), El
Salvador (17), and Ecuador (27) failed to achieve the objective
of raising women's income and expanding employment because the
activities were either inappropriate for the local setting, or
poor women had no time for them, or they were unprofitable in
relation to what women could earn from other sources.

Fortmann's desk review contains a detailed analysis of the
difficulties encountered by the Africa Regional Women in
Development project (05), which served as an umbrella for income-
generating projects in several countries. In Casamance,
Senegal, the income-generating scheme failed because of
inappropriate technology, lack of interest, and lack of a
market. In Kassack Nord, Senegal, handicrafts were promoted
when women wanted palm oil presses and inputs for crops. In
Sierra Leone, a female entrepreneur misappropriated funds
intended to support the group's cloth-making enterprise. On the
whole, the umbrella project's most successful activity was adult
literacy and its least successful activities were income-
generating schemes.

In Burkina Faso, the failure of an income-generating
project for women was attributed to lack of precision at the
design stage about the nature of the activities the project








-21-


would support. TOo much responsibility for the identification
and formulation of income-generating projects was placed on the
women's extension service, which had no expertise in assessing
their economic feasibility and only limited technical
backstopping capability. The evaluation report points out that
"identifying needs at village level is not the same thing as
identifying viable projects."

On the basis of the documentation used for the desk review,
there is no evidence that income-generating projects for women
have actually generated much income. Mainstream credit projects
targeted to microentrepreneurs have been much more successful in
raising women's income and generating employment.


2.3 Education and Training


Cross-national studies show a high correlation between
mothers' education and variables such as fertility, infant
mortality, children's nutritional status, and children's
educational achievements. Female education also correlates with
higher productivity in agriculture and employment. The
relevance of gender for achievement of objectives in education
projects is more direct in basic education projects than for
advanced education and training. Nevertheless, equity in all
human resource development is a desirable goal in and of itself.


2.3.1 Basic Education


A.I.D.'s Blueprint for Development calls for increasing
primary school enrollment to above 90 percent for both boys and
girls and raising adult literacy to above 50 percent for both
men and women. The importance of gender for the achievement of
these targets derives from the low literacy and school
enrollment rates of females relative to males. Women's literacy
rate is only 52 percent of men's in the Indian subcontinent, 57
percent in the Near East, 61 percent in Africa; in the Far East
it is 81 percent and in Latin America it is 94 percent of men's
(Sivard 1985). Primary school enrollment shows a similar
pattern of sex disparities.

Yet few evaluations examined in the desk study sample
considered sex disparities when assessing whether basic
education projects had achieved their purpose. For example,
evaluations rated an education technology project in Colombia
(66) and a primary school reconstruction project in Honduras
(62) as highly successful in achieving their stated purposes,
even though they did nothing to reduce disparities between boys'








-22-


and girls' enrollment. A primary education project in Botswana
(75) was also rated highly successful in achieving its
objectives even though its impact on girls' school enrollment
was extremely modest. In North Yemen, a basic education project
(60) failed to achieve its objectives in spite of an increase in
girls' school enrollment (from under 3 percent to about 12
percent). In light of the Agency's goals, basic education
projects should give more importance to reducing the gender gap
in education and literacy, and evaluation criteria should be
adjusted accordingly.


2.3.2 Technical Education and Training


The implications of gender differences for the success of
technical education and training projects are less easy to
demonstrate.

The case study of the Agricultural College Expansion
project in Botswana (54, see Anderson and McBreen) illustrates
this. Project design documents emphasized that the Botswana
Agricultural College should make special efforts to integrate
women into its certificate- and diploma-level training programs
on the assumption that additional female extension workers would
help improve agricultural production in a situation where 40
percent of the rural households are headed by women. When the
project was designed, there were only six women enrolled in the
agricultural program and only five female extension workers.
Subsequently, female enrollment increased from 20 percent in
1978 to 39 percent in 1984. In 1985, however, the Government
limited the intake of female students to 15 per year (or 10
percent). The stated justification for this change in policy
was that female graduates had proven "difficult to place" as
extension workers in rural areas. In spite of this limitation
on female enrollment, the evaluation team rated the project as
highly successful in achieving its purpose, and did not address
nor analyze the implications for contact with large numbers of
female farmers by predominantly male graduates.


2.4. Energy


The Blueprint for Development states that A.I.D. should
give priority to assisting governments to develop sound energy
policies and to providing technical assistance for research in
fuelwood production and related energy-conversion technologies.
To date, A.I.D.'s energy projects have tended to recognize the
role of women primarily as collectors and consumers of fuel for
domestic use. The desk review by Davenport shows that this view







-23-


is too narrow: gender issues are also significant in the
development and conservation of natural resources, including
afforestation, watershed management, and agro-forestry.

The cases examined in this evaluation represent a small
fraction of A.I.D.'s total energy portfolio because only 10
energy projects in the data base made any reference to "women"
or "gender" in their documentation. In 8 of the projects, women
were expected to participate directly in testing improved
cookstoves (5 cases), as nursery workers (3 cases), as forestry
staff (2 cases), and as laborers on reforestation sites (1
case). In 7 projects, women were also expected to benefit
indirectly from increased supplies of fuelwood.


2.4.1 Cookstove Components


Cookstove components rarely achieved their objectives. The
problem in Burundi (74) was that the stove did not save as much
fuel as anticipated. In Nepal (002, see the case study by
Davenport, Nickell, and Pradhan), the stoves were unsuitable for
many types of cooking, so women continued to use the traditional
hearth alongside the smokeless stove. In Lesotho (72), where
village women were consulted about their preferences, the level
of acceptance was better. In Honduras (64), use of the stoves
did not spread because there were few female extension workers
whom the project could use to disseminate the stoves. The desk
review found that baseline studies on women's actual patterns of
fuel use and preferences (72, 64) led to better design (see
Davenport).


2.4.2 Fuelwood Components


Fuelwood components failed to achieve their objectives for
reasons unrelated to women's participation. In two cases, there
was a problem of drought (71, 73). Also, in the Sahel Regional
Program (71) the wood from project woodlots was too expensive to
compete with alternative sources. And in Senegal (73), where
women were expected to benefit from the creation of nursery
jobs, men got the jobs.


2.4.3 Biogas


Two of the problems encountered with a biogas plant in
Nepal--inadequate supply of manure for the digestor and limited
demand for milling services--are gender-linked. Although women







-24-


provide 90 percent of the labor for feeding manure into the
digestor and are responsible for grinding grains, they have no
voice in the plant's management, cannot afford the plant's
costly milling fees, and give manure to the biogas plant only
very reluctantly because they need it as a source of cooking
fuel (002, see Davenport, Nickell, and Pradhan).


2.4.4 Reforestation


The case study of the Resource Conservation and Utilization
project in Nepal (002, cited above) indicated that gender-linked
labor constraints were probably a factor in the failure to
achieve forestry targets. Although women are the preferred
laborers for nurseries and tree planting, in Gorkha District the
project could only get children to work because the timing of
planting conflicted with women's farm planting.


2.5 Water Supply and Sanitation


In developing countries throughout the world, women and
girls have the primary responsibility for collecting and
utilizing water for domestic purposes. They also have major
responsibilities for family hygiene, household sanitation, and
waste disposal. The desk review sample covered 18 projects in
three regions (Latin America, Africa, and the Near East). Some
projects were rural; others were in urban slums. In most cases,
the project purpose was to design and construct physical,
infrastructure such as village water supply systems, waste
disposal systems, and latrines. In a few cases, the project
purpose was institutional development. In addition to the 17
projects covered by the desk review, 2 of the 10 case studies
examined domestic water supply components in larger rural
development projects (001, 002).

In the design documents for the 18 projects covered by the
desk review, women tended to be viewed as passive beneficiaries
of improvements in water supply and sanitation rather than as
active participants. Although 15 of the 18 water supply
projects included health and hygiene components, these were
usually targeted to "families" or "people," without regard for
women's leading role in these activities.

Likewise, 15 of the 18 projects called for community
participation in self-help construction and maintenance
activities,- but only a few sought to actively involve the end
users--women--in these activities (32, 33, 41). Only one
project (33) made a special effort to increase women's








-25-


participation in, and benefit from, the project by providing
extension workers to work with women on such complementary
activities as health, hygiene, sanitation, and vegetable
gardening (using waste water from the pumps).

In spite of the fact that many projects defined their
purpose in such a way that it could be achieved without direct
contact with the end users, a strong positive correlation was
found between women's level of participation and achievement of
objectives. In fact, all of the projects in which women's
participation was known to be high were highly successful in
achieving their immediate objectives (34, 39, 41, 47). The
project in which women's participation was known to be low
failed to achieve its objectives (37). In this latter case,
project failure could clearly be traced to failure to consult
water users about their preferences.


2.5.1 Design of Water Supply Systems


Findings from the desk review suggest that designs based on
baseline studies of women's actual patterns of water use (34,
41) were more successful than designs based strictly on
technical criteria (37). The same finding applies to the choice
of technology: in two Central American countries, a particular
hand pump's high acceptability to women and its suitability for
use by young girls was a factor in project success (36).


2.5.2 Construction and Maintenance of Water Supply Systems


The desk review found a definite relationship between
women's time constraints and levels of participation in
self-help and water-user groups. Projects that failed to
consider the opportunity cost of women's time were less
successful in mobilizing their labor (see Nieves).

Maintenance of water sources was better when women were
involved in maintenance committees, but there is evidence that
user involvement, by itself, is not enough; it needs to be
complemented by competent technical support if proper
maintenance is to be sustained over time.

The case study of the Arid and Semi-Arid Lands project in
Kenya (001) presents some intriguing findings regarding women's
role in the construction of dams and catchments for village
water supply. Initially, all of the projects carried out by the
Ministry of Water Development (MOWD) used hired male labor,
while water projects carried out by the Ministry of Agriculture








-26-


and Livestock Development (MALD) used women's self-help labor.
Due to labor problems, MOWD was never able to meet its targets.
Comparison revealed that MALD's participatory system was cheaper
and more efficient. As a result, the project is now using
women's self-help groups for all of its water development and
conservation works. Women's self-help labor for the
construction and maintenance of water sources has greatly
reduced Government costs and improved sustainability (see
Carloni and Horenstein).


2.6 Regional Variations


Regional variation in the gender division of labor and
resources is greatest in the agriculture sector. Women's role
in activities related to domestic water and fuel supply is more
constant across regions. In the informal sector of the urban
economy, female entrepreneurs play an important role in all four
geographic regions, in spite of considerable variation among
countries. In the education sector, the gender gap is greatest
in the Indian subcontinent, the Near East, and Africa.

These differences in the baseline situation are only
partially reflected in the projects examined in this
evaluation. Although the sample cannot be considered as
representative of the various regions, some trends are worth
noting. Projects in Africa gave more attention to gender
differences than projects in any other region. Projects in Asia
gave the least attention to gender. Projects in Latin America
and the Near East were somewhere in between.

In the projects in Africa, gender differences were analyzed
mainly in agriculture and energy projects. A few education
projects analyzed gender differences. Water supply projects
gave more attention to gender roles than did such projects in
other regions. Employment projects paid relatively little
attention to gender differences.

As a result of this analysis, resources in agriculture and
education projects were targeted to women through women-only
projects and women's components; in energy projects, resources
were targeted through cookstove components. Employment projects
did not target resources to women. Gender analysis was rarely
followed by adaptation of institutions and delivery systems.
For this reason, access to delivery systems emerged as a major
bottleneck restricting the flow of resources to women.

In the projects in Asia, gender differences were analyzed
in energy projects (63, 68, 69, 70, 002) and a few of the
agricultural projects (92, 94). The energy projects targeted








-27-


cookstove components to women. One of the employment projects
targeted training (98). None of the agricultural projects
targeted resources specifically to women. The assumption seemed
to be that technical advice delivered to the husband would
"trickle over" to the wife. The presence of substantial numbers
of female extension workers may have led project designers and
implementers to overlook constraints affecting outreach to
female farmers. As in Africa, institutional constraints were a
major bottleneck.

In the Near East projects, several projects analyzed gender
differences but only two projects targeted resources to women,
one through a women-only project and one through a women's
component (both in Morocco, 48, 49). Lack of analysis and
adaptation of delivery systems restricted women's actual
participation in, and receipt of resources from, mainstream
projects.

In the projects in Latin America, mainstream projects
analyzed gender differences primarily in the agriculture sector
(06, 09, 13, 22). In three sectors (agriculture, employment,
and education) limited resources were targeted to women through
women-only projects (20, 26, 57, 58) and women's components (27,
09). Rural development projects targeted few resources to
women, and when they did, the emphasis was exclusively on
women's domestic roles. As in other regions, women's
participation and benefit were restricted by inadequate analysis
of institutional and labor constraints and consequent lack of
adaptation of technical packages and delivery systems.


2.7 Conclusions


Evidence from four sectors (agriculture, employment,
energy, and water supply) supports the hypothesis that projects
are more likely to achieve their objectives when the gender of
project participants reflects the baseline division of labor in
project-related activities. The finding holds for
direct-service projects that seek to introduce change at
beneficiary level. It applies to institutional-development
projects only insofar as these projects reach the grassroots
level.

In the same four sectors (agriculture, employment, energy,
and water supply), baseline studies on the division of labor led
to better design. In these sectors, the efficiency of direct-
service projects would clearly be enhanced by institutionalizing
three activities: (1) analysis of gender roles and their
implications for the project strategy, (2) project adaptation in







-28-


the light of analysis of gender roles, and (3) regular
monitoring to ensure that the gender of project participants
matches gender roles in the baseline situation.

The education sector is somewhat anomalous in that the
gender division of labor is largely irrelevant for project
efficiency. Equity--in this case narrowing the gender gap in
basic education and literacy--is the main policy concern. This
concern could be furthered by placing greater emphasis on
identifying institutional and other barriers to women's
participation in education and training activities, removing
these barriers through project adaptation, and monitoring the
gender of participants on a continuous basis.

Findings from all five sectors show that analysis of gender
differences has little effect on the achievement of project
objectives unless women actually participate in and benefit from
the project. Adaptation of institutions, delivery systems, and
technical packages in the light of gender differences appears to
be the key to higher participation and greater benefit.
(Section 4 of this paper examines the process of project
adaptation in greater detail.)


3. THE IMPORTANCE OF GENDER FOR THE
ACHIEVEMENT OF SOCIOECONOMIC DEVELOPMENT GOALS


In each of the five sectors examined in this evaluation,
projects tend to share a common set of goals. The ultimate goal
of most agricultural projects is to improve the social and
economic well-being of small farm households by increasing
production, raising incomes, and (sometimes) reducing
malnutrition. The goal of most employment projects is to create
jobs and raise incomes of poor households. The goal of most
energy projects is to improve well-being by developing new
energy sources and conserving natural resources. The goal of
most water supply and sanitation projects is to improve the
quality of life. The goal of most education projects is to
improve well-being directly, through human capital formation,
and indirectly, by imparting skills relevant to the achievement
of other goals.

This section of the paper takes five goals as its starting
point: increased agricultural production, development and
conservation of natural resources, employment generation, higher
incomes for low-income people, and reduction of malnutrition.
Using evidence from the desk reviews and especially the 10 case
studies, it examines which aspects of gender affect the
achievement of each goal.







-29-


3.1 Increasing Agricultural Production


The importance of gender for agricultural production
derives from the gender-linked division of labor in small farm
households. According to a recent study covering 82 developing
countries, women are 46 percent of the agricultural labor force
in Sub-Saharan Africa, 45 percent in Asia, 40 percent in the
Caribbean, and 31 percent in North Africa and the Near East
(Dixon 1982). The production tasks performed by women and their
share of the total farm work vary widely between countries and
among different social, economic, ethnic, and regional groups
within countries. Women's agricultural role is also affected by
age, marital status, the husband's presence or absence, and the
stage in the household development cycle.

Section 2 showed that in agriculture projects the degree of
match between the gender of project participants and gender
roles in the baseline situation affects the achievement of
immediate project purposes. This section goes a step farther.
It takes four aspects of gender identified by the Women in
Development Policy Paper and examines evidence (primarily from
the case studies) of their effect on the achievement of such
goals as the intensification and diversification of small farm
production and the transfer of technology to low-income
producers. The four aspects are as follows:

-- Access to and control over productive resources
-- Gender-linked labor constraints
Control of crops and income from their sale
Stakes and incentives


3.1.1 Gender Access to Resources as a Factor in Production


Access to Extension Services. Five of the case studies
examined women's access to extension advice and the implications
for agricultural production (Botswana, the Caribbean, Thailand,
Kenya, and Nepal). In all five cases there was evidence that
failure to reach female farmers can have a negative impact on
efforts to increase production. In Kenya, there was a dramatic
increase in outreach to female farmers when extension workers
began contacting local self-help groups instead of individual
contact farmers, which could have equally dramatic effects on
the spread of innovations (see Carloni and Horenstein).

Another finding that emerged from the case studies is that
in this group of countries at least, female extension workers do
not appear to be the major factor in reaching female farmers.
Some extension workers in all five countries were women, but







-30-


their contact with female farmers was not necessarily better
than that of male extension workers. On the contrary, both male
and female extension workers tended to focus on male farmers.
One reason for the lack of contact with female farmers was the
prevailing emphasis on commercial farming and cash crops.
Extension workers had few incentives to spend time with
subsistence farmers. In Botswana, extension workers of both
sexes tended to focus on men because village women's family
responsibilities made it more difficult for them to travel to
extension centers. Direct contact by agents of either sex with
village women was often restricted by inadequate transport (see
Anderson and McBreen).

Access to Land. In Sri Lanka, where outmigration of
daughters from the settlement schemes threatens to undermine the
long-range viability of the Mahaweli program, outmigration seems
to be connected with the fact that daughters are not allowed to
inherit land (see Benson and Emmert). Differences in the status
of women among different categories of settlers could also be
traced to the fact that among local people who were temporarily
moved off the land to make room for the project (referred to as
"former residents" and "evacuees"), women could get irrigated
land allotments in their own name while among new settlers only
men were given allotments.

All Mahaweli scheme households are expected to take out
seasonal production loans, but among new settlers, widows and
wives of migrants are unable to borrow because the land is in
the husband's name. Among former residents and evacuees, the
share of loans disbursed to women was roughly proportionate to
the percentage of women owning land. When land title is a
prerequisite for access to credit and many female farmers are
unable to qualify for loans, allocation of resources is
inefficient and production suffers.


3.1.2 Gender-Linked Labor Constraints as a Factor in
Agricultural Production


Male Migration. As a result of the sex-typing of
agricultural tasks, shortages of male labor for land preparation
can cause serious bottlenecks in production. In Thailand, a
green manure crop intended to fertilize rice did not get planted
because men were not there to prepare the land during the dry
season. In Botswana and Kenya, male migration caused
bottlenecks in land preparation that led to late planting and
lower yields. Migration of young men from the Eastern Caribbean
affected the supply of labor for such heavy work as terracing
and planting tree crops.







-31-


Aggravation of Women's Work Load. The field studies
provide many examples of situations in which the burden of
intensified cropping systems or other new practices fell
primarily on women. In Thailand, this led families to drop out
of on-farm trials (see Blanc-Szanton, Viveros-Long, and
Suphanchainat). In the Caribbean, where the burden of field
packing bananas fell on women, it led them to pressure their
husbands to get out of banana production (see Schminck and
Goddard). In Guatemala, where contract vegetable growing
increased the demands on women's time, they had less time to
earn an independent income from marketing (see Blumberg). This
in turn had implications for family well-being.

Conflict Between Project Activities and Women's Role in
Farming. In Kenya, conflict between the timing of project soil
conservation activities and women's work on the farm slowed
progress until management recognized the need to adjust work
schedules so that they would not interfere with farming (see
Carloni and Horenstein). In Nepal, conflict between the timing
of project tree-planting activities and women's farm planting
may have contributed to the forestry component's failure to
achieve more than 35 percent of its target (see Davenport,
Nickell, and Pradhan). In Ecuador, conflict between the timing
of project citrus-processing activities and the peak season for
female wage employment in the citrus harvest led to a low rate
of participation in one project enterprise (see Flora).

Women's Domestic Responsibilities. In several cases, the
projects' impact on production was lowered because women's
domestic responsibilities interfered with their role in
agriculture. On the Mahaweli irrigation/settlement schemes in
Sri Lanka, lack of extended family networks for the care of
small children has hampered wives' availability for work on the
irrigated allotments (see Benson and Emmert). In Kitui
District, Kenya, one reason women cannot plough and plant at the
onset of the rains is that half of their productive time is
taken up fetching water from distant sources (see Carloni and
Horenstein).

The desk review of water supply and sanitation projects
found that reduction of time spent carrying water did not
increase time spent on production unless income-earning
opportunities already existed or were introduced (see Nieves).
To have a positive impact on production, water supply projects
must be linked to income-earning opportunities in agriculture or
nonfarm production. Only one of the water supply projects in
the desk review sample included an agricultural component (33).







-32-


3.1.3 Gender Differences in the Control of Product as a Factor
in Agricultural Diversification


In the Eastern Caribbean, where many countries give
diversification high priority as a way of reducing dependency on
one or two export crops, extension agents say that women are
more responsive to diversification possibilities than men. The
reason for this difference may be that men control the large
monetary aspects of the farm (including the main cash crops)
while women control the minor crops. Men have a stake in
perpetuating the emphasis on the traditional export crops that
they control, whereas women have a stake in expanding production
of minor crops such as vegetables, which are part of the
diversification strategy (see Schminck and Goddard).

In Sri Lanka, one of the main problems faced by the
Mahaweli irrigation and settlement program has been lack of
diversification of the farming system. In most places there is
not enough water to grow two irrigated crops per year. Reliance
on a single rice crop has created seasonal cash shortages and
malnutrition. Recent efforts to diversify food and income
sources focus on intensifying production of fruit, vegetables,
and spices on the homestead plot and integrating them with
animal production. Because the homestead garden and livestock
care are the wife's domain, recognition of gender differences in
the control of different crops and activities could have
important consequences for diversification (see Benson and
Emmert).


3.1.4 Gender Differences in Incentives as a Factor in Production


Women's incentives to provide unpaid family labor for crop
and livestock production are closely linked to the control of
crops. When women control the disposal of the product and the
earnings from its sale, their incentive to intensify labor
inputs is far greater than it is for crops whose disposal is
controlled by their husbands. This factor is more important for
cash crops and sidelines that men and women treat as a source of
personal income than for staples that all family members
consume.

In Guatemala (004), a 1980 comparison of three poor Indian
villages growing broccoli, cauliflower, and snow peas on
contract to an agribusiness firm revealed a significant
pattern. The level of yields and the production of
first-quality produce appeared to be related to the extent to
which women participated in--and benefited from--the project.
In the first village, women did not help in the fields, and the








-33-


project did poorly for lack of sufficient labor. In the second
village, women had been pulled into field work on these very
labor-intensive vegetables. But this reduced their time
available to earn independently controlled incomes as market
venders. Moreover, payment came in the form of a check made out
exclusively to the husband, a check that had to be cashed in a
town some distance from the village. Project performance was
better than in the first village, but not outstanding. In the
third village (where the growers were organized into a
cooperative), however, women not only worked in the fields, they
benefited directly. They were relatively full partners in the
contract farming, and the co-op paid both husband and wife
individually in cash. In this village, yields and quality of
produce were by far the highest. Interestingly, village women
fared best working in the agribusiness firm's processing plant,
where the vegetables were frozen and packed for export. They
worked long shifts at the government minimum wage and earned
unusually high incomes over which they retained full control.
Both job satisfaction and productivity were extremely high.
Thus it appears that the level and quality of production were
related not only to the extent to which women participated in
project activities but also to the extent to which they were
given incentives and allowed direct control over remuneration.


3.1.5 Gender as a Factor in the Development and Transfer of
Technology to the Rural Poor


The experience of the Northeast Rainfed Agricultural
Development project in Thailand (94) illustrates why adapting to
the gender of project participants is important for technology
transfer to the poor. About 10 percent of the households in the
project area were permanently headed by women, and in addition,
men left the district during the slack season in search of wage
employment. Until now the project has concentrated on
"specialist" farmers who are selected to carry out crop trials.
These farmers have all been men, and what is more, they have
more land, labor, and capital than other farmers. By
concentrating project resources on the better-off specialist
farmers, the project ran a serious risk of developing modified
cropping systems that are beyond the means of the target group.
Since so many poor households depend on women for farm
management, efforts to involve poor women would have been useful
as a source of feedback about their reaction to project packages
and their constraints to adopting them (see Blanc-Szanton,
Viveros-Long, and Suphanchainat).

The experience of the Arid and Semi-Arid Lands project in
Kitui District, Kenya (001) illustrates how targeting extension
to women can dramatically improve technology transfer to the







-34-


poor. In an area where up to 60 percent of the adult males were
working outside the district, the households selected as contact
farms when the Training and Visit System was first introduced
were atypical in two ways: they had more land and assets, and
the husbands were full-time farmers. A farm survey conducted by
the project indicated that the resources concentrated on contact
farmers were not paying off because innovations were not
spreading to the average farmer. Consequently, indigenous self-
help groups (80 percent female) were substituted on a pilot
basis.

Instead of contacting individual farmers, extension workers
made direct contact with groups as large as 30 members. The
pilot experience was so successful that the practice is being
extended to the rest of the district. It has greatly improved
the extension system's contact with women and the poor and is
laying the groundwork for widespread technology transfer. The
project will use these self-help groups to test and disseminate
a wide range of technical innovations including
drought-resistant varieties, water-saving tillage methods,
disease-free cassava cuttings, farm implements, tree seedlings,
and improved beehives (see Carloni and Horenstein).


3.2 Developing and Conserving Natural Resources


The direct connection between women's stakes and incentives
and conservation of natural resources is illustrated by the Arid
and Semi-Arid Lands project in Kenya (001). Initially, project
planners feared that women's self-help (mwethya) groups might
not supply sufficient labor for construction of terraces, dams,
and catchments without cash compensation. When mounting
financial pressures forced the Government to cancel plans for
payment to mwethya groups, they discovered, to their surprise,
that lack of payment was not a constraint. Village women were
willing to provide unpaid labor for soil conservation works
because terracing allowed them to make better use of scarce run-
off for crop production; they were willing to provide labor for
construction of water points because they would benefit directly
from not having to haul water long distances (see Carloni and
Horenstein).

The case study of the Resource Conservation and Utilization
Project in Nepal (002) illustrates how a project's impact on
resource conservation might have been greater if the women's
component had not focused so narrowly on women's domestic roles
cookstovess, kitchen gardens, and sewing). The stove component
appears to have diverted attention from involving women in the
project's broader resource conservation activities, including
afforestation, watershed management, and soil conservation. The








-35-


impact of the women's component's on resource conservation has
been less than anticipated because the new stoves are not as
versatile as the traditional hearth; women who have them use
them for only a part of their cooking.


3.3 Generating Employment


Many projects in this sector achieved their immediate
purpose but stopped short of achieving the ultimate goal of
expanding employment. Of the three types, credit projects were
the most successful in expanding female employment. Most female
job training projects and community based income-generating
projects encountered difficulties. Projects that provided
skills training for self-employment without providing seed
capital so that trainees could set themselves up in business
also failed to generate female employment.

In Morocco, a project designed to integrate women in the
Ministry of Labor's commercial and industrial job training
program (49) was one of the few of its type that was successful
in achieving its goals. Not only did it expand female
employment at favorable wage levels, but the entry of female
graduates into the sector also eased bottlenecks at a time when
the supply of skilled labor was a constraint for industrial
expansion. Female trainees performed well in courses, and the
implementing agency had no difficulty placing graduates in
private sector jobs. In fact, employers were so pleased with
the performance of female graduates that they expressed a
preference for women in the future. And finally, as a result of
the project, women's participation in the Ministry of Labor's
national system of vocational schools has been institutionalized
(see Lycette).

In the Dominican Republic microenterprise project (003),
Blumberg found, after disaggregating project statistics by sex,
that the women microentrepreneurs created more jobs than the
men. The contrast was most dramatic in the large and important
clothing sector (one-half of all female microentrepreneur loan
clients were in the clothing sector, and one-half of the
businesses in the sector were female owned). Women's clothing
enterprises have created an average of 1.4 new jobs each,
compared with 0.64 for male-owned clothing businesses.
Disaggregating data for the clothing sector by sex also showed
that the women's enterprises were growing faster than the men's
on five of six standard business indicators (e.g., sales,
profits) tracked by the project. Thus, achievement of project
goals for stimulating jobs and economic activity in the informal
sector was clearly linked to the level of female participation.







-36-


3.4 Raising Incomes of Low-Income People


Among poor households in both rural and urban areas,
diversification of income sources is crucial for survival.
Family members engage in a wide range of income-earning
activities in different seasons to spread risks. The share of
total income earned by wives and daughters can be major.

Case studies revealed that where male migration is
widespread, as in Botswana (54), Kenya (001), and the Caribbean
(09), women's earnings are the mainstay for everyday subsistence
expenditures, especially food. Growing dependence on women's
earnings was also cited in urban Morocco (49) and the Dominican
Republic (003). In one community in Bolivia, women's earnings
from handicraft production were crucial for tiding their
families over during a drought year when no earnings were coming
in from agriculture (21, see Flora). In Sri Lanka's wet zone,
women's sale of eggs and milk, as well as fruit, vegetables, and
spices from the homestead garden, provides a steady trickle of
cash year-round. On the Mahaweli settlement schemes, where
serious difficulties with the seasonal cash flow threaten to
undermine development, expansion of women's homestead garden
production may provide solutions (91, see Benson and Emmert).

The proportion of their income that different earners pool
with the rest of the family or keep for their personal use
varies widely among socioeconomic groups and household types
within countries. In Bolivia, Ecuador, Guatemala, the Eastern
Caribbean, Kenya, Sri Lanka, and Thailand, the case studies
confirmed that husbands treat earnings from certain crops or
activities as their personal spending money. If the earnings
from the sale of a cash crop are customarily treated as the
husband's personal income, a project that raises production may
well have the effect of merely increasing the husband's
nonessential consumption without improving family well-being.

In Thailand, Kenya, and the Eastern Caribbean, women
generally manage whatever money is pooled for everyday family
expenses. In Ecuador, where husbands usually manage household
cash, wives have difficulty keeping control over how their own
earnings are spent. Women in one village preferred to disburse
profits from the group enterprise once a year rather than
monthly, and in kind (food, household utensils). They stated
they were afraid that if they received cash, their husbands
would take the money and spend it on ritual drinking (see
Flora). (This suggests one reason why the profitability of
women's income-generating projects is hard to document: women
may feel the need to conceal their earnings from their husbands.)








-37-


As noted above, case studies in Kenya, Botswana, and the
Caribbean found that it was usually the wife's earnings rather
than the husband's that were used to buy food for the family.
The baseline study in St. Lucia found that women were solely
responsible for paying for the family food in 37 percent of the
households (food represents 40 percent of total expenditures)
(see Knudson and Yates). Implications for nutrition are
discussed below.


3.5 Reducing Malnutrition


A growing body of evidence suggests that there is no
necessary connection between increases in aggregate household
income and improvements in children's nutritional status. There
is increasing evidence that it is the mother's income-earning
activities that show the higher correlation with children's
nutrition rather than the father's (Kumar 1977; Carloni 1983).

Women's importance for nutrition extends far beyond their
traditional roles in food preparation and infant feeding. They
influence household food availability directly through their
role in food production, processing, storage, and marketing and
indirectly through their activities as income earners. In some
countries, women bear the entire responsibility for providing
the family food. In other countries, husbands are expected to
provide certain foods (such as staple grains) for family
consumption, while wives provide complementary foods. A
thorough understanding of gender roles in food production and
consumption is needed when the goal is to improve nutrition.

The case study of the Mahaweli irrigation and settlement
program in Sri Lanka (91) illustrates the connection between
women's role in food systems and efforts to reduce malnutrition.
The Mahaweli authorities attribute malnutrition to lack of
differentiation in the diet caused by lack of diversification in
the farming system. Before the Mahaweli Authority began to
develop the area, the traditional diet was much more varied.
Tank-irrigated paddy cultivation on the lowlands was
complemented by shifting (chena) cultivation in upland areas,
where finger millet, grains, pulses, and vegetables were grown.
Women assisted men with weeding and.harvesting paddy, and men
assisted women with clearing and burning the upland area in
preparation for planting, but paddy production was regarded as
men's responsibility and chena cultivation as women's.
Curtailment of chena cultivation (which the Government sees as a
cause of environmental degradation) has unbalanced the food
system. A possible solution would be to intensify production of
complementary foods on the homestead plot and to introduce
livestock into the farming system. This would provide a steady







-38-


supply of food throughout the year, diversifying the diet and
counteracting the effects of the seasonality of paddy
production. Both of these activities are traditionally women's
domain, and efforts to improve nutrition would entail supporting
women's productive activities (see Benson and Emmert).

The case study of an agribusiness project in Guatemala
(004) found evidence that women's loss of decision-making power
within the family can adversely affect nutrition. Comparison of
three contract grower villages revealed that women's control of
an independent income has a great impact on their voice in
decisions regarding food and nutrition. When the introduction
of the labor-intensive vegetables reduced the time available for
women to sell in the market, they lost their voice in decisions
regarding which crops to grow. Men over-invested in the
agribusiness company's crops to the detriment of other crops.
When a crisis of over-production hit in 1980 and contract buying
was suspended, families had nothing to fall back on and little
money to buy food. In another village, where a recent (1984)
change in the mode of payment undermined women's access to and
control over income from contract buying, women lost much of
their voice in deciding how much of which crops the family
should grow. In addition, they lost much of their voice over
how the money was spent. Aggregate incomes rose sharply, but
co-op staff suggested that nutrition was negatively affected
when expenditure shifted in favor of male-determined decisions
(see Blumberg).


3.6 Conclusions


The lack of relationship between the achievement of
immediate project purposes and long-term goals continues to be a
sensitive point for development assistance. Many projects that
are successful in their own terms have a limited socioeconomic
impact. Gender analysis is a tool that can shed light on the
causes of this problem.

The case study of the Caribbean Regional Agricultural
Extension project (09) provides a good illustration. Although
the project was rated as highly successful in achieving its
immediate purpose of building national extension systems, it is
not clear whether the project will achieve one of its ultimate
goals, the diversification of small farm production. Even a
rigorous preproject analysis of women's productive roles in
island agriculture, and a women's component emphasizing
awareness training in gender issues, did not translate into
field-level adaptations in extension. Implicit in the strategy
to achieve import substitution through diversifying and
intensifying vegetable crop production is the need to engage








-39-


female farmers in the effort. Yet to date, the extension
workers are not being trained in techniques or equipped with
incentives for women farmers as a particular target group.
While extension agents, ministry, and project staff all
recognize the key role women will play in achieving
diversification targets, the next steps (adapting extension
agent work plans and technical packages, developing
communication networks, scheduling and locating demonstrations
for gender-specific audiences) have yet to be taken. The case
study suggests that without these specific gender-responsive
adaptations, the improved extension systems may fall short of
achieving the national agricultural development goals the newly
improved extension systems are pursuing (see Schminck and
Goddard).

The case study of the Agricultural College Expansion
project in Botswana (54) provides another illustration. The
project was rated as highly successful in achieving its
immediate purpose of developing an agricultural training
institution suited to the rural sector's needs. However, it may
not achieve its ultimate goal of improving the social and
economic well-being of the vast majority of farm families if the
extension service continues to bypass female farmers whose role
in agriculture is major (see Anderson and McBreen).

These two examples have a common thread. Both projects
achieved their immediate purpose, but the lack of match between
gender roles in the baseline situation and the ultimate
recipients of extension services provided by the improved
institutions could prevent the projects from achieving their
goals. In both cases, gender analysis provides important
insight, but gender adaptations are crucial to success.

Another set of examples illustrates a different lesson. A
project can be successful in achieving its immediate objectives
of raising productivity and expanding employment; it can also be
successful in raising beneficiary incomes; but it can still fail
to achieve the ultimate goal of improving the social and
economic well-being of low-income households if gender variables
are overlooked. The case study of the ALCOSA Agribusiness
Employment/Investment Promotion project in Guatemala (004) shows
that increases in one family member's income do not necessarily
result in better nutrition or greater well-being for the rest of
the family. When women's control of earnings dropped and their
voice in household decisions and expenditures fell, there was
evidence that nutrition deteriorated even in the face of overall
abundance (see Blumberg).

In short, the relationship between achievement of immediate
project purposes and long-term socioeconomic goals is neither
linear nor automatic. It can never be assumed that








-40-


institutional development or delivery of services will
automatically increase production or generate employment, or
that increased production or higher employment automatically
means higher incomes for low-income people, or that increases in
one family member's income automatically are translated into
better nutrition and improvements in the whole family's well-
being. Gender variables intervene at every step in the chain in
critical ways:

Planning for gender factors in production--including
access to and control of resources, labor constraints,
and incentives--is particularly useful for development
and successful transfer of technology.

-Planning for differential male and female income
sources is particularly useful for efforts to raise the
level of living of the poor.

Planning for gender roles in consumption, including
whose income is used to buy food, can be crucial for
ensuring that higher incomes result in better nutrition
and family well-being.

The evaluation findings support the assertion in the Women
in Development Policy Paper that "misunderstanding of gender
differences, leading to inadequate planning and designing of
projects, results in diminished returns on investment." The
evidence shows that resources invested in improving gender
analysis, project adaptation, and monitoring of female
participation can not only help projects to achieve their
immediate purposes more efficiently, but more important, it can
help the Agency to ensure that resources invested in projects
contribute to the achievement of broader development goals.


4. PRACTICAL SUGGESTIONS FOR PROJECT
DESIGN AND IMPLEMENTATION


The evaluation shows that understanding gender differences
has a visible payoff in better projects. The challenge now is
to translate that understanding into action. This section gives
practical suggestions on how to carry out gender analysis and
how to adapt projects in the light of gender differences.


4.1 Gender Analysis: Timing in the Project Cycle


What we have come to understand as gender analysis is an
activity that should extend throughout the life of the project,
from design to implementation and evaluation. At the design








-41-


stage, gender analysis should be done by the person responsible
for the project's economic and social analysis. This person
should be a member of an interdisciplinary project design team.
During project implementation, gender analysis can be handled by
project management. At the evaluation stage, a social
scientist's expertise is desirable. There is no need for the
gender analyst to be a women.

At the design stage, gender analysis ought to be an
integral part of both the social and the economic analyses of
projects.

-Economic analyses should specify who (men, women)
control the main factors of production (land, labor,
and capital); consider labor requirements separately
for male and female tasks; specify the differential
off-farm earning capacity of men and women; and
calculate whether direct returns to labor provide
adequate incentives for increased production. The
composition of household income by source and earners
should also be examined.

Social analyses should examine gender roles in the
baseline situation and how these intersect with project
goals and activities; identify target groups for each
activity in the light of the division of labor and
gender roles; examine the outreach of existing
institutions and delivery systems to the target group;
assess the appropriateness of proposed technical
solutions to the needs and resource base of each
category of participant; and specify how to adapt
delivery systems and technical packages in the light of
gender differences.

During project implementation, internal reporting systems
should provide feedback on the relative proportion of project
resources that are going to men and women and to various
socioeconomic groups. Project management needs this feedback to
make certain that the project is reaching the right people. The
definition of what mix of men and women is "right" depends on
existing gender roles.

Reporting forms should provide separate totals for male
and female participants in project-organized activities
and recipients of inputs such as equipment and loans.

Periodic meetings should be held with beneficiaries to
sound out their reactions to project interventions;
separate meetings with women may be necessary to
encourage them to voice their point of view.







-42-


Gender analysis should also be an integral part of
evaluation activities. Numerical indicators of outputs and
achievements (e.g., number of trainees) should be complemented
by information on participants' gender and socioeconomic
status. Cases of unsuccessful technology transfer should be
examined to clarify whether gender differences in access to
delivery systems, labor constraints, or incentives stemming from
control of crops could explain why innovations have not spread.


4.2 Data Requirements and Costs


To design, implement, and evaluate projects, two things are
needed: (1) at least a minimum of information on gender
differences in labor allocation and access to and control of
resources and (2) the analytic ability to relate this
information to project design and implementation. The
information can be obtained by low-cost, rapid techniques;
formal surveys are usually unnecessary. Cost should not be an
obstacle because findings show that resources spent on
understanding the baseline situation have a direct payoff in
project efficiency.


4.3 The Process of Gender Analysis: Ten Steps2


Step 1: Clarify gender roles and their implications for
project strategies. The starting point should be to clarify the
project strategy. For example, what does the project propose to
do to improve agriculture? What activities will be affected by
project interventions? What is the existing division of labor
in these activities? How do these activities fit in with the
total pattern of household productive and domestic activities?
What innovations are being proposed? What are their behavioral
implications for different household members?

Step 2: Analyze eligibility to receive project inputs.
Start by examining what inputs the project intends to provide,
and identify which household member should receive them, in
light of the existing division of labor. For example, if
livestock is women's responsibility and grain is men's, inputs
for livestock should go to women and inputs for grain to men.
If women are responsible for an activity slated for project



2Gender analysis needs to be undertaken for projects in every
sector. The 10 steps used here illustrate the process for
agriculture projects.







-43-


intervention, can they qualify to receive inputs in their own
name? What are the prerequisites for eligibility, and how many
households in the target group can meet those criteria?

Step 3: Define prerequisites for participation in project
activities. In the light of the division of labor, which
household member should participate in activities such as soil
conservation, water user groups, training, and extension? Even
if there is no formal discrimination against women, how will the
location and timing of activities affect their participation?
Does the proportion of women in the pool of eligible
participants match the division of labor?

Step 4: Examine outreach capabilities of institutions and
delivery systems. If analysis of the division of labor shows
that an activity slated for project intervention is women's
responsibility among smallholders, to what extent do existing
institutions and delivery systems have direct contact with
female smallholders, or with any women (e.g., the well-to-do).

Step 5: Assess the appropriateness of proposed technical
packages. Are the technical packages applicable to all
households or only to those with certain types of resources,
such as irrigated land, several head of cattle, or a labor
surplus? How many households in the target group have the right
kind of land? How many, given the sex-typing of tasks and male
migration, can meet the additional labor requirements? How many
can raise the necessary cash? What implications do gender
differences have for the spread of technical innovations to poor
households?

Step 6: Examine the distribution of benefits and its
effect on incentives. Given the gender division of labor and
the control of income from different crops by men and women,
what interest would women have in intensifying production? Do
the direct returns to women outweigh the additional effort? If
the project affects marketing, are women likely to lose an
independent source of income?

Step 7: Consider the reliability of feedback mechanisms.
If women play a major role in project-related activities such as
farming vegetables, how will the project find out whether the
proposed technical innovations are acceptable to them? What
provisions are made for local women and men's participation in
selecting and testing technologies; in evaluating results? Do
monitoring and reporting systems distinguish male and female
participants?

Step 8: Anticipate likely changes in the roles and status
of women. How will the project affect women's access to and
control over land, labor, capital, and expertise? Will women's







-44-


workload increase or decrease? What will happen to their
independent income, to their control of crops and the income
from their sale, and to their voice in household decision-making
on expenditures and other issues?

Step 9: Link changes in the roles and status of women with
the expected project impact. How will changes in women's access
and control of land and productive resources affect food
availability? How will changes in women's ability to earn an
independent income affect household cash flow? How will it
affect their ability to provide for their families? How will
women's workload affect such things as child care and family
nutrition?

Step 10: Identify needed adaptations. Using the previous
steps as a guide, specify what changes are needed in
institutions, delivery systems, technical packages, and feedback
mechanisms to overcome the barriers to women's access to project
inputs and their ability and incentive to participate.


4.4 From Gender Analysis to Adaptations in Project Design and
Implementation


The results of the evaluation show clearly that analysis of
gender differences alone has little effect on project outcomes
unless institutional and other barriers to participation are
identified and overcome. This section examines a variety of
adaptations that can be useful in removing barriers affecting
women's access to project resources, participation in
activities, and the distribution of benefits.


4.4.1 Women-Only, Women's Component, and Integrated Approaches


One way of overcoming barriers to women's access to
development assistance is to design a women's project. Another
alternative is to insert a component for women in mainstream
projects. A third alternative is to integrate women throughout
mainstream projects without a component for women. Since the
Percy Amendment, A.I.D. has learned a great deal about the
advantages and disadvantages of each approach, and each is
discussed in detail below.

Women-only projects are designed exclusively for women
participants and beneficiaries. Those examined by the
evaluation ranged from institutional-development projects aimed
at training staff of women's bureaus to community-based
income-generating schemes. The main advantage of women-only







-45-


projects is that they are highly visible efforts to explicitly
improve the social and economic well-being of women. The main
disadvantage is that their development impact, with few
exceptions, tends to be minimal. Some of the reasons for low
impact are institutional: tiny budgets, low government
priority, and lack of leverage (especially when projects are
located in under-funded social welfare ministries or private
voluntary organizations). Other reasons are technical:
women-only projects are often administered by people with
inappropriate technical skills, their design is
management-intensive, and their backstopping tends to be
disproportionately costly in staff time.

Findings from the desk review indicate that the most
successful women-only projects were located in mainstream
institutions--ministries of agriculture (10), labor (49), and
education (57) or credit banks (26)--rather than in women's
bureaus or private voluntary organizations (17, 55, 85, 86).
Impact was greatest in Malawi and Morocco, where women-only
projects successfully reoriented mainstream institutions to make
them more responsive to women's needs. In general, women-only
projects have been more successful in delivering training than
in raising production or generating income. They can be an
effective way of training women in nontraditional skills or
training women-in-development personnel.

A women's component is a women-focused activity within a
larger project. As part of a larger project, women's components
have access to greater resources and are able to borrow
technical expertise. Nevertheless, the budgets of these
components tend to be smaller (usually no more than 5 percent of
a project's total budget), and like women-only projects they
have positive and negative features.

In Nepal (002) and Ecuador (27), women's components
emphasized women's domestic roles and overlooked their economic
roles. In Nepal, Mauritania (83), and the Eastern Caribbean
(09), women's components were implemented somewhat in isolation
from the rest of the project and lacked technical input. In the
same three projects, women's components diverted attention from
gender issues in the main project components. In Niger (82),
the women's component was successful because of better
integration with the rest of the project and a good balance
between agriculture and home economics. Women's components can
be an effective way of benefiting women in larger projects;
alternatively, they can lead to tokenism.

An integrated project by A.I.D.'s definition is any
mainstream project that "integrates" women without a women-only
design or a women's component. Yet this definition is
misleading: 83 percent of all projects in the sample would








-46-


qualify as "integrated" projects regardless of actual female
participation. In reality, 55 percent of the mainstream
projects are "gender blind" (low levels of gender analysis
coupled with lack of adaptation) and only 25 percent are "gender
sensitive" (specific gender analysis coupled with adaptation);
another 20 percent are somewhere in between (analysis without
adaptation). Only the gender-sensitive approach is truly
"integrated" in the sense that gender differences are dealt with
systematically and design is adapted in the light of these
differences. The evaluation shows that gender-sensitive design
is correlated with achievement of objectives, while gender-blind
design is correlated with failure to achieve objectives. Of the
three project types, gender-sensitive mainstream projects appear
to be the most effective way of promoting and utilizing women's
contribution to socioeconomic development.


4.4.2 Targeting Resources to Women


Targeting is one approach to trying to ensure that some of
the resources of integrated projects actually reach women. The
basic strategy is to earmark a share of such project inputs as
commodities, training, credit, and employment "for women."
Another is to establish quotas for women's participation in
project activities. The underlying premise is that such
earmarking will make project management accountable for
delivering resources to women.

Cloud's desk review in the agriculture sector found that
there was a positive relationship between resources targeted to
women, resources actually received by women, and achievement of
project purposes. Yet in general, the evaluation revealed that
earmarking resources for women alone may not affect project
outcomes when technical and institutional constraints to female
participation are not identified and removed. For example, if
there were no women in the pool of eligible participants, funds
earmarked for women could not be utilized. Thus, earmarking
resources for women cannot be considered a substitute for
gender-sensitive adaptations in projects as a whole.

Mainstream agricultural projects that earmarked a share of
the budget for extension work with village women invariably
channeled resources through a separate women's extension service
(01, 12, 83, 82). This was sometimes counterproductive, because
the women's extension services tended to focused on domestic
tasks rather than agricultural production. The emphasis on a
women's extension program diverted attention from involving
female farmers in agricultural extension per se.








-47-


None of the mainstream credit projects earmarked a share of
the funds for women, but two mainstream agricultural projects
were to have set up a special fund for women's income-generating
activities (82, 83). Neither was established. Women benefited
more from mainstream credit projects targeted generally to the
poor (see Blumberg's ADEMI case study and Lycette's desk study)
than they did from separate funds established by women-only
projects and women's components. This suggests that in credit
projects, adaptation of mainstream delivery systems may be a
more fruitful strategy than earmarking separate funds for women.

There is also a need to distinguish between resources such
as grassroots training and credit earmarked for village women
and resources such as personnel slots, vehicles, and overseas
training earmarked for female professionals. It is village
women's actual receipt of project resources that is correlated
with achievement of objectives.


4.5 Project Adaptation


It is possible to adapt mainstream projects to gender
without designing a women-only project or a women's component or
earmarking a share of the project budget. This can be done by
adjusting such things as the focus of project activities and
their location, timing, and support services. This section
examines some of the gender-responsive design adaptations that
can improve projects by improving their outreach to women.


4.5.1 Change in the Focus of Project Activities


To a large extent, women's participation in mainstream
projects is influenced by the focus of project activities.
Women's participation was higher in energy projects with
cookstove components than in those without them; it was higher
in water supply projects that provided training in health and
hygiene than in those that did not. In general, it was higher
in all those projects in which one or more components focused on
activities typically performed by women, such as cooking,
sewing, child care, nutrition, or home economics. Thus, one way
of increasing women's participation in development projects has
been to add on a small component focused on women's household
and family roles. The shortcomings of this approach have
already been mentioned: they can divert attention from women's
economic roles and their implications for the success of the
project's main components.







-48-


Cloud's desk review in agriculture shows that women's
participation is influenced by the relative importance given to
"women's crops" versus "men's crops." Lycette's desk review on
employment reveals that women's participation in credit projects
is influenced by the relative emphasis on informal sector
microentrepreneurs. However, the experience of the village
livestock project in Burkina Faso (84), the seed multiplication
project in Cameroon (78), and the agriculture project in
Thailand (98) shows that focus on a "women's crop" or activity,
in itself, is no guarantee that women will actually
participate. When the focus is on a directly productive
activity rather than on a domestic activity, involvement of
women is not automatic and special efforts are needed to ensure
their participation.


4.5.2 Change in the Number of Women in the Pool of Eligible
Participants


In several cases, active efforts to increase women's
participation in projects were frustrated because of the
relatively low number of women in the pool of eligible persons
from which project participants were drawn (53, 70, 01). This
was especially true for institution-building projects, whose
participants were drawn from the staff of government ministries.
In a few other cases, the opposite happened: although no effort
was made to include women, their participation was fairly high
because a substantial number of women were already in the pool
of eligible participants in a population targeted for other
reasons.

In Botswana (04) and Thailand (95), professional women's
participation in agricultural planning projects was fairly high
because 30-40 percent of the staff in the planning unit from
which participants were selected were women. The high par-
ticipation of women in an English language training project in
Syria (61) was not due to special efforts to recruit women but
to the high proportion of women among eligible persons. The
differences among countries reflect differences in women's
access to higher education and the openness of government
bureaucracies to their recruitment.

Although in such cases it might appear that women's
participation is outside the control of project planners, this
is rarely true. When there are few women in the pool of
eligible people, three adaptations are possible: (1) eligi-
bility criteria or institutional procedures can be changed so
that women qualify, (2) special programs can be launched to
train more women up to standards, or (3) male staff can be
trained to work with village women in the absence of female
staff.








-49-


4.5.3 Adaptation of Credit Components


Projects with credit components offered some of the
clearest illustrations of problems and adaptations related to
eligibility. Whether in agriculture, nonfarm production, or
urban informal sector enterprises, the main determinants of
women's participation in credit components were as follows:

The focus of lending (microenterprises versus larger
firms and male- versus female-dominated sectors)

Minimum size of loans (the smaller the minimum, the
higher the female participation rate)

-- Collateral requirements (group liability can remove the
obstacle of women's lack of land title and fixed assets)

The hidden costs of borrowing (reducing the cost in
time and money of trips to credit outlets for
application and repayment increases participation)

Bank's incentive to service small loans (innovations
such as group lending can increase outreach to women by
reducing overhead costs to the bank)

When the terms of lending are conducive, women will
constitute a high proportion of loan recipients. When lending
terms are adverse, few women will receive loans, regardless of
efforts to target them. Thus analysis and adaptation of
eligibility criteria and delivery systems are the key to
increasing women's participation in credit programs and the
productive activities that those programs support.


4.5.4 Outreach of Existing Delivery Systems


The outreach of existing delivery systems strongly affects
projects' ability to reach and benefit women. A few of the
mainstream projects with the best gender analysis had limited
success in transferring technology to women because of weakness
in delivery systems (83, 64, 01). In some geographic areas,
where existing delivery systems already had contact with women,
projects were successful in transferring technology to women in
spite of weak gender analysis (16, 80). A social forestry
project in India (70), a resource conservation project in
Honduras (64), and a village development project in Tanzania
(01) all reported limited uptake in districts not covered by
existing women's extension programs. When ceilings on
government expenditure make it difficult for the project to








-50-


recruit additional staff to work with women, the programs'
outreach depends on their ability to utilize whatever
village-level staff (male or female) is already in the area.


4.5.5 Location of Project Activities and Services


The location of training facilities influenced women's
participation in training of all types: because of their family
responsibilities, women were less likely to participate in
out-of-country than in-country training and were more likely to
prefer day training to residential training (see Anderson).

In Kenya (001), project staff discovered that women's
participation was much higher when user committee meetings were
held directly at the water source. When meetings were held in
the village, men dominated the discussion and the primary water
users--women--remained on the sidelines.


4.5.6 Timing and Duration of Activities


Because women's time constraints differ from those of men,
the timing and duration of project activities affect women's
participation differently from men's.

In Niger (82), wives were unable to participate fully
in farmer training courses because the timing of part
of the program conflicted with meal preparation
responsibilities; grain mills were introduced to reduce
meal preparation time so that women could attend (see
Cloud).

In Bolivia and Ecuador (21), sensitivity to women's
time constraints was a key factor explaining the
success of some income-generating projects and the
failure of others (see Flora).

Adaptation of the Arid and Semi-Arid Lands project in
Kenya (001) to women's seasonal time constraints was
crucial in securing women's participation in soil and
water conservation activities (see Carloni and
Horenstein).


4.5.7 Facilities for Sleeping and Child Care


When training requires women's absence from home for
extended periods, the availability of facilities for sleeping
and child care greatly influences women's ability to









-51-


participate. At the Botswana Agricultural College (54),
construction of a women's dormitory facilitated the expansion of
female enrollment (see Anderson and McBreen). In Nepal, where
the planned women's dormitory at the Institute of Agriculture
and Animal Science (90) was-cancelled for lack of funds, only
four women were able to enroll (see Cloud). At the Institute of
Renewable Natural Resources in the same country (002), the
construction of a separate dormitory, with a resident female
warden and a guard, successfully overcame parents' reluctance to
send daughters for forestry training (see Davenport, Nickell,
and Pradhan). In Mali (56), child care facilities facilitated
women's participation in residential training; in Kenya (001)
lack of child care facilities hindered it.


4.5.8 Choice of Language and Communication Network


Finally, the choice of language and communication network
also influences outreach to women. Because of their greater
contact with the world beyond the village, men are more likely
than women to speak the national language. In highland
communities in Bolivia, Ecuador, and Guatemala, where men speak
Spanish but women are monolingual in their indigenous dialect,
Spanish-speaking extension workers could only talk to male
farmers. The solution is to recruit bilingual extension agents
directly in the local area. Communication networks among
village women also differ from those of men. Outreach to women
can be improved by identifying the times and places where
different groups of women get together and then using these
settings as entry points.


4.6 Conclusions


Evaluation findings show that the effort spent on improving
gender analysis has paid off: it has benefited women, and it
has made development assistance more effective. In the future,
the payoff could be much greater if two conditions are met:
institutional barriers to women's participation and benefit need
to be analyzed, and projects need to be adapted to overcome
barriers. While women-only projects and women's components can
be useful in specific contexts, they are not the only
alternatives. Adaptation of gender-sensitive mainstream
projects appears to be a very successful way of promoting and
utilizing women's contribution to development.









APPENDIX A


FINDINGS1


Using information from the desk review, the relationships
between the variables outlined in the conceptual framework were
examined for three types of projects: women-only projects,
projects with a women's component, and mainstream projects. At
a later stage, a further distinction was made between mainstream
projects having direct contact with people at grassroots level
(referred to as "direct service" projects) and those having no
grassroots contact (referred to as "institutional development"
projects). This appendix presents the results of the analysis
of each of the variables in the conceptual framework for all
three project types.


1. GENDER ANALYSIS


Between 1972 and 1984, the quality of gender analysis in
A.I.D. project design documents steadily improved. However,
even in the 1980-1984 period, less than 30 percent of the sample
projects for which design documents were available (89 out of
98) identified the gender of the intended beneficiaries with any
specificity (see Table A-l).

Table A-1. Level of Gender Analysis in
Project Design Documents, 1972-1977 to 1980-1984
(n = 89 projects with information)


Period Level of Gender Analysis No. %


1972-1977 No information/mention of women 18 64.3
Boilerplate 3 10.7
Specific analysis 7 25
Total 1972-1977 28 100.0


1978-1979 No information/mention of women 13 38.2
Boilerplate 14 41.2
Specific analysis 7 20.6
Total 1978-1979 34 100.0


1980-1984 No information/mention of women 11 40.7
Boilerplate 8 29.6
Specific analysis 8 29.6
Total 1980-1984 27 100.0


the
in the


iReaders should consult Appendix B for an explanation of
values assigned to the relationships among the variables
conceptual framework.








A-2

Reference to women's directly productive roles has been
more frequent than reference to their household and family
roles, but less than a third of the projects analyze either in
any detail (see Table A-2). After 1980, there was a sharp
reduction in the number of projects that overlooked women and a
corresponding increase in boilerplate references to their role
in production and reproduction. More projects refer to women's
productive roles but fewer projects analyze that role with any
specificity.

Table A-2. Reference in Project Design Documents to
Women's Directly Productive Roles and to Their
Household/Family Roles, 1972-1977 to 1980-1984
(n = 91 projects with information)



Date/Level of Directly
Analysis Productive Roles Household/Family Roles
No. % No. %

1972-1977 (n=29)
No Mention 15 51.7 17 58.6
Boilerplate 6 20.7 8 27.6
Specific Analysis 8 27.6 4 13.8


1978-1979 (n=35)
No Mention 10 28.6 17 48.5
Boilerplate 10 28.6 9 25.7
Specific Analysis 15 42.6 9 25.7


1980-1984 (n=27)
No Mention 7 25.9 9 33.3
Boilerplate 12 44.4 11 40.8
Specific Analysis 8 29.6 7 25.9



The aspect of women's directly productive roles that has
received the most attention is the gender division of labor
(mentioned by 45 percent of the projects with design documents
and specifically analyzed by 27 percent). Gender differences in
access to and control of resources were mentioned by 31 percent,
but only 22 percent analyzed them with any degree of specificity.

More design documents emphasized how women could benefit
from the project. Only a few emphasized that the project could
also benefit from women's participation (see Table A-3).








A-3


Table A-3. Project Design Document References to
Project Benefits to Women or Women Benefiting the Project,
1972-1977 to 1980-1984
(n = 91 projects with information)


Date/ Women Benefit Project Project Benefits Women
(No. of Projects) No. % No. %


1972-1977 (n=29) 5 17.2 12 41.4

1978-1979 (n=35) 9 25.7 17 48.6

1980-1984 (n=27) 9 33.3 21 77.8

Total (n=91) 23 25.3 50 54.9



1.1 Mainstream Direct-Service Projects


The study analyzed the relationship between gender analysis
and each of the variables in the conceptual framework--the base-
line situation, gender analysis, project adaptation, participa-
tion, achievement of purposes, impact on women, and achievement
of goals. A moderate relationship was found between the quality
of gender analysis and adaptation, in the following sense: if
gender roles were not analyzed or if the analysis was only super-
ficial, no project adaptation occurred; if gender roles were
analyzed, some projects were adapted to increase women's
participation or benefit and others were not (see Table A-4).


Table A-4. Relationship Between Gender Analysis and
Project Adaptation: Mainstream Direct-Service Projects
in All Sectors
(n = 37 projects with information)


Level of Gender Level of Project Adaptation
Analysis Low (n=27) Medium (n=5) High (n=5)


Low (n=20) 19 1 0


Medium (n=9) 5 2 2


High (n=8) 3 2 3







A-4


The relationship between gender analysis and women's actual
level of participation was weaker, but the relationship was in
the same direction. Failure to analyze gender roles was
strongly associated with exclusion of women, but analysis of
gender roles was only weakly correlated with women's
participation. This finding supports the conceptual framework's
hypothesis that gender analysis would improve participation only
when institutional barriers were identified and removed.

No direct relationship was found between gender analysis
and achievement of project purposes, except in cases where
recognition of gender differences led to project adaptation and
greater female participation. In the absence of adaptation and
participation, gender analysis had no effect on the achievement
of purposes. For the same reason, there was no direct
relationship between gender analysis and projects' impact on
women.


1.2 Mainstream Institutional-Development Projects


Few mainstream institutional-development projects adapted
technical messages or delivery systems in response to gender
differences. For this reason, lack of gender analysis was
correlated with lack of adaptation, but there was little
evidence that at higher levels of gender analysis adaptation
occurred more frequently. No relationship was found between
gender analysis and levels of female participation. Likewise,
no relationship was found between gender analysis and
achievement of project purposes. Moreover, projects that lacked
analysis of gender differences tended to have a negative impact
on women, but projects that did analyze gender differences did
not necessarily have a positive impact on women.


1.3 Women-Only Projects


The overall level of gender analysis was higher for
women-only projects, and a positive relationship was found
between the quality of analysis and project adaptation. Good
gender analysis was also correlated with (1) higher levels of
participation (i.e., more effective outreach to large numbers of
women), (2) achievement of project purposes, and (3) a positive
impact on women. Superficial gender analysis was correlated
with lower levels of participation (often a sign of lack of
interest in what the project offered), failure to achieve
purposes, and a corresponding failure to have much impact on
women.







A-5

1.4 Women's Components


No conclusions can be drawn because of the small number of
cases, but the quality of gender analysis appears to have little
relationship with the implementation of women's components. In
one case with gender analysis, the women's component was never
implemented. In another case, the analysis was among the best,
but the woman selected to implement the component had no
qualifications. The Caribbean Regional Agricultural Extension
project (09), which was the subject of one of the 10 case
studies (see Schminck and Goddard), also had high-quality gender
analysis. However, because the project's main thrust is
institutional development, its impact at the farm level is still
very limited, and it is thus premature to judge the impact of
the gender analysis on project beneficiaries.


2. PROJECT ADAPTATION


2.1 Mainstream Direct-Service Projects


The analysis revealed that project adaptation is a key
variable in the conceptual framework. There is a strong
positive correlation between project adaptation and women's
actual level of participation. Little or no adaptation to
gender is associated with low levels of participation, whereas
explicit adaptation to gender significantly improves women's
participation (see Table A-5).


Table A-5. Relationship Between Project Adaptation
and Women's Participation: Mainstream Direct-Service
Projects in All Sectors
(n = 26 projects with information)



Level of Level of Women's Participation
Adaptation Low (n=9) Medium (n=9) High (n=8)


Low (n=16) 9 5 2


Medium (n=5) 0 4 1


High (n=5) 0 0 5






A-6


Even more important, the level of project adaptation shows
a positive correlation with the achievement of project purposes
(see Table A-6). Projects that adapted institutions, delivery
systems, and technical packages to deal with gender differences
were more likely to achieve their stated purposes (34, 41, 70,
72, 80) than projects that made no adjustment to deal with gender
differences (03, 07, 13, 28, 37, 38, 71, 73, 87, 88, 94, 98).


Table A-6. Relationship Between Project Adaptation
and Achievement of Project Purposes: Mainstream
Direct-Service Projects in All Sectors
(n = 36 projects with information)



Level of Achievement of Objectives
Adaptation Low (n=13) Medium (n=9) High (n=14)


Low (n=26) 12 7 7


Medium (n=5) 1 2 2


High (n=5) 0 0 5



Seven projects failed to fit the predicted pattern (i.e.,
they were highly successful in achieving their purposes in spite
of a lack of adaptation). Five of these were concerned with
constructing water supply and sanitation systems; a sixth was
concerned with delivery of education technology such as
computers and software (66). Since their purposes could be
achieved without much contact with villagers, these cases do not
contradict the hypothesis. Such projects share more in common
with institutional-development projects than with direct-service
projects. The seventh project that failed to fit the predicted
pattern was a small farmer credit project (76). (It is
discussed in detail in Section 2.1.1 of the main body of the
paper.)

The relationship between project adaptation and impact on
women was weaker: lack of adaptation was associated with a
limited or negative impact on women, but adaptation, by itself,
did not always lead to a positive impact on women (see Table
A-7). This finding was expected.







A-7

Table A-7. Relationship Between Project Adaptation and
Impact on Women: Mainstream Direct-Service Projects
in All Sections
(n = 27 projects with information)



Level of Impact on Women
Adaptation Low/Negative (n=14) Mixed (n=7) Positive (n=6)


Low (n=18) 13 2 3


Medium (n=5) 1 3 1


High (n=4) 0 2 2



2.2 Mainstream Institutional-Development Projects


The overall level of project adaptation was very low. As a
consequence, no relationship was found between adaptation and
women's actual level of participation. There was no correlation
between adaptation and achievement of project purposes. Nor was
there any relationship between project adaptation and impact on
women.


2.3 Women-Only Projects


In spite of the fact that the participants in women-only
projects are all female, the quality of the gender adaptation in
these projects varied. Those projects with good adaptation
performed better in every way: they reached more women; they
were more likely to achieve their purposes; and they were more
likely to have a positive impact on women.


3. PARTICIPATION


3.1 Mainstream Direct-Service Projects


Table A-8 reports the most important finding of the study,
namely, that for mainstream direct-service projects in all five
sectors, the degree of match between gender roles in the







A-8


baseline situation and the gender of project participants shows
a strong positive correlation with achievement of project
purposes. In other words mainstream projects that reach women
and ensure substantial participation are more likely to achieve
their purposes; projects that bypass women or ensure only token
participation are less likely to achieve their stated purposes.
Outreach to women can be a crucial factor in project success.
It effects efficiency in addition to equity. (Section 2 of the
main body of this paper examines this important finding in
greater detail.)


Table A-8. Relationship Between Level of Female Participation
and Achievement of Project Purposes: Mainstream Direct-Service
Projects in All Sectors
(n = 28 projects with information)



Level of Female Achievement of Project Purpose
Participation Low (n=9) Medium (n=6) High (n=13)


Low (n=9) 5 4 0


Medium (n=8) 4 2 2


High (n=ll) 0 0 11



Table A-9 reports another important finding: the level of
female participation also shows a positive correlation with
achievement of project socioeconomic goals. In other words,
projects that reached women and ensured substantial participa-
tion were more likely to have the desired impact on broader
socioeconomic goals (15, 47, 77, 80, 89); projects that failed
to reach women and ensure their participation were less likely
to achieve these broader goals (13, 28, 37). (This important
finding is explored in Section 3 of the paper.)

There were no major anomalies in the pattern: not even one
project that was known to have low levels of female
participation was highly successful in achieving its goals; only
three projects were moderately successful in achieving their
goals in spite of women's lack of participation: two were
concerned with physical construction of buildings and irrigation
systems on large-scale settlement schemes (91, 93) and the other







A-9


with seed multiplication (78). In two of these cases (78, 91),
there is evidence that higher levels of female participation
would have improved achievement of goals.


Table A-9. Relationship Between Level of
Female Participation and Achievement of Project Goals:
Mainstream Direct-Service Projects in All Sectors
(n = 18 projects with information)



Level of Female Achievement of Goals
Participation Low (n=6) Medium (n=5) High (n=7)


Low (n=6) 3 3 0


Medium (n=5) 3 0 2


High (n=7) 0 2 5



3.2 Institutional-Development Projects


There was no correlation between women's level of
participation and achievement of either immediate project
purposes or long-range goals.


3.3 Women-Only Projects


In spite of the fact that nearly all beneficiaries are
women, the correlation between participation and achievement of
project purposes is moderately positive: the more women reached
by a project and the greater their decision-making role, the
higher- the likelihood of achieving the project purposes.
However, in spite of the positive relationship between
participation and achievement of immediate project purposes, for
women-only projects there was little correlation between female
participation and achievement of long-term goals. The reason
seems to be that many women-only projects that achieved their
immediate purpose of training women or launching community
development activities have by their nature had a limited impact
on broader socioeconomic goals.







A-10


4. ACHIEVEMENT OF PROJECT PURPOSES


4.1 Mainstream Direct-Service Projects


It was expected that, other things being equal, projects
that achieved their immediate purposes would be more likely to
have a positive impact on women than projects that failed to
achieve their stated purposes. Table A-10 shows that although
the relationship is in the expected direction, achievement of
project purposes in itself is not sufficient to ensure a
positive impact on women.


Table A-10. Relationship Between the Achievement of
Project Purposes and Projects' Impact on Women:
Mainstream Direct-Service Projects in All Sectors
(n = 30 projects with information)



Achievement of Impact on Women
Purposes Low/Negative (n=14) Mixed (n=8) Positive (n=8)


Low (n=10) 8 2 0


Medium (n=7) 5 1 1


High (n=13) 1 5 7



4.2 Mainstream Institutional-Development Projects


There was no correlation between achievement of project
objectives and impact on women. Evaluation teams assessed
project performance on other grounds. Impact on women was not
even considered.


4.3 Women-Only Projects


The correlation between achievement of project purposes and
impact on women was positive and nearly perfect (see Table
A-11). This was expected because women-only projects are
designed to improve women's well-being.







A-11


Table A-ll. Relationship Between Achievement of
Project Purposes and Impact on Women:
Women-Only Projects in All Sectors
(n = 10 projects with information)


Achievement of Impact on Women
Objectives Low/Negative (n=3) Mixed (n=2) Positive (n=5)


Low (n=4) 3 1 0


Medium (n=0) 0 0 0


High (n=6) 0 1 5



5. ACHIEVEMENT OF SOCIOECONOMIC GOALS


5.1 Mainstream Direct-Service Projects


The study revealed that projects that failed to achieve
their purposes almost never achieved their goals but that
projects that achieved their purposes were not necessarily more
successful in achieving long-range goals. To shed light on why
projects that achieve their immediate purposes might fail to
have the desired socioeconomic impact, we looked at projects'
impact on women as an intervening variable. The findings
suggest that gender variables could be of major importance in
explaining projects' socioeconomic impact.

The relationship between achievement of immediate project
purposes and achievement of goals is neither direct nor
automatic. However, projects that achieve their purposes and
have a positive impact on women are much more likely to achieve
socioeconomic goals than projects that achieve their purposes
but have a low or negative impact on women (see Table A-12).
What is more, the correlation between impact on women and
achievement of goals is stronger than the correlation between
achievement of project purposes and achievement of goals.







A-12


Table A-12. Relationship Between Projects' Impact
on Women and Achievement of Socioeconomic Goals:
Mainstream Direct-Service Projects in All Sectors
(n = 21 projects with information)


Impact on Achievement of Goals
Women Low (n=7) Medium (n=6) High (n=8)


Low/Negative (n=10) 6 3 1


Mixed (n=5) 1 1 3


Positive (n=6) 0 2 4



Six projects that had a low or negative impact on women
failed to achieve their goals. In two cases, the connection
between gender variables and failure to achieve goals was
clear. One was a water supply project that developed water
sources so far away from where people lived that there was no
reduction in women's workload (37). The other was a livestock
project that initially targeted resources for women's activities
exclusively to men (84). (The livestock project is discussed in
detail in Section 2.1.1 of the paper.) In the other four
projects, there was not enough information to determine whether
the project's negative impact on women was a primary cause of
failure to achieve its goals or merely a contributing factor.

Four of the projects that had a positive impact on women
achieved their goals, but there is only one case with sufficient
information to show a direct link between the project's impact
on women and achievement of goals. This was a watershed
management project in Cape Verde (80), where completion and
maintenance of soil conservation works depended on female
labor. In spite of gaps in documentation, the findings from the
desk review are strongly supported by evidence from the case
studies. (These findings are reported in Section 3 of this
paper.)

The one project that was highly successful in achieving its
goals in spite of a neutral to negative impact on women was the
education technology project mentioned earlier (66); it was
successful in its own terms, but failed to reduce the low school
enrollment of girls. The team that evaluated the project as
highly successful in achieving its stated goals did not consider







A-13


the project's impact on school enrollment of girls, despite
A.I.D.'s explicit policy to reduce sex disparities in
enrollment.


5.2 Mainstream Institutional-Development Projects


The relation between projects' impact on women and
achievement of goals is less clear for institutional-development
projects than for direct-service projects. Four projects that
had a low or negative impact on women failed to achieve their
goals (02, 60, 79, 90). However, in each of these cases other
factors were more important in explaining the failure than the
project's impact on women.

One project was rated highly successful in achieving its
goals in spite of a negative impact on women. It was a project
concerned with rebuilding primary schools after a hurricane
(62); like the education technology project cited above, it was
successful in technical (i.e., school construction) terms, but
failed to improve girls' school attendance. In this case, the
location of the schools had an adverse effect on girls'
attendance because many girls' families oppose their traveling
long distances from home. The case illustrates the need for
bringing evaluation criteria in line with A.I.D. policy.


5.3 Women-Only Projects


A moderately positive correlation was found between
projects' impact on women and achievement of goals. Given the
fact that a positive impact on women was almost synonymous with
achievement of the projects' goals, this was expected.


6. CONCLUSIONS


Recognition of gender differences is important for equity,
but it is equally important for project efficiency. It is at
least as important for mainstream projects as it is for
women-only projects and for women's components of projects.


6.1 Mainstream Direct-Service Projects


Projects that deliver goods and services directly to people
at grassroots level are more likely to achieve their purposes
when the gender of project participants reflects the division of
economic responsibilities in the baseline situation.







A-14


Failure to analyze gender differences is associated with
lack of project adaptation. Lack of adaptation is associated
with low levels of female participation. Low levels of female
participation show a strong correlation with failure to achieve
project purposes. The consequence of inadequate gender
analysis, lack of project adaptation, and low levels of female
participation is failure to achieve project purposes and broader
development goals.

Recognition of gender differences is only the first step.
It increases the probability that projects will be adapted to
increase female participation, but by itself it is not enough.
Project adaptation emerges as a key variable. Without
adaptation, gender analysis has little or no effect on project
outcomes.

High levels of female participation show a strong positive
correlation with achievement of project purposes. Levels of
participation are strongly influenced by project adaptation.
These findings explain why project adaptation is correlated with
achievement of objectives.

In short, projects that recognize gender differences, adapt
delivery systems accordingly, and ensure substantial female
participation are more efficient and are more likely to achieve
their purposes. They are also more likely to achieve broader
socioeconomic goals. This holds true for mainstream
direct-service projects in all five sectors--agriculture,
education, employment/income generation, energy/natural resouce
conservation, and water supply/sanitation.


6.2 Manstream Institutional-Development Projects


The fundamental difference between institutional-
development projects and direct-service projects is that the
former do not directly reach people at grassroots level. For
this reason, the gender division of labor and responsibilities
in the baseline situation is not as important for the projects'
achievement of their immediate purposes. This explains the lack
of correlation between the level of female participation and
achievement of project purposes.


The number of mainstream insititutional-development
projects that adapted design in response to gender analysis was
t'oo small to draw any conclusions about the relationship between
project adaptation and female participation. Gender analysis
without adaptation has no effect on project outcomes.







A-15


No conclusions could be drawn about the importance of
gender variables for achievement of socioeconomic goals because
of the small number of cases with information (12) and the large
number of projects that failed to achieve their goals (9).


6.3 Women-Only Projects and Women's Components


The conceptual framework is also supported by women-only
projects. A positive relationship was found between the quality
of gender analysis and project adaptation. Projects with good
adaptation performed better in every way: they reached more
women; they were more likely to achieve their objectives; and
they were more likely to have a positive impact on women.

No conclusions could be drawn about women's components
because of the small number of projects with information. The
findings did not seem to fit any particular pattern.







APPENDIX B


METHODOLOGY


This study is one of the Center for Development Information
and Evaluation (CDIE) efforts to help the Agency learn from its
experience. The primary objective of the study was to
synthesize Agency experience in the field of women in
development in light of the 1982 Policy Paper emphasis on the
importance of gender to project efficiency. Unlike previous
CDIE studies that focus on particular sectors, this study
examines issues of women in development in five
sectors--agriculture, education, energy, employment/income
generation, and water/sanitation.

This study was organized in three phases: (1) sectoral desk
reviews of project documents, (2) project field studies, and (3)
synthesis of findings from the two earlier phases. Each study
phase is briefly described below.


1. DESK REVIEWS


One objective of the desk reviews was to assess Agency
experience since the beginning of its women in development
program in 1973. A second objective was to use available
project documentation to assess project implementation in light
of the major assertion of the 1982 Women in Development Policy
Paper: that gender is a key variable for the efficient
achievement of objectives in mainstream projects.

Ninety-eight projects were randomly selected for the desk
review from the approximately 4,000 on-going and completed
projects contained in the CDIE automated project data base. The
sampling procedure consisted of (1) identifying projects that
had "women", "female," or equivalent descriptors in the
abstracts of project documents; (2) reviewing this list of
projects with the Office of Women in Development and regional
bureau Women in Development Officers; (3) assigning random
members to the final list of 416 projects; and (4) sampling
projects for each sector. The number of projects for four of
the sectors was determined by the number of projects each
consultant could review for this study. The energy sector was
the exception, because all 10 projects that appeared in the list
of 416 projects were sampled. Table B-l shows the distribution
of projects by sector.







B-2


Table B-1. Number of Projects in Total Listing and in
Sample, by Sector



Listing Sample
Sector No. % No. % Ratio


Agriculture 181 43.5 40 40.8 .22099

Education 105 25.2 19 19.4 .18095

Employment 58 13.9 11 11.2 .18965

Energy 10 2.4 10 10.2 1.0

Water/Sanitation 62 14.9 18 18.2 .29032

Total 416 100 98 100 .23558



The project sample is not representative of all A.I.D.
projects but rather of those that had a gender descriptor in the
CDIE project data base. The assumption that a gender descriptor
in project abstracts reflects greater emphasis on gender issues
was not supported by the review of project documents. Thus,
while the sample may overemphasize the importance of gender for
some A.I.D. projects, a bias toward a more favorable assessment
of women-in-development policy implementation is unlikely.

Table B-2 shows the distribution of projects sampled by
sector and type of project (women-only, women's component, or
integrated project). Eighty-two percent of the total number of
projects are mainstream projects that aimed at working with both
men and women, as opposed to projects with components
specifically designed for women or projects that are designed
exclusively for women.

The largest number of projects sampled were from African
(40) and Latin American (LAC) countries (32); 14 projects were
sampled in Asia and 12 in the Near East (NE) region. The
regional distribution of projects sampled is similar to the
pattern for all projects Agencywide (see Table B-3).








B-3


Table B-2. Number and Percentage of Projects by
Type of Project and Sector


Type of Project

Women-Only Women's Component Integrated
Sector No. % No. % No. % Total


Agriculture 5 41.7 3 50.0 32 40.0 40

Education 5 41.7 2 33.3 12 15.0 19

Employment 2 16.7 1 16.7 8 10.0 11

Energy 0 0 0 0 10 12.5 10

Water/Sanitation 0 0 0 0 18 22.5 18

Total 12 100 6 100 80 100 98



Table B-3. Number of Projects Sampled by Region and by Sector



Africa Asia LAC NE
Sector No. % No. % No. % No. % Total


Agriculture 19 47.5 7 50.0 11 34.4 3 25.0 40

Education 7 17.5 2 14.2 5 15.6 5 41.7 19

Employment 3 7.5 1 7.1 7 21.9 0 0 11

Energy 5 12.5 4 28.6 1 3.1 0 0 10

Water/Sanitation 6 15.0 0 0 8 25.0 4 33.3 18

Total 40 100 14 100 32 100 12 100 98



Desk reviews for projects in the five sectors were conducted
by consultants experienced in their respective sectors and with
A.I.D. projects and processes. Two consultants carried out the
desk reviews for the agriculture sector and one consultant for
each of the other sectors. Agriculture projects were subdivided







B-4


according to region: one consultant reviewed projects from
Asia, North Africa, the Sahel, and the Middle East; the other
reviewed projects from Latin America, the Pacific Islands, and
Sub-Saharan Africa.

Of the 98 projects in the sample, 37 were completed and 61
were on-going at the time of the study (see Table B-4). The
documents available for each project varied considerably by
phase of the project and by sector. Documentation was most
readily available for the design phase. Information on the
implementation phase was lacking for about one-third of the
projects, and evaluation information was available for only 40
projects (see Table B-5).


Table B-4.


Project Completion Status by Sector


Active Completed
Sector No. % No. % Total


Agriculture 27 44.3 13 35.1 40

Education 10 16.4 9 24.3 19

Employment 4 6.6 7 18.9 11

Energy 9 14.7 1 2.7 10

Water/Sanitation 11 18.0 7 18.9 18

Total 61 100 37 100 98


Table B-5. Documentation Available by Project Phase and
Sector for All Sampled Projects


Design Implementation Evaluation
Sector Yes No Yes No Yes No


Agriculture 36 4 23 17 13 27

Education 18 1 13 6 14 5

Employment 11 0 9 2 5 6

Energy 10 0 7 3 5 5

Water/Sanitation 17 1 9 9 3 15

Total 92 6 61 37 40 58








B-5


Consultants used a common questionnarie to code information
found in the documentation for each project. Some problems were
encountered at this stage because of the complexity of the
questionnaire and the limited information available in project
documents. To fill the gaps in the documentation, consultants
relied on their own first-hand knowledge of the projects and on
interviews with project officers. The list of projects included
in the desk review is in Appendix C.

Because the results of the desk reviews did not provide
enough information to assess the importance of gender to project
efficiency, a field study phase was planned and carried out
during 1985. This study phase is described in the following
section.


2. FIELD STUDY PHASE


The objective of this phase of the study was to gain better
understanding of the relationship between gender and the
achievement of project objectives and greater insight into the
relationship between gender issues and socioeconomic development.

As desk reviews were completed, findings were examined by
project officers in CDIE. Consultants recommended projects for
inclusion in the field study phase, as did central and regional
bureau Women in Development Officers. An initial list of 23
projects was compiled, and USAID Missions were asked to approve
fieldwork on the projects, which would be conducted for 2 weeks
between March and May 1985. Eleven Missions agreed to this
request. Only one of these projects was not chosen for
fieldwork because of its delayed implementation. All projects
selected for fieldwork had either been completed or were at
least in their second year of implementation.

Six of these ten projects had been examined during the desk
review; others, which had been eliminated at random from the
desk review, were selected because they were mainstream projects
in which women's participation was considered particularly
important. The projects represented 11 countries or subregions
distributed within four regions. Table B-6 presents summary
information on the 10 projects selected for field study.

Fieldwork was conducted during April, May, and June of
1985. In all, 16 consultants participated in the field study
phase. In-country professionals were added to the study team
for some projects.








B-6


to
4-
C
0)




















01
81



















0



4-1
0
.,
0







44

























0








M 4
U)
41









































&1


0
84











a01
C


C).'-

































eO


a
4-'


-p
0


C)
0







44
0
C)<








z *c 0 0
0



4-1.4:

0
3 i










O4 r,: -A
'400
H













S -0 0z
C) 4J (V
4 4
o a -P






4J 0 41 -4 *.0




z c
44:
(1) 0) 4U 4) *.-

(n 04o4




04 'd H0 0 Ori
0 0 4 10



U ^- C 3 -
00 0 0
(1) Z V 10
.0 C) i-I () o
4J ^ 04-H40'
+J 4: -J C: +J
, C )41 C) $ 4)
*ifl 4)H 0' C )^ <
C) $ 4 C:.$4 C:

o C) C: C:.X: 4
4-. Z C) C 4 'C


N

0
1-4
0 n4 -.
4-1 C t)
4J rC) 4
0 0


4J 4 J
r0 0






4p
*- 4 .I 44


4-0
C) i a4) 0
4- I ) (44 $4
'-4 4.4 )
a)) 04)
u *E P
4- 3O 0) 0 0)
+V > -P U _l1r-
U 0$4C 3 '40
3 0)o' *. 4J
4 C) *H 4) g C: 10
c4-' *H en S E0
W) 0 W 4J $4 5 It
o u o- 0
I 440 > -.0>
H t C:

r f o a > 0>
r- s >1 .14 e 1
4)4. 4-> ) eH 4)4-o
3 : a ) 4- :) a ) Hn
S0 w4U) m .0 u
.H 4- ;C m 9
w >m f: -
-Tm -P c 4.
.4 0 0 0

00 3 0a 4
*HU^ c t

-w o C o i> cf

0 $4 10 $4 1')r
o -i a) 4 0
E. *H- (' '-4 (U 4-U 0


odo
-- **
.C)C
QC CD


* 0 4
'CC --
* 4 t


1
44

S)

c
4


-0 C)




s m


10
joo
C) .4 .-
-COCC
n


4) I

'44 0-H
0 0
>

0


eN "
4- 0 2 10
>1 tF, g1
:H C
( 0 0




wJ a) 4-1
M 3



l 0o
s) ro
A E M
0





1-4 m0 V-
0 --4 44
. 0





p) '414
0 0
C)C)

'OC)




o u -c:
0 -'
4)HH

04)


C
*H

0

a c
CB

H 0




40
c.4 >C10







0
0
a a P


C 0
0 w
-H C:
0 4 -0- P 'l l



n r ( 0 0 a ,
4* 4 4 M r 4 4
V) 4-4-.. 4 01 04
U 4-))XC4:H)40C *4-4 $4
c ciM -R OCH 03<
S*4 0 r *H s 0 0
cllS '0 tp H4 ~-1 Q lO 0<
4.-4 U 0 D 0 '4 4 $ 4 4. ) M
c: 4 0-4 0 4-1 a
4 -1 C- C 4- d 0 C r 0
0 A 0 c 4 rc 444-' e

0 C r r mC
u cmoB fu ^r

$4 4)4 4- 0 0 -4 0l 4 r
$H > -'---4 -'-'4-4


WOM .- OMU24C:
r-- ) W r 4 1" W W 0C4
04: 'C)X '40.4O'~ 0

S0 4* 0 00-0
4)4)) 4- OC) C:,-
0- 4 k C) O ) 0 '0 'M14 0
> C) 04)44 0 4-i/ '44240
:s 'a o 4 o 0,4- r o U, C
*4 0C) -4 :0-aH
o m >' m -4o"c 0 O
O o tn o C C O C U Ml-
0CO44. )44:4) :0)-H#
1-4 C) O d P ) 0,4' C).- 4 l -i









8
a)







H

4-
*M








tr
*I






co
.1-)




a,
04:
H $4





co

a,
-C
'4^
0
0-
CO
Hf
0-


U) 0
(N U

'44
0 0



0 Q







co
a)
HT
0-


I
* !
>




04
8 H



UC 4
o







,I'


0
o


N N.
C C
0 0
o o
1. -4

) 0 C: 4)0 a
-w C:) 4I-1 C w
So0g 1m0 o
$4 'O 0 p 0 I
44J 0 4 41
4V 4 4 4:1 -4)V
H F1 0 H to








B-7


0
v



IC


. ,0
I
0 0



4) 0
t x

44m 0) v
"4 4)





o 4)4 0 >,
r4 4





8 0 > *
00 04













V
I4
a
,-1


0:




* r
HQ






.0
< "








0



-4

I
(0
4rc











4) WV1


-4 C0 4-
>o I 4m 0 (4
SCD C: 0 $4 :t B e r l
0 m 04 E C-4 -4 4
co V o o m > W 0 0 0
m 14 1 >0 M > 1 4 4V U 0

-1 0 S$4 4J r. $4 1 0i t 0 V
( W 4 OK 4Vl P
0 0 ) 4 n > 4 0 4 9 o
M *( >, 4) 0 > $4 > $4 4P V 4 )


0a M O O O X -c4M-4 4* -O a
- 0 0 z :t n C *T l rt S' M 4 O OW r
S0 xm 0 0 O O
404-4 ) o 4m. C 4 H a444a -d p u c

,q C: 4 -H (n 0o 4 V .0 M 10
C4 -H 4J 10E 044)00 W a(V


A4;4
C** 444



* 0 0 4 d 14 (4
* 4 m0 >
* (4 ( sH
S0. 0
11 8 MO H


*4
>
z v
4'- 0
44 W 4) -. c

4) u ) N
3 HU
MUM1 >U>
C *M -
+1 RC *-


4 M m
I 0

c a 4

4 0 4
WOn H 34)







o o


)n uin


(4
m I
1J







0"
S00











0
e
8I)







B-8


0
4J









0






4)


.)














o
&



















0
-4
P4
Sp





0
0
-.4









In


-H


C









a















1
0



2
X
b:


01




4)













u I c En
-.- L H I
C O o 0 i
S0 k 1 4J 4 a,
> W 0 --- u 0
too M n V)
0 0
c t) e-- 9 '0 i i u ,C H

no 4 I ) 4o
MO 40 V 'V O O a -f, W
> W .-I
r U -A1 V > C, 0 Z
S. -H r r -H I0. r

C 40 0 1
C 4 O 0 0 0J E

0 0 0 0 0 0 0 c 0 0














tp
r.- 0.- H -
0.'! -) O 0 0.4) .- uJ4




























*0









--1
00 4 4 0


C
0 0 N
I 0 *4 U

0 -- 0 0 40
U0 00-H0
ct tp rl u4 U

0 u 0u C 4



.r 0 0 4
4J0 t W 44 0 -H
Z0 0 O0

*HZ0 U 0I C 4
44 -H 0 +4 0 44
r 0 4 0 -H r
c o o*H 0
*-4 P4 ) > 0
0 (U 04 -r r(
, 0 C 0 1 1

0 5 0 H
* 00 C 3
i-l B 'O l *-
S4 C 00 0 0
w r-4 *- 0
0 > E 04 0 4

- 0 4 V' d 0 '
0- M C 4O 9) 44






H C M

m u ** v

m 4 'O 1 )
Vs C C 0 C




co
coo
0
40-


0-I
4-1

d Q
.0 OH 00 r












In
-4
1-4
Cl





'0
0 0

-4
nm



'--4
O
0>
4a4





B) t-H
O 03 0









0'
my


0
u -r4
14 w
cH 0

00



.0

0 0C
0 4


H 0
M 0*

OC
0c00

0 0 )
us&o
Omc


*I
C C
C0
004
* o

"4
C, 0

0o

-4 o




















0 0
0& 0











0 '4 O 5
m0 0 -










r >OO 0 E4
0 .C 0





C 0
C> >0 e 0-
0 0 0. 0 -H



0 0 0

HO >u 0 -4
00>400
CO .- C 0-



00 0. C
0 0 4 --4










C,
''0











04
















0 >-






Cr -
00
























tal >1
01
























0.








-L
ID



cro
40















CO
40
0n

















04
0




















0
I$
























0


0















40








.H





4J


















0 4-
0
0n)










-4

0
0






04 4
0 )
4- 4d
U






8R .8










S
0A











S
Oi


0

04

4 -d

V 0404
.u 8~


B-9


*ri


0
o .
0 4








4- 4 0
Dc0
S-4












) ~ 0






.i





















0)0--t 0~-.
4 4 0 W 4




0 vO 0
oU W l


$4
0 1 0 0


440 m *t ) m C
h) 4) 0 .
0 00 ) 0 .1
S om c
S H W 4J 0


0 00 0 M (
0 O *r H c w*
ej -4 4. 4 0 4



















N


0


H





a





.0





0
-4


r-
0!
0nt ^?4 lc r
0 T
+J44 u-- ^ 0 6
0


'0
04
4-)


0

H








oC
. 0 t; '
C C C


400) m 4
U o -d 0,

- -4 *d 4) Z 0 N 04



49 U e a *





4) 0 Q 04 > 4
W > o 0 4.) 0




- M 4 O 0





E0o 4 V c 4 >-
co &MO C U l LO
c 4 40000 a 400











C 0
0444.00 nie 0'0 uo M


4-<4






044.)
040
>00




O)-4



>4 0 4

04 -H-

0 0 1
4 04
H 4
0D4100
10 4 i-










B-10
















o o
0 0 I
4J M




-O 4J U
0





S0

qo 4 4
0 40 O

U 4U 0-- U ,- M 0
0 -44 0 -3 ) i



0 0 4 r Q C n C O )






ea
S3


0 0 0
3 u r w -N -4 n
0. .H i to -


S0 0 4 0 0




* U 34 C
0- Ho 0'o OUH g0















4 --4 0
0 0
U U 400'H40 >4 -'


























0 OH OU
D 4'HU- UV 0* U

4-4






*0



o I r










0
I0 C
0 0 H
o 0'co r






El 0
00
0


0 W

0 H








H 4) 0
S .1 0al

















GClu 1
>.0 I n














0 0














4 H
00
0 M






I 0
sa >i 0






0 4.
~ HO U
0'

4.4H 0








B-l1


All team members participated in a training session before
departure for the field, to discuss the general study questions
and their specific project scope of work. Also, a general
report outline to guide report preparation was distributed.
Study team members used several variations of a rapid
reconnaissance methodology ranging from semistructured
questionnaires to unstructured individual, household, and group
interviews. Field data analysis was complemented with secondary
project data sources.

Interim reports were completed during July and August, and
reports were revised and edited during early 1986 to increase
the uniformity of presentation.


3. SYNTHESIS PHASE


The major objective of this study was to produce a
synthesis of findings that would summarize the project
experience of 10 years of implementation of A.I.D.'s Women in
Development program. The analysis for this phase included data
collected from the desk reviews and the field study phase. Two
factors complicated the synthesis phase:

1. The newness of the state of the art in the field of
women in development. Although the literature on women in
development has grown considerably in recent years, much work
remains to be accomplished, both at the theoretical level
(refining terms, concepts, and hypotheses) and at the practical
level (collecting data from projects that have economic
development objectives). During the synthesis of the study
results, considerable effort was spent clarifying concepts and
developing a conceptual framework within which the data already
collected could be analyzed. Thus, this synthesis represents a
process of inductive theory building. This process was
facilitated by the continuous informal exchange of ideas among
the study participants and through several, more structured
seminars organized for this purpose.

2. The intention to complete the synthesis document for
the Nairobi Conference concluding the U.N. Decade for Women. In
trying to meet a very tight schedule, work on the synthesis
began before consultants had prepared their final versions of
the field reports. Although in the long run this process
strengthened the analysis for individual field reports, it also
delayed the completion of the synthesis report because
individual authors had to be consulted on the accuracy of
interpretation. This interactive dialogue produced a very
useful refinement of the concepts and of the theoretical
framework used in the synthesis analysis.








B-12


The synthesis report includes both qualitative and
quantitative information on projects. The sectoral desk reviews
and the 10 field study reports provided the qualitative data.
The quantitative information resulted from the processing of
project data in a computerized data base.

The data base included the 98 projects originally sampled
for the sectoral desk reviews. Project information included
general characteristics (such as funding, length of project,
completion status, sectoral activities, type of project, and
achievement of project purposes and project goals) and more
gender-specific project characteristics identified through the
questionnarie coded by consultants for the sectoral desk reviews.

After the conceptual framework was refined into the model
presented in the synthesis paper (Section 1.4), questionnaire
items were recorded according to the seven major variables of
this model, using all the information available from desk
reviews and field reports. A description of the indicators used
to measure each variable is presented below.

1. Gender analysis, defined for this study as the
intersection between gender variables and project purposes and
goals, was measured using the following questionnaire items:

a. Are females mentioned in design documents?

0. No answer
1. No
2. Yes, general boilerplate/very scanty
3. Yes, at least some specificity/discussion
4. Yes, more

b. Do design documents mention women's directly productive
roles (subsistence/market)?

0. No answer
1. No
2. Yes, general boilerplate/very scanty
3. Yes, at least some specificity/discussion
4. Yes, more

c. Do design documents mention females' indirectly
productive role (household/nonmarket/human capital
formation)?

0. No answer
1. No
2. Yes, general boilerplate/very scanty
3. Yes, at least some specificity/discussion
4. Yes, more








B-13


(Note: Questions a, b, and c were also used to code information
from implementation and evaluation documents.)

d. Is there any information on women's directly productive
roles related to the project?

0. Not Applicable
1. No
2. Yes, general boilerplate/very scanty
3. Yes, at least some specificity/discussion
4. Yes, more

e. Is there any information on women's indirectly
productive roles related to the project?


Not Applicable
No
Yes, general boilerplate/very scanty
Yes, at least some specificity/discussion
Yes, more


Responses to these questions were combined and classified
into low, medium, and high ratings of gender analysis, as
follows:



Rating Question Category Within Question


Low a 1 or 2 (boilerplate)
b 1 or 2 (no or boilerplate)
c 1 or 2 (no or boilerplate)
d 1 or 2 (no or boilerplate)
e 1 or 2 (no or boilerplate)

Medium a 3 (some)
b 3 (some)
c 3 (some)

High a 4 (more specificity)
b 4 (more specificity)
c 3 or 4 (some or more)
d 3 or 4 (some or more)


2. Adaptation. Questionnaire items did not reflect this
concept. Adaptation within individual projects was therefore
rated as low, medium, or high according to whether project
components (i.e., implementing institutions, delivery systems,
or technical packages) were designed or redesigned to improve
women's participation in project activities.








B-14


For example, a training project in Nepal received a high
rating because it built a dormitory for female participants. A
water project in Tunisia received a low rating because women
were not consulted about the location of the wells nor were they
involved in the maintenance of the water system.

3. Participation. Following the conceptual framework,
participation was defined in terms of three factors: presence
of women in project activities, match between female
participation in specific project activities and gender division
of labor in the project area, and the importance of the activity
for the project.

Four items in the questionnaire were related to women's
participation in project activities:

a. Do the documents mention that women in the area are
participating as active agents in implementation?

0. Not Applicable
1. No
2. Yes

b. Do the documents mention that women in the target area
are participating in decision-making concerning the
project?

0. Not Applicable
1. No
2. Yes

c. Are women receiving any resources earmarked
for/targeted to them in the project design?

0. Not Applicable
1. No, they are not receiving any
2. Yes, but they are receiving less than the
targeted amounts
3. Yes, they are receiving approximately the
amount targeted
4. Yes, they are receiving more than the amount
targeted

d. Are any resources being delivered to/used by women that
were not specifically targeted to them in project
design?

0. Not Applicable
1. No
2. Yes

Based on responses to these questions, women's
participation was rated as low, medium, or high, as follows:








B-15


Rating Question Category Within Question


Low a 1 (no)
b 1 (no)
c 1 (no)
d 1 (no)

Medium a One of the four questions
b had a positive response.
c
d

High a Two or more of the four
b questions had a positive
c response.
d



The rating of women's participation in project activities
was supplemented by a qualitative assessment of the degree to
which women's participation matched the gender division of labor
in the project area and the importance to the project of the
activity in which women participated.

4. Achievement of project objectives was measured by the
responses to the following questions:

a. Do the implementation documents provide an overall
assessment of how the general project is faring?

0. Not Applicable
1. No
2. Yes, and it is characterized as proceeding very well
3. Yes, and it is characterized as proceeding
moderately well/mixed
4. Yes, and it is characterized as proceeding
marginally/not well

b. Do the results1 documents give an overall assessment
of project effectiveness/success (especially the extent
to which it accomplished the outcomes targeted in the
logical framework)?



1"Results documents" refers to mid-term and final evaluation
documents.








B-16


0. Not Applicable
1. No
2. Yes, characterized as very successful/very high
3. Yes, characterized as very high
4. Yes, characterized as moderate/somewhat mixed
5. Yes, characterized as marginal/very mixed/
unsuccessful

c. Can you give an overall assessment of project
effectiveness/success?


Not Applicable
No
Yes, very successful/very high
Yes, high
Yes, moderate/somewhat mixed
Yes, marginal/very mixed/unsuccessful


Projects received ratings of low, medium, or high for
achievement of objectives, as follows:



Rating Question Category Within Question


Low a 4 (marginal)
b 5 (marginal)
c 5 (marginal)

Medium a 3 (moderate)
b 4 (moderate)
c 4 (moderate)

High a 2 (very well)
b 3 or 2 (high or very high)
c 3 or 2 (high or very high)



5. Impact on women was measured by the responses to the
following questions:

a. Do the results documents assess the project's overall
impact on women?


Not Applicable
No
Yes, positive
Yes, mixed
Yes, negative