WORKING PAPER
An Analysis of USAID Assistance to Basic Education 
in the Developing World, 1990-2005
Foreword from USAID
It is important for USAID to undertake serious, rigorous reviews of its programs so that 
we can learn from our collective experiences and make improvements and adaptations to 
our programming approaches accordingly. This document is one of several publications 
by the Education Quality Improvement Program 2: Policy, Systems, and Management 
(EQUIP2) that attempts to provide useful analyses to support improved programming. 
This study is unique and valuable because it is a meta-evaluation of all USAID projects 
in basic education between 1990 and 2005. Whereas other EQUIP2 research takes 
a more in-depth look at USAID education experience in particular countries, or in 
particular technical areas of programming (e.g., complementary models of education), 
this meta-evaluation reviews all USAID basic education programming across this 15-year 
time frame, and across all regions. 
There are many useful findings in this study:
•	 Setting goals and targets: This evaluation found that projects often tended to be 
overly ambitious in setting project goals, objectives, and targets. This finding, and 
the actual historical data and trends noted in the study, provide a useful basis for 
setting realistic objectives and targets.
•	 Education quality and assessment of learning outcomes: Results from the meta￾evaluation indicate that USAID has had a fairly strong emphasis on quality and 
learning assessment for a number of years. Not all USAID interventions are intended 
to directly result in measurable changes in learning in the short-term (e.g. support 
for Education Management Information Systems or the provision of policy advisors); 
however there were fewer instances of reporting on changes in learning outcomes, 
including comparisons against baseline data, than were anticipated. This suggests 
that there should be increased attention to sound assessment of changes in learning 
outcomes in relevant projects.
•	 Incomplete archives: One of the most surprising findings is the significant number 
of gaps in USAID’s publicly accessible records for the education sector. There 
needs to be much greater attention by USAID staff and its partners to the universal 
requirement that all regular reports and evaluations be submitted to the USAID 
Development Experience Clearinghouse (DEC) so that the documentation is 
publicly accessible.
The Comments section at the end of this document provides additional contextual 
information and commentary that may be useful in reviewing the findings of this report. 
This section addresses issues such as broader changes in education in target countries 
during this time period, methodology, and the nature of project versus program support 
in USAID.
2
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
There has been remarkable overall progress in improving international basic education 
in recent decades and there is much that we can be proud of. At the same time, great 
challenges—and opportunities—remain, and this study demonstrates that there is still 
much that we can learn from our collective experience, and much that we can do to 
improve the effectiveness of our efforts. I am extremely grateful to the authors, David 
Chapman and Jessica Jester Quijada, for their initiative and perseverance in carrying out 
this important work. Together with other EQUIP2 publications, we hope that this study 
will be useful as a tool for reflection and action.
Patrick Collins
Senior Education Specialist, USAID/EGAT/ED 
CTO for the EQUIP2 Project 
3
Abstract
Between 1990 and 2005, the United States Agency for International Development 
(USAID) committed over $2 billion to encourage and support basic education systems 
in the developing world. What has been learned from this investment in education 
systems that might guide continued educational development efforts of the U.S. and 
other donors? To address this question, the authors reviewed USAID basic education 
documents for 33 projects across 23 countries implemented between 1990 and 2005. 
Key findings include the following:
•	 USAID can be proud of some excellent work and documented successes. Projects 
reviewed in this study contributed to raising education quality, extending access, 
improving persistence, and strengthening public and private institutions in numerous 
countries. However, a substantial number of projects had little documented evidence 
of program outcomes and impacts, and many projects of the period had limited 
readily accessible documentation of any kind. Fewer than half the projects reviewed 
had complete documentation that included initial, interim, and final reports; final 
evaluations were available for only a third of the projects. The lack of adequate 
documentation constrains USAID’s efforts to learn how to improve the quality and 
focus of its interventions.
•	 For the most part, USAID project designs were formulated in terms of education 
system outcomes (e.g., student achievement, access, persistence, learning) but most 
project documents available for this review concentrated on the extent project 
outputs (e.g., number of textbooks printed, number of teachers trained) were 
delivered. In some respects, this was a reflection of USAID operating practices 
during the period, as well as the nature of documentation reviewed. Regardless, while 
it is undoubtedly important to monitor the delivery of project-level outputs, it is not 
necessarily a guarantee that these inputs to the larger education system assure the 
success of a larger education enterprise.
•	 A review of anticipated outcomes and targets found unrealistically high levels of 
initial expectations, which made even reasonable levels of accomplishment appear to 
be inadequate. Results sometimes were over-promised at the project design stage and 
modest evaluation findings were sometimes overstated. Data indicating that projects 
had an impact on student learning were available in nine of 33 projects; where project 
students out-performed non-project students, gains ranged from 0.04 to 38 percent. 
Documentation of project impact on access, retention, and graduation was available 
in five of the 33 projects. Of these, there was evidence of improved persistence 
in three countries and improved grade retention in two, though the magnitude of 
impact was often modest.
•	 Findings regarding frequency and magnitude of learning gains and improvements 
in student flow are consistent with those found in World Bank primary education 
projects during this same period. Based on USAID and World Bank data, this paper 
offers estimates of gains in learning and persistence that would be reasonable to use in 
planning future education projects.
4
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Development assistance projects operate in complex political and economic 
environments, so success of even the best-designed project can never be assured. Results 
of this study indicate that USAID projects during the 1990s and early 2000s made 
important contributions to strengthening education systems, extending education access, 
and increasing student learning. USAID itself was also undergoing almost constant 
reorganization and reform during this period, which affected systemic continuity and 
documentation practices. If anything, the results of this review represent a conservative 
estimate of USAID successes since impacts may have emerged that were not indicated 
in the documents available for this study. Findings highlight the considerable challenge 
posed by efforts to influence the operation of national education systems. 
Introduction
Between 1990 and 2005, the United States Agency for International Development 
(USAID) committed over $2.3 billion to support and strengthen basic education in the 
developing world (Table 1).1
 This investment reflects both a widely held belief that 
basic education is a right, and recognition of the centrality of education to the economic, 
social, and political development of nations. Education has been linked to, among other 
things, better family health and nutrition, enhanced capacity of the poor to participate 
in the political process, and higher lifetime income (Birdsall, Levine & Ibrahim 2005, 
World Bank 2004a, Lockheed & Verspoor 1992).
Table 1: USAID Funding to Basic Education 1989–20052
Year Investment in $ millions
1989 64
1990 67.1
1991 67.1
1992 135
1993 135
1994 0 3
1995 135
1996 108
1997 98
1998 98
1999 98
2000 103
2001 102.8
2002 150
2003 216.6
2004 382.7
2005 412.8
1 Not all of the funding shown in Table 1 was for projects completed in or prior to 2005 – some of these projects were 
ongoing and extended beyond the scope of this review.
2 Figures for years 1989-1999 were provided by the Basic Education Coalition (through George Ingram) and represent 
appropriated funding to Basic Education; data for years 2000-2005 was provided by the Congressional Research Service and 
indicates actual USAID expenditures to Basic Education.
3 Lack of allocation in 1994 was due to delays in Congressional appropriations that year.
5
Despite the widely recognized importance of international development assistance, 
the effectiveness of that assistance has been the subject of controversy (Easterly 2007; 
Economist 1999; World Bank 1999, 2006). While much of that criticism is directed at 
international aid generally, a generous share of criticism has concerned the effectiveness 
of international assistance to education more specifically (World Bank 2006; Al￾Samarrai, Bennell & Colclough 2002; Government Accountability Office 2007; Nkansa 
& Chapman 2006). Critics question the success of past interventions to improve 
education and point to urgent needs in other sectors where the returns on investment 
are more certain and immediate. These concerns have led many development assistance 
agencies to undertake retrospective assessments of the effectiveness of their work. For 
example, recent studies have examined the impact of education projects sponsored by 
the World Bank (Nielsen 2006), UNICEF (Chapman 2002; Chapman et al. 2004), and 
CARE (Goldenberg 2001). 
The present study examined evidence from 33 USAID projects implemented between 
1990 and 2005 as the basis for addressing three questions: What interventions have 
USAID projects employed to strengthen and improve basic education in the developing 
world? How well have these education projects performed? And, what insight can 
be gained from these investments that might help shape future USAID funding of 
education systems?
The study was grounded in two conceptual models. At the individual project level, 
the analysis employed a goal attainment model (Rossi, Lipsey & Freeman 2003; 
Fitzpatrick, Sanders & Worthen 2003). Questions addressed within this perspective 
concern the extent to which intended goals and objectives were actually achieved, a 
framework consistent with the outcomes orientation advocated by USAID itself. At 
the organizational level, the analysis is grounded in Galal’s (2002, 2007) framework of 
educational reform, elaborated by Welmond (2004). 
Galal and Welmond posit three approaches to improving education practice and argue 
that effective reform requires adopting measures under all three approaches. Each 
approach places a somewhat different constellation of demands on education decision￾makers. In the engineering approach, education is seen as a production function 
in which the quantity, quality and mix of inputs determine educational outputs and 
longer-term outcomes. When outputs or outcomes are unsatisfactory, interventions 
involve increasing the quantity of inputs, improving their quality or changing their mix. 
In the organizational approach
4
 education is seen as a principal-agent problem. The 
principal (e.g., a ministry official) is interested in particular outcomes (such as good 
quality education), but has to rely on an agent (e.g., teachers) to obtain these outcomes. 
The challenge, then, is for individuals at one level of the system to get individuals at a 
different level of the system to act in desired ways. The political accountability approach
concerns the relationship between citizens and policymakers. This approach posits that, 
if citizens can influence decision-makers, politicians, and education managers in their 
formulation of education policies, setting education priorities and allocating resources, 
they can improve educational outcomes. This framework was used to assess USAID’s 4 Galal (2002) refers to this as the industrial organization approach.
6
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
breadth of approach to national educational reform.
The study contributes to international development in three ways:
1. First, though USAID currently emphasizes outcomes-oriented design and results￾based program management, it is often not clear what level of outcome represents 
a realistic target to be accomplished. An analysis of past experience can provide an 
empirical basis for assertions about what future projects can be expected to achieve. In 
the absence of empirical data, targets get set by guesswork that can easily be influenced 
by wishful thinking. Levels of prior success provide a useful indicator of reasonable 
levels of accomplishment in future projects. 
2. Second, the effectiveness of U.S. foreign assistance has been widely questioned. 
Within the constraints of the data available, this study provides a basis for assessing aid 
effectiveness and the extent to which such criticism is warranted. 
3. Third, because USAID project design is somewhat decentralized, often outsourced, 
and USAID personnel rotate jobs frequently, managers do not always have a clear 
sense of what specific development strategies are most widely employed or which ones 
proved to be most effective.
The context of USAID support for education
Any analysis of USAID accomplishments during the 1990-2005 period must be 
understood within the larger context of factors that influenced USAID decision-making 
during that era. Four factors are particularly important in understanding results of this 
study. 
First, the 1990 World Conference on Education for All (EFA) focused and galvanized 
world support for the goal of universal primary education within a framework that 
emphasized equity and highlighted the importance of learning outcomes. The 
international consensus emerging from this conference (represented in the EFA 
movement) was a powerful influence that led to dramatic increases in total donor aid to 
basic education during the 1990s and early 2000s. USAID funding for basic education 
more than doubled between 1992 and 1993, where it remained at relatively consistent 
levels until the early 2000s. USAID contributions to basic education then quadrupled 
between 2001 and 2005.5
Second, in 1987, USAID introduced program assistance (also called non-project 
assistance, or NPA) as a mechanism for distributing aid funds to recipient countries 
(Chapman & Dykstra 2006). NPA emerged in response to criticism in the late 1980s 
and early 1990s that “projectized aid” failed to promote local ownership, lacked 
sustainability, and fostered aid dependency.6
 Consequently, assistance to education 
was provided through a combination of program funds given directly to the receiving 
government provided they meet a set of conditions, and project assistance that typically 
5 Much of this increased funding was for education-related programs in Afghanistan, Iraq, and Pakistan, Most of these 
projects were not completed by 2005 and so were not included in this analysis.
6 The Development Fund for Africa, for example, was an entirely different structure for organizing aid to Africa which 
led directly to non-project assistance in most of the Sub-Saharan countries reviewed here, including Benin, Ethiopia, Ghana, 
Guinea, Lesotho, Malawi, and Namibia.
7
provided technical assistance to help the receiving government accomplish the larger 
reform. In 1998 Congress halted the use of NPA, largely over concerns that there was 
insufficient accountability for how NPA funds were used once they were transferred to 
recipients. 
Third, during the 1990s, influenced in part by the total-quality management movement 
(Deming 1984), USAID introduced internal reforms aimed at emphasizing outcomes￾based project design and results-based project management. Each USAID Bureau 
and Mission developed a set of strategic objectives and all projects and programs 
were justified and eventually evaluated in terms of their alignment with (and eventual 
attainment of) those objectives. Among other things, this led to new reporting protocol 
(referred to within USAID as the Results Review and Resource Request, or R-4, process), 
which was not always clearly linked to the project documentation, and a new system of 
archiving USAID documents (discussed later). 
Finally, USAID went though a period of ambivalence regarding the usefulness of 
program evaluation. For a period of three years in the mid-1990s it stopped funding for 
evaluation of its field projects. This left holes in the documentation during this period.
Method
This study was based on a review of USAID documents prepared as part of country￾based education programs implemented between 1990 and 2005 that could be 
reasonably retrieved through an aggressive search of USAID’s primary archival system, 
the Development Experience Clearinghouse (DEC). The sample included projects that 
started earlier but which ended within this timeframe. Documents were identified with 
the assistance of the Development Information Service (DIS), which from 1985 through 
2006 served as USAID’s principle research arm.7
 The DEC repository is the most 
comprehensive archive of USAID documents available. For the purpose of this study, 
the types of documents retrieved included design documents, intermediate and final 
reports, formative and summative evaluations. Table 2 reports the location, dates, and 
dollar value of projects reviewed for this study. Content analysis was used to categorize 
the nature and frequency of themes.
Note that during this era USAID provided both project and non-project (e.g., program) 
assistance.8
 In all instances examined here, countries that received USAID program 
assistance also received project assistance and available data did not support a separate 
analysis of the outcomes that could be associated with these different mechanisms for 
awarding aid. Since NPA was reported and tracked differently, this analysis focuses on 
the contributions of project assistance, as documented through available project reports. 
7 DIS was first known as the Research and Reference Service (R&RS); the name changed in 1995, under its third 
contract renewal.
8 For a further discussion of this distinction, see Chapman & Dykstra 2006.
8
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Table 2: Summary of projects reviewed for this study9
Country Project Name Acronym Dates Amount
Benin Children’s learning and equity foundations I & II CLEF 1995-2001 $11,169,442
Primary Education NGO project I & II PENGOP 1994-2003 $5,000,000
Bolivia Interactive radio learning project IRLP 1991-1996 $5,000,000
Botswana Basic education consolidation project BEC 1992-1995 $6,950,000
Cambodia Cambodian assistance to primary education CAPE 1995-2000 $30,000,000
Dom. Rep. Private initiatives in primary education PIPE 1990-1997 $5,850,000
Egypt Basic education Project BEC 1981-1994 $190,000,000
El Salvador Early Childhood Family Education Activity EDIFAM
1998-2002 $6,000,000
2003-2005 $2,600,000
Strengthening achievement in basic education SABE 1990-1998 $33,000,000
Ethiopia Basic education system overhaul project BESO 1994-2001 $30,000,000
Ghana Quality improvement project QUIPS 1990-1996 $35,000,000
Guatemala Basic education strengthening project BEST 1993-1998 $16,681,520
Guatemala Education Finance Policy Dialogue EFPD 2004-2005 $650,000
Guinea Fundamental Quality and Equity Levels Project FQEL 1995-2000 $20,000,000
Structural Adjustment of Education Programs PASE 1990-1995 $22,300,000
Haiti Incentives to improve basic education IIBE 1986-1993 $20,000,000
Haiti education 2004, results package ResPack 1996-2001 $20,000,000
Honduras Basic education and skills training BEST 1995-2000 $11,552,189
Primary education efficiency project PEEP 1986-1995 $23,878,000
Jamaica Uplifting Adolescents Project UAP 1996-2000 $7,063,563
New horizons for disadvantaged youth N/A 1997-2004 $10,200,000
LAC Center of Excellence for Teacher Training CETT 2001-2005 $20,000,000
Lesotho Primary education Project PEP 1992-1996 $4,444,134
Malawi
Girls attainment in basic literacy and education GABLE 1991-1998 $20,500,000
Education Sector Policy, Planning, EMIS Support 
Activities in Higher Ed Strategic Plan ESPP 2003-2005 $1,985,620
Mali Basic Education Expansion Project in Mali BEEP 1989- $21,000,000
Namibia Basic Education Support BES 1991-1996 $35,000,000
Nicaragua Basic education, Nicaragua BASE 1992-1998 $30,000,000
South Africa South Africa basic education reconstruction SABER 1992-2002 $55,300,000
Swaziland Education Policy, Management and Technology EPMT 1989-1996 $7,102,000
Uganda Support to Uganda Primary Education Reform SUPER 1992-2002 $25,000,000
Uganda Strengthening EMIS EMIS 1999-2004 $0
Zambia Communities supporting Health, HIV/AIDS, Gender, 
and Equity Education in Schools CHANGES 2000-2005 $0
TOTAL $733,226,468
9 CETT is the one regional project included in this study as it was seen as a direct intervention in support of teacher profes￾sional development. Dates and funding amounts are based on what was recorded in the project documents reviewed. The authors recognize 
that these figures may not be consistent with Agency totals. For example, based on data from the Africa Bureau (2001), $107 million was 
allocated to Benin as of FY2000 since 1990. A review of project documents only accounted for approximately 65% of this funding, or $69 
million for Benin during this period. Financial figures were not reported at all for either Uganda EMIS or Zambia CHANGES, which 
were both multi-million dollar interventions. While acknowledging the limitations of this approach, for consistency across all projects, the 
authors have included here the figures reported in the project documents available.
9
The document sample was limited to those that involved interventions aimed at 
strengthening the design or delivery of basic education in a specific country. Documents 
reporting on activities undertaken by centrally funded, multi-country initiatives that did 
not involve the introduction of specific educational interventions in specific countries 
were excluded.10,11 For projects included in the sample, document selection was further 
limited to those related to design, project descriptions, evaluations, project reports, 
and research studies. Hence, the sample excluded “products”, such as user manuals, 
textbooks, and training materials prepared for use with project participants. Overall, 
286 documents covering 33 country projects, with a total value of nearly one billion 
dollars, were reviewed. Of the documents reviewed, 11 were external final project 
evaluations, 17 were midterm evaluations or audits, 15 were administrative project 
completion reports, and 243 were design, contractual, or intermediate administrative 
documents (see Annex 2). Of the 33 projects reviewed, only 30 percent had a 
summative final evaluation that explicitly sought to answer questions as to whether 
the project achieved its objectives, or to capture substantive learning about effective 
education interventions. Half of the projects (17) completed a midterm evaluation, and 
only 15 projects (45 percent) had a final report on file in the USAID archives.
To assess the adequacy of the document sample, it is necessary to understand how 
USAID documents are archived. During 1990-2005, the U.S. government maintained 
two contracts aimed at creating, maintaining, and accessing information on USAID￾funded projects. The first, the Development Experience Clearinghouse (DEC), created a 
database that collected and archived project documents. The Development Information 
System (DIS) was developed under a separate contract to serve as a research arm of 
USAID. As USAID staff needed information on past USAID-funded projects they 
could request DIS to search the DEC archives. In conducting computer searches of 
information in the DEC system, DIS had privileged access to DEC and was able to 
search on keywords and search criteria not available to outside researchers. 
The DEC information retrieval system was constrained by three factors: (1) While 
USAID required that Mission and project staff submit documents to the DEC, it was 
enforced inconsistently; consequently, many documents were not submitted resulting 
in incomplete documentation for some projects and no documentation for others. (2) 
During the 1990s USAID suspended funding for project evaluation for a period of three 
years (discussed earlier) so the effectiveness of projects conducted during this time was 
not consistently assessed. Finally (3), in 1996 project-based coding was replaced with 
the current system of organizing and tracking based on USAID’s strategic objectives. 
Since (a) USAID missions had different strategic objectives and (b) basic education is 
not addressed by any single USAID strategic objective, the strategic objectives in which 
education projects would be located was not always clear. Together these constraints 
meant that USAID’s primary retrieval system was limited in both comprehensiveness of 
10 For example, documents reporting results of the multi-country research studies sponsored ABEL, EQUIP, and IEQ 
projects were not included in this analysis. However, documents from those centrally funded projects reporting on country-spe
-
cific buy-ins were included. 
11 It is important to note that the centrally funded projects USAID sponsored during this time produced some excellent 
research and conceptual work that is not included in this study since the current focus is limited to projects that were aimed at 
introducing a specific set of interventions aimed at changing the course of an education system.
10
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
the material it contained and the ability of users to access what was there. Even so, DEC 
remains the single best source of past project documentation available. DIS, on the 
other hand, no longer remains. In mid-2006 USAID discontinued the DIS contract and 
no research service is now available to mine the DEC data system. 
While this study did not include all USAID project documents produced by projects 
during this 15-year period, it did include all that were reasonably available through 
an aggressive search of the USAID archives. The strength of this approach was that 
it protects the replicability of the study.12 An independent search of USAID archives 
should yield the same sample of documents. 
Alternative sources for project documents are limited. Project documentation is 
generally not published in a format that gets retained by libraries. Such documents are 
produced in limited quantities and when the supply is exhausted, access is effectively 
ended. Documents may be retained on a more informal basis by the contractors 
involved in project implementation but these contractors have little reason to maintain 
long-term archives. These documents rather quickly enter the realm of “fugitive 
literature,” available only on the personal bookshelves of a few individuals who were 
interested in the project at the time. Had efforts been made to capture documents 
by searching the offices and bookshelves of development specialists, the yield might 
have been greater but the sample would have been ad hoc and impossible to replicate. 
Nonetheless, the difficulty encountered in document collection is the basis for one of the 
key findings from this study.
The study comes with six caveats. First, it is probable that more good evaluation was 
done than was documented. Second, some of what was well documented at the time 
of the original project may not have been contained in the DEC archive system and, 
consequently, would not have been included in this study. Third, within individual 
projects, evidence of impact and judgments of success may have varied over different 
stages of the project. Documentation about individual projects that was available to this 
study was, in some cases interim reports and presumably would not have captured the 
full range of project accomplishments. Fourth, project- and program-level data represent 
a snapshot of development impacts over a particular period of time. In many of the 
countries other donor interventions as well as other USAID projects preceded those 
included in this analysis, and may well have contributed to results evident during this 
era. Impacts (or lack thereof) often cannot be clearly attributed to any particular project. 
Fifth, in some cases the political and economic context of the countries in this study 
changed dramatically over the 15-year frame. Projects designed for one context ended 
up operating in a different context which sometimes affected their effectiveness. Sixth, 
it should be acknowledged that the importance assigned to long-term knowledge 
development within these projects varied and was not necessarily an explicit goal of 
USAID’s monitoring, evaluation, and reporting mechanisms. USAID’s focus was usually 
on immediate decisions for project management. Still, an analysis of experience across 
12 The weakness of this approach is that some well-known programs are not included, as the official documentation was 
insufficient.
11
projects across countries across years can yield some useful clues about effective practices. 
For all these caveats, the documentation reviewed in this study represents the most 
comprehensive basis for reviewing USAID project accomplishments that is reasonably 
available to any researcher using the USAID document retrieval system. 
Findings
Part I: USAID project goals and strategies 
What goals and objectives did USAID seek to achieve?: The stated goals and objectives 
of nearly all of USAID programs and projects during this period were (a) to improve 
education quality, (b) increase access, retention and graduation rates, (c) improve equity, 
and (d) do these things in a sustainable way. Other frequently stated objectives were 
to promote decentralization, strengthen non-governmental organizations (NGOs) and 
promote donor coordination.13 Since these represent the overarching outcomes that 
USAID sought to achieve, the analysis and presentation of findings were organized to 
examine the data with respect to each of these. 
Project design: USAID project design typically was based on considerable needs 
assessment and problem identification, usually in consultation with governments, 
Ministries of Education and other donors. Most projects approached project design 
from a systems perspective and were designed as multiple interwoven interventions. 
Projects were grounded in a belief that meaningful education reform depends on 
simultaneously addressing multiple constraints. For example, teacher training is largely 
wasted if newly trained teachers do not have the textbooks and instructional materials 
they need to be effective in their classrooms. These in turn depend on both the 
development of a curriculum, a textbook development process, and an effective textbook 
distribution system. Projects, then, were typically multi-component interventions 
in which the success of each component often depended on the success of other 
components.
What strategies did USAID fund?: Table 3 reports the level of the education system, 
primary intended beneficiaries, subject matter focus and strategies employed within 
these projects. For the most part, projects focused on the general population of students 
at the primary school level, though nearly a third had a special emphasis on promoting 
girls’ access achievement and completion. This emphasis reflects the consensus within 
the development community during the 1990s that a focus on primary education and, 
within that, on increasing access, was the most appropriate investment in education 
(Heyneman 2006). At least half of the projects involved a curriculum or instructional 
materials component that included a focus on math or language.
The 33 projects reviewed in this study employed a wide variety of interventions to 
accomplish their objectives, as illustrated in Table 3. The solutions that projects 
employed offer an insight into what USAID thought was the underlying problem they 
13 In some projects, decentralization and NGO development were presented as overarching project objectives; in others 
they were presented as strategies for accomplishing other objectives, such as improving quality or access. 
12
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
were trying to solve.14 The ten interventions most widely employed in the USAID 
projects fall into three categories: (1) capacity development (usually focused on 
teachers and MOE officials), (2) development or revision of curricula and provision of 
instructional materials, and (3) improving decision support mechanisms (e.g., EMIS 
systems, test results, monitoring, evaluation, operations research) (Table 4 and Annex 3). 
This suggests a widespread belief that the absence (or low quality) of inputs, and these 
inputs in particular, was the root source of constraints on education access and quality, a 
point discussed later.
How did education projects assess their success?: The dominant measure used as the basis 
for judging project success was the extent to which promised inputs were delivered. 
Annex 4 reports evaluation criteria by project while Table 5 provides a cross-project 
summary. Thirteen of the 33 the projects also examined criteria related to how teachers’ 
or students’ classroom behavior changed as a result of project inputs or activities (e.g., 
change in enrollment patterns, student flow through the system, teachers’ instructional 
practices, or changes in actual student learning).
Table 3: Summary of project activities by country (33 country projects reviewed)
Activity Frequency Activity Frequency 
Level
Early childhood 3 Primary 24
Secondary 4 Adults 3
Vocational 1 Higher/tertiary education 2
Population
General 17 Girls 8
Out-of-school youth 4 Ethnic minorities 3
Private schools 2 Special education 2
Subject matter focus
Math 13 Language 13
Social studies (& civic ed) 4 Health/HIV/AID 3
Science 3 Environmental education 2
The review of documents indicates that many USAID and project staff were more 
oriented toward tracking project outputs (which generally were system inputs) than 
documenting system outcomes.15 To a considerable extent this was to be expected, 
since USAID and project staff were required to track inputs as part of their ongoing 
monitoring and reporting functions. Nonetheless, attention to outputs was thin. 
One seeming reason for this input-orientation is that converting project funds into 
educational inputs (e.g., textbooks, teachers, data management systems) was most 
14 The identification and rationale for the larger problems that USAID sought to address through these interventions is 
often contained in planning documents and annual reports prepared at the USAID Mission level, which were not incorporated 
into the scope of this study.
15 In some cases, there appeared to be a confusion of inputs with outputs. This occurred when the number of teachers 
trained, textbooks delivered, or workshops conducted were presented as outputs rather than being treated as inputs toward a larger 
goal. For a discussion of inputs versus outputs, see Windham & Chapman 1990.
13
directly under the manageable control of project staff. The extent that these inputs 
subsequently performed effectively once deployed was viewed as beyond their ability 
to control. Nonetheless, this orientation led some Mission and project staff to declare 
project success without really assessing the consequences of the training, materials, and 
structures on teachers’ classroom practices or students’ learning. For the most part, 
USAID project designs were formulated in terms of outcomes, but most monitoring and 
evaluation efforts concentrated on inputs. These findings must be interpreted within the 
context of the work and the times; the focus on project outputs rather than on education 
system outcomes was, in part, a response to USAID reporting requirements and, even 
more broadly, Congressional interest.
Table 4: Ten most widely used strategies in USAID education projects, 1990-2005
Focus of projects Number of projects
Teaching materials
Curriculum development 21
Instructional and materials development 19
Education materials distribution 15
Capacity development
Teacher training 20
Capacity development – MOE 16
Data for decision making
Operation/policy research 17
Achievement testing 16
Test development 15
EMIS development 15
Monitoring and evaluation 14
Table 5: Criteria used to judge project success
Criteria used to judge project success Number of Projects
Extent of promised inputs delivered 22
Change in enrollment 9
Change in student flow (repetition, drop-out) 7 
Change in student or teacher behavior 6
Student learning 10
One consequence of the dynamic nature of the political and economic environment 
of the countries in which USAID worked was that the logic of initial project designs 
sometimes needed to be revisited once a project was underway. Occasionally, project 
designs seemed to be grounded in questionable logic that linked the project intervention 
to the intended goal. For instance, in Honduras (PEEP) there was disagreement between 
the evaluation team and the USAID Mission about the logic of the assumptions 
14
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
underpinning the project design (USAID/Honduras, 1993). In the view of the 
evaluators, the pre-specified indicators of project success were misaligned with the actual 
project activities. In particular, the evaluation team asserted that the design assumption 
that improving textbooks would improve efficiency of education system was faulty. 
They argued that success in the textbook component would not necessarily affect those 
indicators, a view rejected by the USAID staff.
In still other cases there appeared to be a loose linkage between initial design and 
subsequent implementation. In Malawi, an evaluation attributed increases in girls’ access 
to schooling to attitudinal changes created by the project, but seemingly ignored the 
possibility that the increase resulted from a recent government policy abolishing school 
fees which was widely credited by other observers for the enrollment increases (Hebert, 
Millsap & El-Sanabary 2002). In Ethiopia, the evaluation team found a disconnect 
between the BESO reform objectives and elements of the BESO project (Sommers 
1996). The evaluation concluded that the allocation of BESO program (NPA) funds and 
project resources were not always clearly linked to BESO purposes and objectives. While 
at times these issues might reflect weak project design, they also highlight the efforts 
of USAID to assess the success of its efforts in ways that would allow for mid-course 
correction.
Part II: Were goals and objectives achieved? 
As stated earlier, the most frequent goals and objectives of USAID projects were: (a) to 
improve education quality; (b) to increase access, retention and graduation rates; (c) to 
improve equity; and (d) to do these things in a sustainable way. USAID often expressed 
a desire to promote decentralization (e) and strengthen NGOs (f). In some cases, 
promoting donor coordination (g) was also indicted as a desired outcome of the project. 
The next section examines the extent to which these were achieved.
A central output of most projects was the delivery of educational materials and training 
that, in turn, represented inputs to the country’s education system. The following quick 
sample of project output measures illustrates the nature and scale of some of the USAID 
efforts in this regard:
•	 Honduras PEEP project: 5,000,000 textbooks developed and distributed; 25,000 
teachers trained; 600 classrooms constructed and equipped and another 500 
renovated. 
•	 Honduras BEST: 223,897 students attended alternative schools; 29 NGOs 
strengthened in collaborative work. 
•	 Swaziland EPMT: 144 publications created and distributed, 6 master’s degrees, 7 
study tours, 121 trainers trained, 4,500 school personnel trained. 
•	 Dominican Republic PIPE: 4,466 teachers trained, 70,300 textbooks and 
instructional sets distributed, 329 school directors trained.
•	 Benin PENGOP: 1,300 PTAs supported, 635 school infrastructure projects. 
•	 Bolivia IRL: increased access to math programs in 700 schools, 150,000 students 
served, health programs in 345 schools serving 60,000 students. 
15
The ultimate goal of USAID education activities, however, needs to be assessed in terms 
of the extent these system inputs were used and made a difference in the experience of 
teachers and students.
(a) Improving Education Quality—Achievement testing: In its most common use, 
education quality refers to the extent that an education system is able to achieve 
the generally accepted goals of education, central to which is knowledge and skill 
development (Adams 1998; Anderson & Burns, 1989; Chapman & Adams, 2002; 
Chapman et al. 2005; Craig et al. 1998; Fuller 1987; Lockheed & Verspoor 1991). 
Across the 15 years of USAID education projects, improving education quality was 
the stated priority of nearly all projects. As Table 4 indicates, improving quality was 
generally defined in terms of increasing student learning, though some projects defined 
it in terms of improved system efficiency, including management or administrative 
structural changes or training, or increased capacity for actors within or external to 
the educational system. Where student learning was the explicit goal, achievement 
testing was the most widely advocated way of measuring whether such learning actually 
occurred. Based on available data, testing systems were a part of over half of the projects 
reviewed (Table 4).16
Project efforts to assess student learning are summarized in Annex 5. Of the 33 projects 
reviewed, 28 identified raising educational quality as a stated goal. Of those, 17 
proposed student learning as a key indicator of educational quality. Of these, available 
data suggest that 12 actually measured student learning, of which 11 used pre- and 
post-testing to assess change over time. Of these eleven, nine designed the testing in 
a way that would support claims that student learning changed over time as a result of 
the project intervention (e.g., used a comparison group). Of these, five projects found 
meaningful increases in learning as a result of the project intervention. Another three 
found mixed results, in which project students out-performed non-project students 
in some subject areas and under-performed non-project students in others. In one 
additional case the results were not clearly enough reported to interpret.
This summary suggests that, though raising education quality was the main goal of 28 of 
the 33 education projects reviewed, only 12 actually assessed the extent to which project 
activities led to desired student learning outcomes and only six were able to demonstrate 
clear success in increasing student learning (three others yielded mixed results). While 
it is possible that a larger number of these 28 USAID projects resulted in student 
learning gains, documentation to support that speculation is not available. Development 
specialists wishing to learn from past experience are limited to the documentation that is 
available.
The limited attention to learning outcomes, despite assertions about the importance of 
increasing the amount of student learning, has defenders. Testing is expensive. Setting 
up special testing systems to assess the impact of a project takes money away from 
direct delivery of services. There is often a tension between spending project resources 
16 Readers are reminded that testing systems might have been employed in other USAID projects but the results not 
archived in DEC.
16
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
to assess progress and spending project resources to make progress. There also may 
be a perception that donor-created testing systems are a wasteful duplicate of existing 
government efforts, even when government tests do not specifically test what the 
interventions were designed to teach. 
Where achievement was assessed with some degree of rigor, results indicate that a 
number of projects achieved remarkable gains in student learning (Annex 6). Ghana 
QUIPS provides an excellent example of a well-designed study of achievement gains 
using a pre-post comparison group design that supports claims that the gain was due 
to the project intervention. In the first year of the project, in both English and Math, 
students in project schools out-performed students in comparison schools by about 2 
to 4 percent. Over two years that achievement gain for projects students grew to 38 
percent in Math and 15 percent in English. Moreover, the rate of learning was faster 
in project schools than in non-project schools. In most situations, students in project 
schools maintained or even improved on achievement gains made as a result of project 
interventions even after leaving the program.
The evaluation of the Guatemala BEST radio Math project indicated a 16 percent gain 
in test scores (a proxy for the amount learned) in the first year of the project. Similarly, 
the El Salvador Strengthening Achievement in Basic Education (SABE) project also 
demonstrated respectable achievement gains over the life of the project (Table 6). In 
reporting on a project in Swaziland, evaluators observed that in Grade 1 Math the 
median score rose from 34 to 37 between 1992 and 1994, with the number of students 
achieving mastery (defined as 80 percent correct) rising from 36 percent to 51 percent 
over three years. In Grade 1, English mean scores rose from 19 to 24 over that same 
period while the proportion of cases in the two highest score categories increases from 12 
to 34 percent. Results from Grade 2 Math and English show similar increases in mean 
scores and in proportions of student scores at the high-end of the score distribution. The 
evaluation team concluded that the introduction of continuous assessment in Grades 
1 and 2 made a difference in learning outcomes for students in Swaziland (Clark & 
Pearson 1996).
Table 6: El Salvador SABE project: National achievement gains over life of project17
Subject 1994 1995 1996 1997
Third grade language 19 20 19 28*
Third-grade math 38 38 40 50*
*smaller sample used in 1997
A number of projects start with a clear intention to monitor student achievement gains 
only to encounter difficulty, usually around collection of baseline data. For example, 
in the Honduras BEST project, the project team acknowledged the importance of 
achievement testing, but documentation suggested that no testing was done. In the 
Honduras PEEP project, the plan called for student achievement testing as a basis 
for project evaluation, but this was not accomplished. The project team encountered 
17 Source: AED Final Report: SABE Project (1998)
17
a problem when initial student tests were administered too late to provide baseline 
measures thereby losing the opportunity to assess achievement gains over time. As the 
project had not carried out the proposed standardized testing of students, it lacked data 
on the impact and quality of the project (USAID/Nicaragua 1995: xv).
In some cases, tests were developed as deliverables under the project, but then not used 
in the evaluation of the project’s success, for example Botswana, Dominican Republic, 
and Lesotho. When this occurred, it was often because the time needed to develop the 
test resulted in missing baseline data collection.
Evaluation reports sometimes emphasized gains even when they were small and even 
when they did not consistently surpass non-project student gains. Similarly, some 
presented data in ways that implied greater success than was actually achieved. Authors 
tended to “round up”, putting the most positive spin on their results. While that may be 
understandable, it sometimes is misleading. In a Jamaica project, one report concluded 
that all students in project and non-project schools showed gains during the life of 
the project. The evaluation report concluded that students in project schools made 
considerable improvement (Dye, Jennings, Lambert, Hunt & Wein 2002:35). Yet, in 
fact, students in project schools did not necessarily outperform non-project schools. 
Some might view this claim as misleading since it implied project success even when 
there was no evidence that project students did better. Another interpretation of the 
results is that project students would have done equally well without the project.
In a different Jamaica (UAP2) project, evaluators noted that achievement data indicated 
that most project students improved by at least one grade level in Math (Simpson, 
Harris, Schubert & Yusuf-Kalil 2003). While true, that minimizes the finding that, 
while 55 percent of the project students increased at least one grade level, 63 percent of 
the non-participants increased at least one grade level. Such claims may be accurate but 
are misleading and are not helpful in creating a solid knowledge base about interventions 
that work. Projects that yield minimal, though positive, achievement gains beg the 
question of whether the payoff was worth the cost, an issue virtually never addressed in 
project documents.
In Malawi, achievement in project schools in English, Math, and Chikewa appeared 
to be impressive. In Math, scores of student in project schools increased by 17 percent 
over the course of the project. However, when comparisons were made to non-project 
schools, the interpretation was more mixed. For example, in Grade 5, project schools 
out-performed comparison schools in Math, only slightly out-performed them in 
English, and were comparable in Chikewa (Hebert, Millsap & El-Sanabary 2002). The 
evaluation report emphasized that project schools out-performed non-project schools in 
most grades and years, but minimized the narrow margin of difference at some of those 
grades. 
Finally, lacking achievement data, several projects fell back on anecdotal and subjective 
judgments of teachers about whether students were learning. While such judgments may 
18
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
or may not be accurate, the validity of such claims is often suspect since those making 
the judgments often had self-interest in the project appearing to be successful. One 
example of the fall-back to subjective judgments is the South Africa DDSP project. 
Overall, the experience of the projects included in this study indicates a wide variation 
in success of efforts to improve student learning outcomes. While USAID projects 
did contribute to meaningful achievement gains in some countries, that was not the 
dominant outcome. The message of experience is to be “modest in your expectations,” 
more rigorous in assessing learning gains, and more candid in reporting the results.
One way to help interpret these findings is to compare student learning gains in USAID 
projects to those of other international development organizations. In particular, 
between 1963 and 2004 the World Bank undertook 188 projects in which at least half 
of the funds were allocated to primary education (World Bank, 2004b, 2006). An in￾depth analysis of a stratified random sample of 33 completed primary education projects 
yielded findings similar to those found in the present analysis of USAID projects (Table 
7). About the same proportion of World Bank education projects identified education 
quality as a goal, though fewer of those projects selected student achievement as an 
indicator of project success (11 vs. 18), actually measured student learning (7 vs.12), and 
or utilized a comparison group. Of the 27-28 projects in each of the two studies which 
sought to improve education quality, about the same number of USAID and World 
Bank projects found evidence of increased student learning. 
Table 7: Assessment of learning outcomes in World Bank primary education projects18
Extent that learning was assessed # of USAID projects # of WB projects
Number of projects reviewed 33 33
→ Of those, the number in which raising educa￾tional quality was a stated goal
27 28
→ → Of these, the number in which student learn￾ing was proposed as a measure of project success
17 11
→ → → Of these, the number which actually 
measured learning outcomes 
12 7
→ → → → Of these, the number in which
•	 a comparison group was used
•	 only at gains over time of project students was 
measured
9
1
2–3
4–5
→ → → → → Of these, the number of projects 
that found 
•	 an increase in student learning among project 
schools
•	 mixed results
6
3
7
0
18 Adapted from World Bank (2004). Primary Education Portfolio Review, Background paper for the evaluation of the 
World Bank’s support to primary education, Washington DC, p. 25.
19
The magnitude of the achievement gains in the World Bank projects is similar to 
those in USAID projects (Annex 7). In three of the five World Bank projects (China, 
Mexico, Uruguay) achievement gains over three to seven years of project activity ranged 
from 0.7 to 6.0 percent. In one country (Chile), gains ranged between 3.4 and 19.2 
percent, depending on the subject area and type of school. A project in China showed 
more dramatic gains (16-40 percent) but methodological issues make interpretation 
problematic (World Bank 2002, 2006). In general, findings from both organizations 
suggest that a realistic estimate of multi-year achievement gains in a basic education 
project is around 2 to 6 percent, depending on context, subject matter, and grade level.
(b) Improving student access, retention and graduation rates: Along with improving 
quality, increasing education access was a central goal of eight projects. Retention and 
graduation rates were recognized as important extensions of access and tracked in seven 
countries (see Annex 4).
 
•	 In El Salvador (SABE) primary net enrollment increased from 74 to 79 percent 
between 1991 and1995 while repetition dropped from 8.03 to 6.10 percent. 
•	 In Guatemala, both Nueva Escuela Unitaria and Intercultural Bilingual Education 
(IBE), project schools had a higher percentage of first graders continue to Grade 
2 than in the comparison schools. In both cases, this effect was sustained and 
translated into higher percentages of student persistence to Grade 6. The difference 
favored NEU children entering Grade 6 by 13.4 percent for boys and by 7.9 percent 
for girls; in IBE, these differences were 4.2 percent for boys and 2.9 percent for girls. 
•	 For Honduras PEEP, primary graduation graduates rose from 50 percent of the 13-
year-old population in 1985 to 70 percent in 1994, a change that is probably due in 
part to the efforts of the PEEP project. 
•	 Yet the BEST project, also in Honduras, yielded mixed findings. Retention rates met 
project targets at existing centers; however rates were 17 percent below anticipated 
targets for new centers.
Results were mixed in other projects as well. In Ethiopia, gross enrollment rate (GER) 
in the two target regions was higher than the national average. In one of those regions, 
the increase was meaningful but in the other the increase hardly surpassed the national 
average (Caribbean Resources International Consult, Inc 1998). A summary of these 
past accomplishments provides a basis for establishing benchmarks for future project 
efforts to improve student flow (Annex 8).
Grade repetition rates in Malawi declined from 18 percent to 15 percent between 1994 
and 2000, though the overall primary dropout rate actually increased by 1 percent. 
Further, while the education portion of national budget increased by 15 percent between 
1992 and 1995, expenditure per child decreased due to dramatic enrollment increases 
(Hebert, Millsap, & El-Sanabary 2002). The gains in student flow were impressive 
but the overall trends might still be seen to be downward. Yet the conclusion of the 
evaluation was that “these results demonstrate that the original goals and objectives 
of [the project] have been addressed and met to an appropriate and adequate measure 
20
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
of success. As such, an important impact on the whole primary education sector and 
system has been affected” (Hebert, Millsap, & El-Sanabary 2002:11). Some might see 
this conclusion as overstated.
While increased access, persistence and graduation were goals in their own right, 
several projects explicitly used enrollment growth as a proxy for improved educational 
quality. The assumption was, presumably, that enrollment is quality sensitive. When 
families believe that instructional quality is improving, they are more willing to absorb 
the direct and indirect (e.g., opportunity) costs associated with having their children in 
school. This argument is complicated by three factors. First, greater access can depress 
student flow statistics as more academically average students enroll, which sometimes 
lowers average achievement rates and drives up grade repetition rates. The experience of 
Malawi and Uganda following the introduction of universal free education illustrate the 
downward effect on quality posed by rapidly increasing access. Second, governments 
can manipulate graduation rates by altering the pass rates on school leaver examinations, 
as was the case in pre-civil war Liberia (USAID 1988).19 Finally, progression and 
graduation rates depend on the extent that tests used for grade promotion and cycle 
completion are valid and reliable measures of what the curriculum actually teaches.
Again, a comparison with World Bank experience provides a benchmark useful in 
assessing the USAID experience (Annex 7). Across the USAID and World Bank 
projects, declines in multi-year grade repetition rates ranged from 2.3 to 7 percent (with 
the exception of the USAID project in Honduras where it was reported to decline by 55 
percent) while declines in student dropout rates ranged between 1.5 and 9.2 percent. 
Findings from both USAID and World Bank projects suggest that a realistic estimate of 
a multi-year improvement in progression (or conversely, a decline in dropout) rate might 
be in the 4 to 8 percent range, depending on context and grade level. 
Improved completion rates across the USAID and World Bank projects ranged from 2.9 
percent to 13.4 percent. This suggests that promises during project design of completion 
rates above about 14 percent are likely to be wishful thinking. While these estimates are 
offered only as general benchmarks, grounding promises of future project achievements 
in evidence of past accomplishments offers a sensible approach to project design. 
(c) Changes in teacher or student behavior: While increasing learning may be the desired 
outcome, most projects try to do that by changing teachers’ and students’ classroom 
behavior. Consequently, some argue that measuring teacher and student behavior is a 
more valid indicator of project success than is actual student achievement, since these 
behaviors are more directly linked to project activities. Achievement, they argue, is 
influenced by other factors beyond the direct control of the project, such as student 
aptitude and family circumstances. Yet few projects tracked changes in teacher or student 
behavior to ensure that project activities were having their intended impact. One likely 
reason is that assessing teachers’ and students’ classroom behavior is labor intensive and 
expensive. Nonetheless, some projects were successful in doing this. For example, in 
19 Note that the Liberia example was not drawn from the documents reviewed for this study, but from a broader review 
of the literature conducted as part of this research.
21
Guatemala, evaluators found improved turn-taking and collaborative behavior among 
students (Nieves, Lundgren, Bezmalinovic, & Matute 1994).
In Jamaica, the project was successful in changing classroom environments to facilitate 
learning. However, pedagogical approaches that emphasized participation of the child 
in a variety of learning opportunities had not been implemented. Instructional delivery 
in project schools remained highly traditional. The concentrated effort by the Jamaica 
project to provide hands-on professional development and other technical assistance at 
the school level, at the point of the midterm evaluation had not yet shown an impact on 
student performance or teacher behavior (Chesterfield, Enge, & Simpson 2002). 
In Nicaragua, a midterm evaluation found that teachers were in favor of the child￾centered curriculum but did not know how to apply it within a decentralized setting 
(USAID/Nicaragua 1996). The guides and manuals for first and second grades had 
been distributed to a large number of schools but the evaluation team was doubtful that, 
if teachers were not shown how to use them, they would ever be used. The evaluators 
observed that the 50,000 participant days of training were single-goal oriented (to train 
teachers in the use of the guides, etc.), sometimes theoretical, and quite expensive. It 
was one-shot training which lacked follow-up. (While documentation from subsequent 
evaluations was not available, anecdotal information suggests that the project made 
necessary adjustments and was ultimately successful.) Across projects, the main lesson 
from experience is that changing teachers’ and students’ classroom behavior is difficult. 
(d) Sustainability: A criterion of success in many projects, in addition to positively 
influencing access, retention, achievement, and graduation, was that the interventions 
be sustainable. The emphasis within USAID on sustainability is, in part, an offshoot 
of its increased focus on outcome-based funding, e.g., outcomes are of little value if 
they are only transient. It is also a reaction to the criticism that there has been little to 
show from the U.S. investment in assistance to education systems over the last 30 years. 
Indeed, the lack of sustainable impact is widely seen as a key threat to continued flow 
of international development assistance (Picard & Garrity 1997; Nkansa & Chapman 
2006). 
However, one of the challenges in assessing sustainability is the diversity of views about 
what should be sustained (Nkansa & Chapman 2006). What is meant by sustainability 
in international assistance projects is often poorly defined. Hence, it is not clear whether 
the goal is that: (a) specific organizational structures established by a project (e.g., parent￾teacher associations, interactive radio instruction) remain after the funding ends; (b) 
capacity of participants increases, regardless of the fate of particular structures; or (c) 
the overall economic health of a country improves due to the aggregate impact of donor 
assistance.20 Without a clear definition, it is difficult to determine the extent that donor 
projects have actually been successful in terms of sustainability. 
20 For further discussion of the meanings of sustainability see Nkansa and Chapman 2006. Authors discussing sustain￾ability as capacity development include Dahl 1995; Bhat, Cheria and Edwin 1999; Harris 2000; Picard and Garrity 1997). 
Capacity viewed as the overall economic health of a country is discussed by Center of Excellence for Sustainability 2001; Harris 
2000; Daniels 2000. 
22
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
In its simplest terms, sustainability is the ability of an activity or system to persist 
(Dempster 1998). Within some USAID projects sustainability is defined in terms of the 
extent that specific activities and structures created during the project continue beyond 
the life of the funding (Nkansa & Chapman 2006).21 The most persuasive assessment of 
sustainability would emerge from longitudinal or retrospective analyses of what actually 
persevered after external funding ended. 
Since USAID project designs and evaluations have generally failed to either define or 
measure sustainability, no substantive conclusion can be made about program success in 
this area. However, the judgment of project staff during implementation offers a proxy 
measure of what is likely to follow. Across many projects, there were strong doubts 
expressed by project personnel about the prospects for sustainability without continued 
external donor funding, though the reasons tend to differ slightly from country to 
country (Annex 9). Availability of local funding to pick up costs was the largest (but 
not the only) factor inhibiting continuation of these projects. Management capacity of 
government and lack of clear ownership of project activities were also cited. The overall 
conclusion from this review is that the likelihood of sustainability was not seen as an area 
of particular success.
(e) Decentralization: Greater decentralization of responsibility and authority for 
education decision-making (and often financing) has been an increasingly popular 
component of USAID (and other donor) projects. During this 15-year period, six 
projects incorporated some explicit attention to decentralization (Annex 3). Since final 
reports of these projects were not always available, the overall success of decentralization 
efforts is difficult to determine. However, among the observations offered in the available 
project documentation were the following:
•	 In Ethiopia, BESO project staff observed that decentralization improved the planning 
capacity of the MOE departments and improved the management functions of target 
regional education offices (Caribbean Resources International Consult, Inc 1998). 
•	 In Uganda the decentralization of primary education created a cleavage between 
the still centralized teacher training system and the teacher related functions of the 
district education offices (Moulton 2000). 
•	 In Nicaragua, a midterm evaluation found that after two years of implementation 
of the autonomous school approach, the reforms had reached fewer than 500 
schools and had not yet created solid community support for basic education 
(USAID/Nicaragua 1995); although the subsequent decade of intensive USAID 
and World Bank support for the reforms was not captured in the available USAID 
documentation. 
•	 The midterm evaluation of the SABE project in El Salvador indicated that the 
decentralization plan was not working as intended. Establishment of regional and 
sub-regional ministry offices may have instead reduced administrative efficiency as 
bureaucratic structures became obstacles for principals and teachers in making their 
voices heard (Management and Business Associates, Inc. 1994). How this concern 
21 The Nkansa and Chapman (2006) citation, while based on a USAID project during this period, is not drawn from 
the documents reviewed for this study but from a broader review of the literature conducted as part of this research.
23
was addressed is not indicated in the documentation that was available to this study. 
In some countries, the fate of decentralization depended on those at decentralized 
levels with new-found decision-making powers to assign priority to education over 
needs in competing sectors. The evaluation team in Uganda observed that the push 
for decentralization was threatened by the sustainability of the textbook component of 
the SUPER project. At the point of the formative evaluation, the evaluators concluded 
that the fate of textbooks would depend on how well the local priorities match national 
policies. If the finance ministry were to allocate block grants to districts, the district 
councils would have the power to allocate funds within the granted amount. However, 
“there is no guarantee that the district council will award priority to primary education, 
or within primary education to instructional materials. This could seriously jeopardize 
SUPER’s policy objectives” (Guild, Moulton, Wolf, DeStefano, J., & Wandira 1995:48). 
Further documentation was not available to determine how USAID and the Uganda 
Ministry of Education dealt with these challenges.
In other countries, USAID projects included strategies closely linked to decentralization 
which were intended to strengthen local involvement in decision-making. For example, 
Ethiopia, Ghana, and Haiti all introduced mini-grants programs, in which schools 
could compete for funds for innovative activities aimed at improving school practice. 
The midterm evaluation in Ethiopia indicated that the mini-grant program appeared to 
have been well received (Caribbean Resources International Consult, Inc. 1998), but 
an adequate assessment of the design and effectiveness of mini-grant programs was not 
documented in the USAID evaluations reviewed. 
The important point in all of these cases is that substantive education reform programs 
require on-going adjustment and continuing problem-solving. None of the conclusions 
reached in the above examples represent final judgments about the effectiveness of either 
the project or the reform, but rather represent findings that should inform subsequent 
reform support strategies. The lack of on-going assessment and documentation severely 
limit the interpretation and utilization of the reports.
(f) Non-Governmental Organizations (NGOs): During this era, USAID projects sought 
to strengthen the role of NGOs within the education sector. In some countries, there 
were political reasons for wanting to work through NGOs. For example, virtually 
all USAID project work in South Africa during apartheid was channeled through 
NGOs as a way to avoid working with the apartheid government. Often, however, 
the interest in working with NGOs was because they generally had strong ties to the 
local community. Hence, strengthening NGOs was viewed as a way to simultaneously 
promote decentralization and community participation. For example, the BEST project 
in Honduras worked with 24 NGOs and up to 12 municipalities to offer 250,000 out￾of-school youth and young adults basic education programs through existing Ministry 
of Education adult education centers. The IIBE project in Haiti sought to strengthen 
private provision of primary education as an alternative to a low-quality public system.
24
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
A number of projects reported success in strengthening NGOs but, in many cases, 
expressed doubt that they were yet at a stage where those NGOs could operate 
independent of external funding. For example, in Benin the evaluation found that 
national NGOs were meaningfully strengthened, but project staff still expressed concern 
about the capacity of the NGOs to carry the work forward on their own, due largely to 
uncertainty about financial capability (Mamadou 2003). Similarly, in the Dominican 
Republic the NGO funded in this project was not operating strategically and was 
uncertain of how to replace funds when the project ended (Bernbaum & Locher 1997). 
The evaluation concluded that the NGO was seen as an elitist organization by local 
stakeholders. 
(g) Donor coordination: In virtually all countries, USAID was only one of many 
international organizations working to improve education. It was not unusual for the 
World Bank, DfID, JICA, UNICEF, CIDA and others to have projects underway in 
the education sector that sometimes overlapped or even competed with USAID for 
Ministry attention. Evaluation reports seldom documented positive experiences in donor 
coordination, perhaps because when things were working as planned it was considered 
unremarkable; or perhaps because donor coordination was not an explicit objective at 
the project level. Often donor coordination is the responsibility of entities at a higher 
level than projects. While USAID is philosophically committed to effective coordination 
among these organizations, the operational experience of project staff was sometimes one 
of frustration. 
An evaluation of a project in Benin for example found that USAID and the World 
Bank were running competing projects, with the USAID effort in direct competition 
with the World Bank-initiated parastatal. USAID staff complained of confused lines of 
authority and competition for MOE attention. In Ghana, the World Bank was declaring 
considerable success and impact in solving educational problems while the main USAID 
project was designed under the premise that these problems were practically unaddressed 
(Education Development Center 1994). Project staff in Honduras (PEEP) expressed 
concern that the project was vulnerable to derailment when the USAID project ended 
and other donors such as the World Bank and Germany replaced USAID inputs 
(Creative Associates International, Inc. 1995). In Malawi, donor coordination was not 
regarded by USAID project staff as being successful (Hebert, Millsap & El-Sanabary 
2002), and in Lesotho, initial donor cooperation ebbed over the course of the project 
(Reece & Kilpatrick 1996). 
Discussion
USAID projects have both a political and technical purpose. In broadest terms, the 
political purpose is to build goodwill and promote U.S. strategic interests. The technical 
purpose is to improve quality, access, and sometimes efficiency, of education systems, 
hopefully in ways that are sustainable over time. A project can be technically flawed 
and still yield important political payoffs, just as a project can succeed technically but 
be judged a failure politically. The balance between political and technical purposes 
differs across countries and within countries over time. This study examined evidence 
25
related only to the technical aspects of project success. From that perspective, this review 
supports six cross-cutting observations.
USAID has accomplished some excellent work and documented successes. USAID 
projects contributed to raising education quality, extending access, improving 
persistence, and strengthening public and private institutions in several countries. 
However, some projects appeared to have little retrievable documentation indicating 
that stated goals and objectives had been achieved. It is still possible, indeed probable, 
that many projects that lack evidence of having achieved their goals still were positively 
regarded by governments and citizens of the countries in which they operated. It is 
not unusual for beneficiaries to be positive about the source of the benefit stream. 
However, positive regard is not a substitute for improving education systems in ways that 
demonstrably raise access and quality. Within the goal attainment framework advocated 
by USAID and given the limitations of the data available to this study, a number of 
projects cannot be counted as successes. The lack of adequate documentation constrains 
USAID’s efforts to learn how to improve the quality and focus of its interventions.
The pattern of interventions suggests a balanced approach to education development. 
A broader lens through which to assess USAID project success is offered by the Galal 
(2002, 2007) and Welmond (2004) framework (discussed earlier). Did USAID 
projects rely solely on the provision of inputs to drive education reform (an engineering 
approach) or did they also address the need for incentives to support implementation 
(an organizational approach) and wider public participation in the activities of education 
reform (a political accountability approach)? Galal and Welmond argue that effective 
reform requires adopting measures under all three approaches. 
Table 8: Distribution of USAID project strategies within Galal Model
Production function Incentives Political accountability
Teacher training
Commodities
School construction
Textbooks
Teacher salaries
Teacher deployment
NGOs
School grants
Decentralization
EMIS
Table 8 illustrates the strategies employed within USAID education projects when 
organized within the Galal framework. At an organizational level, the distribution of 
USAID project strategies during this 15-year period, taken together, reflects a balanced 
approach. At a project level, while many projects had some focus on provision of inputs 
(training, textbooks, etc), they also generally included some attention to incentive 
systems, and many included components aimed at improving wider community 
participation and political accountability.
26
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
USAID project designs are formulated in terms of education system outputs and 
outcomes, but most monitoring and evaluation efforts concentrated on system-level 
inputs. Evaluations and reports were mostly focused on documenting the delivery of 
project-level outputs (e.g., number of textbooks produced, teachers trained, etc.) which 
serve as system-level inputs. In some cases this appeared to be an effort to focus on those 
elements most directly under the manageable control of project staff. While project 
funds may be used to produce trained teachers, textbooks, and classrooms, these are not 
desired ends of the education process, but only intermediate inputs toward the goals 
of improving access and learning. The real value of the inputs can only be established 
through their program of use. While a project that does not deliver needed project-level 
outputs is likely to be judged a failure, providing those project-level outputs does not 
necessarily ensure success. It is important to monitor the degree of project-level outputs 
delivered but a mistake to assume that these inputs to the larger education system 
guarantee the success of the larger education enterprise.
Little attention was given to assessing project accomplishments against stated goals 
and objectives. As noted earlier, the goals of nearly all USAID projects implemented 
between 1990 and 2005 were to increase student learning, increase access, progression 
and graduation rates, do these in a sustainable way and, in some cases, strengthen 
NGOs. Yet, with notable exceptions, these were not the focus of most project-level 
monitoring or evaluation. More attention was given to the extent to which clients were 
satisfied and promised inputs were delivered. It is quite possible that the assessment of 
progress toward stated goals was included in broader USAID Mission- or Bureau-level 
studies. However a review of project-level documents did not reveal this.
Results of many projects were over-promised at the design stage and overstated in 
subsequent evaluations. Findings suggest that results sometimes were over-promised 
at the project design stage and modest evaluation findings were sometimes overstated. 
Small differences were used to support broad claims of success. While this may be 
good politics and generate positive public relations, such practices contribute little to 
informing new project design.
Although projects at times fell short of promised accomplishments, the problem may 
be more with the expectations than the results. Small gains in student achievement 
emerging from a multimillion dollar project may be the hard truth of development. 
Unrealistically high expectations for what externally funded projects can accomplish lead 
to projects being perceived as failures, not because the intervention was inadequate, but 
because the project could not achieve such unrealistic levels of anticipated success. As 
long as inflated promises are seen as a necessary strategy for securing funds and project 
approval, and to the extent that there is no serious consequence for consistent failure to 
achieve inflated promises, such strategies will likely persist.
USAID is limited in its ability to learn from its experience. A key finding is that USAID 
is largely unable to access the findings and experience gained from much of its work 
or to learn from its own experience. While USAID has invested considerable resources 
27
in documenting, monitoring, and evaluating projects, much of that information is 
lost to outside researchers, and perhaps to USAID staff as well, due to past archiving 
practices. This history highlights a choice for the future: If USAID seeks to learn from 
its experience it may need to reconsider how project experience is captured and retained. 
Alternatively, if USAID is not going to seek lessons from project level documentation, 
it may not be necessary to continue to invest so heavily in the production of that 
documentation.
USAID personnel have multiple sources of informal and formal information on how 
well projects are performing. They may not feel they need formal documentation to tell 
them what they already know. Beyond that, there is often a tension between investing 
time and money in monitoring, testing, and evaluation to assess program success versus 
investing those same resources in direct delivery of services. Based on the documents 
available to this study, USAID and project staff appeared to believe that there were 
greater consequences for failure to meet production schedules than for failure to fully 
satisfy evaluation needs. They have a point. If the project is not fully implemented, 
the evaluation will not be very meaningful. This frustration is mentioned in several 
documents. For example, in Honduras (PEEP), the analytic requirements of the project 
were greatly overshadowed by the production demands. Project staff observed that, 
while production schedules were met, there tended to be a lack of technical analysis 
which would have sharpened the focus on utility of project efforts (Creative Associates 
International, Inc. 1995).
Conclusion
Results of this study indicate that USAID project assistance during the 1990s and early 
2000s made important contributions to strengthening education systems, extending 
education access, and increasing student learning. If anything, the results of this study 
represent a conservative estimate of USAID successes since impacts may have emerged 
that are not indicated in the documents available for this study. That said, findings 
highlight the considerable challenge posed by efforts to influence the operation of 
national education systems. 
Development assistance projects always operate in a complex cross-current of forces 
posed by, among other things, national history; the vision, administrative capacity, 
and political will of the recipient governments; and the economic environment of 
the country. Many desired outcomes depend for their success on factors outside the 
manageable interest of USAID and project personnel. Even so, results of this study can 
contribute to the ongoing reconsideration of development assistance strategies within 
USAID and similar organizations committed to improving education systems through 
development assistance.
28
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Comments from Patrick Collins, EQUIP2 Cognizant Technical 
Officer (CTO)
Following numerous reviews of this important evaluation, as CTO I feel it is important 
to provide the following comments; these comments do not necessarily reflect the views 
of USAID or the United States Government.
Contextual Changes in Target Countries. One aspect that the meta-evaluation did not 
address is the historical context of education progress in the target countries. Since the 
study relies on project documentation, which does not always capture national trends, 
it can leave the false impression that there was little progress from 1990–2005. This 
period was one of considerable national and international emphasis and greatly increased 
investment in education. Multiple global initiatives including Education for All, the 
Dakar Framework, the Millennium Development Goals, and the Fast Track Initiative, 
were coupled with rising national commitments to education. USAID’s investment 
in the education sector alone increased six-fold over the period. Whereas it is difficult 
to attribute the changes directly to donor support, it is reasonable to conclude that 
the major international effort has contributed to the historic progress in educational 
achievement. 
Progress against key indicators was most notable during this period (see related charts 
in Annex 1). Primary school enrollment, for example, increased substantially; of the 
23 countries in the study, gross enrollment (GER) improved in 17 countries by an 
average of 32 percent, with the largest increases of 67 and 54 percent in Ethiopia and 
Malawi respectively. The net enrollment rate (NER) at the primary level improved 
in 14 of the countries by an average of 26 percentage points. The most dramatic 
improvements in access were in Guinea (25 to 65 percent), Malawi (50 to 94 percent), 
and Ethiopia (23 to 61 percent). Data is more limited on other measures of education 
system efficiency and quality, and in many cases the story is mixed. While five countries 
showed measurable improvements in pupil-teacher ratio (PTR), the rapid increase in 
access created by universal primary education (UPE) policies resulted in overcrowded 
classrooms in some countries. The most striking examples of deterioration in education 
quality as measured by PTR were a 100 percent increase in Ethiopia, an 82 percent 
increase in Uganda, and a 63 percent increase in Cambodia. It is notable that of the 
18 countries in the sample with sufficient data on PTR in 1990 and 2005, only three 
exceeded the FTI target of 40 pupils per teacher in 1990, but eight had PTRs higher 
than 40 in 2005. 
The data for survival to Grade 5 are mixed but with an overall positive trend. Of the 11 
countries with sufficient data on this measure, 7 had improved the survival measure by 
between 5 percent (Lesotho) and 35 percent (Namibia). Survival to Grade 5 decreased 
in 4 countries: –5 percent in Benin, –8 percent in Botswana, –12 percent in Uganda, 
and –35 percent in Malawi. For the 13 countries with data on primary completion 
rates, 12 showed improvements between 7 percent (Swaziland) and 253 percent (Mali). 
Overall, 5 of the 13 countries— El Salvador, Malawi, Benin, Guinea, and Mali—
improved primary completion by over 100 percent during the period. 
29
Methodological Challenges. In order to have a replicable methodology, this meta￾evaluation is based only on those documents that were publicly available from USAID’s 
official archive, the Development Experience Clearinghouse (DEC). Unfortunately, 
because the documentation is incomplete, conclusions could only be drawn from limited 
information reflecting curtailed observations at a fixed point in time, and therefore may 
be misleading. For example, in Nicaragua, based on the midterm evaluation, the project 
did not seem to be doing well. According to the documentation available, after two 
years of implementation of the autonomous school approach, the reforms had reached 
fewer than 500 schools and had not yet created solid community support for basic 
education (USAID/Nicaragua 1995). The subsequent decade of intensive USAID and 
World Bank support for the reforms and their commensurate impact was not available 
in the DEC, and consequently not included in the study. However, although the project 
evaluation and final reports were not properly archived, these documents indicate that 
the project was able to correct the initial problems and was perceived as a great success. 
Substantive and meaningful education reform is, after all, a long-term process. The 
methodological approach therefore potentially led to conclusions that at least in some 
cases, when viewing programs in their entirety would seem incomplete or imprecise. 
USAID Program vs. Project Interventions. It is also important to make a distinction 
between education programs and projects. The meta-evaluation attempted to mine 
lessons learned from project experience through project-related documentation. While 
beneficial, it should be noted that USAID projects typically operated within a broader 
Mission-level programmatic rationale, much of which would not have been reflected in 
project-level reports. As a result, there is a potential misalignment between the project￾level goals and objectives as reported by the implementing organization and the broader 
programmatic strategic objectives that the Mission was aiming to achieve through 
funding multiple discrete projects. One of the challenges then in interpreting the results 
of the meta-evaluation is that it did not situate the findings of individual projects within 
the larger programmatic context in which USAID Missions were operating.
USAID support to education in countries generally takes the form of a USAID Mission 
with education staff. This usually consists of a long-term, three to five year program 
that addresses high-level—meaning broad Mission-level policy objectives—education 
needs in the country. Typically, several projects make up the program. In developing 
the program, USAID staff coordinate with the host country government, in-country 
stakeholders, and other donors to determine the strategic direction of the program 
in support of the country education plan. Program goals and objectives are generally 
articulated on a more general level (e.g., support for improved equitable access to quality 
basic education) and may focus on certain grade levels, geographical areas, subject 
matter, or other thematic areas. 
Based on this strategic program focus, USAID will then design specific projects, 
generally lasting two to five years, with lower-level—meaning technical and/or 
management related—objectives, activities, and indicators within the manageable 
30
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
interest of the project (e.g., number of teachers trained in a new curriculum, or number 
of books distributed). These projects are designed to support the achievement of the 
higher-level USAID program and, in turn, the higher-level national education plan. The 
achievement of these higher-level national objectives (e.g., gender parity in enrollment, 
targets for increased enrollment or completion rates) generally requires a longer time 
frame—five to twenty years or more. For USAID, this means that multiple projects 
are often designed as part of a sequence within a larger USAID program that in turn 
supports the achievement of significant national improvements over time. 
It is also important to note that USAID functions as a decentralized agency. During 
this period, Missions had substantial, yet perhaps varying, latitude to design their own 
programs; one Mission might have only one large, long-term project contributing to 
one strategic objective and implemented by only one contractor, whereas other Missions 
developed multiple education projects, sometimes as a combination of short- and long￾term interventions implemented by various contractors, and contributing to one or 
multiple strategic objectives. This variation across Missions, countries, programs, and 
projects over time further complicates the analysis of individual projects as presented in 
the meta-evaluation.
In addition, education spending is typically one of the largest portions of national 
budgets, and even when USAID has robust education programs, they are generally only 
a tiny percentage—1 percent or less—of total education spending in countries. Thus, 
USAID education programs do not always provide direct services; rather, programs are 
often strategically designed in coordination with the host country and other donors to 
support collective achievement of national goals through, for example, capacity building 
(e.g., training of teachers, curriculum developers, and district education officers) and/or 
program support in certain thematic or geographic areas. 
In other words, while USAID projects may support the provision of specific inputs—
such as teacher training or learning materials—the sheer size of the education sector in 
countries is such that there is virtually always shared attribution with other stakeholders 
in terms of achieving national outcomes such as improved learning, gender parity, and 
enrollment rates. Moreover, making improvements in education requires far more than 
just the provision of inputs, such as schools and books, it involves behavioral changes 
on the part of thousands of individuals in each country, including students, parents, 
teachers, principals, and district and national level officials.
This is a critical point to understand. This evaluation correctly notes that many project 
documents mainly reported on inputs vs. system level outcomes or impact (though 
program level documents not included in this study often did provide such higher level 
data). The issue, however, is that even where national, system level results are available, 
they are invariably due to numerous interventions and actors, and thus success or 
failure cannot solely be attributed to USAID; indeed it is often difficult to disaggregate 
the unique contributions of USAID support. An ongoing question then, is what are 
reasonable levels of targets and objectives for USAID support, and what is a reasonable 
31
balance - for targeting, reporting and performance management purposes - between 
input results that can clearly be attributed to USAID, and the higher, national level 
results that we all desire, but which are well beyond USAID’s manageable interest?
It is important to understand the difference between projects and programs and the 
reality of shared attribution in reviewing the findings of this meta-evaluation since, 
unfortunately, the publicly available documentation of program-level planning and 
reporting is very poor. This has occurred for a variety of reasons. First, USAID project 
and program planning and reporting systems changed several times during the period of 
this study, complicating or eliminating the opportunity to assess the overall progress of 
country programs and projects over time using publicly available information. Second, 
as someone working inside USAID, this study is also a reminder that there is a wide 
range of “grey literature,” such as regional trends analyses, program concept papers, and 
other key strategic analysis and planning documents that are not included in official 
archives because they are internal documents or procurement sensitive. 
Conclusions. As a result of all the factors noted above, publicly available information 
on USAID education efforts during this time period is less complete than would be 
desired. Thus, the study is based on available project-level data rather than Mission- or 
country-level and long-term program-level information. This suggests the value of using 
this meta-evaluation as a basis for undertaking additional analyses using an expanded 
methodology.
USAID has devoted significant resources to supporting international basic education 
over the years. Our sincere hope is that the results of this analysis will be useful 
in informing decision-making and in highlighting areas where additional actions, 
information, or analysis are needed.
32
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Comments from John Gillies, EQUIP2 Project Director (CTO)
The EQUIP2 Leader with Associates Award on Education Policy, Systems, and Manage￾ment has explored numerous issues affecting education quality including decentraliza￾tion, school effectiveness, school fee abolition, student assessment, business partnerships 
in education, education data and indicators, monitoring and evaluation, and education 
system reform. This paper, An Analysis of U.S. Assistance to Basic Education, is an impor￾tant addition to the EQUIP2 portfolio. The study is the most comprehensive review of 
USAID education projects available, analyzing patterns and findings of 33 projects over 
a 15-year period. Some of the findings provide fresh insights into education program￾ming. Other findings are not new, but bear repeating and emphasizing. This study will 
complement a broad range of EQUIP2 research activities looking at effective strategies 
for improving education systems over the long term.
The study is important not only for the questions that it can answer, but equally impor￾tant for those it cannot. A reader who is familiar with USAID education projects will 
recognize a lot of truth in the details, but will also note that the report does not speak 
to the very positive impact and influence of USAID assistance over time. The serious 
limitations of documentation and evaluation, a finding in itself, left the big questions 
unanswered: Was aid to education effective? Which strategies result in the best outcomes? Are 
the programs sustainable? Does decentralization work? Even some basic questions could 
not be answered definitively: How many of the projects achieved their intended outputs and 
did those outputs have an impact?
USAID’s reporting, monitoring, and review procedures should serve two functions: to 
provide accountability for the use of public funds to Congress and the American people 
and to enable organizational learning to improve effectiveness at the project/country level 
as well as for the overall Agency. This study shows that the system did not meet either 
challenge very well. 
Evaluation
The study found that only 11 of the 33 projects had a final evaluation, and only two of 
these were considered highly rigorous, with a clear purpose, using multiple methods, 
sufficient time, and addressing policy relevant issues. Although nine of the evaluations 
concluded that the project was successful, neither the definition nor evidence of success 
was always clear. Most of the evaluations focused on compliance and accountability for 
results, but they did not always report against the targets. More importantly, the evalu￾ations failed to ask policy-relevant questions about the validity and execution of the 
underlying strategy and development hypothesis that a given set of activities will result 
in better education systems. Every project is based on implicit or explicit policy assump￾tions, which were infrequently evaluated. Therefore, the evaluation conclusions too of￾ten consist of generic insights: projects should be flexible; more time is needed; projects 
should consult with stakeholders; or projects should plan better.
A substantive evaluation should distinguish between theory failure and implementation 
failure, and should identify which aspects of strategies work or do not work, and under 
33
what circumstances. Most of the USAID evaluations reviewed for this study do not do 
this. The project evaluations had a “black box” approach—looking at activities and, 
sometimes, outcomes, but seldom analyzing the relationship between the two. Conclu￾sions about the relative effectiveness of alternative strategies cannot be made, in part 
because the data is not there and in part because the questions are never asked. 
External evaluations were undermined by the shortage of robust project monitoring and 
evaluation (M&E) systems, which aside from missing or poor quality data (which in 
part reflects weak capacity in developing countries), lacked internal formative evaluation, 
technical expertise, and a project evaluation framework. Only two evaluations identified 
the project M&E system as a particular strength, and five studies found this component 
deficient. While projects usually collected data for reporting and performance monitor￾ing, this is not a substitute for substantive M&E. This demonstrates a deeper problem 
than the limited ability to report on accomplishments; it shows that M&E was not an 
integral part of the project design and strategy. 
Mission and Project Reporting
The EQUIP2 study captured all of the project-level documents available in the archives 
for the target projects. A possible weakness of the methodology is that it does not cap￾ture the program-level reporting through which Missions formally reported results. In 
the context of assessing the effectiveness of aid to education, it is important to note that 
the education portfolio consisted of multiple projects, both concurrently and over time. 
Given documentation weaknesses, the study was not able to include all of the projects or 
program support in any of the countries studied. While some of the program outcomes 
might have been captured in program or country level reports, such reports were not 
reviewed in the study. Since the program reporting is sometimes only loosely coupled 
to specific project interventions, it presents a different set of challenges in terms of both 
attribution and its analytical power in understanding the impact of particular strategies. 
The program reporting (Results Review and Resource Request or R-4, and Annual Re￾ports) contains information that was not used in this study, but does provide additional 
perspective on some issues. A limited review of R-4s and Annual Reports found that 
they can provide a rich narrative about such issues as engagement in donor coordination, 
influence on national policy dialogue, and long-term strategic vision that are not cap￾tured in the project documents reviewed in this study. Moreover, the R-4 reports usually 
provided a transparent “targeted vs. achieved” format and sometimes included important 
measures of access and quality, such as: leaving exam pass rate; gross enrollment ratio; 
curriculum developed and used; percentage of teachers with qualifications; percentage of 
students performing at standards; active parent associations; promotion rates; or policy 
frameworks adopted.
The program reporting was by no means problem-free. The best R-4 reports provided a 
longitudinal picture of country progress against stable indicators. Unfortunately, the R-4 
reporting in some countries suffered from lack of continuity and coherence; indicators 
could appear and disappear quickly or be redefined in the middle of the reporting pe￾riod, which made it difficult to assess results over time. The indicators being reported by 
34
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
the Missions were not always clearly related to the project activities in a way that change 
could be attributed. 
Consistency in results reporting suffered from a series of organizational reforms and 
changing requirements, procedures, and indicators, efforts to find common indicators, 
and changing reporting formats. In FY 2002, the R-4s (often over 100 pages long) were 
replaced by very short Annual Reports that dramatically disrupted the data consistency. 
In 2002, the main indicator appearing in Annual Reports was “number of children 
enrolled in schools affected by USAID programs,” eliminating sometimes a dozen spe￾cific measures. Since that time, the Agency reporting requirements and guidelines have 
changed several times. 
The Challenge of USAID Monitoring and Evaluation of Aid Effectiveness
The procedural and reporting reforms in USAID have been undertaken with the best of 
intentions: to better capture the Agency’s activities and successes in a way that informs 
Congress and the American people about the impact of foreign aid on education. There 
is an undeniable challenge in capturing and rolling-up a diverse array of locally relevant 
programs into a metric that is meaningful to Congress. Unfortunately, as the study 
demonstrates, it appears that the reforms during 1996–2002 did not solve the prob￾lem. The reporting, evaluation, and documentation formats are neither informative as a 
reporting tool, nor do they have any great value for organizational learning and program 
improvement. It is unclear whether or not this exercise is meeting Congressional needs 
for program oversight.
These issues are symptoms of an underlying institutional problem. The problem of inad￾equate evaluation and organizational learning is not new; studies have repeated this find￾ing for decades. The problem is not due to a lack of understanding, but rather the lack 
of institutional incentives for and value of evaluation. USAID under-invests in learning 
in part because of structural problems—a classic case of externalities. The benefits of 
substantive evaluation accrue to the organization at large, or to future projects, but the 
funding and priorities are controlled by project managers whose performance is mea￾sured not by learning, but rather by production and activities. Project managers under￾invest in knowledge because they cannot capture the benefits of evaluation. 
A similar structural constraint results from the changing guidelines about program 
reporting (i.e., common indicators and Performance Monitoring Plans or PMPs). The 
frequent changes in reporting requirements absorb management attention, and distract 
it from the programmatic goal of gaining substantive knowledge for program improve￾ment. This process has undermined organizational learning both because of the lack 
of continuity about measures, and because it undervalues long-term evaluation. High￾level managers appear to focus on reporting rather than program improvement. When 
USAID temporarily eliminated the requirement for midterm and final evaluations in the 
mid-1990s, it was not done out of ignorance, but rather because the evaluations were 
pro-forma—and expensive—exercises that were not being used. Although the intent was 
35
to encourage evaluations that would be used, the impact was to lessen attention to evalu￾ation in general.
Furthermore, the question of effectiveness of aid to education is more difficult to answer 
when the primary unit of analysis and programming is the project. A project is a time￾limited intervention with defined objectives that is taking place in the context of long￾term national institutional and political change. The project approach has both manage￾ment and implementation advantages to providing assistance, but creates distortions and 
limitations as the unit of analysis for evaluating impact. As the study aptly points out, 
the institutional pressures to over-promise on sustainable outcomes in a short period 
are counterproductive. This is a real problem even with the historical norm of five-year 
projects. It will be greatly exacerbated if the trend continues to shorter planning and 
implementation periods of one to three years, in which case meaningful evaluation must 
be able to place the project into context. 
Although the need for meaningful assessment and learning about effective strategies 
for education is clear, the solution is less so. The challenges are lack of capacity in the 
Agency for framing and conducting evaluations, lack of institutional incentives, and the 
lack of an overarching framework for program assessments. If the challenge of organiza￾tional learning is to be addressed, some of the following elements must be in place:
 
•	 Clear and unambiguous value in the system for substantive evaluation and learning. 
This may come from Congress, USAID leadership, or others but it is unlikely to be 
sustained and systemic if left to individual project managers.
•	 A coherent set of framing questions at the Agency level that addresses the broad 
policy-relevant questions that constitute the sector concerns. Such framing questions 
will be an important factor in moving beyond the sole focus on accountability and 
compliance.
•	 The process must first and foremost be utilization-focused. Evaluation needs to 
be a central part of project design and implementation rather than a documentary 
requirement.
•	 If evaluations are to be useful for management purposes, they need to be professional 
and high quality, with clear questions, solid methodology, and a coherent use plan. 
36
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
List of References22 Academy for Educational Development (AED). (1998). Final Report: Strengthening 
Achievement in Basic Education (SABE) Project. Washington D.C.: AED. USAID 
document # PD-ABR-636. 
Academy for Educational Development. (1994). Indicators of Democratic Behavior in 
Nueva Escuela Unitaria (NEU) Schools. Washington D.C.: AED. USAID document # 
PN-ABX-940. 
Adams, D. (1998). Defining Educational Quality: Educational Planning. Educational 
Planning, 11(2), pp. 3–18.
Al-Samarrai, S., Bennell, P., & Colclough, C. (2002, September). From Projects to 
SWAPs: An evaluation of British Aid to Primary Schooling, 1988-2001, Evaluation Report 
EV639. London: Department of International Development.
Anderson, L. & Burns, R. (1989). Research in Classrooms: The Study of Teachers, Teaching, 
and Instruction. Oxford, England: Perganon.
Bernbaum, M., & Locher, U. (1997). EDUCA Evaluation. Washington D.C.: Creative 
Associates International, Inc. USAID document # PD-ABQ-048. 
Bhat, M.K., Cheria, A., & Edwin. (1999). Life goes on: Sustainability of community 
development programs and the withdrawal of NGO support: an enquiry into expectations 
and implications. Bangalore, India: The Center for Innovation in Voluntary Action 
(CIVA), Bangalore Consultancy Office.
Birdsall, N., Levine, R., & Ibrahim, A. (2005). Toward Universal Primary Education: 
Investments, incentives, and institutions, UN Millennium Project, Task Force on 
Education and Gender Equality. Sterling, VA: Earthscan, p. 25.
Caribbean Resources International (CRI) Consult, Inc. (1998). Basic Education System 
Overhaul (BESO) Midterm Assessment. Miami, FL: USAID document # PD-ABT-531. 
Center of Excellence for Sustainability. (2001). Overview of sustainable development 
definitions and principles. Retrieved in 2006 from http://www.sustainable.doe.gov/
overview/definitions.html.
Chapman D.W. (2002). A Review of Evaluations of UNICEF Education Activities (1994-
2000), UNICEF Working Paper prepared for the Office of Evaluation and Program 
Planning. New York, NY: UNICEF.
22 A complete list of the documents reviewed for this study is available from the authors.
37
Chapman D.W. (2001). A Review of Evaluations of UNICEF Education Activities (1994-
2000), UNICEF Working Paper prepared for the Office of Evaluation and Program 
Planning. New York, NY: UNICEF.
Chapman, D.W. and Adams, D. (2002). The quality of education: Dimensions and 
Strategies, Education in Developing Asia, Vol. 5, Hong Kong, China: University of Hong 
Kong Comparative Education Centre.
Chapman, D.W., with Coulibaly, N.C., Coyne, B., Emert, H., Diallo, A., Hickson, K., 
Issayas, S., Kyeyune, R., Lexow, J., Lokkesmoe, K., & Osei, J. (2004). Changing Lives 
of Girls: Findings, conclusions and lessons from the external evaluation of the African Girls’ 
Education Initiative. New York, NY: Evaluation Section, UNICEF. 
Chapman, D.W. & Dykstra, A. (2006). From Money to Action: Strategies in U.S. 
support to education in the developing world. Journal of International Cooperation in 
Education, 9(1), 27–43.
Chapman, D.W., Weidman, J., Cohen, M., & Mercer, M. (2005). The search for 
quality: A five country study of national strategies to improve educational quality in 
Central Asia. International Journal of Educational Development, 25(5), pp. 514–530.
Chesterfield, R., Enge, K., & Simpson, H. (2002). Formative Evaluation 2002. 
Washington D.C.: Juarez and Associates, Inc. USAID document # PD-ABY-194. 
Clark, L. P. & Pearson, P. (1996). Final Evaluation, Educational policy management and 
technology project. Washington D.C.: CAII and USAID/Swaziland. 
Community School Alliance (CSA). (2001). Education in Ghana: The need for reform. 
Retrieved in 2006 from http://www.edc.org/CSA/ed.htm.
Congressional Research Service. (2006). Response to special request for information.
Craig, H., Kraft, R., & Plessis, J. (1998). Teacher Development: Making an Impact. 
Washington, D.C.: World Bank.
Creative Associates International, Inc. (1995). Final Evaluation of the USAID Primary 
Education Efficiency Project (Honduras). Washington D.C.: Creative Associates 
International, Inc. USAID document # PD-ABM-539. 
Dahl, A.L. (1995). Towards indicators of sustainability. Geneva, Switzerland: United 
Nations Environment Programme. 
Daniels, M.R. (2000). Introduction: What are sustainable community programs? In 
M. R. Daniels (Ed.), Creating sustainable community programs: Examples of collaborative 
public administration. (pp.1–9). Westport, CT: Praeger Publishers.
38
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Dempster, Beth M. (1998). A self-organizing systems perspective on planning for 
sustainability. http://eserver.uwaterloo.ca/jjkay/grad/bdempster/index.html.
Dye, R., Hunt, B., Meza, M., Broehl, A., Kissam, E., & Valverde, G. (2000). Project 
BEST Evaluation. USAID: Washington D.C. USAID document # PD-ABT-051 (in 
English and Spanish). 
Dye, R., Jennings, J., Lambert, C., Hunt, B., & Wein, G. (2002). Evaluation and 
Recommendations for strengthening and extending the new horizons for primary school project 
in Jamaica. Washington D.C.: Aguirre International. USAID document # PD-ABY-203. 
Easterly,W. (2007). The White Man’s Burden: Why The West’s Efforts To Aid The Rest Have 
Done So Much Ill And So Little Good. New York, NY: Penguin Press.
The Economist (1999, June 26). How to make aid work. The Economist. pp. 23-27.
Education Development Center, Inc. (1994). Midterm Evaluation: Ghana Primary 
Education Program Intervention through Conditionalities. USAID document # PD-ABI￾547. 
Fitzpatrick, J.L., Sanders, J.R., & Worthen, B.R. (2003). Program Evaluation: Alternative 
Approaches and Practical Guidelines, 3rd Edition. Boston, MA: Allyn & Bacon.
Fuller, B. (1987). What Factors Raise Achievement in the Third World? Review of 
Educational Research, 57 pp. 255–292.
Galal, A. (2007). The Road Not Traveled Education Reform in the Middle East and North 
Africa. Washington D.C.: World Bank.
Galal, A. (2002). The Paradox of Education and Unemployment in Egypt. Working Paper 
No. 67. Cairo, Egypt: Egyptian Center for Economic Studies.
Government Accountability Office (GAO) (2007, March). Foreign Assistance: Enhanced 
coordination and better methods to assess the result of U.S. international basic education 
efforts are needed. Report to Congressional Committees. Washington D.C.: GAO.
Goldenberg, D.A. (2001). The MEGA Evaluation, Meta-Evaluation of Goal Achievement 
in Care Projects: A review of Findings and Methodological Lessons from CARE Final 
Evaluations, 1994-2000. Atlanta, GA: Program Division, Atlanta, CARE USA.
Guild, P., Moulton, J., Wolf, J., DeStefano, J., & Wandira, A. (1995). Final Report 
Support for Ugandan Primary Education Reform: A Formative Evaluation. Washington 
D.C.: DevTech Systems, Inc. USAID document # PD-ABP-618. 
39
Harris, J. M. (2000). Introduction: An assessment of sustainable development. In J. M. 
Harris (Ed.), Rethinking sustainability: Power, knowledge and institutions. (pp. 1-10). Ann 
Arbor, MI: The University of Michigan Press. 
Hebert, P., Millsap W., & El-Sanabary, N. (2002). Summative evaluation of USAID 
Malawi’s Girls attainment in basic literacy and education (GABLE). Washington D.C.: The 
Mitchell Group.
Heyneman, S. (2006). The effectiveness of development assistance in education: An 
organizational analysis. Journal of International Cooperation in Education, 9(1).
Lockheed, M. & Verspoor, A. (1992). Improving Primary Education in Developing 
Countries. Washington D.C.: World Bank.
Mamadou, G. (2003). Project Evaluation Primary Education NGO Project (PENGOP). 
USAID document # PD-ABY-587. 
Management and Business Associates, Inc. (1994). Midterm Evaluation of the 
Strengthening of Achievement in Basic Education Project (SABE). Washington D.C.: 
USAID document # PD-ABM-868. 
Moulton, J. (2000). Support for Ugandan primary education reform. Final Report. 
Washington D.C.: USAID.
Nieves, I., Lundgren, R., Bezmalinovic, B., & Matute, J. (1994). First Phase: Lessons 
Learned from a Pilot Program, Second Phase: Assessing the Efficiency and Comparing the 
Performance of Girls in a Pilot Scholarship Program with that of other School Girls. USAID 
document # PN-ACA-136. 
Nkansa, G.A. & Chapman, D.W. (2006). Sustaining community participation: What 
persists after the money ends? International Review of Education, 52(6). 
Picard, L., & Garrity, M. (1997). Institutional development revisited. In L. Picard and 
M. Garrity (Eds.) Policy reform for sustainable development in Africa: The institutional 
imperative. (pp. 151–168). Boulder, CO: Lynne Rienner Publishers. 
Reece, G. & Kilpatrick, M. (1996). External Final Evaluation of the Lesotho Primary 
Education Program. Washington D.C.: AED. USAID document # PD-ABP-351. 
Rogers, Everett M. (1995). Diffusion of innovations. New York, NY: The Free Press. 
Rossi, P.H., Lipsey, M.W., & Freeman, H.E. (2003). Evaluation: A Systematic Approach. 
Thousand Oaks, CA: Sage Publications Inc.
40
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Simpson, H., Harris. A., Schubert, J., & Yusuf-Kalil, Y. (2003). Uplifting adolescents 
project study. Washington, D.C.: American Institutes for Research. 
Sommers, M. (1996). Decentralizing Education: The BESO/Tigray Case Study.
Washington D.C.: USAID document # PN-ACA-337. 
USAID (1988). Education and Human Resources Sector Assessment. Monrovia, Liberia and 
Washington, D.C.: United States Agency for International Development and Liberian 
Ministry of Education.
USAID/Honduras (1993). A.I.D. Evaluation summary. Washington D.C.: USAID.
USAID/Nicaragua (1995). A.I.D. Evaluation Summary – Part I. (1996). BASE Project 
AID evaluation summary. Washington D.C.: USAID # PD-ABN-402.
Welmond. (2004). Concept paper: Education Policy in the Middle East and North Africa. 
Washington D.C.: World Bank.
World Bank. (1999). Assessing Aid: What works, what doesn’t, and why. A World Bank 
Policy Research Report. Oxford, UK: Oxford University Press.
World Bank (2002). Implementation completion report on a credit in the amount of 
SDR 67.3 million to the People’s Republic of China for a Third Basic Education Project. 
Washington, D.C.: World Bank.
World Bank. (2004a). 2004 Annual Review of Development Effectiveness; The Bank’s 
contributions to poverty reduction. Washington D.C.: World Bank.
World Bank. (2004b). Primary Education Portfolio Review. Background paper for 
the evaluation of the World Bank’s support to primary education. Washington D.C.: 
Operations Evaluation Department, World Bank.
World Bank. (2006). From schooling access to learning outcomes: An unfinished agenda. An 
evaluation of World Bank support to primary education. Washington D.C.: Independent 
Evaluation Group, World Bank.
41
Annex 1
Primary Gross Enrollment Rate (1990-2005)
0
20
40
60
80
100
120
140
160
Namibia
Jamaica
South Africa
Botswana
Honduras
Egypt
Swaziland
Zambia
Dominican Republic
Nicaragua
Bolivia
Lesotho
Ghana
El Salvador
Guatemala
Benin
Mali
Guinea
Uganda
Cambodia
Malawi
Ethiopia
1990
2005
target
Primary Net Enrollment Rate
0
20
40
60
80
100
120
Namibia
Jamaica
South Africa
Botswana
Honduras
Swaziland
Bolivia
Zambia
Egypt, Arab Rep.
Lesotho
Nicaragua
Ghana
El Salvador
Dominican Republic
Guatemala
Mali
Cambodia
Benin
Guinea
Malawi
Ethiopia
1990
2005
target
42
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Annex 1. (continued)
Primary Pupil Teacher Ratio
0
10
20
30
40
50
60
70
80
Ethiopia
Uganda
Cambodia
Benin
South Africa
Zambia
Mali
Guinea
Ghana
Egypt, Arab
Rep.
Nicaragua
Swaziland
Bo ivia
Guatemala
Honduras
Jamaica
Botswana
Lesotho
1990
2005
target
Grade 5 Survival Rate
0
20
40
60
80
100
120
Malawi
Botswana
Uganda
Benin
Lesotho
South Africa
Nicaragua
El Salvador
Mali
Guinea
Namibia
1990
2004
target
43
Annex 2. Types of documents included in this review
Country Project Interim 
Documents
Final 
Report
Mid-term 
Evaluation
Final 
Evaluation
Benin CLEF √ √ Audit
PENGOP √ √ √
Bolivia RLP √ √ √ √
Botswana BEC √ √ √
Cambodia CAPE √
Dominican Rep. PIPE √ √ √ √
Egypt BEC √ √
El Salvador EDIFAM √ √
SABE √ √ √
Ethiopia BESO √ √
Ghana QUIPS √ √
Guatemala BEST √ √ √
EFPD √
Guinea FQEL √
PASE √ √
Haiti IIBE √
ResPack √
Honduras BEST √ √
PEEP √ √ √
Jamaica UAP √ √ √
New Horizons √ √ √
LAC CETT √
Lesotho PEP √ √ √ √
Malawi GABLE √ √
ESPP √ √
Mali BEEP √ √
Namibia BES √ Audit
Nicaragua BASE √ √
South Africa SABER √ √ √
Swaziland EPMT √ √ √ √
Uganda SUPER √ √
EMIS √
Zambia CHANGES √
44
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Annex 3. Summary of Project Activities by Frequency (33 country projects 
reviewed)
Activity Frequency Activity Frequency 
Improving infrastructure -- cluster-based 2
-- Commodities (furniture, computers) 10 -- distance 2
-- School construction 6 -- cascade 3
Strengthening systems -- master teacher training 2
-- Capacity development – MOE 16 -- Headmaster training 8
-- EMIS development 15 -- Radio instruction 7
-- Monitoring & evaluation 14 -- Television instruction 1
-- TA to support MOE operations 10 -- Distance education 3
-- Teacher training coll support 8 -- Multi-grade classrooms 4
-- Short term participant training 6 -- Demonstration schools 2
-- Long term participant training 6 -- Laboratory schools 1
-- Improve financial management 6 -- Training resource teachers 1
-- NGO development 5 Developing curriculum/materials
-- Strategic planning 5 -- Curriculum development 21
-- Personnel management 4 -- Instructional materials devel 19
-- Governance/leadership training 4 -- Active student learning 5
-- Teacher redeployment 3 -- Self instructional materials 3
-- Devel school quality standards 3 -- Education materials distribution 15
-- Training school board members 3 -- Teacher/curriculum guides 10
-- TA support to NGOs 3 Increasing public support for Education
-- School grants 2 -- Social marketing/Advocacy 10
-- Reorganize teacher colleges 1 -- Community participation 9
-- Expenditure tracking system 1 -- Parent committees/PTAs 6
-- School mapping 1 -- Decentralization 6
Measuring learning outcomes -- Parent education 5
-- Achievement testing 16 -- Scholarships for girls 4
-- Test development 15 -- Community devel seed funds 3
Educational policy Other
-- Policy/operations research 17 -- Early childhood family education 1
-- Policy development 13 -- Deworming 1
-- Special studies 13 -- School feeding program 1
Instructional quality -- Reduce # of subjects taught 1
-- Teacher training 20 -- Promoted community schools 1
 -- direct 2
45
Annex 4. Predominant evaluation criteria employed
Country Project Inputs 
tracked
Change in 
enrollment
Change in student 
flow (repetition, 
attrition, graduation)
Change in 
behavior
Change in 
learning
Benin
CLEF √ 
PENGOP √
Bolivia RLP √ √ See Note 1
Botswana BEC √ 
Cambodia CAPE
Dom. Rep. PIPE √ √
Egypt BEC √ √ 
El Salvador
EDIFAM √ √
SABE √ √ √ √
Ethiopia BESO √ 
Ghana QUIPS √ √ 
Guatemala
BEST √ √ √ √ √ 
EFPD √
Guinea FQEL
PASE √ √
Haiti
IIBE
ED 2004
Honduras
BEST √ √ 
PEEP √ √ √
Jamaica
UAP √ √ √ 
New 
Horizons
√ √ √ √ 
LAC CETT √ √ √ 
Lesotho PEP √ 
Malawi
GABLE √ √ √ √ √ 
ESPP √
Mali BEEP √ √
Namibia BES √ √ 
Nicaragua BASE √ √ 
South 
Africa
SABER √ 
Swaziland EPMT √ √ √ √ √ 
Uganda
SUPER √ √ 
EMIS √
Zambia CHANGES √ √ √ √ √ 
Note 1: Measured, but results are hard to interpret
46
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
 Annex 5. Did projects that aimed to improve education quality result in increased 
student learning?
Extent that learning was assessed Number of projects
Number of projects reviewed 33
→ Based on the data available to this study, the number of 
these projects in which raising educational quality was a 
stated goal
28
→ → Of these, the number in which student learning was 
proposed as a measure of project success
17
→ → → Of these, the number in which testing was used to 
assess student learning
12
→ → → → Of these, the number in which pre-testing was 
used to assess change over time
11
→ → → → → Of these:
– The number in which a comparison group was used to 
assess impact of project vs. naturally occurring change
9
– The number that looked only at gains over time of 
project students
1
→ → → → → → Of these:
– The number of projects that found a significant 
increase in student learning among project schools/
students
6
– The number of projects that found mixed results, with 
comparison groups outperforming project groups in 
some subjects at some grade levels
3
– Can’t tell from the information reported 1
47
Annex 6. Magnitude of project learning gains compared to comparison group (Projects 
in which some learning gains were reported)
Country Project Magnitude of project learning gains compared to 
comparison group Citation
Ghana QUIPS 38% (Math, over 2 years) 
15% (English over 2 years)
Amuah, 
Gyamera, & Vor￾mawor (2000)
Guatemala BEST 16% (radio Math) Enge, Etheridge, 
et al. (1993)
Jamaica 
(Horizons)
New Hori￾zons
Mixed results, over 3 years: In Language Arts, 
project girls out-performed comparison group 
girls by 3.3%; project boys out-performed non￾project boys by 0.4%. In Math, non-project girls 
out-performed project girls by 4.0 %>; non-proj￾ect boys out-performed project boys by 4.4%.
Chesterfield, 
Enge, & Simpson 
(2001)
LAC CETT CETT teachers did better than non-CETT teach￾ers
Magnitude of difference not provided
Chesterfield, 
Culver, Hunt, & 
Linan-Thompson 
(2005)
Malawi GABLE Grade 3: Results favored project schools all three 
subjects. 
Grade 4: Project schools out-performed non￾project schools in Math, only marginally outper￾formed comparison schools in English, under￾performed comparison schools in Chichewa. 
Grade 5: Project schools outperformed in Math, 
only slightly outperformed in English, essentially 
comparable in Chichewa.
Herbert, Millsap, 
& El-Sanabary 
(2002)
Namibia BES Grade 1 & 2 – project students out-performed in 
Math and Environmental Studies, no differences 
in English achievement
Snyder (2000)
Swaziland EPMT Grade 1 Math, increase of 3% in mean score over 
2 years, with number of students achieving 80% 
mastery increasing from 36% (1992) to 40% 
(1993) to 51% (1994). 
Grade 1 English, mean scores rise from 19 (1992) 
to 24 (1994) while the percent of students scoring 
in the two highest categories increases from 12% 
(1992) to 29% (1993) to 34% (1994). 
Grade 2 Math and English show similar increases 
in mean scores and in proportions of student 
scores at the high-end of the score distribution. 
Clark & Pearson 
(1996)
Zambia CHANGES Significant increase in cognitive assessment scores 
for 2017 pupils. Children who received treatment 
improved significantly more than children in 
control schools.
Graybill (2004)
48
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Annex 7. Magnitude of learning gains and changes in student flow in World Bank 
projects
Country Name of 
project
Magnitude of project learning gains 
compared to comparison group
Magnitude of changes in student 
flow indicators (internal efficiency)
Chile1 Primary School 
Improvement Gain in learning score between 1991-1996a
Spanish 4th 8th
Private-paid schools 7.1% 3.4% 
Private-subsidized 16.2 9.1 
Municipal schools 13.1 9.0 
Math
Private-paid schools 13.3 4.0 
Private-subsidized 18.8 8.8 
Municipal schools 19.2 7.5 
Between 1991-1996 average repetition 
rates for the primary cycle dropped 7% 
(from 12.8 to 5.8). 
Average years to completion of primary 
school dropped 2.7 years (from 12.4 
to 9.7).
China2 Third Basic 
Education
Between 1993 and 2001, grade 6 Chinese 
language pass rates increased by 3.8% (from 
95.4% to 99.2%) and math pass rates in￾creased by 5.5% (from 94.6% to 99.1%).f
Between 1993 and 2001: Grade 1 stu￾dent dropout rate declined 1.5% (from 
3.0% to 0.5%). Grade 1 repetition rate 
fell 7.3% (from 8.4% to 1.1%). The 
completion rate for 15 year olds in￾creased 9.2% (from 90.4% to 99.2%).
At lower secondary, dropout rates 
declined 4.9% (from 6.8% to 1.9%) 
and completion rates increased by 12% 
(from 86% to 98%).
India3 District Primary 
Education
During the first three years of implementation 
(1997-2000), achievement scores increased 
between grade 1 and grade 4 by 40% in math 
and 23% in language. For grades 3 & 4, there 
was an 18% increase over baseline in math 
and 16% increase over baseline in language. a During 2000-2003, grade 1 scores increased 
by 13% in math and 12% in language. Grade 
3 & 4 scores increased by 26% in math and 
16% in language.a Whether findings were 
based on cross-sectional or longitudinal analy￾sis could not be determined. No comparison 
group was used. A different test was used at 
baseline and at final-point testing.
16% of districts reduced dropout to the 
10% target but baseline data was not 
available.
Mexico4 Second Primary 
Education
Although project schools still lagged behind 
the national average in student academic 
achievement, there was a 2.1% reduction in 
the score gap in Spanish ( 6.9% to 4.8%) and 
a 0.7% reduction in mathematics (3.5% to 
2.8%) between 1996 and 2000.
Dropout rate in project schools declined 
from 3.9% (6.0% to 2.1%), to the same 
level as the national average. Repetition 
rates fell 2.3% (from 10.4% to 8.1%).
Uruguay5 Basic Education 
Quality 
Improvement
From 1996 to 1999, cognitive assessment of 
grade 6 students indicated that the percentage 
whose performance was deemed satisfactory 
increased 6% in mathematics (from 35% to 
41%) and 4% in English (57% to 61%).
From 1996 to 1999, repetition rates for 
primary grades declined by 2% (from 
12% to 10%).
1. World Bank. (1999). Implementation completion report, Republic of Chile, Primary Education Improvement project.
2. World Bank. (2002). Implementation completion report on a credit in the amount of SDR 67.3 million to the People’s 
Republic of China for a Third Basic Education Project.
3. World Bank. (2003). Implementation completion report on a credit in the amount of SDR 291.7 million to India for a 
District Education Project II.
4. World Bank. (2002). Implementation completion report on a loan in the amount of US$412 million to the United 
Mexican States for a Second Primary Education Improvement Project
5. World Bank (2002). Implementation completion report on a loan in the amount of US$31.5 million to the Republic 
of Uruguay for a Basic Education Quality Improvement Project
a. Lack of comparison group makes it difficult to assess contribution of project over alternative instructional treatments 
students would have received.
49
 Annex 8. Changes in student flow as reported by projects
Country
Change 
reported 
by project
Context Target Group Citation
Change in Persistence
El Salvador 5% Increase over 5 years in 
primary net enrollment
National AED (1998:7)
Guatemala￾NEU
13.4% 
(boys)
7.9% 
(girls) 
Increase in persistence 
between Gr 1 and Gr 6.
NEU project 
schools vs. com￾parison group
Chesterfield & 
Rubio (1997)
Guatemala￾DIGEBI
4.2% 
(boys) 
2.9% 
(girls)
Increase in persistence 
between Gr 1 and Gr 6.
DIGEBI project 
schools vs. com￾parison group
Chesterfield & 
Rubio (1997)
Honduras 3.0%
-17.0% 
Increase in anticipated 
target for existing training 
centers
< Shortfall in anticipated 
target for new centers>
CADERH basic 
education train￾ing centers for 
adults
Dye, Hunt, et al. 
(2000:25).
Honduras 
(PEEP)
12% Increase over 5 years in 
primary net enrollment
National Seelye, Munger, et al. 
(1995:6).
Malawi “lower”
1%
Change in Grades 1-4
<Increase in overall pri￾mary dropout rate during 
this time>
National Hebert, Millsap, & El￾Sanabary (2002:10).
Benin 26.1% 
(total)
16.65% 
(girls)
National change in GER 
over 8 years (1990-1998); 
increase in girls’ GER 
from 1995-2001
National Williams, V. (2001:5).;
The Mitchell Group 
(2001: 5).
Change in Repetition
El Salvador 2.07% Over 5 years National AED (1998:7).
Honduras 
(PEEP)
55% 
reduction
Over 5 years National Seelye, Munger, et al. 
(1995:6).
Malawi 3.0% Grade 1-4 over 6 years National Hebert, Millsap, & El￾Sanabary (2002:11).
50
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
Annex 9. Observations about sustainability from project evaluation documents
Country Project Assessment of sustainability
Benin PENGOP No evidence of project being taken to scale with local funds.
Bolivia IRLP Sustainability is in question. Government is not capable of manag￾ing project (interactive radio learning) on its own. Sustainability 
depends on subsequent funding from the World Bank and other 
sources.
Ethiopia BESO Evaluation report does not provide strong (or much of any) evi￾dence that project activities will be sustained.
Honduras PEEP Ongoing mechanisms for quality control were not institutionalized. 
Analytic technology was not transferred.
Mission expressed doubts that training centers could be financially 
self-sustaining; training centers felt the increasing demands by 
USAID did not consider their financial and technical capacity to 
provide quality services.
Malawi GABLE Low likelihood of sustainability.
South Africa SABER Evaluation team questions the sustainability of the project in 
its present form. Evaluation team had serious doubts about the 
DDSP’s sustainability and replicability due to lack of integration of 
the project.
Namibia BES There are real causes for concern about costs and sustainability. 
There are grounds for suspecting that the system as it currently 
exists involves a level of cost that is probably not sustainable in the 
long run.
Swaziland EPMT Evaluation report questioned sustainability of the project because 
most of the project accomplishments were undertaken by expatriate 
advisers.
Uganda SUPER The sustainability of the supply of textbooks to primary schools 
hinges on the evolution of the macro economic situation, the extent 
that textbook provision is pursued at the expense of other recurrent 
investments in the sector, the extent to which the cost of materials 
are shared with parents, whether liberalization of procurement lead 
to cost savings, and how decentralization will impact allocation 
decisions.
Zambia CHANGES For the second year running, the project was not linked to work 
with the MOE when it was developing its annual work plan and 
budget. This is regrettable because the only way project activities 
can be maintained is if the development of annual work plans and 
budgets are conducted in a collaborative manner.
51
Acknowledgements
This paper was written for EQUIP2 by David Chapman (University of Minnesota) and Jessica Jester 
Quijada (Academy for Educational Develoment), 2008.
EQUIP2: Educational Policy, Systems Development, and Management is one of three USAID￾funded Leader with Associates Cooperative Agreements under the umbrella heading Educational 
Quality Improvement Program (EQUIP). As a Leader with Associates mechanism, EQUIP2 
accommodates buy-in awards from USAID bureaus and missions to support the goal of building 
education quality at the national, sub-national, and cross-community levels.
The Academy for Educational Development (AED) is the lead organization for the global EQUIP2 
partnership of education and development organizations, universities, and research institutions. The 
partnership includes fifteen major organizations and an expanding network of regional and national 
associates throughout the world: Aga Khan Foundation, American Institutes for Research, CARE, 
Center for Collaboration and the Future of Schooling, East-West Center, Education Development 
Center, International Rescue Committee, Joseph P. Kennedy, Jr. Foundation, Michigan State 
University, Mississippi Consortium for International Development, ORC Macro, Research Triangle 
Institute, University of Minnesota, University of Pittsburgh Institute of International Studies in 
Education, Women’s Commission for Refugee Women and Children.
For more information about EQUIP2, please contact:
USAID
Patrick Collins
CTO EGAT/ED
USAID Washington
1300 Pennsylvania Ave., NW
Washington, DC 20532
Tel: 202-712-4151
Email: pcollins@usaid.gov
AED
John Gillies
EQUIP2 Project Director
1825 Connecticut Ave., NW
Washington, DC 20009
Tel: 202-884-8256
Email: equip2@aed.org
Web: www.equip123.net
An Analysis of USAID Assistance to Basic Education in the Developing World, 1990-2005
52
This paper was made possible by the generous support of the American people through the United States Agency for 
International Development (USAID) under Cooperative Agreement No. GDG-A-00-03-00008-00. The contents are 
the responsibility of the Academy for Educational Development (AED) through the Educational Quality Improvement 
Program 2 (EQUIP2) and do not necessarily reflect the views of USAID or the United States Government.