This publication was produced at the request of the United States Department of Agriculture. It was prepared 
independently by Abt Associates. 
Foreign Agricultural 
Service, United States 
Department of Agriculture
McGovern-Dole Mozambique 
Impact Evaluation
June 15, 2022
Submitted by: Amy Power
Abt Associates
6130 Executive Boulevard
Rockville, MD 20852
DISCLAIMER: The author’s views expressed in this publication do not necessarily reflect the views of 
the United States Department of Agriculture or the United States Government.
McGovern-Dole Mozambique Program Impact Evaluation
The United States Department of Agriculture’s (USDA) McGovern-Dole International Food for 
Education and Child Nutrition (McGovern-Dole) Impact Evaluation in Mozambique was 
conducted by Abt Associates from July 2, 2021 to July 1, 2022. This report aims to evaluate the 
impact of two McGovern-Dole supported projects, one implemented by Planet Aid Inc. in 
Maputo province from 2015 to 2020 and one implemented by World Vision Inc. in Nampula 
province from 2015 to 2021. 
Agreement Number: GS-OOF-252CA/12314421F0453 
Project Duration: July 2, 2021 to July 1, 2022 
Implemented by: Abt Associates 
Evaluation Authored by: Abt Associates 
• Catherine Darrow
• Daniel Litwok
• Marius Meijerink
• Feliciano Chimbutane 
• Jillian Ouellette
• Megan Bogia 
• Adi Greif
• Amy Power
DISCLAIMER: The author’s views expressed in this publication do not necessarily reflect the views of 
the United States Department of Agriculture or the United States Government.
CONTENTS
ACKNOWLEDGEMENTS ........................................................................................................................ i
LIST OF EXHIBITS................................................................................................................................... ii
ABBREVIATIONS..................................................................................................................................... v
EXECUTIVE SUMMARY ........................................................................................................................ 1
1. OVERVIEW OF EVALUATION AND KEY FINDINGS ........................................................ 6
2. BACKGROUND ............................................................................................................................ 7
2.1 Benefit of this Evaluation....................................................................................................... 7
2.2 Mozambique Context ............................................................................................................. 7
2.3 McGovern-Dole International Food for Education and Child Nutrition Program ................. 8
2.4 Summary of Other Literacy and School Feeding Projects Implemented in the Maputo 
and Nampula Provinces ......................................................................................................... 9
2.5 Baseline Assessment............................................................................................................ 10
2.6 Other Evaluations Conducted by Implementing Partners .................................................... 11
2.7 Prior Evidence on School Feeding in Mozambique (or in the Sub-Saharan Region).......... 11
2.8 Prior Evidence on Teacher Pedagogy, Training, and Remedial Education in 
Mozambique (or in the Sub-Saharan Region) ...................................................................... 12
2.9 Quasi-Experimental Impact Evaluation ............................................................................... 12
2.10 Differences Between Baseline and Endline Approaches ..................................................... 13
2.11 COVID-19 Effects on Schooling ......................................................................................... 14
3. EVALUATION APPROACH..................................................................................................... 15
3.1 Impact and Implementation Focus ....................................................................................... 15
3.2 Evaluation Questions ........................................................................................................... 15
4. METHODOLOGY ...................................................................................................................... 18
4.1 Sampling Frame ................................................................................................................... 18
4.2 Qualitative Sampling Design ............................................................................................... 18
4.3 Outcomes ............................................................................................................................. 19
4.4 Impact Analysis Models....................................................................................................... 21
4.5 Implementation Fidelity Analyses....................................................................................... 24
4.6 Analysis of Stakeholder Perspectives .................................................................................. 24
4.7 Analysis of Impacts in Adequate Fidelity Sites................................................................... 25
5. DATA COLLECTION................................................................................................................ 27
5.1 Overview .............................................................................................................................. 27
5.2 Participant Contact and Information Collection................................................................... 27
5.3 Enumerator Recruitment and Training ................................................................................. 27
5.4 Data Quality and Security .................................................................................................... 28
5.5 Data Management and Security ........................................................................................... 29
5.6 COVID-19 Specific Precautions.......................................................................................... 29
6. CHARACTERISTICS OF POPULATION............................................................................... 31
6.1 Sample Description.............................................................................................................. 31
6.2 Student Demographic Characteristics.................................................................................. 31
DISCLAIMER: The author’s views expressed in this publication do not necessarily reflect the views of 
the United States Department of Agriculture or the United States Government.
6.3 Language of Literacy Instruction ......................................................................................... 32
6.4 Teacher Characteristics........................................................................................................ 32
6.5 School Characteristics.......................................................................................................... 33
6.6 Parent Focus Group Characteristics..................................................................................... 34
6.7 Key Official Characteristics................................................................................................. 35
7. MAIN FINDINGS........................................................................................................................ 36
7.1 Summary of Findings ........................................................................................................... 36
7.2 Impacts on Student Reading Performance ........................................................................... 36
7.3 Impacts on Student Attendance ............................................................................................ 38
7.4 Impacts on Student Attentiveness ........................................................................................ 40
7.5 Impacts for Student Subgroups ............................................................................................ 41
8. SECONDARY FINDINGS.......................................................................................................... 44
8.1 Impacts on Instructional Inputs.............................................................................................. 44
8.2 Impacts on Feeding Outcomes and Nutrition Knowledge ................................................... 45
8.3 Impacts on Student Knowledge and Use of Recommended Health Practices ..................... 47
9. FINDINGS ON IMPLEMENTATION...................................................................................... 49
9.1 Summary of Findings ........................................................................................................... 49
9.2 Determination of Implementation Adequacy ....................................................................... 49
9.3 Implementation of School Feeding ...................................................................................... 50
9.4 Implementation of Literacy Instruction................................................................................ 51
9.5 Implementation of Teacher Training ................................................................................... 51
10. FINDINGS ON IMPACT CONTROLLING FOR IMPLEMENTATION............................ 53
10.1 Summary of Findings ........................................................................................................... 53
10.2 Impact Findings for Schools with Combined Literacy and Feeding Interventions .............. 53
10.3 Impact Findings for Schools with Feeding Interventions Only ........................................... 53
11. PROJECT ADAPTATIONS RESULTING FROM COVID-19.............................................. 57
11.1 COVID-19 and School Operations ...................................................................................... 57
11.2 COVID-19 and School Feeding Interventions ..................................................................... 58
11.3 COVID-19 and Health Services ........................................................................................... 59
12. EVALUATION CONSTRAINTS AND LIMITATIONS......................................................... 60
12.1 School Closures and Gaps in Project Delivery .................................................................... 61
12.2 Sample Size Added Challenges to Detecting Impacts of School Feeding Only .................. 61
12.3 Multiple Comparisons.......................................................................................................... 61
13. CONSIDERATIONS WHEN CONDUCTING EVALUATIONS DURING 
PANDEMICS AND NATURAL DISASTERS....................................................................................... 62
14. CONSIDERATIONS FOR MCGOVERN-DOLE PROGRAM SUSTAINABILITY 
BY STAKEHOLDERS............................................................................................................................. 63
14.1 Considerations for School Feeding ...................................................................................... 63
14.2 Considerations for Literacy.................................................................................................. 64
14.3 Considerations on Sustainability.......................................................................................... 65
DISCLAIMER: The author’s views expressed in this publication do not necessarily reflect the views of 
the United States Department of Agriculture or the United States Government.
15. RECOMMENDATIONS FOR FUTURE MCGOVERN-DOLE PROGrAMMING............ 66
15.1 Expand Awareness of McGovern-Dole Program Theory of Change and Impact 
Evaluation Criteria ............................................................................................................... 66
15.2 Improving Teachers’ Capacity to Support Student Development of Early Reading 
Skills .................................................................................................................................... 66
15.3 Language of Early-Grade Literacy Development ................................................................ 67
15.4 Increase Gender Sensitivity and Improve Instruction Focusing on Girls ............................ 67
15.5 Ensure the Sustainability of McGovern-Dole Programming ............................................... 68
16. REFERENCES............................................................................................................................. 69
Appendix A: Supporting Exhibits........................................................................................................... 72
Appendix B: Data Collection Instruments ............................................................................................ 132
Abt Associates June 2022 ▌i 
ACKNOWLEDGEMENTS
This impact evaluation would not have been possible without the financial support of the United 
States Department of Agriculture – Foreign Agriculture Service (USDA/FAS). We extend 
special thanks to the USDA/FAS team including Delphine Hamlin, who served as our 
contracting officer; Traci Johnson, Amy Elhadi, and Molly Kairn, who provided ongoing 
technical guidance and program support; and Almeida Zacarias, who supported our team from 
the USDA office in Maputo. 
We extend our gratitude to the Ministry of Education and Human Development and its 
Provincial and District Directorates for their contributions and advice, specifically during the 
design phase of the study. Our data collection efforts would not have been successful without 
their support. We thank all the school management teams, community members, and local 
government officials for the support they extended to our team, with a special recognition to all 
the students and teachers who voluntarily participated in the study. 
We deeply appreciate the support of numerous staff at World Vision, Planet Aid Inc., ADPP, 
RTI, and Counterpart International who provided ongoing valuable insights and advice 
throughout the evaluation. We acknowledge the dedicated work of the entire Ernst & Young 
team led by José Diogo, who were responsible for data collection, using safe data collection 
practices during the COVID-19 pandemic. We greatly appreciate the work of Maria Boyle, Ben 
Johns, and Christine Ohresser-Joumard and all Abt staff who contributed to the evaluation. We 
are profoundly grateful for the technical expertise and guidance provided throughout the entire 
evaluation by Barbara Goodson and David Judkins, Senior Advisors with Abt Associates. Lastly, 
we extend thanks to Aster Charmine Sitoe for supporting the data collection process and special 
thanks to Feliciano Chimbutane for providing invaluable technical guidance, critical insights, 
and program support throughout every step of the evaluation. 
Abt Associates June 2022 ▌ii
LIST OF EXHIBITS
Exhibit 1. Evaluation Questions.................................................................................................................. 16
Exhibit 2. Score Ranges and Implementation Adequacy Thresholds ......................................................... 26
Exhibit 3. Staff Hired for Data Collection .................................................................................................. 27
Exhibit 4. Baseline and Endline Evaluation Sample ................................................................................... 31
Exhibit 5. School Characteristics, by Study Arm and Time ....................................................................... 34
Exhibit 6. Parent Focus Group Sample with Literacy and School Feeding Treatment............................... 34
Exhibit 7. Parent Focus Group Sample with School Feeding Treatment Only ........................................... 34
Exhibit 8. Summary of Impact on Average Mean Reading Performance, Attendance, and Attentiveness 36
Exhibit 9. Impacts on Average Mean Reading Comprehension Scores ..................................................... 37
Exhibit 10. Impacts on Student Attendance (Measured by Any Absence Last Week) ............................... 39
Exhibit 11. Impacts on Teacher-Reported Student Attentiveness............................................................... 40
Exhibit 12. Impacts on Student Reading Performance and Attendance for Boys and Girls ....................... 42
Exhibit 13. Impacts on Instructional Inputs ................................................................................................ 44
Exhibit 14. Impacts on Feeding Outcomes ................................................................................................. 46
Exhibit 15. Impacts on Student Knowledge and Use of Recommended Health Practices ......................... 47
Exhibit 16. Proportion of Adequate Implementing Schools Included in Analyses .................................... 50
Exhibit 17. Number of Schools Without a Feeding Intervention at Endline .............................................. 50
Exhibit 18. Average Fidelity Scores Across School by Intervention Component and by Treatment Arm . 51
Exhibit 19. Impacts on Average Mean Reading Comprehension Scores (Adequately Implementing 
Treatment Schools Only) ............................................................................................................................ 54
Exhibit 20. Student Attendance (Measured by Any Absence Last Week; Adequately Implementing 
Treatment Schools Only) ............................................................................................................................ 55
Exhibit 21. Impacts on Teacher-Reported Student Attentiveness (Adequately Implementing Treatment 
Schools Only).............................................................................................................................................. 56
Exhibit 22. Percentage of Treatment Schools Where Students Received a Daily Meal During COVID-19
.................................................................................................................................................................... 58
Exhibit 23. Percentage of Treatment Schools Where School Garden Was Cared For During COVID-19 58
Exhibit 24. Percentage of Schools Where Families Received Take-Home Rations During COVID-19 .... 59
Appendix A
Exhibit A1. McGovern-Dole Results Framework (SO1)............................................................................ 72
Exhibit A2. McGovern-Dole Results Framework 2 ................................................................................... 73
Exhibit A3. Program logic model with representation of project implementation adequacy and impacts . 74
Exhibit A4. School Feeding Fidelity Index ................................................................................................ 75
Exhibit A5. Literacy Fidelity Index ............................................................................................................ 76
Exhibit A6. Head Teacher and Teacher Training Fidelity Index ................................................................ 77
Exhibit A7. Sample Size, Data Source, and Level of Analysis by Outcome Indicator .............................. 78
Exhibit A8. Student Responses on Household Demographic Characteristics ............................................ 79
Exhibit A9. Language of Literacy Instruction and Difference with Language Spoken at Home ............... 80
Exhibit A10. Teachers Characteristics........................................................................................................ 81
Exhibit A11. Reading Comprehension Estimates for School Feeding Only Schools ................................. 82
Exhibit A12. Reading Comprehension Estimates for School Feeding and Literacy Schools ..................... 82
Exhibit A13. Listening Comprehension Estimates for School Feeding Only Schools ............................... 82
Exhibit A14. Listening Comprehension Estimates for School Feeding and Literacy Schools ................... 83
Exhibit A15. Letter Identification Estimates for School Feeding Only Schools ........................................ 83
Exhibit A16. Letter Identification Estimates for School Feeding and Literacy Schools ............................ 83
Exhibit A17. Syllable Identification Estimates for School Feeding Only Schools..................................... 84
Exhibit A18. Syllable Identification Estimates for School Feeding and Literacy Schools ......................... 84
Exhibit A19. Oral Reading Fluency Estimates for School Feeding Only Schools ..................................... 84
Abt Associates June 2022 ▌iii
Exhibit A20. Oral Reading Fluency Estimates for School Feeding and Literacy Schools ......................... 85
Exhibit A21. Word Dictation Estimates for School Feeding Only Schools ............................................... 85
Exhibit A22. Word Dictation Estimates for School Feeding and Literacy Schools ................................... 85
Exhibit A23. Impact Estimates on Student Attendance (Student Self-Reported Absence At least One Day 
Last Week) in School Feeding Only Schools ............................................................................................. 86
Exhibit A24. Impact Estimates on Student Attendance (Student Self-Reported Absence At least One Day 
Last Week) in School Feeding and Literacy Schools ................................................................................. 86
Exhibit A25. Impact Estimates on Student Attentiveness as Perceived by Teachers in School Feeding 
Only Schools ............................................................................................................................................... 86
Exhibit A26. Impact Estimates on Student Attentiveness as Perceived by Teachers in School Feeding and 
Literacy Schools.......................................................................................................................................... 86
Exhibit A27. Teachers' Perception on Student Attentiveness..................................................................... 87
Exhibit A28. Impact on Teacher-Reported Instructional Input in School Feeding Only Schools .............. 87
Exhibit A29. Impact on Teacher-Reported Instructional Input in School Feeding and Literacy Schools .. 88
Exhibit A30. Impact on Feeding Outcomes in School Feeding Only Schools ........................................... 89
Exhibit A31. Impact on Feeding Outcomes in School Feeding and Literacy Schools ............................... 89
Exhibit A32. Impact on Student Knowledge and Health Practices in School Feeding Only Schools ........ 90
Exhibit A33. Impact on Student Knowledge and Health Practices in School Feeding and Literacy Schools
.................................................................................................................................................................... 90
Exhibit A34. Impact on Students’ Reading Comprehension in School Feeding Only Schools, by Grade 
and Gender .................................................................................................................................................. 91
Exhibit A35. Impact on Students’ Reading Comprehension in School Feeding and Literacy Schools, by 
Grade and Gender ....................................................................................................................................... 92
Exhibit A36. Impact of Adequate Implementing Programs on Student Reading Comprehension ............. 93
Exhibit A37. Impact of Adequate Implementing Programs on Oral Comprehension ................................ 93
Exhibit A38. Impact of Adequate Implementing Programs on Letter Identification.................................. 94
Exhibit A39. Impact of Adequate Implementing Programs on Syllable Identification.............................. 95
Exhibit A40. Impact of Adequate Implementing Programs on Oral Reading Fluency .............................. 95
Exhibit A41. Impact of Adequate Implementing Programs on Word Dictation......................................... 96
Exhibit A42. Impact of Adequate Implementing Programs on Student Attendance .................................. 96
Exhibit A43. Impact of Adequate Implementing Programs on Student Attentiveness ............................... 96
Exhibit A44. Impact on Student Reported Access to School Supplies and Study Materials ...................... 97
Exhibit A45. Impact on Teacher Access to Supplies and Learning Materials ............................................ 98
Exhibit A46. Percentage of Teachers Reporting Access to School Supply Items ...................................... 98
Exhibit A47. Impact on Head Teacher Reporting Access to Library/Book Banks and Materials .............. 99
Exhibit A48. Percentage of Head Teachers Reporting Availability of Specific Types of Print Materials . 99
Exhibit A49. Impact on Teacher Reporting on Teacher Training ............................................................ 100
Exhibit A50. Teachers Reporting on Teacher Training ............................................................................ 101
Exhibit A51. Teacher Knowledge and Skills ............................................................................................ 102
Exhibit A52. Impact on Head Teacher Reported Teacher Training ......................................................... 103
Exhibit A53. Head Teacher Knowledge and Skills .................................................................................. 104
Exhibit A54. Percentage of Head Teachers Reporting Teachers Attended Literacy Training and WI or 
PAI Training ............................................................................................................................................. 104
Exhibit A55. Impact on Head Teacher Reported Number of Male and Female Teachers who Attended 
Literacy Training ...................................................................................................................................... 105
Exhibit A56. Impact on Head Teachers’ Reported Reading Coaches Training and Supervising Teachers, 
and Number of Trainings Reading Coaches Provided .............................................................................. 106
Exhibit A57. Percentage of Head Teachers Reporting on Frequency Reading Coaches Supervise Teachers
.................................................................................................................................................................. 106
Exhibit A58. Impact of Head Teachers’ Reported Teacher Attendance and Head Teacher Attendance.. 107
Exhibit A59. Head Teacher Reporting on How Class is Handled when Teacher is Absent..................... 107
Exhibit A60. Impact on Students Reporting School Serving Schools Meals, Current Hunger, and 
Household Hunger .................................................................................................................................... 108
Abt Associates June 2022 ▌iv
Exhibit A61. Percentage of Students Reporting that School Served School Meals Yesterday, and 
Frequency of Household Hunger .............................................................................................................. 109
Exhibit A62. Head teacher Reporting on School Feeding Program ......................................................... 109
Exhibit A63. Impact on Teacher-Reported Nutrition and Hygiene Teaching .......................................... 110
Exhibit A64. Teachers' Perception of Student Knowledge of Nutrition, Hygiene, and Sanitation........... 110
Exhibit A65. Impact on Head Teacher Reported Teacher Training on Nutrition, Hygiene, and Health .. 111
Exhibit A66. Head Teacher Reported Frequency of Nutrition, Hygiene, and Sanitation Lessons ........... 111
Exhibit A67. Impact on Students Receiving Deworming Medication ...................................................... 112
Exhibit A68. Head Teacher Reported Cleanliness of latrines ................................................................... 112
Exhibit A69. Cooks Reporting on School Infrastructure .......................................................................... 113
Exhibit A70. Cook Training and Knowledge ........................................................................................... 114
Exhibit A71. Percentage of Schools that Adequately Implemented Literacy, Training, and Feeding 
Program Items........................................................................................................................................... 115
Exhibit A72. Impact on Reading Comprehension by Literacy Item Adequately Implemented by Schools 
in the Treatment Arms.............................................................................................................................. 116
Exhibit A73. Impact on Reading Comprehension by Training Item Adequately Implemented by Schools 
in the Treatment Arms.............................................................................................................................. 117
Exhibit A74. Impact on Reading Comprehension by Feeding Item Adequately Implemented by Schools in 
the Treatment Arms .................................................................................................................................. 118
Exhibit A75. Impact on Student Attendance by Literacy Item Adequately Implemented by Schools in the 
Treatment Arms........................................................................................................................................ 120
Exhibit A76. Impact on Student Attendance by Training Item Adequately Implemented by Schools in the 
Treatment Arms........................................................................................................................................ 121
Exhibit A77. Impact on Student Attendance by Feeding Item Adequately Implemented by Schools in the 
Treatment Arms........................................................................................................................................ 122
Exhibit A78. Impact on Student Attentiveness by Literacy Item Adequately Implemented by Schools in 
the Treatment Arms .................................................................................................................................. 124
Exhibit A79. Impact on Student Attentiveness by Training Item Adequately Implemented by Schools in 
the Treatment Arms .................................................................................................................................. 125
Exhibit A80. Impact on Student Attentiveness by Feeding Item Adequately Implemented by Schools in 
the Treatment Arms .................................................................................................................................. 126
Exhibit A81. Outcome Levels at Baseline and Endline by Treatment Arm for Main Outcomes............. 129
Exhibit A82. Outcome Levels at Baseline and Endline by Treatment Arm for Instructional Inputs ....... 129
Exhibit A83. Outcome Levels at Baseline and Endline by Treatment Arm for Feeding Outcomes and 
Nutrition Knowledge ................................................................................................................................ 130
Exhibit A84. Outcome Levels at Baseline and Endline by Treatment Arm for Student Knowledge and Use 
of Recommended Health Practices........................................................................................................... 130
Abt Associates June 2022 ▌v 
ABBREVIATIONS
ADPP Ajuda de Desenvolviemento de Povo para Povo 
BL Baseline
CapED Capacity Development for Education Programme 
C Control Group Schools 
EGRA Early Grade Reading Assessment
EL Endline 
EY Ernst & Young 
FAS Foreign Agriculture Service
IFP Teacher Training Institution
McGovern-Dole McGovern-Dole International Food for Education and Child Nutrition 
Program
MINEDH Ministry of Education and Human Development 
PAI Planet Aid Inc.
PFG Parent Focus Group 
PRONAE Programa Nacional de Alimentação Escolar
SF School Feeding 
SF+L School Feeding and Literacy 
TTC Teacher Training College
UNESCO United Nations Educational, Scientific, and Cultural Organization 
USAID United States Agency for International Development 
USDA United States Department of Agriculture
WASH Water, Sanitation, and Hygiene 
WFP United Nations World Food Programme 
WV World Vision 
Abt Associates June 2022 ▌1 
EXECUTIVE SUMMARY
The United States Department of Agriculture’s (USDA) McGovern-Dole International Food for 
Education and Child Nutrition Program (McGovern-Dole) uses commodities grown by American 
farmers to support education, child development, and food security in low-income, food-deficit 
countries around the globe. McGovern-Dole projects support activities to address two critical 
issues for children in Mozambique: (1) food insecurity and hunger and (2) inadequate literacy 
skills for school achievement and life success. In an effort to reduce hunger and improve literacy 
among primary school children in Mozambique, McGovern-Dole, together with implementing 
partner organizations, provides food in schools, teacher training in effective literacy instruction, 
and other resources. To ensure sustainability of these project activities, as well as the targeted 
literacy, health, and nutrition outcomes, McGovern-Dole works to build the capacity of 
participant communities to sustain food sources and instructional quality on their own or by 
working with government agencies and local organizations. 
McGovern-Dole Projects in Mozambique
USDA/Foreign Agricultural Service (FAS) supported a first-of-its-kind evaluation of the 
effectiveness of the McGovern-Dole interventions across multiple contexts within Mozambique. 
The impact evaluation, conducted by Abt Associates Inc., focused on schools in six districts: 
four in Maputo province (Magude, Manhiça, Matutuine, and Moamba) and two in Nampula 
province (Muecate, Nacaroa). 
The evaluation focuses on two McGovern-Dole projects in Mozambique, one implemented by 
Planet Aid, Inc. in Maputo province from 2015 to 2020 and one implemented by World Vision, 
Inc. in Nampula province from 2015 to 2021. These projects were the focus of a baseline 
assessment conducted by RTI International in 2017. McGovern-Dole supported activities 
expected to result in improved student performance in school. The McGovern-Dole projects 
supported increased school success by providing food at school and through nutrition and health 
interventions. Reducing hunger and food insecurity were expected to increase student attendance 
and student ability to pay attention during school. Some project schools also participated in 
interventions designed to increase student success directly, through higher quality literacy 
instruction, by providing teacher and school administrator training, curricular support, and access 
to school supplies. Improved attendance and attentiveness were expected to result in students 
having greater exposure to the higher-quality instruction. 
In Mozambique, the McGovern-Dole projects consisted of two different sets of interventions: 
schools either received school feeding plus literacy or they received school feeding only. In the 
four districts in Maputo province, both sets of interventions were implemented in the schools; in 
the two districts in Nampula province, only the school feeding plus literacy interventions were 
implemented. The decisions about which interventions were implemented in which districts and 
schools was made in 2012, before the decision to conduct an impact evaluation was made. 
Assignment to the interventions was therefore non-random. 
Evaluation Objective 
The purpose of this evaluation was to provide USDA/FAS with findings on the impacts of the 
two McGovern-Dole projects in Mozambique upon project completion. Through this impact 
evaluation, the study team was able to understand the role of different strategies used for 
McGovern-Dole interventions; strategies including community engagement, access to clean 
Abt Associates June 2022 ▌2 
water and sanitation services, supportive community-based literacy, teacher training and support, 
and feeding activities. The findings in this report can be used to better understand the 
implementation of school feeding and literacy projects in Mozambique and will inform future 
McGovern-Dole programming in Mozambique, and, potentially, programming more broadly 
across countries.
Evaluation design
This evaluation assesses the impact of the two sets of interventions of the McGovern-Dole 
projects after four years of implementation. The evaluation focused on the outcomes for students 
in third grade (early-grade students), who had the opportunity for up to two years of exposure to 
the intervention(s), and students in fifth grade (mid-grade students) who had the opportunity for 
up to four years of exposure to the intervention(s). The impact evaluation was conducted across 
two projects in two provinces and was not designed to compare impacts by province or 
implementing partner. 
The impact evaluation compared average school-level performance at RTI International’s
baseline assessment to average school-level performance at endline, four years later. Abt 
Associates used the same evaluation design, instruments, key outcomes, and research questions 
as in the baseline report to compare changes on key indicators from the project activities. 
The evaluation is quasi-experimental because schools were not randomly assigned to a version of 
the intervention. The ramification of this design versus a randomized design is that we cannot be 
certain that differences between intervention and control schools is entirely due to the difference 
in programming but may also be affected by other differences that might be related to the 
outcomes, some of which could be unplanned and unobservable, such as varying COVID-19 
transmission rates across school communities or differential changes in school staff composition 
brought about by COVID-related turnover. 
The primary impact evaluation questions for the endline evaluation are about the effects of the 
McGovern-Dole projects in Mozambique on student outcomes: literacy, attendance, and 
attentiveness. Secondary evaluation questions ask about effects on teachers and schools. These 
questions are answered by separately comparing the students and teachers in schools in each of 
the two intervention arms, with the control arm. There are also exploratory impact evaluation 
questions about (1) intervention effects in a subset of children attending adequately 
implementing schools and (2) intervention effects for younger versus older students and for boys 
versus girls. 
Key findings on student outcomes
• Effect on reading comprehension: Students in schools implementing the combined literacy
and school feeding intervention improved their average reading comprehension scores
relative to students in comparison schools. Early-grade students experienced an increase of
4.8 percentage points and mid-grade students experienced an increase of 7.6 percentage
points in their reading comprehension. Students in schools implementing the school feeding
only intervention did not improve average mean reading comprehension relative to students
in comparison schools.
 Impact analyses using only students attending adequately implementing treatment schools
 found stronger impacts on student reading comprehension for both early-grade (0.6
 percentage point increase) and mid-grade (3.3 percentage point increase) students compared
 to impacts in all treatment schools.
Abt Associates June 2022 ▌3 
• Effect on student attendance: No impact was observed on student attendance (absences) for
either early-grade or mid-grade students who experienced either the combined school feeding
and literacy intervention or the school feeding intervention alone. Unlike impacts on reading
comprehension, adequately implementing schools in either treatment conditions (school
feeding plus literacy and school feeding alone) did not improve average mean student
attendance relative to comparison schools.
• Effect on student attentiveness: Neither school feeding only nor school feeding and literacy
improved student attentiveness for either early- or mid-grade students relative to comparison
schools. For mid-grade students in school feeding only schools, teacher-reported student
attentiveness declined by 7.2 percentage points relative to comparison schools. Adequately
implementing schools in either treatment conditions (school feeding plus literacy and school
feeding alone) did not improve average mean student attentiveness relative to comparison
schools.
Implementation findings
• On average, McGovern-Dole projects were implemented at adequate levels of fidelity across
intervention components, although there was variation in the fidelity scores across the
treatment schools.
• Overall, the schools in Nampula were rated as having better implementation of their
interventions, compared to the schools in Maputo. Nampula schools implementing school
feeding and literacy interventions received the highest scores on the evaluation’s
implementation fidelity scales.
• There was evidence that alternative programming of a similar nature to the McGovern-Dole
program was concurrently being implemented in comparison schools. Because McGovern￾Dole schools had comparatively higher implementation scores, the contrast between
treatment and comparison schools was detectable. More robust implementation in
McGovern-Dole schools could produce a larger contrast to alternative programming, greater
improvements in student outcomes, and stronger impacts overall.
Considerations
The positive findings on student literacy are very promising. Despite the diminished exposure on 
students and weakened training and supports for teachers around literacy instruction caused by 
COVID-related disruptions, the evaluation detected some positive findings. This is evident by 
the substantial positive impacts detected for teacher training on the fundamentals of reading 
instruction (15.3 percentage points for teachers of early-grade students and 27.9 percentage 
points for teachers of mid-grade students). These results suggest that teacher training is a 
plausible mechanism for explaining the impacts on student reading performance. Even in times 
of great disruption for schools as was the case between 2020 and 2021, training may be 
associated with changes in teacher practice, which is linked to increases in student literacy 
development. 
The lack of effects on attendance and attentiveness suggests that the impact on literacy was not 
because students were in school more often or more attentive but that the instruction was more 
effective. It is possible that the effects on student literacy would have been larger if the feeding 
intervention had resulted in improving these aspects of students’ readiness to learn. As discussed 
below, there were limitations affecting our ability to estimate the combined effectiveness of the 
two interventions delivered in combination. 
Abt Associates June 2022 ▌4 
Limitations
Interpretation of the evaluation findings is limited by several key factors. These include factors 
related to outcome measures, precision of the impact estimates, gaps in implementation, and 
disproportionate sample size. 
Outcome measures: Two of the primary outcome measures were based on survey responses that 
are subject to potential bias. Student attendance was measured with self-reported data from 
young children, and student attentiveness was measured using teacher recall in the teacher 
survey. In both cases, these measures could suffer from bias due to recall and measurement error.
However, given that the survey included a substantial number of students and teachers, the effect 
of the expected errors could be minimized. For future evaluations, it is advisable to consider 
alternative measures of outcomes like student attentiveness to capture impacts at the student 
level more precisely. Relatedly, teacher-reported levels of student attentiveness were very high 
(greater than 90 percent in all arms) at baseline. This implies there is limited room for growth in 
this measure at endline. 
Precision of impact estimates: The evaluation interprets the estimates from the quasi￾experimental difference-in-differences analysis as the causal impact of the McGovern-Dole 
projects. The difference-in-differences design assumes the observed trend in the comparison 
schools can be used to project what the outcome would have been in treatment schools in the 
absence of treatment. This typically implies comparison schools do not experience any treatment.
However, our assessment of implementation fidelity found that many comparison schools had 
some evidence of comparable feeding, literacy, and teacher training activities, although levels of 
implementation were determined to be low. For these schools, the implied comparison is not of 
McGovern-Dole interventions to no intervention, but of McGovern-Dole interventions to other 
programs being implemented. This could either attenuate or exacerbate the treatment effect and
complicates interpretation of the average effect; however, if the conditional parallel trends
assumption holds, it does not threaten the causal interpretation of the findings. 
Gaps in implementation: In Maputo, Planet Aid was contracted to implement the McGovern￾Dole project from 2015 to 2020. Counterpart was then awarded the contract in Maputo from 
2020 to 2025. However, at the time of endline data collection in October 2021, Counterpart was 
still preparing to implement the project. School feeding and literacy interventions were not yet in 
place, creating a situation in which the study team was evaluating an intervention that had phased 
out a year before. Schools themselves were also functioning at a reduced level. When reopening 
in March 2021, after a year of COVID-19-related closures, many students were going to school 
only two or three times a week. 
Across provinces, a prevailing issue with the impact evaluation is that timing of data collection
in 2021 proved problematic in detecting delivery of school feeding and other project elements.
Even if project delivery such as the provision of school meals, supplemental role of school 
gardens, access to clean water, and use of literacy materials was well established up until school
closures in March 2020, their presence diminished in Nampula and disappeared almost entirely 
in Maputo by endline. Because of this, intervention impacts may be difficult to detect, especially 
when capturing outcomes of interest like access to meals and student short-term hunger. 
When conducting a fair test of a project’s effectiveness, as was the goal of this evaluation, it is
clear that project implementation matters. The lack of project implementation makes it difficult 
to understand impacts, as a lack of impact may be more likely due to the absence of the project
rather than a flawed theory of change or ineffective project design. In this evaluation, given the 
Abt Associates June 2022 ▌5 
unprecedented disruptions caused by the COVID-19 pandemic, we suspect that intervention 
impacts were likely weakened by project disruption and less than ideal implementation. 
Disproportionate sample sizes. Because it was only tested in Maputo, the school feeding only 
treatment arm had a smaller quantitative sample than the literacy and school feeding arm or the 
comparison group. This smaller sample implies impacts had to be larger in that treatment arm 
than in the literacy and school feeding arm to achieve statistical significance.
Recommendations
Given the interest in understanding the effects of the McGovern-Dole projects and the aim of 
sustaining project implementation, the study team offers the following recommendations: 
Expand stakeholder awareness of the McGovern-Dole program theory of change and 
engagement in identifying student and school outcomes of interest. Understanding the final 
goal of the McGovern-Dole program and the interconnection among the different components of 
the projects including desired outcomes would help ministry officials, school managers, teachers, 
and parents better monitor progress toward that goal. This understanding at the local level could 
encourage parents to advocate for stronger implementation and collaborate with school leaders 
and other stakeholders in meeting those goals. 
Improve teachers’ capacity to support student development of early reading skills. The 
study team recommends that teacher training increase focus on early-grade literacy instruction, 
allowing teachers to better understand the casual relationship between lower- and high-order 
reading skills and enhance their capacity to help children develop these two levels of reading 
growth. We also suggest strengthening the connection between the Pedagogical University and 
teacher training activities with local schools. 
Expand dual language of early-grade literacy development. McGovern-Dole projects could 
further support the scaling of dual language early literacy instruction inclusive of languages 
familiar to the child (their home languages). McGovern-Dole projects may help local education 
authorities to expand the bilingual program to more schools or to cover more classes in schools 
where it is already in place. 
Increase gender sensitivity and improve instruction focusing on girls. McGovern-Dole 
projects work to improve the environment that allows girls to have equal opportunities for 
learning in schools, homes, and larger communities. We suggest looking more closely at the 
implementation of programming for girls to determine if community members understand the 
importance of girls’ education, and training teachers on gender-sensitive classroom interaction 
and reading practices. We also suggest enhancing the gender-sensitivity approach with the 
production of teaching and learning materials and in the selection and hiring practices for school 
leadership positions. 
Ensure the sustainability of McGovern-Dole programming. The study team encourages 
USDA/FAS, implementing partners, ministry officials, and others to put in place strong exit 
strategies to allow sustainability of the McGovern-Dole interventions after the end of designated 
projects, including through increased involvement of local education authorities and 
establishment of linkages with similar government programs/interventions. 
Abt Associates June 2022 ▌6 
1. OVERVIEW OF EVALUATION AND KEY FINDINGS
The United States Department of Agriculture – Foreign Agriculture Service (USDA/FAS) 
supported an evaluation of McGovern-Dole International Food for Education and Child Nutrition 
(McGovern-Dole) in Mozambique. This evaluation, conducted by Abt Associates Inc., focuses 
on impacts from 2015 to 2020 in four districts of Maputo province (Magude, Manhiça, 
Matutuine, and Moamba) and two districts of Nampula province (Muecete and Nacaroa). These 
projects were the focus of a baseline evaluation conducted by RTI International in 2017. The 
report aims to evaluate the impact of two McGovern-Dole interventions: literacy and school 
feeding interventions combined and a school feeding intervention alone. The evaluation 
estimated impacts at endline, after four years of implementation, and reports program impacts 
between baseline and endline. The evaluation focused on the outcomes for students in grade 3 
(early-grade students) who had two years of exposure to the intervention and students in grade 5 
(mid-grade students) who had experienced four years of the intervention. The evaluation was 
conducted in six districts: four districts in Maputo province, where both interventions were 
implemented, and two in Nampula province, where only the school feeding intervention was 
implemented.
Key findings of this evaluation include:
• Effect on reading comprehension: Students in schools implementing the combined literacy 
and school feeding intervention improved their average reading comprehension scores 
relative to students in comparison schools. Early-grade students experienced an increase of 
4.8 percentage points and mid-grade students experienced an increase of 7.6 percentage 
points in their reading comprehension. Students in schools implementing school feeding 
alone did not improve average mean reading comprehension relative to students in 
comparison schools. 
• Effect on student attendance: No impact was observed on student attendance (absences) for 
either early-grade or mid-grade students who experienced either the combined school feeding 
and literacy intervention or the school feeding intervention alone.
• Effect on student attentiveness. Neither school feeding nor school feeding and literacy 
improved student attentiveness for either early- or mid-grade students relative to comparison 
schools. For mid-grade students in school feeding only schools, teacher-reported student 
attentiveness declined by 7.2 percentage points relative to comparison schools. 
• Impacts for schools with adequate fidelity of implementation. On average, McGovern￾Dole projects were implemented at adequate levels of fidelity, although there was variation in 
the fidelity scores across the treatment schools. Impact analyses using only adequately
implementing treatment schools found stronger impacts on student reading comprehension 
for both early grade (0.6 percentage point increase) and mid-grade (3.3 percentage point 
increase) students compared to impacts in all treatment schools. 
• Impacts on teacher training. Substantial positive impacts were detected for teacher training 
on the fundamentals of reading instruction (15.3 percentage points for teachers of early-grade 
students and 27.9 percentage points for teachers of mid-grade students) suggesting that 
teacher training is a plausible mechanism for explaining the impacts on student reading 
performance.
Abt Associates June 2022 ▌7 
2. BACKGROUND
2.1 Benefit of this Evaluation 
The purpose of this evaluation is to provide USDA/FAS with findings on the impacts of two of 
its McGovern-Dole projects in Mozambique upon project completion. Through this endline 
impact evaluation, the study team was able to understand the role of different strategies used for 
McGovern-Dole activities; strategies including community engagement, access to clean water 
and sanitation services, supportive community-based literacy, teacher training and support, and 
feeding activities. The findings in this report can be used to better understand the implementation 
of school feeding and literacy projects in Mozambique and will inform future McGovern-Dole 
programming in Mozambique, and, potentially, programming more broadly across countries. 
2.2 Mozambique Context 
Mozambique has made progress over the past decade in improving both the literacy and nutrition 
of its population. Nevertheless, rates of illiteracy and malnutrition remain high, particularly in 
rural areas and among women. In addition, during the period of the McGovern-Dole projects
being assessed by this impact evaluation (2015–2021), the country was hit with major events, 
including the COVID-19 pandemic and Tropical Cyclone Eloise. These events had the potential 
to negatively influence progress toward improving literacy and nutrition of school-age children. 
Education in Mozambique
In most provinces in Mozambique, schools are overcrowded, student/teacher ratios are high, and 
reading and math test scores are low. One study on school operations found that teacher 
absenteeism, limited instructional time, and other factors limited Mozambican schools to, on 
average, 30 days of actual instructional time per the 193-day school year in 2010 (USAID, 
2021a). Furthermore, the study found that 59 percent of third grade students in the 49 schools 
studied could not read a single word per minute nor recognize letters, and that students who 
could read were able to read, on average, only five words per minute. More than one-third of 
students drop out before Grade 3 and less than half complete primary school (Global Partnership, 
2021). Mozambique’s overall adult literacy rate is 47 percent but with a significant gap between 
male (60 percent) and female (28 percent) rates (USAID, 2021a). 
Nutrition in Mozambique
Many Mozambicans, particularly children, suffer from food insecurity, including poor diet 
quality. Diets in Mozambique consist mainly of starchy staple foods, which results in limited 
intake of key micronutrients. This contributes to high levels of stunting. In rural areas in the 
northern part of the country – which also has the highest levels of stunting – the World Food 
Programme (WFP) found that households consume mostly maize, with additional food groups 
such as oils/fats and green leafy vegetables consumed as infrequently as 2–3 days per week. In 6 
out of Mozambique’s 11 provinces, more than 40 percent of children under 5 are stunted 
(USAID, 2021b), and Nampula province has the highest level of stunting, at 55 percent (USAID, 
2014). 
Water, Sanitation and Hygiene (WASH) in Mozambique 
Mozambique has made significant strides in terms of access to water, sanitation and hygiene 
(WASH) services, although the level of deprivation continues to be a matter of concern. Only 
61% of Mozambicans have access to improved water supply and only 36% have improved 
sanitation facilities (WHO/UNICEF, 2021). Access to water supply improved by 12 percentage 
points between 2015 and 2020 and access to improved sanitation facilities by 8 percentage points 
Abt Associates June 2022 ▌8 
in the same period. According to World Bank (2018: 27), “about 20 percent of the substantial 
disease burden in Mozambique is linked to poor sanitation and unsafe water.” WASH-related 
diseases include diarrhea, dysentery, cholera, respiratory infections, and parasitic infections, 
which are responsible for most of the deaths in Mozambique, particularly among children. 
Assuming that WASH is an integral part of learning environment and key to school achievement, 
the McGovern Dole Program includes initiatives that aim at improving WASH infrastructures 
and practices in target schools in Mozambique.
2.3 McGovern-Dole International Food for Education and Child Nutrition Program 
The McGovern-Dole program is structured around two strategic objectives. Strategic Objective 1 
represents a theory of change linking the improvement of student literacy to improvements in
reading instruction, student attentiveness in class, and student attendance at school (Exhibit A1). 
Improvements in teacher behaviors, instructional materials, student hunger, economic incentives, 
and school infrastructure are some program components hypothesized to be causally linked to 
student literacy, attentiveness, and attendance. Strategic Objective 2 proposes use of health and 
dietary practices as a pathway to reduce health-related absences, which improves school 
attendance (Exhibit A2). As a result, the McGovern-Dole program takes steps to improve student 
literacy by (1) enhancing teacher and school administrator training, curricular support, and 
access to school supplies; (2) reducing short-term hunger and improving student attendance and 
attentiveness by ensuring greater access to food; (3) supporting student attendance and 
attentiveness through nutrition and health interventions. 
Projects in Maputo and Nampula provinces were constructed around this theory of change 
represented in the strategic objectives. In Maputo province, where the project was implemented 
by Planet Aid, Inc. (PAI) in partnership with Ajuda de Desenvolviemento de Povo para Povo 
(ADPP), four districts were included in the McGovern-Dole project. For literacy instruction, 
PAI/ADPP conducted early-grade reading and literacy interventions, including providing 63 
schools materials in either Rhonga or Changana, to help the schools provide improved 
instruction in local languages. For attentiveness and attendance, PAI/ADPP provided daily 
meals, school gardens, safe water supply, and nutrition education activities. In Nampula 
province, where the project was implemented by World Vision, two districts were included in the 
McGovern-Dole project. World Vision’s literacy activities emphasized the use of literacy 
assessments and curriculum analysis; access to materials such as children’s books and other 
supplies; and enhanced community engagement in literacy activities, such as reading clubs. For 
attentiveness and attendance, World Vision provided health and nutrition training, supported the 
construction of water infrastructure, helped with medical delivery and services, and constructed
savings and lending groups to invest in school feeding activities. To reduce short-term hunger in 
students, World Vision provided school meals. World Vision activities also included training of 
teachers, school administrators/education officers, construction of classrooms and latrines, 
formation and training of farmer groups, and formation of Savings and Loans Associations. 
This evaluation offers a robust assessment of the hypothesis above – that is, whether these 
interventions indeed improved conditions for literacy development and students’ literacy in 
general. We use a mixed methods design to quantitatively examine the causal impact of the 
intervention on desired outcomes, as well as qualitatively understand the “what” and “why” of 
implementation and outcomes in schools during the intervention period. This provides a clearer 
window into the effects on literacy development, not only because it isolates the causal impact in 
general, but also because this analysis allows us to examine the effects of school feeding 
interventions alone – for instance, food or public health programming (in the school feeding only 
Abt Associates June 2022 ▌9 
arm) – on literacy development, as well as the effects of combining this programming with more 
direct literacy support (in the school feeding plus literacy arm). Exhibit A3 presents the 
associated logic model with inclusion of project-specific components. This model incorporates 
project implementation adequacy with resulting impacts on key outcomes of interest. 
This evaluation focused on the Phase II interventions of both projects, which were implemented 
from 2015 to 2020 in Maputo and 2015 to 2021 in Nampula. As such, it is also the first of its 
kind to evaluate the impact of McGovern-Dole projects across multiple contexts within 
Mozambique. 
2.4 Summary of Other Literacy and School Feeding Projects Implemented in the Maputo 
and Nampula Provinces
In August 2021, the study team conducted a desk review of existing documentation and 
conducted stakeholder interviews to better understand both the needs and support projects
currently in the Mozambique region. Multiple projects are implemented in the region; some are 
associated with McGovern-Dole while others have alternative host agencies and funding sources. 
The review and interviews revealed a few other literacy and school feeding projects in the target 
provinces. The initial delivery and sustainment of services is a benefit to the local communities 
but may complicate the detectable contrast between McGovern-Dole schools and comparison 
schools involved in this evaluation. These alternative projects include the USAID-funded Vamos 
Ler! project that was implemented in the provinces of Nampula and Zambezia, the Capacity 
Development for Education (CapED) Programme, under UNESCO, that is being piloted in 
Mozambique to develop a new curriculum for primary education for youths and adults, and the 
United Nations WFP take-home rations program. These initiatives are described below. 
Vamos Ler! (Let’s Read!) was a five-year project (2016–2021) funded by USAID to develop 
bilingual education pedagogical tools and activities, improve national early grade literacy 
policies and delivery and monitoring systems, enhance school leadership and increase parental 
and community engagement in early grade literacy. Let’s Read! assisted the government in 
delivering a quality bilingual education program for first, second and third grade students in the 
provinces of Nampula and Zambezia. The project worked alongside government counterparts to 
develop instructional tools and learning materials in the local languages (Emakhuwa, Elomwe, 
and Echwabu) with oral language development and transition to Portuguese reading and writing 
in the fourth grade. Creative Associates International implemented the program in partnership 
with World Education, Inc., American Institutes for Research, Overseas Strategic Consulting, 
and Blue Tree Group. 
CapED is working alongside the government to develop a new curriculum for primary education 
for youths and adults and builds the capacities of teacher-trainers who will go on to instruct 
teachers of non-literate adults and youths. CapED is also supporting the Family Learning 
Programme, which was originally developed by the UNESCO Malala Fund for Girls’ Right to 
Education. CapED enriched the program by developing the content of its learning manual, and 
supported the training of literacy teachers to use this manual. The program aims to equip non￾literate families with literacy, numeracy, and life skills, such as hygiene techniques, nutrition, 
and parenting. The program also aims to address the scarcity of early childhood education in 
Mozambique, as only 4 percent of under 5-year-olds receive this type of education. The program 
therefore equips parents with the skills they need to support the early development of their 
children and prepare them for primary school education through joint learning. The Family 
Learning Program is currently being rolled out in two provinces, Nampula and Maputo. 
Abt Associates June 2022 ▌10
The WFP’s take-home rations program has reached over 100,000 beneficiaries in Nampula with 
take-home rations in the districts of Malema, Nacala-a-Velha, and Ribaue. This assistance is part 
of a five-years partnership WFP established with the company Nacala Logistics to expand school 
feeding in the province. The baskets are directed at the students from these schools and their 
families as an adaptation to the closure of schools. WFP works closely with implementers of 
school feeding in Mozambique and the Ministry of Education and Human Development 
(MINEDH) to discuss the sustainability of its national home-grown school feeding program 
(Programa Nacional de Alimentação Escolar - PRONAE) through the creation of specific 
legislation and a budget line from the national budget. 
2.5 Baseline Assessment
In 2016, USDA contracted RTI International to design an independent impact evaluation of 
McGovern-Dole Mozambique and to conduct a baseline assessment. The baseline assessment 
provides a snapshot of the situation in 2017 after some programming began and was not an 
evaluation of project impact. Rather, the baseline assessment was designed to support this quasi￾experimental impact evaluation which is intended to capture changes in student and school 
outcomes between 2017 (baseline) and 2021 (endline). Its results were intended to inform
decision makers about the situation in Nampula and Maputo and share with the implementing 
partners a wider vision of the overall status of student reading ability, student attendance, student 
nutrition practices, student nutrition and health knowledge, and teacher and head teacher 
attendance in these provinces. This quasi-experimental evaluation only captures impacts between 
baseline and endline; a resulting limitation is that any impacts of programming that occurred 
prior to the baseline assessment are not captured in this evaluation’s impact estimates. 
The school-based survey used to obtain the results was conducted in the third trimester of the 
school year to allow for adequate exposure of the students to standard education and school 
meals. The instruments were developed to address the components of the theory of change, 
calculate the relevant indicators, address the research questions for the evaluation, and assess the 
similarities and differences between the intervention and comparison schools at baseline. To 
assess literacy, enumerators administered an Early Grade Reading Assessment (EGRA) tool, a 
15-minute oral assessment that measures the basic skills a child must have to eventually be able
to read fluently and with comprehension. A student survey was conducted that included
questions on student attendance, school feeding activities, and nutrition and hygiene knowledge
and practices. Teachers, head teachers, and school cooks were interviewed to obtain information
on teacher attendance, school amenities and the school feeding activities. The survey was paired
with qualitative data collection, which provides supplemental insight from parents about
McGovern-Dole.
In February 2018, the baseline report was completed. It included student reading ability, student 
attendance, student nutrition practices, student nutrition and health knowledge and teacher and 
head teach attendance. The baseline report included the following recommendations for future 
evaluations: 
• Continue tracking activities in both intervention and comparison schools within sampled
districts
• Invest additional effort in matching intervention and comparison schools, students, and
teachers in future analyses
• Conduct a meta-analysis to account for variance across implementers and provinces
Abt Associates June 2022 ▌11
• Use the proposed difference-in-difference methodology to account for the significant
differences between the comparison and intervention arms at baseline and track the rate of
progress of these schools over time
2.6 Other Evaluations Conducted by Implementing Partners
Implementing partners have conducted evaluations of the McGovern-Dole projects in 
Mozambique with markedly promising results. Evaluations of phases I and II of the McGovern￾Dole projects in Nampula province conducted by COSDER Consultants (Phase I) and 
International Food Policy Research Institute (Phase II), for instance, examined World Vision’s 
implementation of the McGovern-Dole projects in 2012–2015 and 2015–2021, respectively. 
They found that the participation in McGovern-Dole resulted in a 6.5 percentage point increase 
in third-graders’ performance in the Portuguese language. Moreover, Dr. Simone Doctors 
evaluated PAI’s implementation of the McGovern-Dole project from 2015 to 2020 in Maputo 
(Doctors, 2021). Again, by midline, over one-quarter of students met the literacy benchmark. 
Most recently, Counterpart International is implementing the McGovern-Dole project in Maputo 
from 2020 to 2025, with results to be assessed by Maraxis. World Vision also implemented the 
Local and Regional Food Procurement Program in Mozambique from 2017 to 2019. While this 
did not assess literacy outcomes, it was reported that “the program was able to achieve most of
its output targets on an aggregate level” (ICED, 2019).
All four evaluations used a similar mixed methods approach to this study – most typically, a
combination of semi-structured interviews, focus groups, quantitative surveys, and EGRA 
scores. However, they differ in analysis. International Food Policy Research Institute and 
Maraxis’s evaluations both use multi-level modelling when evaluating EGRA scores, allowing 
them to assess causality on assessment outcomes, whereas other evaluations did not. The midline 
report for Phase II of World Vision’s implementation notes, for instance, that while the results 
may be positive, “student performance is not only explained by [the program] but also by other 
variables” not accounted for the model. (p. 44). Phase I of World Vision’s implementation of the 
McGovern-Dole project was evaluated through cross tabulation and correlations, rather than 
regression models that could control for other factors. The remainder – including the Local and 
Regional Food Aid Procurement Program and Food for Education Phase II – are only evaluated
through comparing endline outcomes across certain indicators, but not necessarily their statistical 
significance.
2.7 Prior Evidence on School Feeding in Mozambique (or in the Sub-Saharan Region) 
The endorsement of PRONAE, the national school feeding program, by the Mozambican 
Council of Ministers in 2013 represented a key stage in the institutionalization of school feeding 
and local food purchases in the country’s education system (Milhorance de Castro, 2018). 
PRONAE’s pillars emphasized the responsibilities of the education sector, with less emphasis
placed on coordination with agricultural production and local purchases. These pillars included: 
(1) improving the nutritional and health status of students; (2) food and nutrition education in
schools; and iii) developing students’ agricultural production skills. The assessment of the pilot
phase of PRONAE conducted in 2018 indicated that the introduction of a school feeding
program in Nampula province appeared to have a positive influence on the increased enrollment,
attendance, and retention of primary school learners. Schools without school feeding programs
failed to retain learners (Sitao, 2018).
The McGovern-Dole projects in Mozambique began in late 2012 supporting projects in Maputo 
and Nampula provinces. The first phase of the project was from 2012 to 2015 in Nampula and 
Abt Associates June 2022 ▌12
2012 to 2016 in Maputo. The evaluation report on the project in Nampula done at the end of 
Phase I reported that school feeding contributed to increased enrollment and retention rates and 
improved health and dietary practices (World Vision Mozambique, 2016). 
As a result of the COVID-19 pandemic, the final evaluations for the second phase (2015–2021) 
of the McGovern-Dole projects in Maputo and Nampula provinces encountered challenges 
collecting data and interpreting results. However, the mid-term evaluations for both projects 
reported the projects were beginning to have impact. The project in Nampula reported notable 
achievements in increasing literacy, enhancing student and teacher performance, and reducing 
hunger at the point of the midline evaluation (Diogo, 2018; Doctors, 2019).
School feeding programs in sub-Saharan Africa have been in place for decades and are currently 
embedded in national policies. The Global School Feeding Sourcebook, developed in 2016, 
includes case studies on several school feeding programs in sub-Saharan Africa including 
Botswana, Cote d’Ivoire, Kenya, Mali, Namibia, and Nigeria. Although quantitative data were
not collected for evaluating the impact of most of these programs, anecdotal evidence from 
stakeholders indicated growth in enrollments and school attendance rates that are highly 
associated with the availability of food at school. During the past five years, evaluations have 
been conducted on McGovern-Dole projects in sub-Saharan Africa, including Tanzania 
(Vyamanam, Ntunda, and Dulla, 2017), Malawi (Tirivayi, 2019), Rwanda (Ravesloot et al., 
2021), and Mali (Sarfarha et al., 2020). Although not all the evaluations were able to assess 
improved literacy, attentiveness, and attendance, most of the evaluations were able to establish 
that project implementation contributed to ongoing student enrollment and attendance. 
2.8 Prior Evidence on Teacher Pedagogy, Training, and Remedial Education in 
Mozambique (or in the Sub-Saharan Region)
According to the midline results of the USAID-funded Vamos Ler! program, implemented from 
2016 to 2021 in Nampula and Zambezia provinces, the program improved student reading skills 
(Turney et al., 2020). Compared to the program baseline, student performance improved 
substantially for all EGRA subtasks. This was true for all three Mozambican languages assessed 
(Emakhuwa, Elomwe, and Echwabu). The improvements were in large part due to a substantial 
reduction in the percentage of students registering zero scores on the EGRA subtasks. The 
largest improvements were seen in letter name and letter sound identification subtasks, with 
more modest gains in word and text reading. The midline results also noted positive reception of 
bilingual education in classrooms and communities. 
Prior evidence in Mozambique showed that bilingual education has been transformative.
Chimbutane (2011, 2015) shows that using local languages for teaching and learning has enabled 
classroom interaction and contributed to bridging the gap between rural bilingual schools and 
pupils’ communities in Mozambique.
Consistent with the United Nations’ Sustainable Development Goal 4, improving literacy 
outcomes in sub-Saharan Africa has been a central focus of national governments, donors, and 
nongovernmental organizations. Although significant international development funding has 
been devoted to teacher training in the region, little evidence is available on how teachers 
improve their literacy instruction in practice (Lee, J., D'sa, N., and Zuilkowski, S. S. (2020). 
2.9 Quasi-Experimental Impact Evaluation 
In late June 2021, USDA/FAS contracted with Abt Associates to conduct an endline quasi￾experimental design impact evaluation of McGovern-Dole projects in Mozambique. Abt 
Abt Associates June 2022 ▌13
Associates did not have a role in the implementation of the McGovern-Dole projects. Abt 
Associates mirrored the evaluation conducted at the baseline in order to compare changes on key 
indicators from the project activities. The evaluation design, instruments, key outcomes 
measured, and research questions remained the same as they were at baseline. However, the 
endline evaluation design contains modifications to better reflect and capture the actual project 
activities and the adequacy of the implementation of activities as planned. 
2.10 Differences Between Baseline and Endline Approaches
Target grades
The initial evaluation design was to assess the reading performance of second grade and fourth 
grade students, four years after the start of implementation. At that point, the students in second 
grade would have been in their schools for two of the four years and the students in fourth grade 
would have been in their respective schools for all four years. The COVID-related school 
closures shut down the schools from March 2020 to March 2021. This meant that the second 
grade students experienced only one year and two months of exposure while the students in 
fourth grade experienced only three years and two months of exposure.
With agreement from USDA, the Abt research team assessed students one year older (third 
grade), who had been in their schools for an additional year for a full two years of exposure. 
Instead of assessing fourth graders for the endline assessment, the Abt team assessed students 
one year older (fifth grade), who had been in their schools for an additional year. 
Language of Assessment
Mozambique has both monolingual and bilingual education, depending on the school and 
district, and in some cases, Portuguese does not become the language of instruction until third 
grade. This was not the case during the baseline assessment, when all teaching was done in 
Portuguese. By the end of the implementation activities, all districts in Nampula province 
continued to have language instruction in Portuguese. However, in some districts in Maputo 
province, students were offered bilingual education in both local languages and Portuguese; for 
the midline evaluation, the independent evaluator conducted EGRA assessments in local 
languages or Portuguese, as was most relevant for the intervention. 
In each school where bilingual instruction has been identified, we assessed a random selection of 
eight third grade students in the appropriate local language (Rhonga or Changana). We assessed 
a random selection of eight fifth grade students in Portuguese. Ultimately, we tested 441 students 
in a local language in 31 Maputo schools across 445 sample schools. All but eight of these 
students were in the third grade. Most local language assessments were conducted in Changana 
(323) compared to Rhonga (118). 
Instrument Enhancements
To maintain consistency across baseline and endline evaluations, the study team used baseline 
instruments. To capture more details on implementation including and beyond the effects of 
COVID-related school closures, we added several open-ended questions to head teacher and 
cook questionnaires as well as adding items to parent focus group (PFG) discussions and the 
interviews with the MINEDH. These supplementary items focused primarily on the 
implementation of McGovern-Dole and contextual factors that may have influenced 
implementation. 
Abt Associates June 2022 ▌14
2.11 COVID-19 Effects on Schooling
The planned implementation of the project activities over the four-year period from 2017 
through 2021 was disrupted by school closures related to COVID-19. These closures affected 
nearly 70 percent of students in over 160 countries worldwide.1
 In Mozambique, over 6.9 million 
primary school students were shut out of in-person schooling for a year (March 2020–March 
2021) (UNICEF, 2020). The closures meant that students attending the study schools will have 
had fewer years of exposure to their study condition than was expected. Furthermore, for most 
students in the two mostly rural provinces in this evaluation, alternative delivery of instruction 
via online education was not a possibility in the country due to lack of access to computers and, 
even more basic, to electricity. 
1 Education: From disruption to recovery (unesco.org)
Abt Associates June 2022 ▌15
3. EVALUATION APPROACH
The evaluation design is informed by the McGovern-Dole Results Frameworks (Exhibit A1 and
Exhibit A2), which includes two strategic objectives. Strategic Objective 1 identifies improved 
literacy, quality of literacy instruction, improved attentiveness, and improved attendance as 
interrelated primary objectives. Strategic Objective 2 identifies use of health and dietary 
practices as a pathway to reduce health-related absences, which improves school attendance. 
3.1 Impact and Implementation Focus
The impact evaluation used a quasi-experimental difference-in-difference design, comparing 
average school outcomes at project beginning to average school outcomes at project end, four 
years later. Alternative impact analyses are conducted on a subset of schools with adequate 
fidelity; this analysis is based on an assessment of the level of implementation in each school, 
based on descriptive information on level of fidelity of implementation collected as part of the 
evaluation. This approach allows us to generate exploratory evidence on both the “what” and 
“why” of implementation and outcomes in schools during the intervention time period. This 
approach also supports a clear and robust understanding of what happened as planned in schools 
during the intervention time period, what did not happen as planned, and why. The mixed 
methods approach takes into account the importance of the quality of the implementation when 
testing the impact of an intervention. The results of our approach provide detailed qualitative 
data to help to interpret the data from the impact evaluation, which is important given the four￾year time period from baseline to endline. 
Conducting the impact evaluation across two projects in two provinces allows this report to 
describe common lessons learned across provinces in Mozambique. This report provides 
information on how to estimate effects on student literacy performance, such as the effect of 
COVID-19, which is especially relevant since many project-level evaluations will face 
challenges implementing literacy testing as planned. 
3.2 Evaluation Questions
The primary impact evaluation questions for the endline impact evaluation are about the effects 
of the McGovern-Dole projects on improved literacy of students and increased student 
attendance and attentiveness. Secondary evaluation questions ask about effects on teachers and 
schools (Exhibit 1). These questions will be answered by separately comparing the students, 
teachers, and schools in each of the two intervention arms with the control arm. 
There are also exploratory impact evaluation questions about how the effects vary as a function 
of level of implementation of the McGovern-Dole projects. These questions will be addressed by 
conducting exploratory analyses estimating impacts of the two treatment arms including only the 
subset of schools showing adequate fidelity of implementation of their assigned project(s). 
Finally, there are also descriptive evaluation questions about implementation, parent perceptions, 
as well as sustainability and recommendations that will be answered descriptively, using 
qualitative data. 
Abt Associates June 2022 ▌16
Exhibit 1. Evaluation Questions
Primary Evaluation Questions about Impacts on Students
How does reading ability differ for Grade 3 and 5 students who attend:
1. schools with the McGovern-Dole Literacy and School Feeding activities, 
2. schools with the McGovern-Dole School Feeding Only activities, and 
3. schools without any McGovern-Dole activities (Comparison)?
How does attendance differ for Grade 3 and 5 students in: 
1. schools with the McGovern-Dole Literacy and School Feeding activities,
2. schools with the McGovern-Dole School Feeding Only activities and 
3. schools without any McGovern-Dole projects (Comparison)?
How does attentiveness differ for Grade 3 and 5 students in:
1. schools with the McGovern-Dole Literacy and School Feeding activities, 
2. schools with the McGovern-Dole School Feeding Only activities, and
3. schools without any McGovern-Dole projects (Comparison)?
Secondary Impact Evaluation Questions about Mediators 
Factors relating to student hunger and diet
How does the level of short-term hunger differ for Grade 3 and 5 students who attend: 
1. schools with a McGovern-Dole School Feeding project, and 
2. schools without any McGovern-Dole projects (Comparison)?
How does dietary intake/access to food differ for Grade 3 and 5 students who attend: 
1. schools with a McGovern-Dole School Feeding project, and
2. schools without any McGovern-Dole projects (Comparison)? 
How does access to food preparation and storage facilities differ for: 
1. schools with a McGovern-Dole School Feeding project, and 
2. schools without any McGovern-Dole projects (Comparison)? 
Is there a reduction in health-related absences in: 
1. McGovern-Dole schools and
2. non McGovern-Dole schools.
Factors relating to student reading ability
How does availability of literacy materials and resources differ for Grade 3 and 5 students who attend: 
1. schools with the McGovern-Dole Literacy and School Feeding activities, 
2. schools with the McGovern-Dole School Feeding Only activities and
3. schools without any McGovern-Dole projects (Comparison)?
How does teacher training availability and attendance differ for Grade 3 and 5 teachers in: 
1. schools with McGovern-Dole Literacy and School Feeding activities, 
2. schools with the McGovern-Dole School Feeding Only activities and 
3. schools without any McGovern-Dole projects (Comparison)? 
How does teacher attendance of Grade 3 and 5 teachers differ across:
1. schools with the McGovern-Dole Literacy and School Feeding activities, 
2. schools with the McGovern-Dole School Feeding Only activities and
3. schools without any McGovern-Dole projects (Comparison)?
Factors Related to Student Knowledge/Use of Recommended Health and Nutrition Practices
Abt Associates June 2022 ▌17
How does knowledge of recommended health and nutrition practices differ for Grade 3 and 5 students who attend: 
1. schools with a McGovern-Dole School Feeding project, and
2. schools without any McGovern-Dole projects (Comparison)?
How does use of recommended health and nutrition practices differ for Grade 3 and 5 students who attend: 
1. schools with a McGovern-Dole School Feeding project, and
2. schools without any McGovern-Dole projects (Comparison)?
Implementation Evaluation Questions
To what extent were McGovern-Dole project activities in schools in the intervention arms implemented as planned? 
To what extent were school operations interrupted by external school events? 
• How did these interruptions affect the project implementation (i.e., especially COVID-19) that may have affected a 
change in dosage of intervention received by schools in the two intervention arms?
What are parents’ perceptions of the McGovern-Dole School feeding and literacy programs? 
• How did the project affect what parents did at home, if at all? Were the issues or concerns identified by parents at 
baseline addressed by the project and were there changes in parental perceptions by endline?
When project activities were modified, to what extent were these adapted activities considered successful by project 
participants?
In what ways could provincial and local efforts aid in program sustainability?
Differential Impacts by Level of Implementation
Are impacts greater among the subset of schools assessed as having adequate fidelity of implementation of their assigned 
McGovern-Dole programs? 
• Among students attending schools assigned to implement the McGovern-Dole School Feeding only, are there 
differences in the reading ability, attentiveness, and attendance among Grade 3 and 5 students who attend schools 
with high implementation adequacy and students who attend schools with low implementation adequacy?
• Among students attending schools assigned to implement the McGovern-Dole Literacy and School Feeding 
activities, are there differences in the reading ability, attentiveness, and attendance among Grade 3 and 5 students 
who attend schools with high implementation adequacy and students who attend schools with low implementation 
adequacy?
Abt Associates June 2022 ▌18
4. METHODOLOGY
4.1 Sampling Frame 
The sampling frame for the treatment arms was developed based on school and enrollment data 
provided by the implementing partners of the McGovern-Dole projects. The sample of schools in 
the evaluation were selected at the time of the baseline evaluation. At the time of the baseline 
assessment, the McGovern-Dole projects were implemented in 273 schools across four districts 
in Maputo and in 150 schools across two districts in Nampula (RTI Baseline Report, 2018). 
This evaluation uses a three-stage stratified clustered sampling design. At baseline, the provinces 
and districts were selected in a non-random process to represent treatment arms based on 
services schools were receiving. Within the province of Maputo and Nampula, schools were 
stratified by treatment arm and sampled with equal allocation. As noted in the RTI Baseline 
Report (2018), the use of equal allocation rather than proportional allocation was used because 
school enrollment data were missing for many of the schools in the frame and the enrollment 
data were believed to be inaccurate after comparison with other sources.
The sampling frame for the comparison group was developed using Education Management 
Information System data. The selection of comparison districts was based on a number of criteria 
including: location in the province, no known literacy intervention, geographically close to 
districts with intervention schools, similar urban-rural divide as the intervention districts; similar 
number of students and schools; similar languages spoken at home. The schools in the districts 
that met these criteria were automatically selected into the comparison group. Note, the city of 
Maputo was not selected because it is primarily urban (RTI Baseline Report, 2018). Classrooms 
and students within schools were sampled as part of the data collection in schools. Within each 
school, one second grade and one fourth grade classroom were selected. Selection of the study 
classrooms was random if the school had more than one classroom in either second grade or 
fourth grade. Because of COVID-19-related school closures from March 2020 to March 2021, 
students missed a year of instruction. With agreement from USDA, the Abt research team 
collected endline data from third grade students for the impact evaluation. By October 2021, 
third grade students had been in their schools for an additional year, for a full two years of 
exposure. Abt also collected endline data from fifth grade students for impact evaluation. These 
students had also been in their schools for an additional year, mitigating the loss of a year of 
schooling. At endline, we used the same approach to select eight students in third grade 
classrooms and eight students in fifth grade classrooms from available daily rosters. 
4.2 Qualitative Sampling Design
The endline evaluation used the same qualitative sample for PFG discussions that was selected at 
baseline. Schools where PFG discussions were conducted were selected from a subset of schools 
in the quantitative sample where survey data collection was also completed. Schools were 
stratified based on the number of second grade classes (as a proxy for school size), head teacher 
satisfaction with parental involvement, and rural and urban locations. Within these strata, schools 
were randomly selected within each province (RTI Baseline Report, 2018). 
We used purposive sampling to select key stakeholders at the MINEDH for interviews. The rest 
of this report refers generally to these respondents as “key officials”; the group includes 
representatives from the district, provincial, and national levels. 
Abt Associates June 2022 ▌19
4.3 Outcomes
Primary Student Outcomes
Primary outcomes of interest include student reading ability, student attentiveness, and student 
attendance (absences).
Reading Ability
We used the EGRA tool to determine whether students in the early and middle grades are 
developing the fundamental reading skills. The EGRA tests students’ competency on a number 
of subtasks, including oral comprehension, letter and syllable identification, word dictation, oral 
reading fluency, and reading comprehension. The ability to read and understand a simple text is 
one of the most fundamental skills a student can learn. In this report, we discuss the impact on 
the reading comprehension subtask in Section 7: Main Findings, and have included exhibits with 
the impacts for the other subtasks in Appendix A. Below we briefly describe the subtasks and 
how we measured student scores.
Reading ability was measured with six subtasks: 
• Oral comprehension: measures the ability to understand basic oral vocabulary. The
enumerator reads the student a short text and when finished asks the student a few questions
to test their understanding of the text. In the Portuguese EGRA, the enumerator asks the
student five questions about the text; in the local language EGRA, the enumerator asks the
student four questions about the text. Outcome measure: percent zero scores, percent of
questions correctly answered.
• Letter identification: assesses the ability to provide the names of the letters of the alphabet.
The student is given a letter card with 100 letters (10 letters on 10 rows with letters in
uppercase and lowercase) and is asked to name as many letters on the card as they can within
one minute. The order of letters differed between the Portuguese and local language EGRAs,
and the local language EGRA contained a number of different letters. Outcome measures:
percent zero scores, percent of letters correctly identified, a score that measures the correct
letters read per minute.
• Syllable identification: assesses the ability to correctly pronounce syllables. The student is
given a syllable card with 50 syllables (five syllables on 10 rows) and asked to pronounce as
many syllables on the card as they could within one minute. The order of the syllables
differed between the Portuguese and local language EGRAs. Given that the syllable structure
of the local languages is different, some of the syllables differed from the Portuguese EGRA.
Outcome measures: percent zero scores, percent of syllables correctly pronounced, a score
that measures the correct syllables read per minute.
• Word dictation: measures the ability to correctly write words that are dictated. The
enumerator reads five words and after each word the student is asked to write the word on a
piece of paper. Portuguese EGRA has different words than the local language EGRA.
Outcome measure: percent zero score, percent of words correctly written.
• Oral reading fluency: measures the ability of students to read passages fluently. In this
subtask, the students were given a short story and were asked to read the story aloud in one
minute. The story in the local language EGRA was a little shorter than the store in the
Portuguese EGRA. Outcome measure: percent zero score, percent of words correctly read.
• Reading comprehension: measures the student’s textual understanding. After the student
finished reading the passage in the oral reading fluency subtask, students were orally asked
Abt Associates June 2022 ▌20
five questions about the text in the Portuguese EGRA and four questions in the local 
language EGRA. Outcome measure: percent zero score, percent of questions correctly 
answered. 
Student attendance
To estimate student attendance, we asked students whether they were absent at least one day last 
week. Students answered this question with a simple yes or no response. The results from this 
question are presented in this report. Findings from other variables that measured student 
attendance (e.g., teacher perception on average percent of students absent in a classroom) is also 
included in this report. 
Student attentiveness 
To measure how student attentiveness has changed, we asked teachers about their perception on 
whether students are attentive in class. Teachers were asked to answer a question with a simple 
yes or no response. In this report, we show the results from this question. Findings from other 
variables that provide insight into how student attentiveness changed (e.g., student self-reported 
attentiveness, teacher perception on approximate number of students who are attentive in class) 
are also included in this report. 
Secondary Outcomes
Impacts on instructional input
• Teacher-reported receipt of training on the fundamentals of reading instruction: teachers
were asked if they ever received training on the fundamentals of reading instruction, which
they could answer with yes or no response.
• Teacher-reported absence of at least one day last week: teachers were asked if they were
absent from school any day last week. Teachers answered this question with a yes or no
response.
• Teacher has access to school supplies: to measure access to teaching and learning materials,
we asked teachers if they have access to school supplies which was answered with a yes or
no response.
Feeding outcomes
• Short-term hunger: to measure the McGovern-Dole impact on short-term hunger, we asked
students how hungry they felt during the time of the interview. Students could choose from
the following response options: very hungry, somewhat hungry, just right, somewhat full, or
full. We report on use a binary outcome that shows the percentage of students who felt
hungry (very hungry or somewhat hungry).
• Dietary intake: as McGovern-Dole aimed to improve access to school meals and use of
dietary practices, we measure the impact on student’s receiving a minimum acceptable diet.
The minimum acceptable diet indicator is based on both the minimum meal frequency and
minimum dietary diversity. Minimum meal frequency is defined as three or more meals per
day. Minimum dietary diversity is defined as the daily consumption of four or more of seven
food groups (grains, roots, and tubers; legumes and nuts; dairy products; flesh foods; eggs;
vitamin A-rich fruits and vegetables; and other fruits and vegetables). If a student meets the
Abt Associates June 2022 ▌21
minimum feeding frequency and minimum dietary diversity for their age group, then they are 
considered to receive a minimum acceptable diet.2
• Student nutrition knowledge score: students were asked eight questions to test their
knowledge of nutrition based on the curriculum provided by the implementer partners. These
nutrition questions focused on student’s knowledge of food groups, nutrients, and
consumption of a balanced diet.
• Food preparation: cooks were also surveyed about the conditions and supplies available for
food preparation at the school, including the availability of pots and utensils, access to clean
water, and the presence of a kitchen. Cooks were also asked a series of questions to evaluate
their knowledge of food safety and hygiene.
Student knowledge and use of recommended health practices
• Hygiene and sanitation knowledge score: students were asked seven questions to test their
knowledge of nutrition based on the curriculum provided by the implementer partners. The
hygiene and sanitation questions tested students’ knowledge of basic hygienic and sanitation
practices, including prevention of diarrhea and when and how long they should wash their
hands.
• Health-related absence: students who said they were absent from school any day last week,
were also asked for the reason they were absent. To assess the impact on health-related
absences, we show the percentage of students who self-reported that they were absent due to
illness.
• Improved latrines: head teacher were asked to report on whether the school has improved
latrines.
• Tippy-taps or hand-washing stations: head teacher reports that school has tippy taps or hand￾washing stations.
• Tippy taps or hand-washing stations have ash or soap: head teacher reports that tippy taps or
hand-washing stations have ash or soap.
For information on the sample sizes by treatment arm for each of the primary and secondary 
outcomes, we refer Exhibit A7. 
4.4 Impact Analysis Models 
The quasi-experimental estimates use data on the same outcomes for schools at baseline and 
endline to support difference-in-differences analyses. Difference-in-differences analysis
implicitly projects what the endline outcome for schools that received the intervention in 
question would have been “absent treatment” using the observed difference between baseline and 
endline in comparison schools. The difference between the observed endline outcome for 
treatment schools and this “absent treatment” projection is an estimate of the treatment effect. 
Specifically, our estimates of impact for outcomes based on student or teacher survey responses 
come from the following regression model: 
2 The Minimum Acceptable Diet (MAD) for children 6–23 months old, is one of the core indicators for assessing infant and 
young child feeding (IYCF) practices developed by the World Health Organization (WHO, 2008). The minimum acceptable 
diet indicator is not validated for school-age children.
Abt Associates June 2022 ▌22
The variables in the model are as follows: 
• represents the outcome measure for student/teacher i in grade g in school s in province 
p in year t. 
• is a binary variable indicating whether the outcome is for students in Grade 4 
at baseline and Grade 5 at endline (and equal to zero if the outcome is for students in Grade 2 
at baseline and Grade 3 at endline).
• is a binary variable indicating whether the year is 2021 (and equal to zero if the 
year is 2017). 
• 1 is a binary variable indicating whether school s received the school feeding 
intervention only (and equal to zero if the school is either a comparison school or received 
the combined school feeding and literacy interventions). 
• 2 is a binary variable indicating whether school s received the combined school 
feeding and literacy interventions (and equal to zero if the school is either a comparison 
school or received the school feeding intervention alone). 
• 4 is the effect of the school feeding intervention by itself for students in who entered Grade 
1 in 2019 (and were thus in Grade 3 in 2021). 
• 5 is the effect of the combined intervention for students who entered Grade 1 in 2019. 
• 6 is the additional effect of the school feeding intervention by itself for students who 
entered Grade 1 in 2017 and therefore had two additional years of exposure to the 
intervention. 
• 7 is the additional effect of the combined intervention for students who entered Grade 1 in 
2017 and therefore had two additional years of exposure to the intervention. 
• is a school fixed effect.3
• is a school-by-grade fixed effect.4
• is a province-by-time fixed effect.
• is the error term. 
Unbiased estimation of the impacts requires us to assume “conditional parallel trends” between 
the treatment and comparison schools (Lechner, 2011). Conditional parallel trends implies that, 
net of the factors we control for in the model, the trend between baseline and endline in 
comparison schools is parallel to what would have been observed in treatment schools had they 
3 The evaluation design plan described as a school random effect. We revised this to a school fixed effect to relax 
assumptions about the nature of school-specific effects.
4 “Grade” refers to either early-grade students (in grade 2 at baseline and grade 3 at endline) or mid-grade students (in grade 4 
at baseline and grade 5 at endline).
 = 0 + 1 + 2 + 3 × 
+ 4( 1 × ) + 5( 2 × )
+ 6 1 × × + 7 2 × × 
+ + + + 
Abt Associates June 2022 ▌23
not been treated. To make this assumption more plausible, the model includes a series of fixed 
effects. The fixed effects aim to control for unobserved factors that are unchanged between
baseline and endline. For instance, school fixed effects control for unobserved factors common 
to schools, such as school size or quality of school leadership. School-by-grade fixed effects 
control for any unobserved factors common to early grades and middle grades within schools 
that are unchanged between baseline and endline, such as curriculum or teacher quality for those 
grades. The province-by-time fixed effects control for unobserved factors common to the 
province and year such as hurricane damage and COVID closures.5
We used software that calculates standard errors and associated p-values in a manner that reflects 
the extra uncertainty caused by testing eight students per school instead of spreading the students 
out across the entire province.6 After fitting the model, we also used software to predict 
regression-adjusted average outcomes at endline for both intervention groups and the comparison 
group. These predictions are calculated for both treatment groups and the comparison group by 
summing together weighted baseline outcomes and estimates from the regression model. 
For outcomes based on head teacher and cook survey responses, where there is only one 
response from each school, the regression model is modified as follows: 
The new variables in this model are as follows: 
• represents the outcome measure for head teacher/cook i in school s in province p in 
year t. 
• 4 is the effect of the school feeding intervention by itself for head teachers/cooks. 
• 5 is the effect of the combined intervention for head teachers/cooks. 
• is a province-by-time fixed effect.
• is the error term. 
As with students and teachers, we calculated standard errors and associated p-values in a manner
that reflects the extra uncertainty caused by testing a random subset of schools.7
 
5
 With data only analyzed at baseline and endline, formal testing for conditional parallel trends was not feasible. We explored
the possibility of placebo testing using outcomes that, in theory, were unlikely to be influenced by the treatment; but no such 
outcomes were available for analysis. Given the careful selection of comparison schools (RTI Baseline Report, 2018) as 
well as our series of fixed effects to control for unobserved factors, we felt comfortable with the conditional parallel trends 
assumption.
6 This clustering makes for far more efficient data collection, but it also results in higher standard errors than would be seen 
on an evenly dispersed student sample of the same total size. We adjust the standard errors by using regression software that 
allows for survey analysis of the data. We declare provinces as strata and schools as primary sampling units. We also apply 
weights that reflect the inverse probability of respondents’ selection for the survey (these probabilities are different for 
students and teachers). 7 We declare provinces as strata and schools as primary sampling units. We also apply weights that reflect the inverse 
probability of school selection for the survey.
 = 0 + 1 + 2 1 + 3 2 + 4( 1 × ) + 5( 2 × )
+ + 
Abt Associates June 2022 ▌24
The relative size of the quantitative sample across treatment arms has important implications for 
statistical power. Because it was only tested in Maputo, the school feeding only treatment arm 
has a smaller quantitative sample (178 schools across baseline and endline) than the literacy and 
school feeding arm (360 schools) or the comparison group (355 schools). The impact analyses 
separately compare literacy and school feeding to comparison and school feeding only to 
comparison. The smaller quantitative sample in the school feeding only treatment arm implies 
impacts must be larger in that treatment arm than in the literacy and school feeding arm to 
achieve statistical significance.
Conducting statistical tests of many hypotheses creates a “multiple comparisons” problem and 
increases the likelihood of a false discovery. To mitigate this concern, the study team specified a 
single primary outcome in each of three separate domains: reading comprehension, attendance, 
and attentiveness. The key findings presented in the report are based on the results for these three 
primary outcomes. Impacts for secondary evaluation questions use the same methodological 
approach and are included in the report but are considered to be exploratory. 
4.5 Implementation Fidelity Analyses
Given the scope and length of the McGovern-Dole projects in Mozambique, it is important to 
identify variation in implementation across schools, provinces, treatment arms, and project
components. A deeper understanding of implementation variation has implications for 
effectiveness, adaptations, and sustainability. 
In order to identify implementation variation, we calculated a quantified level of fidelity for 
every study school, including those served by one of the two McGovern-Dole projects as well as 
comparison schools that may have had school feeding and literacy interventions. The study team 
created an implementation fidelity index informed by survey items within the head teacher, 
teacher, and cook instruments. An index and calculated score represent each of the three primary 
components: (1) School feeding, (2) Literacy instruction, and (3) Head teacher and teacher 
training. Exhibits A4, A5, and A6 represent the fidelity indices and scoring. 
4.6 Analysis of Stakeholder Perspectives
Qualitative data and analysis allowed the study team to provide additional details on McGovern￾Dole intervention activities in order to contextualize findings from the impact evaluation. In 
collecting open-ended responses from stakeholders, the study team can provide illustrative 
examples to complement the quantitative findings. Qualitative data were collected using two 
main instruments, the PFG discussion guide and the key officials’ semi-structured interviews. 
Data from the open-response questions from the head teacher and cook/warehouse manager 
surveys were also used but were not as prevalent in the qualitative analysis. 
This section provides detail on the qualitative data sources and methodology. Findings generated 
from qualitative data are presented throughout this report. 
Qualitative Data Sources
The study team received translated notes of responses from the PFG discussions and transcripts 
from the key officials semi-structured interviews. Copies of the guides are included in Appendix 
B. PFG discussions were held at baseline and endline, but the study team added the key officials
interviews for the impact evaluation at endline. For both sets of data, interviews and notes were 
transcribed and translated by the in-country data collection firm and submitted to Abt Associates 
for analysis. 
Abt Associates June 2022 ▌25
Parent Focus Group discussions: At baseline, treatment schools were randomly selected for 
inclusion in the sample for PFG discussions. At endline, the data collectors used the same sample 
of schools to conduct the PFG discussions. At endline, some questions were added to capture 
parent and family perspectives on events that occurred during the project implementation, 
specifically parent perceptions on the impact of COVID-19. 
Parent participation in the discussions was voluntary. The discussions usually lasted about 60 
minutes. An audio recording and notes were collected during each discussion. The discussions 
covered a common set of topics including involvement at the school, satisfaction and 
expectations with the school, positive and negative aspects of the McGovern-Dole project, at￾home activities, as well as sustainability and recommendations.
Ministry of Education and Human Development interviews: Key officials’ interviews were 
not completed at baseline. The study team for the endline evaluation proposed adding interviews 
with “key informant stakeholders” to provide an administrative and management perspective on 
implementation of the McGovern-Dole projects. There were several levels of perspective 
provided, described in the sample section below. 
Participation in the interviews was voluntary. Interviews usually lasted between 60 and 90 
minutes. Interviewers used a conversational approach to allow probing. The interviews were 
semi-structured and covered a common set of topics: involvement in intervention activities, 
positive and negative aspects of the McGovern-Dole project, goal attainment, the influence of 
COVID-19, as well as sustainability and recommendations. 
Qualitative Analysis Methodology 
The study team received translated notes of responses from the PFG discussions and transcripts 
from the interviews with the key officials. The files were uploaded to Nvivo, a qualitative 
analysis software. The team created a “codebook” of categories using the qualitative evaluation 
questions and based on the major topics covered in the different data collection protocols (e.g., 
types of activities, experiences with the project, and the influence of COVID-19). Text segments 
associated with specific broad categories were coded and analyzed using an inductive thematic 
approach (Guest, Namey, and Mitchell, 2013). Transcripts were analyzed to provide contextual 
understanding of the various stakeholders’ experience. 
Though responses to the broad, open-ended questions provide important contextual information 
to complement the closed-ended responses, the respondents that mentioned particular themes or 
topics cannot be assumed to represent all stakeholders. 
4.7 Analysis of Impacts in Adequate Fidelity Sites
In order to more accurately estimate the size of treatment-control contrast, the team captured 
implementation of similar project elements in all schools, across treatment and comparison 
groups. The team then calculated for each school and for each major component (i.e., literacy 
instruction, school feeding, and teacher training) a score on the adequacy of implementation. 
Based on these scores, each school was identified as either an adequate or inadequate 
implementer for each component. A school is defined as an “adequate implementer” if it scored 
above the adequacy threshold on all three components or a “low implementer” if that score is 
below the threshold. Threshold for each component were set at 50 percent of the total possible 
(Exhibit 2). This determination is the basis of an alternative impact analysis using only the subset 
of schools where implementation of the project activities was assessed to be adequate. 
Abt Associates June 2022 ▌26
Exhibit 2. Score Ranges and Implementation Adequacy Thresholds
 Inadequate Adequate
School feeding 0-7 8-15
Literacy instruction 0-5 6-11
Teacher training 0-5 6-12
For these analyses, we used the same regression model as that used to estimate the primary 
impacts described above. However, we limited the sample to all comparison schools and only 
McGovern-Dole schools identified as an adequate implementer. 
Abt Associates June 2022 ▌27
5. DATA COLLECTION
5.1 Overview
For this evaluation, the Abt team used baseline data collected in 2017 and new endline (post￾project) data. The endline data collection processes were designed to closely match the baseline 
data collection processes. In country, direct data was collected by Ernst & Young (EY) Limitada
located in Maputo, Mozambique. 
5.2 Participant Contact and Information Collection
The mobilization strategy consisted of two phases. In the first phase, EY submitted the 
permissions and authorizations to MINEDH’s Health and Nutrition Directorate and Primary
Education Directorate. After receiving the necessary permissions and authorizations, the EY
mobilization team visited the District Services of Education, Youth and Technology, and the 
selected schools in advance to inform them of the work plan and methodology to be used. 
The mobilization strategy was key to the success of the evaluation. The selected schools were 
involved and understood how relevant the work that Abt Associates and EY were implementing.
This involvement and understanding allowed the work to be conducted under the best conditions 
possible. 
5.3 Enumerator Recruitment and Training
EY recruited the data collectors and prepared their training. EY prepared additional background 
materials and lists of references for the data collection managers and team leads to familiarize 
senior staff with the content. At that time, EY made modifications to the Training Guide/Field
Manual for Enumerators that Abt had developed. 
Enumerator Recruitment
EY recruited and hired data collectors and data quality staff with backgrounds in early child 
development, education, psychology, and social welfare (Exhibit 3). Additionally, EY recruited 
staff with prior experience in national surveys such as Demographic and Health Survey or 
Multiple Indicator Cluster Survey. All data collectors were required to have at least completed
required schooling, and supervisors were required to have a degree in social sciences. Data 
collectors were also required to be fluent in Portuguese and the local languages spoken in the 
select districts of Maputo and Nampula. All hired staff were familiar with using electronic
devices for data collection.
Exhibit 3. Staff Hired for Data Collection
Staff Role Maputo Nampula 
Enumerators 36 24 
Facilitators 2 2 
Note takers 1 1 
Data controllers 15 15
Mobilizers 3 5 
Abt Associates June 2022 ▌28
Enumerator Training 
The training ensured that data collectors were familiar with not only the techniques for 
successful interviewing, but also the content of the tools. 
The training spanned five days in October 2021 and was 
conducted in person by EY using COVID-19 safety 
precautions. 
The training was divided into four parts: 
1. Objectives and scope of the project
2. Training on the questionnaires
3. Fieldwork testing
4. Fieldwork results and administrative issues with the
enumerators
In terms of training methods, EY conducted the testing of 
the tools in simulation form (enumerators interviewing each 
other) and also in real conditions. This method allowed 
interview techniques to be assimilated more effectively.
Pilot Testing 
EY performed pilot testing to ensure that all staff and respondents understood the data collection 
tools. The pilot testing process allowed the team to test abilities of the enumerators in real 
circumstances. As a result of this exercise, refinements and/or adjustments of the tools were 
made based on the situations reported by the field team during the pilot testing. All 
questionnaires and data collection tools were tested. Data collected during the pilot tests were not 
included in the evaluation sample.
There were two phases of pilot testing. The first phase was conducted in two days in September 
and took place in four schools in Maputo. The EGRA instrument and student questionnaire was 
tested with 10 students. The head teacher, teacher, and cook questionnaires were tested with four 
head teachers, four teachers, and four cooks and warehouse managers, respectively. The PFG
discussion guide was also tested with seven parents. 
The second phase was conducted in one day in October and took place in three schools across 
Maputo and Nampula. Each enumerator pilot tested student assessment with the EGRA, 
alongside one of Abt’s in-country consultants. 
After the pilot data collection, the data were shared with the Abt study team for analysis and 
improvement of the scripts. The data from the pilot were analyzed to identify any 
implementation issues. After any needed resolution, EY integrated the revisions into the tools 
and manuals. 
5.4 Data Quality and Security
Checks of Accuracy of Sample
Backchecks were conducted as internal control method to verify if the enumerators effectively 
visited the correct schools and interviewed the correct populations. 
Additional Training Topics
• Background information
• General information for the
enumerators
• Objectives and nature of baseline
evaluation
• The prescribed questionnaires and
listing schedules
• Procedures to be followed
• Analysis of the tools
• Introduction to data entry software
• Practical work
• Editing
• Administrative instructions for
enumerators
Abt Associates June 2022 ▌29
The backcheck questionnaire was implemented by data quality controllers through phone calls to 
teachers, head teachers, and cooks across all districts within the scope of work. A total 218 
backcheck phone calls were conducted: 
• 57 backchecks for cooks (37 in Maputo and 20 in Nampula) representing 22 percent of the 
total cooks interviewed 
• 53 backchecks for head teachers (30 in Maputo and 23 in Nampula) representing 12 percent
of the total head teachers interviewed
• 108 backchecks for teachers (70 in Maputo and 38 in Nampula) representing 13 percent of 
the total teachers interviewed. 
It was found that all schools contacted for the backchecks are on the list and the teachers, head 
teachers, and cooks confirmed that were interviewed. 
5.5 Data Management and Security
Mobilization Strategy 
Relevant information was collected on the target schools. The team contacted the government 
and other relevant stakeholders at central, province, and district levels. The team subsequently 
elaborated on the mobilization plan based on these meetings. 
Data Collection Logistics 
Data collection logistics were managed internally, including transportation, support to field team, 
supply of materials, and other required goods and services. Team members also printed and 
photocopied questionnaires for a back-up option for enumerators to use in the case of survey 
device failure while with a respondent. 
Data Collection
Quantitative data were collected using the EGRA instrument and questionnaires. Any problems 
were resolved as they arose throughout the data collection process. The key informant interviews 
and PFG transcripts were translated into English. The data collection team regularly updated Abt 
Associates on the progress of data collection. 
Data entry was done through the mobile data collection platform. For this project the team used 
“ODK” software, which in addition to typing, the platform allows the verification of the insertion 
to ensure consistency in data entry. In addition, ODK software allows data transfer to different 
statistical analysis packages. The program verifies the consistency of entered data. Violations of 
these checks lead to an alert about data inconsistency or violations. 
5.6 COVID-19 Specific Precautions 
Since the beginning of the state of emergency imposed to limit the spread of COVID-19, EY 
Mozambique implemented a set of measures to protect the health and safety of employees and 
the communities. The entire EY team received a summary of these protocols prior to beginning 
fieldwork, and during training. Prior to data collection staff completed COVID-19 forms. 
Reminder messages of the protocols were sent to staff throughout the data collection period. All 
data collection team members were required to comply with the following measures and other 
measures established by the Government of Mozambique: 
• Mandatory use of face masks 
• Frequent hand hygiene using hand sanitizers or soap and water 
• Practice of physical distance, keeping a distance of two meters apart
Abt Associates June 2022 ▌30
• Holding interactions, such as interviews with respondents, outdoors or in open spaces with 
ventilation
• Daily self-monitoring of symptoms (fever or chills; cough; shortness of breath or difficulty 
breathing; fatigue from their data collection team members each morning)
• Daily collection of COVID self-screening checklists by Field Work leaders
• Avoiding the sharing of tools, phones, or other objects when possible 
• Use of photos to confirm compliance with protocols during events like enumerator training 
and travel
• Use of privately contracted vehicles and drivers for data collection team member travel
Abt Associates June 2022 ▌31
6. CHARACTERISTICS OF POPULATION
This section describes the realized sample used in this evaluation and details the demographic 
characteristics of the students, teachers, and parents who participated in the evaluation. The 
description of characteristics provides context to the findings discussed in the sections below and 
confirms that there are no substantial differences in sample characteristics between base- and 
endline and across treatment arms. 
6.1 Sample Description 
The endline evaluation revisited the same schools as sampled at baseline. At baseline, data were 
collected at 448 schools, while at endline, data were collected at 445 schools (three schools that 
were permanently closed in the implementation period were lost to the analysis sample).
8 The 
goal for the endline data collection was a sample that matched the baseline sample in terms of 
numbers of respondents of each type. The endline samples represent an average of 88 percent of 
the number of participants in the baseline samples (Exhibit 4). 
Exhibit 4. Baseline and Endline Evaluation Sample 
Treatment 
Arm
Phase 
Evaluation Schools Head 
Teachers Cooks
Teachers Students
Grade 
2/3
Grade 
4/5
Grade 
2/3
Grade 
4/5
Maputo
Literacy and 
school 
feeding
BL 90 84 62 83 78 661 638
EL 90 89 68 84 88 690 696
School 
feeding only
BL 90 85 70 82 81 666 616
EL 88 89 81 85 81 670 645
Comparison BL 89 86 n/a 86 83 704 681
EL 88 88 27 90 79 689 676
Nampula
Literacy and 
school 
feeding
BL 90 85 82 79 72 707 685
EL 90 88 88 85 85 717 710
Comparison BL 89 89 n/a 82 80 690 676
EL 89 89 n/a 85 78 712 664
Total sample baseline 448 429 214 412 394 3428 3296
Total sample endline 445 443 264 429 411 3478 3391
Total sample (base- and endline) 893 872 478 841 805 6906 6687
Expected sample (base- and endline) 898 898 538 898 898 7184 7184
Overall response rate 99.4% 97.1% 88.8% 93.7% 89.6% 96.1% 93.1%
Note: The expected sample at endline was based on the realized sample at baseline. BL indicates baseline, and EL indicates endline. 
6.2 Student Demographic Characteristics 
The average student at endline is six to nine months older than the average student interviewed at 
baseline. This reflects third grade and fifth grade students being interviewed at endline while at 
baseline students in second grade and fourth grade were interviewed. On average, students in the 
8 Three schools in the Maputo province that were visited at baseline were found to be permanently closed at endline: two 
schools in the school feeding only group and one school in the comparison group.
Abt Associates June 2022 ▌32
school feeding only intervention were a little younger than the average student in the comparison 
group and the literacy and school feeding intervention. Note that students self-reported their age.
Key findings on student characteristics and differences across treatment arms are outlined below. 
Exhibit A8 presents student responses on household demographic characteristics. 
• Approximately half of the students interviewed were female despite random sampling that 
did not take account of gender. As a proxy for the student’s family household wealth, 
students were asked about their family’s access to water and ownership of household assets. 
A smaller percentage of students at endline said their family sources water from a river, 
stream, or lake, or a well or borehole, while a larger percentage of students indicated that 
their family has a communal tap or a water pipe or tap at home. A larger share of the families 
of students in the comparison group have access to a water pipe or have a tap in their house 
(between 39 and 49 percent), compared to students in the other treatment arms (between 13 
and 28 percent). Families of students in the literacy and school feeding intervention have on 
average less access to a water pipe or a tap at home. 
• Students reported on their family’s ownership of nine household assets (asset index). 
Compared with baseline, the average number of household assets in student’s homes at 
endline was slightly lower. For mid-grade students in the comparison group, however, the 
average number of household assets was a little higher compared to baseline. Families of 
mid-grade students in the comparison group on average also own the largest number of 
household assets (four household assets out of nine). Families of students in the literacy and 
school feeding intervention have on average the lowest number of household assets. 
6.3 Language of Literacy Instruction
As expected, nearly all teachers responded that Portuguese is the language of literacy instruction. 
Local languages are also used as a language of literacy instruction and to facilitate the teaching 
of Portuguese. Rhonga and Changana, languages that are primarily spoken in the south, are used 
by teachers in Maputo province. Few teachers, however, said they used Rhonga to teach literacy; 
a larger percentage of teachers indicated they used Changana. In school feeding only schools, 
Changana is used by approximately 35 to 40 percent of teachers, compared to 12 to 22 percent of 
teachers in comparison and literacy and school feeding schools at endline. About 22 to 43 
percent of teachers at comparison and literacy and school feeding schools in Nampula province 
said they used Macua, a language primarily spoken in the North. 
Students often speak languages other than the language of instruction at home. Between 13 and 
31 percent of students speak a different language at home than the language used for instruction 
(Exhibit A9). This difference is largest among students in the school feeding only intervention. 
6.4 Teacher Characteristics
At endline, just over 60 percent of teachers in the comparison group were female whereas a little 
over 40 percent of teachers at schools in the school feeding only intervention and school feeding 
and literacy intervention were female. Compared to baseline, the percentage of teachers in 
comparison schools and the school feeding only schools who were female was slightly higher. In 
the school feeding and literacy schools, fewer female teachers were interviewed at endline 
compared to baseline.
Teachers at schools in the school feeding only intervention more often teach multi-grade classes. 
While close to 10 percent of teachers in the comparison schools and the literacy and school 
feeding schools say they teach multi-grade classes, roughly a third of teachers in the school 
Abt Associates June 2022 ▌33
feeding only schools say they teach a multi-grade class. Compared to baseline, the percentage of 
teachers teaching classes with students from two or more grades has decreased. 
PAI had supported teacher training in 11 teacher training colleges throughout Mozambique, 
essentially facilitating the development of new teachers who would soon be placed in rural 
primary schools located in several provinces including Maputo and Nampula. Nearly all teachers 
say they attended and finished a teacher training college (TTC). The majority of teachers 
attended the Primary Teacher Training Institute. A relatively small percentage of teachers were 
trained at ADPP and the Pedagogical University with no noteworthy differences between base￾and endline. Similar to baseline, the percentage of teachers who attained a certification was near 
100 percent. The question about teacher’s highest level of education differed between baseline
and endline. The study team suspects that a literal approach to the English to Portuguese 
conversion of the baseline teacher survey failed to take into account education classifications 
commonly used in Mozambique. Baseline reports of teacher qualifications are unclear; therefore, 
responses are reported separately in the table A.10. Most teachers interviewed at baseline said 
that the certificate was their highest level of education (65 to 71 percent). At endline, most 
teachers responded that middle school (equivalent to high school in other contexts) was their 
highest level of education (61 to 73 percent). A smaller percentage of teachers said that 
secondary education (12 to 17 percent) or a bachelor’s degree or higher (14 to 23 percent) was 
their highest level of education (Exhibit A10). Given the comparability of percentages, despite 
the erroneous classification, we feel confident that teachers’ levels of education from baseline to 
endline did not vary dramatically. 
Not all teachers received specific training on how to teach reading and writing to students in 
primary grades during pre-service training. At baseline, between 58 and 61 percent of teachers 
received this training; at endline between 58 and 71 percent of teachers did. 
6.5 School Characteristics
At endline, schools in the school feeding only intervention had the fewest teachers on average 
(6.5 teachers), whereas schools in the comparison group have on average the largest number of 
teachers (14.8 teachers). The number of teachers in the treatment schools remained more or less 
the same between base- and endline, but comparison schools had on average two more teachers 
at endline compared to baseline. Treatment schools have on average fewer female teachers than 
male teachers. Comparison schools, however, have more female than male teachers and showed 
an increase in the number of female teachers when compared to the number of female teachers at 
baseline. 
Most schools in the sample have slightly more teachers teaching early-grade students than 
teachers teaching mid-grade students. The number of early-grade and mid-grade classes 
(streams) the school had mirrored the number of teachers teaching early-grade and mid-grade 
students. Differences between base- and endline in the number of teachers teaching early- and 
mid-grade students and number of early- and mid-grade classes are minimal (Exhibit 5). 
Abt Associates June 2022 ▌34
Exhibit 5. School Characteristics, by Study Arm and Time
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Total number of teachers 12.7 14.8 2.2 6.8 6.5 -0.3 10.0 10.3 0.3
Total number of male teachers 6.1 6.9 0.8 4.5 4.2 -0.3 5.8 5.8 0.0
Total number of female teachers 6.7 8.1 1.4 2.4 2.4 0.0 5.1 4.5 -0.6
Number of teachers currently 
teaching early-grade students 2.3 2.1 -0.3 1.2 1.2 0.0 1.9 1.7 -0.2
Number of teachers currently 
teaching mid-grade students 1.8 1.9 0.2 1.1 1.2 0.1 1.5 1.5 0.0
Number of early-grade classes 
(streams) 2.3 2.1 -0.1 1.4 1.2 -0.2 1.9 1.7 -0.2
Number of mid-grade classes 
(streams) 1.7 1.9 0.2 1.3 1.1 -0.1 1.4 1.5 0.1
Note: BL indicates baseline, and EL indicates endline. 
6.6 Parent Focus Group Characteristics
At baseline and endline, the goal was to complete 30 PFG discussions. Exhibit 6 and Exhibit 7
below provide details on the number and average size of PFG discussions per province and 
district. The sample planned to include a larger number of schools from Maputo to achieve 
adequate representation from both the school feeding only and the literacy and school feeding 
schools. At baseline, 30 PFG discussions were conducted. 
Exhibit 6. Parent Focus Group Sample with Literacy and School Feeding Treatment
Baseline Endline
Province/District
Number of 
Parent Focus 
Groups
Mean Number of 
Parents in the 
Focus Group
Number of 
Parent Focus 
Groups
Mean Number of 
Parents in the 
Focus Group
Maputo 
Magude 1 7 1 7 
Manhiça 4 7 3 10
Matutuine 3 8 2 9 
Moamba 3 7 3 11
Nampula Muecate 6 11 6 10
Nacaroa 6 12 6 10
TOTAL 23 9 21 9 to 10
Exhibit 7. Parent Focus Group Sample with School Feeding Treatment Only
Baseline Endline
Province/District
Number of 
Parent Focus 
Groups
Mean Number of 
Parents in the 
Focus Group
Number of 
Parent Focus 
Groups
Mean Number of 
Parents in the 
Focus Group
Maputo 
Magude 3 11 3 7 
Manhiça 2 8 3 10 
Matutuine - - 1 9 
Moamba 2 12 2 11 
Nampula Muecate - - - - 
Nacaroa - - - - 
TOTAL 7 10 9 10
Abt Associates June 2022 ▌35
There were 289 unique parent respondents (285 at baseline), with an average of 9–10 
participants in each discussion at baseline and endline. The average age of the parents was about 
39 (ranging from 17 to 74 years). The PFG discussions mainly included females (68 percent at 
endline, 57 percent at baseline). The average number of children that the parents had was 2.2 (3.4 
at baseline). On average, PFG discussions were composed of slightly more female parents at 
endline, and the average number of children was slightly lower for families represented at 
endline compared with those who participated at baseline. 
At endline, parents were asked how they were involved at the school. Roughly two-fifths (41 
percent) of the parents said that they did not have any role or involvement at the school. Parents 
were most likely involved as a cook at the school (26 percent of parents), followed by 16 percent 
of parents serving on the school council. Parents also reported being involved with the school by 
helping with the garden, construction, or cleaning. 
6.7 Key Official Characteristics 
At endline, 28 interviews were completed, including 17 representatives from Maputo province, 
nine from Nampula province, and with national representatives. Nearly half (46 percent) of the 
interview respondents held positions in the Department for General Education. Another quarter 
(25 percent) were District Services of Education, Youth and Technology directors, and the 
remaining six (including the national representatives) varied in their positions. We have not 
provided the more detailed breakdown of the respondents by district to protect the privacy of the 
respondents. 
Abt Associates June 2022 ▌36
7. MAIN FINDINGS
7.1 Summary of Findings
This section of the report provides a review of the effects on the three main outcomes: reading 
performance, attendance, and attentiveness. Evidence indicates that the school feeding and 
literacy intervention had a substantial effect on students’ reading performance. Effects on the 
other two outcomes were not statistically significant (Exhibit 8).9
Exhibit 8. Summary of Impact on Average Mean Reading Performance, Attendance, and Attentiveness
........................................................
School Feeding 
Only
School Feeding and 
Literacy
Impact Impact
Early-Grade Students
Student reading performance
(Average Reading Comprehension Scores) 1.2 4.8**
Student reading performance
(Percent Zero Reading Comprehension Scores) -10.6 -10.0**
Student attendance -4.6 -5.7
Student attentiveness -3.5 -4.1
Mid-Grade Students
Student reading performance
(Average Reading Comprehension Scores) -1.8 7.6*
Student reading performance
(Percent Zero Reading Comprehension Scores) -6.5 -14.4**
Student attendance 4.8 3.9
Student attentiveness -7.2* -3.0
Asterisks indicate statistical significance of difference-in-differences estimates. ** p<0.01; * p<0.05.
7.2 Impacts on Student Reading Performance
In schools implementing both literacy and school feeding interventions, the average student
reading comprehension scores for both early-grade and mid-grade students showed significantly 
more improvement over time than did scores in comparison schools. Early-grade students 
experienced an increase of 4.8 percentage points (p<.01, effect size=.16), and mid-grade students 
experienced an increase of 7.6 percentage points (p<.05, effect size=.26). There was no parallel 
difference in the improvement in reading comprehension for students in schools implementing 
only school feeding intervention versus students in comparison schools (Exhibit 9).
10 This
finding held up when the analysis models included covariates describing the difference in (1) 
9 See Appendix A for outcome levels at baseline and endline by treatment arm for all main and secondary outcomes. 
10 Exhibit 9 shows the average reading comprehension score across all treatment conditions at baseline separately for each 
condition and student age group. At endline, the reported reading comprehension score is the sum of the baseline average and 
the analysis model’s predicted change in reading comprehension for each condition and student age group. The difference-in￾differences estimate from the analysis model can be calculated by subtracting the average change in reading comprehension in 
the intervention group under study from the average change in reading comprehension for the comparison group.
Abt Associates June 2022 ▌37
language of instruction and the child’s home language and (2) the language of assessment and 
the child’s home language.11 
Exhibit 9. Impacts on Average Mean Reading Comprehension Scores
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average outcome for each condition at baseline. The depicted outcome at endline is the sum of that 
average at baseline and estimates from the regression model. Y-axis for this exhibit (not shown) runs from 0 to 40 percent.
In addition to improving average mean scores, another way to look at performance on the 
reading comprehension task is to score students on whether they answered at least one test 
question correctly. Combining literacy and school feeding reduced the share of early-grade 
students who could not answer any reading comprehension question correctly by 10.0 percentage 
points (p<.01, effect size = -.21) and of mid-grade students by 14.4 percentage points (p<.01, 
effect size = -.31) (Exhibit A12). As with average mean reading comprehension scores, school 
feeding alone did not impact the share of early-grade or mid-grade students who could not 
answer any reading comprehension question correctly. The magnitude of the impact estimates 
for school feeding alone suggest a reduction in the share of early-grade and mid-grade students 
11 A potential concern for the analysis of reading comprehension is a lack of correspondence between the language spoken at 
home, the language of instruction, and the language of assessment. This potential concern is not addressed by school fixed 
effects, as there is variation in these variables within schools. To test the robustness of the findings to this concern, we re￾estimated the regression model with two covariates—one indicating that the language of instruction differed from the 
language spoken at home and a second indicating that the language of assessment differed from the language spoken at 
home. Results were robust to including these covariates (see Exhibit A7).
Abt Associates June 2022 ▌38
who could not answer any reading comprehension question correctly, but the larger standard 
errors imply these estimates are not statistically significant.12
While it is a primary outcome, reading comprehension is one of six subtasks of the EGRA. The 
other five EGRA subtasks evaluated include oral comprehension, letter identification, syllable 
identification, oral reading fluency, and word dictation. However, findings were consistent 
across all subtasks, with early- and mid-grade students demonstrating statistically significant 
gains across most subtasks (Exhibit A13 though Exhibit A22). 
Stakeholder Perception on Student Learning
Respondents mentioned themes that may contextualize the 
learning environment or educational progress. A common 
theme was that the teachers felt that their training helped 
them with planning strategies and teaching methods. For 
example, the training helped teachers work with children 
according to their individual needs. 
One head teacher described that when bilingual 
teaching was introduced, the children were more at 
ease. Several key official respondents felt similarly 
that expanding bilingual education was important, 
and they described noticing stronger pedagogical 
performance. 
7.3 Impacts on Student Attendance 
Neither school feeding nor school feeding and 
literacy had an impact on average student 
attendance rate (absences) relative to average 
attendance in comparison schools. The magnitude of 
the impact estimates suggests opposite trends in attendance rates from baseline to endline for 
early-grade students (improved attendance) and for mid-grade students (decreased attendance) 
(Exhibit 10). However, none of the estimates is statistically significant (Exhibit A23 and Exhibit 
A24). 
12 Given that the school feeding only program was only implemented in the Maputo province, the number of schools sampled 
into this treatment arm is smaller than the number of schools in the comparison group and the literacy and school feeding 
group. Hence, the probability that we find a statistically significant difference between the school feeding only program and 
the comparison group, if such a difference would exist, is smaller. 
“…there is nothing better than learning 
in your own language… I didn't have 
that opportunity… but if I had learned in 
that way I would have been much
better.”
- Key Official (Maputo)
“Those who were in the bilingual system and those 
who were in the monolingual system had a slight 
difference, so I concluded that learning in their 
own language was better… children who have had 
literacy with bilingual reinforcement, they have 
better results... We were able to reinforce reading, 
writing and basic math in those children, but we 
could have reached more children.”
- Key Official (Maputo)
Abt Associates June 2022 ▌39
Exhibit 10. Impacts on Student Attendance (Measured by Any Absence Last Week)
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average student-reported absence (at least one day last week) for each condition at baseline. The 
depicted outcome at endline is the sum of that average at baseline and estimates from the regression model. Y-axis for this exhibit (not shown) 
runs from 0 to 40 percent.
Stakeholder Perception of Student Attendance
During focus groups and interviews, many parents and key 
officials reported that they noticed an impact on attendance 
because of the project, specifically because of the school 
feeding intervention. Parents noted that their children were 
more excited to attend classes because they were being fed.
The lack of corroborating quantitative evidence that the 
McGovern-Dole project improved attendance for early￾grade and mid-grade students may suggest that reduction in 
attendance was in large part due to COVID-19-related school closures. 
Key official interview respondents provided a 
broader perspective and felt that the projects
impacted attendance, as well as enrollment into 
school. The respondents also observed greater 
retention of students in school because of the school 
feeding intervention or framed the improvement in 
terms of a decrease in dropouts. A few interviews 
specifically mentioned an improvement in the 
enrollment and retention of girls. 
“When there is a meal at school, the 
children are motivated and adhere to 
school… That program contributes 
directly to the reduction of dropouts, the 
children look at that in positive way.”
- Key Official (Maputo)
“We had higher retention because [students]
knew that by going to school they were not only 
going to learn, but they were also going to get 
a meal. So, they had more time in school, less 
truancy, and less lateness… the reinforcement 
of their learning abilities was through literacy 
and also through bilingual education.”
- Key Official (Maputo)
 
Abt Associates June 2022 ▌40
Some of the officials described that the parent and 
community involvement at the school, or parent 
participation in literacy centers, helped parents see the 
value of education for their children. One PFG participant 
mentioned that getting involved at the school motivated 
them to make sure their children were attending school 
and leaving on time for school. A few officials also 
mentioned that students’ positive experience in school, 
including getting meals, encouraged their siblings to attend.
7.4 Impacts on Student Attentiveness
The improvement in the average level of student attentiveness was not significantly better in 
treatment schools than in comparison schools (Exhibit 11). Although there were declines in 
teacher-reported student attentiveness for students across grades and conditions, the impact 
estimates are not statistically significant (Exhibit A25 and Exhibit A26). 
Levels of teacher-reported student attentiveness were very high (greater than 90 percent in all 
arms) at baseline. As a result, there is limited room for growth in this measure at endline. Indeed, 
the apparent decline in student attentiveness is partly explained by the growth in attentiveness in 
comparison schools. It would not have been possible for attentiveness to increase in treatment 
schools at the same rate, as this would have resulted in rates of attentiveness greater than 100 
percent.
Exhibit 11. Impacts on Teacher-Reported Student Attentiveness
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average outcome for each condition at baseline. The depicted outcome at endline is the sum of that 
average at baseline and estimates from the regression model. Y-axis for this exhibit (not shown) runs from 80 to 105 percent.
“The younger ones, who were not yet of 
school age, were already looking forward 
to [school] when their classes would start, 
because they always followed their 
[siblings] to school and [saw their] meals.”
- Key Official (Nampula)
Abt Associates June 2022 ▌41
Stakeholder Perceptions of Student Attentiveness
Concerns about attentiveness were supported by the head teacher and key official responses, with 
some respondents mentioning challenges with student attentiveness related to the timing of the 
school meal. Similar to the findings above, there were some challenges reported with 
attentiveness directly after students received a meal. Head teachers suggested that meals should 
be carefully scheduled so they do not interfere with 
learning activities. 
Respondents generally felt that the net benefit of students 
having a school meal was positive and that it improved the 
academic environment or attentiveness. A theme that 
emerged from head teachers’ open-ended responses was 
that the intervention improved the learning environment.
7.5 Impacts for Student Subgroups
As described earlier, research suggests that literacy rates vary between women and men in 
Mozambique, with 60 percent of males identified as literate compared to only 28 percent of 
women (USAID, 2021a). Given the disparities in literacy and a growing focus on improving the 
educational experiences of girls, it is important to highlight impacts by gender. 
Impacts on student reading comprehension scores did not vary in meaningful ways between boys 
and girls. Average reading comprehension scores improved for both early-grade and mid-grade 
boys in schools implementing the literacy and school feeding intervention (by 5.1 percentage 
points for early-grade boys and 11.1 percentage points for mid-grade boys). While the magnitude 
of impacts for girls is smaller and not statistically significant, the difference in impacts between 
boys and girls is also not statistically significant. As in the assessment of overall impacts, there 
was no difference in the improvement in reading comprehension for boys or girls in schools 
implementing the school feeding only intervention versus boys or girls in comparison schools 
(see Exhibit 12). 
Similarly, impacts on the percentage of students who could not answer any reading 
comprehension question correctly did not vary for boys and girls. Combining literacy and school 
feeding reduced the share of early-grade boys who could not answer any reading comprehension 
question correctly by 9.5 percentage points and reduced the share of early-grade girls who could 
not answer any reading comprehension question correctly by 10.5 percentage points. Combining 
literacy and school feeding also reduced the share of mid-grade boys who could not answer any 
reading comprehension question correctly by 17.9 percentage points and reduced the share of 
mid-grade girls who could not answer any reading comprehension question correctly by 12.9 
percentage points. The differences in these impacts between boys and girls were not statistically 
significant. As in the assessment of overall impacts, school feeding alone did not impact the 
share of early-grade or mid-grade boys or girls who could not answer any reading 
comprehension question correctly (Exhibit 12).
“The children progressed a lot, and the 
children's desire to learn was greater. It 
was possible for the children to retain 
more of what they learned.”
- Head Teacher (Maputo)
Abt Associates June 2022 ▌42
Exhibit 12. Impacts on Student Reading Performance and Attendance for Boys and Girls13
........................................................ School Feeding Only
School Feeding and 
Literacy
Impact Impact
Early-Grade Students
Boys
Student reading performance
(Average Reading Comprehension Scores) 2.9 5.1*
Student reading performance
(Percent Zero Reading Comprehension Scores) -12.8 -9.5*
Student attendance -9.3 -14.8**†
Girls
Student reading performance
(Average Reading Comprehension Scores) -1.3 4.1
Student reading performance
(Percent Zero Reading Comprehension Scores) -6.9 -10.5**
Student attendance -1.4 0.6†
Mid-Grade Students
Boys
Student reading performance
(Average Reading Comprehension Scores) -0.7 11.1*
Student reading performance
(Percent Zero Reading Comprehension Scores) -3.3 -17.9**
Student attendance 3.7 -1.9
Girls
Student reading performance
(Average Reading Comprehension Scores) -2.2 5.5
Student reading performance
(Percent Zero Reading Comprehension Scores) -9.3 -12.9*
Student attendance 7.0 9.9
Sample sizes for analysis of student reading performance: 
Boys: 2,745 comparison; 1,238 school feeding only; 2,737 school feeding and literacy.
Girls: 2,729 comparison; 1,338 school feeding only; 2,767 school feeding and literacy.
Sample sizes for analysis of student attendance:
Boys: 2,743 comparison; 1,236 school feeding only; 2,731 school feeding and literacy.
Girls: 2,730 comparison; 1,338 school feeding only; 2,760 school feeding and literacy.
Asterisks indicate statistical significance of difference-in-differences estimates. ** p<0.01; * p<0.05. Crosses indicate statistical significance of 
difference in impact estimates between boys and girls. †† p<0.01; † p<0.05.
Impacts on student attendance in schools implementing the literacy and school feeding 
intervention varied between boys and girls. Early-grade boys experienced a reduction in 
13 Subgroup analysis of attentiveness is not included as that outcome is based on teacher report and is not able to 
be disaggregated by sex. 
Abt Associates June 2022 ▌43
attendance of 14.8 percentage points while attendance for girls increased by 0.6 percentage 
points (the difference in these impacts is statistically significant). Although not statistically 
significant, the magnitude of impacts for mid-grade students in these schools had a similar 
pattern (the difference between the -1.9 percentage point impact for boys and 9.9 percentage 
point impact for girls had a p-value of 0.0504). There was no evidence that impact on attendance 
varied for boys and girls in schools that implemented only school feeding interventions.
Abt Associates June 2022 ▌44
8. SECONDARY FINDINGS
Analysis of secondary evaluation questions generates evidence to further explore the mechanisms 
through which impacts on primary evaluation questions did (or did not) operate. The subsections that 
follow report estimates of impact for secondary evaluation questions pertaining to instructional 
inputs, feeding, and student knowledge and use of recommended health practices, and food 
preparation facilities. The analysis of each secondary evaluation question is limited to key secondary 
outcome measures; estimates of other related outcomes appear in the appendix. 
8.1 Impacts on Instructional Inputs
Teacher Training on Reading Instruction
Training teachers on the fundamentals of reading instruction is a direct avenue through which 
literacy activities might influence student reading performance. Teachers of both early-grade and 
mid-grade students in school feeding and literacy schools experienced sizeable positive impacts in 
receipt of training on the fundamentals of reading instruction (15.3 percentage points for teachers of 
early-grade students and 27.9 percentage points for teachers of mid-grade students) ( Exhibit 13). 
The positive impacts are strongly consistent with the primary impacts on student reading 
performance, which were both positive and only apparent in school feeding and literacy schools. This
suggests that teacher training on the fundamentals of reading instruction is a plausible mechanism for 
explaining the impacts on student reading performance.
Exhibit 13. Impacts on Instructional Inputs
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
0.3 14.5 14.3 15.5 15.3*
(9.3) (7.0)
Teacher-reported absence of at 
least one day last week -0.6 -12.2 -11.6 -1.7 -1.1
(6.7) (6.6)
Teacher has access to school 
supplies 11.5 0.2 -11.3** 1.5 -10.0**
(3.3) (3.7)
Mid-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
-8.9 -5.0 3.9 19.0 27.9**
(9.1) (7.5)
Teacher-reported absence of at 
least one day last week -1.2 -3.4 -2.2 1.4 2.6
(6.7) (5.8)
Teacher has access to school 
supplies 8.7 3.3 -5.5 0.7 -8.1*
(4.1) (3.9)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 625 Comparison; 310 School feeding only; 632 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌45
Resources for Reading Instruction
The literacy intervention also provided resources to schools to support reading instruction. No 
statistical difference was shown in access to instructional resources for teachers in schools 
implementing literacy interventions and comparison schools or schools implementing only 
feeding interventions (Exhibit 13). Teachers of early-grade students in both treatment arms 
experienced a reduction of 10.0 percentage points in access to school supplies, such as 
pens/pencils, notebooks, chalk, and cardboard, and teachers of mid-grade students in school 
feeding and literacy schools experienced a reduction of 8.1 percentage points in access to school 
supplies. That these findings operate in a different direction of the impacts on student reading 
performance suggests that provision of instructional resources was not a key mechanism for the 
improvement in reading performance. 
Stakeholder Perception on School Materials
Several of the key officials discussed the emphasis during teacher training on the benefits of 
using materials. Overall, key officials thought that materials such as notebooks, writing utensils, 
books, and teacher materials promoted a positive learning atmosphere. However, some of the key 
officials mentioned that access to school materials may have been a challenge for the schools. 
Further, a theme emerged that materials provided 
by donors or grants only last so long, and that the 
material resources provided were negligible. One of 
the key officials pointed out that trainings for 
teachers should challenge teachers to use other 
types of resources. 
On the other hand, 16 PFGs mentioned that they 
had some materials at home for their children to 
use. They were more likely to have writing materials versus drawing materials. A few mentioned 
that the students brought materials back and forth between school. 
Teacher Absence
Another hypothesis is that the feeding and literacy interventions may have increased teacher 
attendance and, as a correlate, student instructional time. However, no statistical difference was 
shown in average level of teacher absence in schools in either treatment arm versus level of 
teacher absence in comparison schools (Exhibit 13).
8.2 Impacts on Feeding Outcomes and Nutrition Knowledge 
The McGovern-Dole program theory of change hypothesizes that the school feeding intervention 
will improve student attendance/health outcomes and attentiveness, which, in turn, may function as 
mediators of or contributors to improvement in reading performance. There were no significant 
impacts on average levels of student hunger and dietary sufficiency in treatment schools versus 
comparison schools (Exhibit 14). Although non-significant, there were consistent negative trends in 
both treatment arms and both student age groups. 
Indeed, some feeding outcomes were worse in treatment schools than comparison schools. In school 
feeding only schools, the share of mid-grade students reporting current hunger at the time of survey 
completion increased by 14.3 percentage points. Across both treatment arms, early-grade students 
experienced a reduction in minimum meal frequency (34.2 percentage points in school feeding only 
schools and 14.7 percentage points in school feeding and literacy schools) and minimum acceptable 
diet (13.8 percentage points in both school feeding only schools and school feeding and literacy 
schools). Mid-grade students experienced similar reductions in dietary outcomes. In school feeding 
“in the trainings I feel that there is [effort] to make 
teachers see that they should not expect 
everything from the project. You can recycle 
cardboard, cans, and make lots of things. So, it 
is not possible for the project to offer everything.”
- Key Official (Nampula)
Abt Associates June 2022 ▌46
only schools, they experienced a reduction in minimum meal frequency of 16.8 percentage points; in 
school feeding and literacy schools, they experienced a reduction of 9.8 percentage points in 
minimum dietary diversity; and in both treatment arms they experienced reductions in minimum 
acceptable diet of 13.9 and 13.1 percentage points, respectively.
Nearly all PFGs stated their household diets and meals mostly stayed the same, regardless of the 
school feeding activity. Nearly half (n=14) of PFGs said that they cooked fewer meals as a result of 
the school feeding activities, and eight PFGs stated that the children ate fewer meals at home.
While the feeding intervention was unable to improve direct measures of hunger and diet, it
succeeded in improving students’ nutrition knowledge. Early-grade students across both treatment 
arms experienced improvements in nutrition knowledge of 0.5 and 0.9, with effect sizes of 0.35 and
0.51 respectively. Similarly, mid-grade students in school feeding and literacy schools experienced 
improvement of 0.8 units (with effect sizes of 0.17 and 0.09).
Exhibit 14. Impacts on Feeding Outcomes
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Student reports feeling hungry now -8.2 0.9 9.1 -15.2 -6.9
(6.2) (4.9)
Student achieved minimum meal 
frequency 7.0 -27.2 -34.2** -7.7 -14.7**
(6.2) (5.1)
Student achieved minimum dietary 
diversity -28.8 -26.5 2.3 -28.0 0.8
(6.2) (4.7)
Student achieved minimum 
acceptable diet -7.0 -20.8 -13.8* -20.8 -13.8**
(5.8) (4.4)
Student's nutrition knowledge 
score -0.4 0.0 0.5** 0.4 0.9**
(0.1) (0.1)
Mid-Grade Students
Student reports feeling hungry now -2.1 12.2 14.3* -2.0 0.1
(6.7) (4.6)
Student achieved minimum meal 
frequency 1.4 -15.4 -16.8** -2.5 -3.9
(4.9) (4.2)
Student achieved minimum dietary 
diversity -24.6 -31.2 -6.7 -34.3 -9.8*
(7.3) (4.6)
Student achieved minimum 
acceptable diet -9.2 -23.1 -13.9* -22.2 -13.1**
(6.7) (4.2)
Student's nutrition knowledge 
score -0.5 -0.1 0.4 0.3 0.8**
(0.2) (0.2)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 2,576 School feeding only; 5,504 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌47
8.3 Impacts on Student Knowledge and Use of Recommended Health Practices
Students’ knowledge of recommended health practices and their translation of that knowledge 
into use of those practices are considered to be potential mediators of improved reading 
performance because good health directly corresponds to school attendance and attentiveness, 
both of which could influence student achievement. There were no significant differences in 
student knowledge and use of health practices across grade levels and treatment arms (Exhibit 
15). There were improvements in hygiene and sanitation knowledge among both early-grade 
students and mid-grade students but the difference from baseline to endline was not statistically 
significant. 
Relatedly, another series of outcomes captured access to clean water and sanitation. Head 
teachers reported whether schools had improved latrines, access to hand-washing stations, and 
access to ash or soap. Neither treatment arm had a favorable impact on access to clean water and 
sanitation, and the school feeding only intervention reduced access to hand-washing stations as 
well as to ash or soap (by 27.7 and 21.0 percentage points, respectively). 
Exhibit 15. Impacts on Student Knowledge and Use of Recommended Health Practices
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Student's hygiene and sanitation 
knowledge score 0.8 1.1 0.3 1.6 0.9**
(0.3) (0.2)
Student reports absence last week 
as a result of illness 0.3 -17.8 -18.1 13.2 13.0
(10.2) (8.2)
Mid-Grade Students
Student's hygiene and sanitation 
knowledge score 1.0 0.9 -0.0 1.3 0.3*
(0.2) (0.1)
Student reports absence last week 
as a result of illness 1.7 -2.1 -3.8 10.6 9.0
(12.1) (9.1)
All Students
Head teacher reports school has 
improved latrines 11.4 11.8 0.3 8.6 -2.8
(7.5) (5.0)
Head teacher reports school has 
tippy taps or hand-washing 
stations 
29.4 1.7 -27.7** 19.6 -9.8
(9.0) (6.1)
Head teacher reports all tippy taps 
or hand-washing stations have ash 
or soap 
26.7 5.7 -21.0** 21.0 -5.7
(7.0) (4.2)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,471 Comparison; 2,576 School feeding only; 5,503 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌48
Stakeholder Perception on Student Knowledge and Health Practices
Parents shed some light on what students were learning and what health practices they were 
applying at home. Parents perceptions of the health education practices were positive. Most PFG
participants mentioned that their children learned the importance of washing their hands at 
school. The next common themes that parents described were that their children learned personal 
hygiene (including bathing, cutting their fingernails, and brushing their hair) and oral hygiene 
(including brushing their teeth). 
Eight of the 30 PFG participants described that 
their children shared their knowledge with the 
parents or their siblings. Parents also described 
that the children’s lessons supported the parent’s 
responsibilities, as students would take more 
initiative with household chores, including 
washing dishes, sweeping, and cleaning their 
clothes. Two parent focus groups mentioned 
specific cleaning practices, such as that the children learned to sweep with the wind, so that they 
do not inhale the dust. Nearly half of the PFG participants mentioned they noticed their children 
washing their hands before or after eating. 
“…they are applying what they have learnt at 
school. Because we can see that what he is 
doing he has not learnt at home but at school… 
we ask ‘where did you learn that?’ and he says 
‘Teacher taught me about hygiene.’”
- Parent Focus Group (Nampula)
Abt Associates June 2022 ▌49
9. FINDINGS ON IMPLEMENTATION
9.1 Summary of Findings
Capturing implementation in October 2021, which for some schools was over a year after 
services were provided, has challenges. Whenever possible, we asked participants to describe 
programming prior to COVID-19, but respondents may have mingled descriptions of 
programming in October 2021 and how it was implemented prior to COVID-19 when 
implementing partners were more active. We believe the overall results, across provinces, are
useful in understanding intervention receipt concurrent with the measure of intervention 
effectiveness.
When considering implementation across treatment schools and provinces, results of the fidelity 
analyses suggest that the school feeding and literacy interventions were adequately implemented. 
The availability of similar feeding and literacy activities in comparison schools was also 
examined. Although there was evidence that alternative programming of a similar nature to
McGovern-Dole was concurrently being implemented in comparison schools,14 McGovern-Dole
schools had higher implementation scores. This suggests that there was a contrast between 
treatment and comparison schools that could lead to greater improvements in the treatment 
schools. On average, the schools in Nampula were rated as having better implementation of their 
interventions than the schools in Maputo. Nampula schools implementing school feeding and 
literacy programming received the highest scores. There is evidence that implementation in 
Maputo schools was severely limited largely due to endline data collection occurring when 
services were disrupted, and the implementing partner was no longer operating in the districts. 
Although we collected data at the same chronological time, the provinces were not at comparable 
stages of implementation. This is likely a key factor in the differences in fidelity scores between 
the two provinces. 
A closer examination of implementation fidelity at the school level reveals that several schools 
were no longer receiving intervention components at endline in 2021. In Maputo, 68 percent of 
treatment schools no longer received the school feeding intervention. This is likely due to 
COVID-related school closures and transitions in implementing partners. 
9.2 Determination of Implementation Adequacy
Across the treatment arms, 32 percent of schools implemented intervention elements with an 
adequate level of fidelity. Schools with feeding and literacy interventions had the highest 
percentage of adequate implementing schools. It should be noted, however, that 7 percent of 
comparison schools were scored as adequate implementers on the collective elements of school 
feeding, literacy instruction, and teacher training (Exhibit 16). This may be an indication that 
alternative programming was being delivered in comparison schools and the treatment￾comparison contrast was smaller than anticipated. Existence of alternative programming in 
comparison schools, though not uncommon in real-life evaluations, can diminish our ability to 
understand intervention effectiveness.
14 In particular, comparison schools across Maputo were implementing literacy instruction and teacher training 
activities that were comparable to the McGovern-Dole although overall less intensive. 
Abt Associates June 2022 ▌50
Exhibit 16. Proportion of Adequate Implementing Schools Included in Analyses
Treatment Condition Inadequate Adequate Missing 
Comparison 165 (93%) 12 (7%) 0 (0%)
Feeding only 51 (58%) 37 (42%) 0 (0%)
Feeding + literacy 84 (47%) 95 (53%) 1 (1%)
Total 300 (67%) 144 (32%) 1 (<1%)
9.3 Implementation of School Feeding
On average, the school feeding intervention was delivered with adequate fidelity in schools in 
Nampula implementing the school feeding and literacy interventions (Exhibit 18). In Maputo, the 
feeding intervention was not being implemented with adequate fidelity at the time of endline data 
collection regardless of whether the feeding intervention was implemented alone or in 
combination with the literacy intervention. In this case, it is better to interpret these results as an 
indication of sustainability and conclude that programming was not sustained in Maputo after the 
withdrawal of implementing partners and upon school reopening in March of 2021. 
The lack of adequate implementation of the feeding intervention may be part of why there were 
no positive impacts on secondary feeding outcomes, for example, on student short-term hunger, 
which is expected to improve attentiveness. The majority of schools in Maputo (both treatment
and comparison) and Nampula (comparison only) did not have a school feeding intervention in 
place during endline data collection (Exhibit 17). Schools served by World Vision in Nampula 
reported that feeding activities continued in the 2021 school year once school reopened and were 
in place when school staff and students were surveyed. 
Exhibit 17. Number of Schools Without a Feeding Intervention at Endline
Study Arm # Schools % Schools
Maputo comparison 61 of 89 68.5%
Maputo food & literacy 65 of 90 72.2%
Maputo food only 56 of 89 62.9%
Nampula comparison 89 of 89 100.0%
Nampula food & literacy 0 of 89 0.0%
Note: Head teacher survey item: Does the school have a school feeding program? [Answer=No]
Stakeholder Perception of School Feeding Implementation
In the cook survey, when asked what changes they observed since they started working at the 
school as a cook, 12 cooks from Maputo and 
8 from Nampula described that there were not 
enough kitchen supplies at endline; more 
specifically they mentioned that there were 
fewer dishes or kitchen utensils. Similarly, 
there were some, but not many, instances of 
cooks describing “improvements” or 
construction to the kitchen infrastructure. 
Sixteen cooks from Maputo and nine from 
Nampula mentioned improvements; some said 
that the cooking space (or food storage space) 
“we know that [the school meal prep] work is not 
easy, [as time went on we realized] that they should 
also have some incentive, right? At the end of the 
month they should have a part of the amount of flour, 
right? This ended up encouraging the community 
itself to continue doing that kind of work.”
- Key Official (Nampula)
Abt Associates June 2022 ▌51
was created or constructed due to the project, or that the cooking area was given more space over 
time. 
Key officials described other challenges with implementing the school feeding intervention. 
Namely, some reported that it was difficult to convince parents or volunteers to prepare the 
school meals if there was a weak incentive to do so. Overall, as described in Section 7, key 
officials still felt that when school feeding was implemented, there was an impact on student 
attendance and attentiveness.
9.4 Implementation of Literacy Instruction
On average, literacy instruction was delivered with adequate fidelity across schools in Maputo 
and Nampula (Exhibit 18). Using the same implementation fidelity measure, there is evidence 
that comparison schools located in Maputo can be considered to have received a literacy 
intervention that met the criterion for adequate fidelity (although the fidelity score, on average, 
was lower than for the McGovern-Dole schools in the province). In comparison schools in 
Nampula that were implementing a literacy intervention, it was delivered with low fidelity 
(average fidelity score across Nampula comparison schools = score 5.7 of 12).
9.5 Implementation of Teacher Training
Since teacher training is a potential mechanism through which students can be exposed to higher 
quality literacy instruction, which could lead to increases in literacy, we explored the degree to 
which teacher training and support was implemented with fidelity. Analyses suggest that
teachers in treatment schools in Nampula and Maputo received training that was delivered with 
adequate fidelity. Nampula McGovern-Dole schools experienced the highest level of 
implementation fidelity (average fidelity score across World Vision treatment schools = score 
10.80 of 12). PAI delivered the literacy intervention with adequate fidelity to McGovern-Dole
schools in Maputo (average fidelity score across PAI food and literacy treatment schools = score 
9.83 of 12; food only = score 9.15 of 12). Using the same fidelity measure, there is evidence that 
comparison schools located in Maputo experienced a lower degree of training than treatment 
schools (average score across Maputo comparison schools = score 5.98 of 12). Comparisons 
schools in Nampula also received inadequate literacy programming (average score across 
Nampula comparison schools = score 5.03 of 12). 
Exhibit 18. Average Fidelity Scores Across School by Intervention Component and by Treatment Arm
Study Arm
Average Feeding 
Implementation Scores
(0-15)
Average Literacy 
Implementation Scores
(0-12)
Average Training 
Implementation Scores
(0-12)
Maputo comparison 3.32 7.08 5.98
Maputo food & literacy 6.14 7.83 9.83
Maputo food only 7.34 7.70 9.15
Nampula comparison 0.11 5.70 5.03
Nampula food & literacy 8.72 9.53 10.80
Note: Intervention elements are considered to have been implemented with “adequate” fidelity if they delivered at least 50 percent of the 
measured items (e.g., received a score of 5 and above out of 10). 
Abt Associates June 2022 ▌52
Stakeholder Perception of Implementation
Reported by parents and key officials, 
qualitative analysis found that community
engagement, specifically local buy-in and 
parent participation, was important to the 
success in implementing the school feeding 
intervention. 
On the other hand, some key official 
respondents noted that the lack of buy-in 
sometimes hurt the implementation. They 
attributed this to an initial lack 
of understanding about the 
benefits of the intervention. 
“For parents, school feeding is a great opportunity to 
participate in school activities. The parents were giving 
their contribution in terms of labor for the construction 
of small infrastructures such as the kitchen, but the 
biggest contribution was in the preparation of meals 
and productive activities.”
- Key Official (Maputo)
“…in the early years, people sometimes refused to go to prepare the 
porridge, so the ministry had to go down there to find out what was 
happening, and we ended up discovering that it was because we had 
trusted the school council and not the local leadership… but [the leader] 
managed to mobilize the community... So, we have to have this 
perspective that if we need to introduce something new, the first people 
to be informed are the leaders.”
- Key Official (Maputo)
Abt Associates June 2022 ▌53
10. FINDINGS ON IMPACT CONTROLLING FOR 
IMPLEMENTATION 
10.1 Summary of Findings
Analyses of the impacts on reading performance were conducted using the subset of treatment 
schools with adequate implementation of their assigned interventions. For the schools 
implementing combined feeding and literacy interventions, the positive and significant impacts 
were stronger for the adequately implementing treatment schools. For the schools implementing 
only feeding interventions, the results were no different for adequately implementing schools 
compared to analyses using all schools. 
10.2 Impact Findings for Schools with Combined Literacy and Feeding Interventions
For average mean reading comprehension scores, early-grade students in adequately 
implementing schools experienced an increase of 5.4 percentage points (p<.05) and mid-grade 
students experienced an increase of 10.9 percentage points (p<.01) (Exhibit 19). When including 
all treatment schools, regardless of implementation adequacy, average mean reading 
comprehension scores for early-grade students increased by 4.8 percentage points (p<.01) and for 
mid-grade students by 7.6 percentage points (p<.05) (see Exhibit A12 for more detailed findings 
on reading comprehension). Limiting this analysis to only adequate implementers demonstrates 
that higher implementation is associated with stronger intervention impacts on student reading 
comprehension. Below, we also visually depict this finding using the average regression-adjusted 
mean reading comprehension score across all adequately implementing schools in treatment 
conditions at baseline and graphically depicting the change over time separately for the same 
schools in each condition at endline (Exhibit 19). 
Though reading comprehension was the main indicator of interest, we also evaluated student 
performance on the other five EGRA subtasks related to literacy (Exhibit A36 through Exhibit 
A41).
10.3 Impact Findings for Schools with Feeding Interventions Only
As was the case for impact analyses using all treatment schools implementing feeding 
interventions, analyses using only adequately implementing schools with school feeding alone 
did not improve average mean reading comprehension relative to comparison schools. While it is 
a primary outcome, reading comprehension is one of six subtasks of the EGRA. Exhibit A36
through Exhibit A41 report findings for the other five EGRA subtasks (oral comprehension, 
letter identification, syllable identification, oral reading fluency, and word dictation). 
Abt Associates June 2022 ▌54
Exhibit 19. Impacts on Average Mean Reading Comprehension Scores (Adequately Implementing Treatment 
Schools Only)
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average outcome for each condition at baseline. The depicted outcome at endline is the sum of that 
average at baseline and estimates from the regression model. Y-axis for this exhibit (not shown) runs from 0 to 40 percent.
Unlike impacts on reading comprehension, adequately implementing schools in either treatment 
conditions (school feeding plus literacy and school feeding alone) did not improve average mean 
student attendance relative to comparison schools (Exhibit 20). At baseline, roughly a quarter of 
students were absent at least once in the prior week. That proportion of absenteeism remained 
relatively stable – if not decreasing slightly – at endline. These findings mirror the lack of 
impacts on attendance (absences) across all treatment schools regardless of implementation 
adequacy (Exhibit A23 and Exhibit A24). 
Abt Associates June 2022 ▌55
Exhibit 20. Student Attendance (Measured by Any Absence Last Week; Adequately Implementing Treatment 
Schools Only)
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average outcome for each condition at baseline. The depicted outcome at endline is the sum of that 
average at baseline and estimates from the regression model. Y-axis for this exhibit (not shown) runs from 0 to 40 percent.
Unlike impacts on reading comprehension, adequately implementing schools in either treatment 
conditions (school feeding plus literacy and school feeding alone) did not improve average mean 
student attentiveness relative to comparison schools (Exhibit 21). These findings mirror the lack 
of impacts on attentiveness across all treatment schools regardless of implementation adequacy 
(Exhibit A25 and Exhibit A26). 
Abt Associates June 2022 ▌56
Exhibit 21. Impacts on Teacher-Reported Student Attentiveness (Adequately Implementing Treatment 
Schools Only)
Notes: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. “C” denotes comparison schools; “SF” denotes school feeding only schools; “SF+L” denotes school feeding and literacy schools. 
Exhibit reports the unadjusted, weighted average outcome for each condition at baseline. The depicted outcome at endline is the sum of that 
average at baseline and estimates from the regression model. For mid-grade students, this depiction results in a rate of student attentiveness 
above 100 percent in school feeding and literacy schools. This estimate should not be interpreted as implying that greater than 100 percent of 
teachers reported their students were attentive at endline, but instead is useful for reproducing the model-based difference-in-differences 
estimate via subtraction. Y-axis for this exhibit (not shown) runs from 80 to 105 percent.
Abt Associates June 2022 ▌57
11. PROJECT ADAPTATIONS RESULTING FROM COVID-19
Parents, key officials, head teachers, and cooks all provided details on challenges and barriers 
related to the COVID-19 pandemic and how the McGovern-Dole projects were adapted in 
response to these. In the endline interviews, surveys, and PFG discussion respondents were asked 
about the impact of the COVID-19. Often, before data collectors even specifically mentioned the 
pandemic, it was clearly “front of mind” for the respondents. The COVID-19 interruptions and 
modifications are categorized as those affecting school operations, school feeding programming, 
and school health services.
11.1 COVID-19 and School Operations 
In March 2020, with the closure of in-person teaching, schools 
switched to remote learning, including the use of community 
radio and television, video conferencing, and WhatsApp to 
conduct remote classes. A few key officials raised concerns 
because of unequal access to the technologies, such as internet 
access. They also noted unequal assistance from parents. Some 
students have parents at home to help them with schoolwork, while others do not. When school 
was the primary source of learning, there was previously more equality of learning opportunities, 
since all students had access to support from teachers. 
Some key officials felt that the inconsistent schooling 
hurt early grades the hardest and argued that “face-to￾face” time with teachers was fundamental to student 
learning. Some referenced that with breaks in their 
learning, students were not as motivated to continue 
with at-home school or to return to school after 
COVID-19 restrictions were lifted. 
Another role that parents were 
expected to play at some schools was 
coming to school to obtain materials 
for children to complete at home, such 
as exercise sheets. Some key officials 
and teachers discussed challenges at 
these schools with parents not picking 
up materials for their children, or the 
school not being able to produce 
enough materials for all students. For example, one respondent mentioned that the school made 
an effort to copy the materials for all students but that there were not enough handouts for 
students to take home. 
Parents also detailed changes in the patterns of their students attending school and changes in 
their motivation. Some attributed this to students prioritizing agricultural work over schoolwork, 
which some parents preferred. One PFG explained 
witnessing differences in the teaching style throughout 
the pandemic, which has interrupted student’s studies 
and progress. However, the parents were hopeful that 
“The handouts were not enough… the 
parents didn't even take them to heart, but 
the teachers, the principals did mobilization 
work, so that the parents came to school to 
pick up the handouts to give their children.” 
- Key Official (Nampula)
“…[children working from 
home] when the network is 
poor…How do you solve the 
problem of exclusion?”
- Key Official
“…[first and second graders have] classes twice a week, and 
the rest of the days they are at home…this negatively affects 
the [learning] process… in these initial classes the physical 
presence of the teacher to follow up is indispensable... there 
was no shortage of appeals to urge teachers to develop [take￾home exercises] from March of last year until January of this 
year. This process did not work very well, there was a setback.”
- Key Official (Nampula)
“This disease has killed the school, they are 
not studying at all.”
- Parent Focus Group (Nampula)
Abt Associates June 2022 ▌58
getting back to school daily would re-motivate the students. 
11.2 COVID-19 and School Feeding Interventions
School closures halted the usual implementation of the McGovern-Dole feeding intervention, 
resulting in fewer students at endline reporting that they were receiving a meal from their school. 
Key officials, cooks, and PFG respondents referenced stopping other activities within the feeding 
intervention, including the school gardens and agricultural classes. 
According to the school cooks from McGovern-Dole schools, most (87 percent) of the schools 
did not provide a daily meal while the schools were shut down (Exhibit 22). Additionally, a 
majority of the cooks reported that the school garden activity was halted (79 percent) (Exhibit 
23). 
Exhibit 22. Percentage of Treatment Schools Where Students Received a Daily Meal During COVID-19
While schools were shut down 
during COVID-19,* did students 
still receive daily meal?
Yes No
# % of respondents # % of respondents
Maputo (n=149) 5 2.1% 144 60.7%
Nampula (n=88) 25 10.5% 63 26.6%
Total (n=237) 30 12.6% 207 87.3%
Source: Cook survey
* In the surveys, “during COVID-19” refers to the period from March 2020 to March 2021.
Exhibit 23. Percentage of Treatment Schools Where School Garden Was Cared For During COVID-19
While schools were shut down 
during COVID-19, did anyone 
take care of the school gardens?
Yes No
# % of respondents # % of respondents
Maputo (n=149) 38 16.0% 111 4.6%
Nampula (n=88) 10 4.2% 78 32.9%
Total (n=237) 48 20.3% 189 79.7%
Source: Cook survey
During school closures, students and families 
may have had access to take-home rations, but 
families were more likely to receive rations in 
the following academic year, when schools 
reopened with adjustments (Exhibit 24).15 Some 
parents mentioned receiving cornmeal and 
soybeans to take home, but in most cases it was 
once, at the beginning of the pandemic when 
schools first closed. Of the 122 head teachers who 
reported students were given take-home rations, 111 (90.9 percent) of them reported that they 
were distributed only once, early in the shutdown. 
15 Exhibit 24 is based on responses to the head teacher survey. In Nampula, World Vision, with agreement from USDA,
provided parents with take-home rations for all students. 
“The food was packaged in plastic bags and 
parents were called in small groups to receive the 
food according to the number of children they had 
at school, and the bags were already identified 
with the names of the beneficiaries.” 
- Key Official (Maputo) “After school closures the children did not 
receive this kit, but this year when the 
schools reopened, each child received a kit 
with five kilos of cornmeal” 
- Parent Focus Group (Nampula)
Abt Associates June 2022 ▌59
Exhibit 24. Percentage of Schools Where Families Received Take-Home Rations During COVID-19
While schools were shut down 
during COVID-19 did students 
receive take-home rations?
Yes No
# % of respondents # % of respondents
Maputo (n=178) 51 19.2% 127 47.7%
Nampula (n=88) 71 26.7% 17 6.4%
Total (n=266) 122 45.9% 144 54.1%
Source: Head teacher survey
Some key officials described that stopping school would 
mean negative impacts on enrollment, attendance and 
attentiveness. They linked receiving meals at school as a 
positive aspect that drew in students to be more engaged 
with school. 
11.3 COVID-19 and Health Services
Key officials and PFGs talked about sanitary practices implemented in schools to minimize the 
impact of the COVID-19 pandemic and allow 
schools to reopen. This included people wearing 
masks, promoting hand washing, and sanitizing 
spaces and providing hand sanitizer. One key 
official emphasized the importance of parents’ 
help implementing new sanitation guidelines 
and getting access to the proper materials.
Some interviews mentioned a lack of resources
or materials to effectively sanitize learning 
spaces, especially as time went on. A few key officials connected the access to sanitation 
materials with their relationships with partners, such as Terra dos Homens and UNIESCO. One 
key official described that after a while they ran out of sanitation material, and that while support 
for buying sanitation materials was helpful, there was never enough. 
Parents were asked about whether their children received deworming medication (albendazole 
or mebendazole) during school closures. A third of the PFG respondents answered “yes,” that 
their children received the medication. Most of the other two-thirds said that the children did not 
receive the medication. Three PFGs were split (with some saying “yes” and others saying “no”) 
or were unsure. 
“…[in many schools] that had no masks suddenly in 
a short time already had masks… helped to clean 
the school, to create conditions for hand washing, 
all this the parents and guardians contributed… 
they clean, they weed… all this to facilitate the 
literacy of their children. The participation of parents 
and guardians is admirable.” 
- Key Official (Maputo)
“The children already had that rhythm 
of going to school to have their 
meals, and I think that interrupted the 
desire they had to stay in school.” 
- Key Official (Maputo)
Abt Associates June 2022 ▌60
12. EVALUATION CONSTRAINTS AND LIMITATIONS
Interpretation of the evaluation findings is limited by several key factors. These include factors 
related to outcomes measure, internal validity of the impact estimates, timing of baseline, and 
precision of the impact estimates. 
Two of the primary outcome measures were based on survey responses that are subject to 
potential bias. Student attendance was measured with self-reported data from young children, 
and student attentiveness was measured using teacher recall in the teacher survey. In both cases, 
these measures could suffer from bias due to recall and measurement error. However, given that 
the survey included a substantial number of students and teachers, the effect of the expected 
errors could be minimized. For future evaluations, it is advisable to consider alternative measures 
of outcomes like student attentiveness to capture impacts at the student level more precisely.
The evaluation interprets the estimates from the quasi-experimental difference-in-differences 
analysis as the causal impact of the McGovern-Dole projects. A strength of the evaluation design 
is that it is based on similar data collection efforts at baseline and endline in the same schools. 
Despite that strength, causal interpretation requires the conditional parallel trends assumption. 
The series of fixed effects included in the difference-in-differences model control for unobserved 
factors at the school, school-by-grade, and province-by-time levels such that conditional parallel 
trends assumption plausible. However, those fixed effects do not capture all unobserved factors, 
including any other unobserved shock at endline that differentially affects treatment and 
comparison schools. The existence of such “unobservables” would threaten the conditional 
parallel trends assumption, and therefore the internal validity of the evaluation. 
The likely contamination of the comparison group complicates interpretation of the evaluation 
findings. The difference-in-differences design assumes the observed trend in the comparison 
schools can be used to project what the outcome would have been in treatment schools in the 
absence of treatment. This typically implies comparison schools do not experience any treatment. 
However, our assessment of implementation fidelity found that many comparison schools had 
some evidence of comparable feeding, literacy, and teacher training activities, although levels of 
implementation were determined to be low. For these schools, the implied comparison is not of 
McGovern-Dole projects to no intervention, but of McGovern-Dole projects to other programs. 
This could either attenuate or exacerbate the treatment effect and complicates interpretation of 
the average effect; however, if the conditional parallel trends assumption holds, it does not 
threaten the causal interpretation of the findings. 
Because program activities began prior to baseline, an evaluation limitation is that the impact 
estimates do not capture the impact of all programming. The impact estimates only capture 
changes between baseline and endline. If the activities that occurred prior to baseline 
measurement were very impactful, this would have the overall effect of biasing our impact 
estimates downward. However, in most cases it seems unlikely that this downward bias would be 
of sufficient magnitude to change our inference. Future evaluations should consider improving 
alignment between the timing of baseline and implementation of program activities. 
Three elements of the methodology for the evaluation merit discussion because of their potential 
to reduce the precision of the impact estimates: 
• School closures and gaps in project delivery
• Sample size concerns for the impact analyses of the school feeding intervention
Abt Associates June 2022 ▌61
• Limitations when conducting multiple comparisons 
12.1 School Closures and Gaps in Project Delivery
In Maputo, PAI was contracted to implement the McGovern-Dole program from 2015 to 2020. 
Counterpart was then awarded the contract from 2020 to 2025. However, at the time of endline 
data collection in October 2021, Counterpart still was preparing to implement the program. 
School feeding and literacy interventions were not yet in place creating a situation in which the 
study team was evaluating a project that had phased out a year before. Schools themselves were 
also functioning at a reduced level. When reopening in March 2021, after a year of COVID-19-
related closures, many students were going to school only two or three times a week. 
The combination of COVID-19-related closures, partial school reopening, and the ending of 
PAI’s implementation in late 2020 could have posed even greater risks to sustainability, making 
it more difficult to capture implementation and effectiveness during endline data collection in 
October 2021. The fact that despite these disruptions the evaluation detects some positive 
impacts reflects positively on the sustainability of the McGovern-Dole program. 
Nampula schools and World Vision implementation was also affected by the COVID-19 
shutdowns. But since the World Vision contract started in 2016 and ended in September 2021 
through a no-cost extension, the implementer was still connected to the study schools when 
schooling resumed in March 2021. 
Across provinces, a prevailing issue with the evaluation is that timing of data collection in 2021 
proved problematic in detecting delivery of school feeding and other program elements. Even if 
program delivery such as the provision of school meals, supplemental role of school gardens, 
access to clean water, and use of literacy materials was well established up until school closures 
in March 2020, their presence diminished in Nampula and disappeared almost entirely in Maputo 
by endline. Because of this, impacts may be difficult to detect, especially when capturing 
outcomes of interest like access to meals and student short-term hunger. 
12.2 Sample Size Added Challenges to Detecting Impacts of School Feeding Only 
Because it was only tested in Maputo, the school feeding only treatment arm had a smaller 
quantitative sample than the literacy and school feeding arm or the comparison group. This 
smaller sample implies impacts had to be larger in that treatment arm than in the literacy and 
school feeding arm to achieve statistical significance.
12.3 Multiple Comparisons 
In this evaluation, we conducted a large number of tests by estimating impacts on several 
outcomes and subgroups. Studies that employ multiple treatment arms and either explore 
treatment effect heterogeneity or use multiple estimators of the treatment effect may choose to 
adjust for these multiple comparisons, as some significance across outcomes is likely to arise by 
chance. We considered multiple comparison adjustments for the primary impact estimates on 
student reading performance, attendance, attention, and hunger. We ultimately decided that for 
these analyses, multiple comparison adjustments were not essential because the primary 
outcomes represent different domains (i.e., academic performance, behavior, and attendance). 
Many of the findings are from exploratory analyses that are meant to help provide context for 
understanding the main effects. 
Abt Associates June 2022 ▌62
13. CONSIDERATIONS WHEN CONDUCTING EVALUATIONS 
DURING PANDEMICS AND NATURAL DISASTERS
This evaluation of the McGovern-Dole projects in Mozambique accounts for project
implementation across a five-year span of time, beginning in 2015 and extending through 2021. 
At the time that the baseline evaluation was designed in 2016, there was no way of predicting the 
extent to which a global pandemic would cause widespread school closures and extensive gaps in 
delivery. When contracted to conduct the endline evaluation in 2021, the world was over a year 
into the COVID-19 pandemic and Mozambique schools were reopening to students for the first 
time in a year. Given these realities, Abt’s study team members considered several enhancements 
and modifications to the evaluation approach. These approaches were developed in response to 
COVID-19 but could be applicable to evaluations conducted during a variety of health- and 
environment-related disasters.
Due to travel restrictions, Abt Associates’ United States-based staff were unable to travel to 
Mozambique for meetings and presentations typically held in person. With the direct connection 
to the focus communities severed, we understood the heightened need for a trusted and expert in￾country local consultant and the importance of quicky establishing an in-person and virtual 
hybrid approach. The role of the consultant was expanded as necessary throughout the contract 
to carry out tasks that would have otherwise been handled by the Abt study team. During the 
endline data collection process, we benefitted from the participation of the in-country consultant 
in establishing and maintaining communication with implementing partners, providing guidance 
and relaying supports to the Mozambique-located data collection during enumerator training and 
direct data collection, reviewing and participating in data analysis and interpretation, and 
attending dissemination events in Maputo and Nampula provinces with stakeholders at the 
conclusion of the study. 
For this impact evaluation, the contracted data collection firm traveled to schools and conducted 
in-person student assessments, school staff and student interviews, as well as PFGs. At this point, 
COVID-19 transmission rates were high and vaccines were not readily available. Person-to￾person proximity and contact remained dangerous to both the data collection team and study 
participants. 
The period of time between baseline and endline data collection was filled with many challenges. 
When endline data were collected in 2021, Mozambican schools were working to reopen and 
were not yet operating at full capacity. In retrospect, a midline evaluation in 2019 with robust 
data collection mirroring the baseline approach could have captured the state of implementation 
and any changes in student and school outcomes at that point. These findings, collected prior to 
pandemic-related school closures, would have informed the endline data collection and analysis 
and helped interpret intervention effectives. Regardless of health crises and natural disasters, 
interim and systematic data collection efforts guard against challenges at endline and provide 
additional evidence of intervention impacts. 
Abt Associates June 2022 ▌63
14. CONSIDERATIONS FOR MCGOVERN-DOLE PROGRAM 
SUSTAINABILITY BY STAKEHOLDERS
Parents, key officials, head teachers, as well as cooks were all asked about their thoughts on 
program sustainability as well as on recommendations they have for consideration. The study 
team documented a preliminary set of themes based on stakeholder responses. 
14.1 Considerations for School Feeding
Meal and nutrition quality: Despite parents largely having positive perceptions of the school 
feeding intervention, parents from half of the PFG discussions raised meal quality concerns and 
recommended that the school meals should provide a greater variety. Some respondents noted 
that sometimes the children did not like the food and wasted it. 
More than one PFG participant suggested 
alternating what was served “porridge 
today, rice tomorrow.” Another suggested 
that a variation in food, like “rice, xima 
[cornmeal porridge], beans and other 
foods” would provide more nutritional 
benefit. From a broader perspective, a key 
official interviewed felt there was a 
contradiction between the nutrition training 
and educational component of the project
compared to the nutrition of the school 
meals being provided. 
Food access: A few PFG discussions made specific recommendations related to the supply of 
food. A few discussants mentioned a need for improved storage and warehouse conditions, 
noting that the soy or flour they received was spoiled. Another PFG participant mentioned it 
would be helpful to have a protocol for what should be purchased with the revenue generated
from the sale of products. A few respondents mentioned needing stronger connections or 
associations with farms and other agricultural stakeholders. A respondent described that a school 
farm is helpful, but there are challenges if it is only seasonal and cannot be a stable supplier for 
school meals. 
School gardens, water, and other agricultural 
inputs: Despite reported challenges with school 
gardens, key officials wanted to continue them and 
had several recommendations. Needing help from 
partners was acknowledged, especially when it 
comes to access to inputs required for making 
gardens successful. Respondents noted that many 
inputs were necessary along the production chain, 
including seeds, water, fertilizer, garden tools, 
machinery (such as motor pumps), infrastructure 
(such as fences to keep out animals), and human 
resources for planning and labor. 
“School vegetable gardens should exist in every 
school. The goal is not only to produce for the 
children to eat, but also to teach the children to 
know how to do it at home and for the future. 
We would need partners to support the School 
Vegetable Garden Program, because we 
already have the land, some schools have 
water, we would need reinforcements in inputs 
and other materials (e.g., electric water pump).”
- Key Official (Maputo)
“They learned a lot about eating habits, how they should 
eat, how to improve their diet, [and] the variety of foods. 
But there was a contradiction, because they prepared soy 
every day, and this aspect is linked to the curriculum, so 
ADPP provided us with some nutrition books/manuals
that were used as support in the teaching of [on] food 
groups… the child learned that he should vary the foods 
to improve the diet, but at school they ate soy every day.”
- Key Official (Maputo)
Abt Associates June 2022 ▌64
Many respondents from PFG discussions and key official interviews cited issues with access to 
water for school gardens. Several PFG discussions recommended the installation of irrigation 
systems, or a borehole, to help extend the growing season. A few also noted that the lack of 
water at schools also prevented the preparation of school meals. 
Several PFG discussants recommended needing a greater 
variety of seeds to provide a greater variety of food for 
school meals. One pointed out that access to the seeds must 
coincide with seasonal changes and planting timelines. 
Role of the cooks: Several respondents mentioned 
formalizing the role of the cooks. Respondents 
recommended doing this through providing stronger 
incentives (including paying/hiring cooks) and formalizing the management of the cooks by 
providing a clear schedule or a role description with clear expectations. One key official noted 
that using volunteers was not always reliable, and “calls into question the expected results and 
more than that, compliance.” Some felt that paying the cooks would strengthen their 
commitment to the helping at the school.
14.2 Considerations for Literacy
Training: Many respondents, including parents and 
key officials, thought that more training was 
necessary to maintain the knowledge and sustain the 
project. A few elaborated that there could be 
stronger follow-up monitoring of the training, 
potentially to identify needs for refresher trainings. 
Respondents recognized the value in training to 
provide teachers with methodologies to improve their teaching. 
School materials: Respondents linked materials and training to greater success and 
sustainability of the project. They noted that teachers needed training, but also the material 
resources (i.e., books) to implement what they are learning. Respondents suggested that materials 
could be recycled and used for future students.
Bilingual teaching: As described earlier, several respondents recommended expanding the 
bilingual education. One key official detailed that he saw the value in bilingual, local language 
education, citing that he didn’t have that experience in school, but he thought he would be much 
better off now if he had had that experience 
at a young age. Further, he noted observing a 
difference in performance between schools 
where education was monolingual versus 
bilingual. 
Literacy teacher/coach incentives: Key 
official respondents felt that teachers should 
be provided stronger incentives and benefits 
for teaching and for attending trainings. 
“We have had difficulties in getting real 
seeds, sometimes they are brought at 
inappropriate time, to sow while others 
have already sown … [we’d like the 
donors] to allocate us early seeds.” 
- Parent Focus Group (Nampula)
“The following year they should be trained 
again… [Partners] must continue to give these 
trainings because there are those with weak 
memories and they don’t take long to forget.” 
- Parent Focus Group (Nampula)
“[Teachers] need water, juice… the project always 
gives them water, some allowance so that the teachers 
can travel from the countryside to the session location 
without having to sacrifice their pockets, or take out 
debts to a stall to get fuel for their motorcycle. These 
resources have to exist, and I would like the project to 
continue to provide them as much as possible.” 
- Key Official (Nampula)
Abt Associates June 2022 ▌65
14.3 Considerations on Sustainability
At least 12 respondents specifically mentioned challenges with continuity of the project, 
suggesting some frustration with the end of the project, and wishing that there was a better plan 
for sustainability.
Local buy-in, support, and involvement: A 
prominent theme across parent and key official 
respondents was the importance of bolstering 
the local community for the sustainability and 
success of the project. Key officials especially 
felt it was necessary to mobilize the community, 
potentially by including them more in the 
planning, or by making the goals of the project
clearer from the start. A key official noted that 
training for, and support from, the local 
government could be especially influential 
given that they can promote and describe the 
purpose of the project to others.
A few respondents felt that incorporating local feedback into the design of the project was 
important. For example, one key official noted that
implementation seemed to be “based on studies carried 
out in a generalized and non-specific way.” They gave 
the example that the enriched porridge served pleased a 
group of children and parents but not other groups. 
Training and capacity building: In relation to the 
literacy intervention, several key officials reported that 
training should continue in order to promote sustainability. In some interviews, respondents also 
felt that training for cooks, and people helping in the gardens or on school farms should also 
continue to get training. Some respondents told us that, beyond training, projects and 
stakeholders should provide “testimonies” documenting successes and the value of the project
and how to carry it forward. A few specifically recommended incorporating stronger capacity 
building plans to do so. 
“The small leaders have the responsibility to monitor 
the activities and make sure they are successful, so I 
think there should be training for them to explain 
what this is, [and the] goal to achieve... In everything 
we do, we have to understand the "why" and "what 
for" so that we can work towards a common goal. 
The idea is very good, but we have to involve 
everyone, and divide up the responsibilities in order 
to have the effect, the impact of our project.”
- Key Official (Maputo)
“It's important that somehow more research 
is done at the community level to reach a 
consensus… I cannot dictate an individual's 
diet based on the experience of another.” 
- Key Official (Maputo)
Abt Associates June 2022 ▌66
15. RECOMMENDATIONS FOR FUTURE MCGOVERN-DOLE 
PROGRAMMING
15.1 Expand Awareness of McGovern-Dole Program Theory of Change and Impact 
Evaluation Criteria
The McGovern-Dole program posits that improved literacy instruction paired with regular school 
meals will lead to both improved literacy and improved health and dietary practices of students 
via various program impact pathways. Evaluations of McGovern-Dole projects in Mozambique 
have shown that while provision of school meals and associated reports of almost zero hunger in 
beneficiary schools as well as availability of WASH (Water, Sanitation, and Hygiene) 
infrastructures and reading materials are the most visible components of the program, the impact 
of these inputs on the students’ literacy skills has not been that substantial. In fact, from 
interviews with key stakeholders, in particular parents and other community members, the school 
feeding component of the program is the most referenced program component, but not its impact 
on students’ learning, or more specifically on literacy development. This seems to suggest that, 
while implementers and education authorities are aware of McGovern-Dole program theory of 
change and impact evaluation criteria, most of the other groups of stakeholders are not, including 
school managers, teachers, parents, and other members of the beneficiary communities. 
Recommendation 1: The study team recommends involving stakeholders in understanding and 
engaging in the McGovern-Dole program theory of change and identified outcomes. By using 
non-technical registers tailored to different audiences, USDA and McGovern-Dole leaders can 
enhance stakeholder capacity to assess the progress and impact of the projects. Parents and other 
community members, particularly through school councils, have been playing a key role in 
monitoring school management, and teacher behavior and performance. Knowing the 
overarching goal of the McGovern-Dole program, the criteria for impact evaluation, and the 
interconnection among the different components of the program would help them better monitor 
progress toward that goal. The same applies to school managers, education authorities, and local 
government authorities. 
15.2 Improving Teachers’ Capacity to Support Student Development of Early Reading 
Skills
Despite the positive impact of the McGovern-Dole programming on literacy development,
EGRA scores also show that most students assessed do not have the foundational skills needed to 
become fluent readers and use their reading ability to learn, even by fourth grade. Studies show 
that there is correlation between lower-order (letter knowledge, vocabulary, familiar words 
reading, oral listening comprehension) and high-order reading skills (oral reading fluency and 
reading comprehension) in such a way that the former predict the latter. Moreover, there is also a 
positive relationship between oral reading fluency and reading comprehension: the speed and 
accuracy at which a person reads correlates closely with reading comprehension (Daane et al., 
2005; Abadzi, 2011). Therefore, the overall weak performance in reading comprehension 
uncovered in this study can be explained by the poor results on ‘letter knowledge,’ ‘familiar 
words reading,’ ‘oral listening comprehension,’ and ‘oral reading fluency.’ Moreover, although 
the study found that teachers in treatment schools in Nampula and Maputo received training in 
literacy instruction that was delivered with adequate fidelity, it is important to consider that only 
about 60 percent of the teachers surveyed reported that, during pre-service training, they received 
specific training on how to teach reading and writing to students in the primary grades. This 
suggests that teachers need continuous training in literacy instruction and highlights the study 
Abt Associates June 2022 ▌67
team’s recommendation to strengthen the connection between teacher training programs and 
local schools. 
Recommendation 2: The study team recommends that teacher training increase focus on early￾grade literacy instruction, allowing teachers to better understand the casual relationship between 
lower- and high-order reading skills and enhance their capacity to help children develop these 
two levels of reading growth. We are not suggesting transforming language arts classes into 
sessions for preparing students to succeed in EGRA tests (i.e., “teaching to the test”), but to 
include in such classes opportunities for them to reflect and practice pre-reading and actual 
reading skills (reading fluency and reading comprehension). Teachers and students must be 
aware that the goal of reading acquisition is to be able to construct meaning out of what one 
reads and use the reading ability to learn. 
15.3 Language of Early-Grade Literacy Development 
Study results indicate that between 13 and 31 percent of students were instructed in a language 
different from the one they speak at home (Exhibit A9), which may have a bearing on the overall 
weak EGRA results obtained. In fact, studies have shown that early literacy development is best 
achieved in the child’s home language (e.g., Bamgbose, 2000; Fafunwa, 1990). When a child 
learns how to read in a language different from the one spoken at home, that child is faced with 
two levels of challenges – to learn the basics of the new language at the same time that he or she 
learns how to read. The co-occurring challenges diminish his or her chances to develop reading 
skills (e.g., Bamgbose, 2000). In interviews with key stakeholders, in particular with parents and 
education authorities, respondents frequently called for the introduction or expansion of bilingual 
education inclusive of local African languages. 
Recommendation 3: The study team recommends support for the scaling up of dual language 
early literacy instruction inclusive of languages familiar to the child (their home languages). 
Mozambique is already implementing mother tongue-based bilingual education in selected 
schools throughout the country, using 17 African languages for early instruction. The 
McGovern-Dole program may help local education authorities to expand the bilingual program 
to more schools or to cover more classes in schools where it is already in place. 
15.4 Increase Gender Sensitivity and Improve Instruction Focusing on Girls 
About half of the students (47 percent) in the school feeding only intervention who benefited 
from McGovern-Dole were female and about half of the students interviewed in this study were 
female. Enrollment and data collection efforts confirm that there is gender parity in access to 
education in the schools targeted by McGovern-Dole programing. However, literacy outcomes 
show that boys performed better than girls. Average reading comprehension scores improved for 
both early-grade and mid-grade boys in schools implementing both literacy and school feeding 
interventions. Impacts on reading for girls is smaller and not statistically significant, although the 
difference in impacts between boys and girls is also not statistically significant. This reading 
achievement gap is consistent with the general pattern in Mozambique, particularly in rural 
settings (MINEDH, 2014, 2017; Bassi et al., 2019). Studies suggest that this pattern has to do, at 
least in part, with rural communities devoting less attention to the education of girls compared to 
boys, lack of gender sensitivity in materials and teaching practices, which tend to privilege 
males, and scarcity of female role models among teaching and school management staff (Bagnol 
et al., 2015). Less than a quarter of head teachers involved in this evaluation are female.
McGovern-Dole programing is already addressing some of these issues, especially through 
Abt Associates June 2022 ▌68
community involvement in education and gender sensitive teacher training interventions, but 
those can still be improved and expanded. 
Recommendation 4: The study team recommends improving the environment to allow girls to 
have equal opportunities for learning in schools, homes, and larger communities. We suggest 
meeting this objective by sensitizing communities about the importance of girls’ education, 
training teachers on gender-sensitive classroom interaction and reading practices, and 
incorporating a gender-sensitivity approach in the production of teaching and learning materials 
and in the selection and hiring practices for school leadership positions. 
15.5 Ensure the Sustainability of McGovern-Dole Programming
When the evaluation took place in Maputo province, the McGovern-Dole program supported 
ADPP/PAI Food for Knowledge project had been phased out a year prior. It is important to note 
that none of the activities implemented by ADPP/PAI remained in place, including activities like 
teacher training in literacy instruction that could have been implemented by education authorities 
with certain modifications to mitigate the spread of COVID-19. Children who had been 
accustomed to receiving at least one meal at school suddenly found themselves without that kind 
of motivation to attend school. School fields and gardens lay fallow. These conditions created a 
situation in which outcomes gained over five years disappeared when the Food for Knowledge 
project ended. The situation in Nampula also suggests that at least the arrangements made for 
continuing the project initiatives, in particular school feeding, may not be sustainable. Farmer 
groups are challenged by relying on rain to water their crops and on very traditional agriculture 
technics. Consequently, they do not have the capacity to supply enough food to schools and 
throughout the school calendar. 
Recommendation 5: The study team encourages USDA/FAS, implementing partners, ministry 
officials and others to put in place strong exit strategies to allow sustainability of the USDA/FAS 
McGovern-Dole program after the end of designated projects, including through increased 
involvement of local education authorities and establishment of linkages with similar 
government programs. For example, Mozambique has the PRONAE national school feeding 
program in place, and it could be expanded to schools immediately after the McGovern-Dole 
projects end. The same can be applied to other project interventions such as literacy and WASH. 
Abt Associates June 2022 ▌69
16. REFERENCES
Abadzi, H. (2011) Reading Fluency Measurements in EFA Partner Countries: Outcomes and 
Improvement Prospects, EFA, FTI Secretariat Update September 2011. 
https://documents.worldbank.org.pdf
Bagnol, B., De Sousa, L., Fátima, F., and Cabral, Z. (2015). As barreiras à educação da 
rapariga no ensino primário na Zambézia. Maputo: IBIS. 
Bamgbose, A. (2000). Language and Exclusion: The Consequences of Language Policies in 
Africa. Hamburg: Lit Verlag Munster. 
Bassi, M., Medina, O., and Nhampossa, L. (2019). Education Service Delivery in Mozambique: 
A Second Round of the Service Delivery Indicators Survey. 
http://documents1.worldbank.org/curated/en/811891562864504006/pdf/Education-Service￾Delivery-in-Mozambique-A-Second-Round-of-the-Service-Delivery-Indicators-Survey.pdf. 
Chimbutane, F. (2011). Rethinking Bilingual Education in Postcolonial Contexts. Multilingual 
Matters.
Chimbutane, F. (2015). Bilingual education: Enabling classroom interaction and bridging the gap 
between schools and rural communities in Mozambique. International Journal of 
Educational Development in Africa, 2(1), 101-120. https://doi.org/10.25159/2312-3540/19
Daane, M.C., Campbell, J.R., Grigg, W.S., Goodman, M. J., and Oranje, A. (2005) Fourth-grade 
Students Reading Aloud: NAEP 2002 Special Study of Oral Reading (NCES 2006-469). U.S. 
Department of Education, Institute of Education Sciences, National Center for Education 
Statistics. Washington, DC: Government Printing Office. 
https://nces.ed.gov/nationsreportcard/pdf/studies/2006469.pdf. 
Diogo, J. (2018). Mid-Term Evaluation of the Mozambique McGovern Dole International Food 
for Education and Child Nutrition Educating Children Together - Phase II (ECT2). USDA.
Doctors, S. (2019). Mozambique McGovern-Dole Food for Education Project Phase 2: Midterm 
Evaluation Report. USDA. https://pdf.usaid.gov/pdf_docs/PA00XF2T.pdf 
Doctors, S. (2021). Mozambique McGovern-Dole Food for Education Project: Final Evaluation. 
USDA. 
Fafunwa, A.B. (1990). Using national languages in education: A challenge to African educators. 
In UNESCO/UNICEF (eds) African Thoughts on the Prospects of Education for All (pp. 97–
110). Dakar: UNESCO Regional Office for Education in Africa. 
Guest, G., Namey, E.M., and Mitchell, M.L. (2013). Collecting Qualitative Data: A Field
 Manual for Applied Research. SAGE Publications.
Global Partnership. (2021, September 10). Mozambique. 
https://www.globalpartnership.org/where-we-work/mozambique
The International Centre for Evaluation and Development. (2019). World Vision Mozambique: 
Final Evaluation Report. 
Lechner, M. (2011). The estimation of causal effects by difference-in-difference methods. Foundations 
and Trends in Econometrics, 4, 165–224. 
Abt Associates June 2022 ▌70
Lee, J., D’sa, N., and Zuilkowski, S. S. (2020). Integrating new knowledge into everyday 
practices: Teacher pedagogy in early grade literacy in rural Mozambique. Literacy, 54(3), 
111-122. https://doi.org/10.1111/lit.12221
Milhorance de Castro, C. (2018). PAA Africa’s Contributions to the Consolidation of PRONAE 
in Mozambique (Working Paper 170). Food and Agriculture Organization of the United 
Nations. 
https://agritrop.cirad.fr/597856/1/WP170_PAA_Africa_s_contributions_to_the_consolidatio
n_of_PRONAE_in_Mozambique%281%29.pdf
MINEDH. (2017). Avaliação Nacional da 3a Classe. Maputo: MINEDH.
MINEDH. (2014). Avaliação Nacional da 3a Classe. Maputo: MINEDH.
Ravesloot, B., Downen, J., Tuyiringire, J., Muteteri, D., Mujawase, J., Kayiraba, C., Mueller, M., 
Banwart, L., and Kaelin, J. (2021). WFP’s USDA McGovern-Dole International Food for 
Education and Child Nutrition Program’s Support in Rwanda 2016-2021: Endline 
Decentralized Evaluation. USDA. 
https://dec.usaid.gov/dec/GetDoc.axd?ctID=ODVhZjk4NWQtM2YyMi00YjRmLTkxNjktZT
cxMjM2NDBmY2Uy&rID=NTk0MDM0&pID=NTYw&attchmnt=True&uSesDM=False&r
Idx=MzEzNjA2&rCFU= 
RTI International. (2018). McGovern-Dole Mozambique Impact Evaluation and Literacy 
Assessment: Baseline Report. USDA. 
Safarha, E., Hoffman, V., Zaas, D., Gulemetova, M., Taube, A., Li, J., and Gulati, K. (2020). 
USDA McGovern-Dole International Food for Education and Child Nutrition Program 
Phase III in Mali Endline Evaluation Report. USDA. 
Sitao, V.J.R. (2018). An assessment of school feeding programme - Pilot phase and its 
relationship with enrolment attendance retention and the local agricultural production in 
Nampula province in Mozambique [Doctoral thesis, University of Pretoria]. UPSpace. 
http://hdl.handle.net/2263/67908 
Tirivayi, N. (2019). Final Evaluation of the School Meals Programme in Malawi with financial 
support from United States Department of Agriculture (USDA)'s McGovern￾Dole Food for Education (Vol. II). USDA. https://pdf.usaid.gov/pdf_docs/PA00XX3F.pdf
Turney, A., Noyes, D., Magaia, F., Dhliwayo, T.M., Machkasov, Y., Zacarias, E.C.D., and Hua, 
H. (2020). Effectiveness Evaluation: Early Grade Reading Assessment & Supplementary 
Tools. USAID. https://pdf.usaid.gov/pdf_docs/PA00X21D.pdf 
UNICEF. (2020). The Impacts of COVID-19 on Children in Mozambique. COVID-19 Policy 
Note. https://www.unicef.org/esa/media/6536/file/The%20Impacts%20of%20COVID￾19%20on%20Children%20in%20Mozambique.pdf
USAID. (2021a, December 1). Education – Mozambique. 
https://www.usaid.gov/mozambique/education
USAID. (2021b, May). Mozambique – Nutrition Profile. 
https://www.usaid.gov/sites/default/files/documents/tagged_Mozambique-Nutrition￾Profile.pdf
USAID. (2014, June). Mozambique – Nutrition Profile. 
https://www.usaid.gov/sites/default/files/documents/1864/USAID-Mozambique-Profile.pdf
Abt Associates June 2022 ▌71
Vyamana, V., Ntunda, P., and Dulla, R. (2017). End-Line Evaluation report for FEE II Project. 
USDA. 
https://dec.usaid.gov/dec/GetDoc.axd?ctID=ODVhZjk4NWQtM2YyMi00YjRmLTkxNjktZT
cxMjM2NDBmY2Uy&rID=NTgwMTY4&pID=NTYw&attchmnt=True&uSesDM=False&r
Idx=MzA0NjUy&rCFU= 
World Bank Group. (2018). Findings of the Mozambique Water Supply, Sanitation, and Hygiene 
Poverty Diagnostic. Washington, DC. 
World Health Organization. (2008). Indicators for assessing infant and young child feeding 
practices: conclusions of a consensus meeting held 6–8 November 2007 in Washington D.C., 
USA.
World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF). (2021).
Progress on household drinking water, sanitation and hygiene 2000-2020: Five years into the 
SDGs. Geneva.
World Vision Mozambique. (2016). End of Program Evaluation Report: Educating Children 
Together Project I. USDA. https://pdf.usaid.gov/pdf_docs/PA00XKT4.pdf
Abt Associates June 2022 ▌72
APPENDIX A: SUPPORTING EXHIBITS
MCGOVERN_DOLE RESULTS FRAMEWORKS
Exhibit A1. McGovern-Dole Results Framework (SO1)
Abt Associates June 2022 ▌73
Exhibit A2. McGovern-Dole Results Framework 2
Abt Associates June 2022 ▌74
Exhibit A3. Program logic model with representation of project implementation adequacy and impacts
Abt Associates June 2022 ▌75
FIDELITY INDEX
Exhibit A4. School Feeding Fidelity Index
Data 
Source Survey Item Per Respondent 
Rescoring Across Respondent Scoring Decision
Head 
Teacher
Survey 
Does the school have a school 
feeding program? Yes = 1; No = 0 1 = 100% responded yes 
How often do students receive the 
school meals? 
(1) Daily, (2)A few times
a week = 1; else=0
1 = 100% reported at least a few times a 
week 
Does the school have a garden? Yes = 1; No = 0 1 = 100% responded yes 
How often is the garden used to 
supply food for school meals? 
(1) Every day, (2) 3-4
times per week, (3) 1-2
times per week = 1; else
=0
1 = 100% reported at least 1-2 times a 
week 
Cook 
Survey
Does the school have a kitchen 
and storeroom for the school 
feeding program? 
Yes = 1; No = 0 1 = 100% responded yes 
Are cooking pots and cooking 
utensils (such as stirrers or 
scoops) available? 
Yes = 1; No = 0 1 = 100% responded yes 
Do you supplement the corn-soy 
blend with other foods? 
Yes = 1; No = 0 1 = 100% responded yes 
Does the school have bowls and 
spoons for students? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there soap available to clean 
the bowls, spoons, pots, and 
cooking utensils? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there a supply of clean water 
(from a borehole or rainwater 
tank) available at the school? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there a place for children to 
wash their hands before eating? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there soap available for children 
to wash their hands before 
eating? 
Yes = 1; No = 0 1 = 100% responded yes 
Did you receive any training for 
being a school cook? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there a place for cooks to wash 
their hands? 
Yes = 1; No = 0 1 = 100% responded yes 
Is there soap for cooks to wash 
their hands? 
Yes = 1; No = 0 1 = 100% responded yes 
School Feeding Fidelity Score Range per school 0-15
Abt Associates June 2022 ▌76
Exhibit A5. Literacy Fidelity Index
Data Source Survey Item Per Respondent 
Rescoring Across Respondent Scoring Decision
Head Teacher 
Survey 
Do you have a school literacy 
program? 
Yes = 1; No = 0 1 = 100% responded yes 
Does the school have a library/book 
bank? 
Yes = 1; No = 0 1 = 100% responded yes 
Does the school have a literacy 
afterschool program? 
Yes = 1; No = 0 1 = 100% responded yes 
Are there other types of afterschool 
learning clubs? 
Yes = 1; No = 0 1 = 100% responded yes 
Teacher 
Survey 
Do you have a Teacher Guide for 
reading instruction?
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Do you use the Teacher Guide for 
reading instruction? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Do each of your students have 
exercise books that can be used 
during reading instruction? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Do your students use the exercise 
books during reading instruction? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Do you have access to school 
supplies? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Do you use any school supplies in 
your classroom during reading 
instruction? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1)
Is there an afterschool literacy club 
in this school?
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1)
Do your students attend an 
afterschool literacy club at this 
school? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Literacy Fidelity Score Range per school 0-12
Abt Associates June 2022 ▌77
Exhibit A6. Head Teacher and Teacher Training Fidelity Index
Data Source Survey Item Per Respondent 
Rescoring Across Respondent Scoring Decision
Head Teacher
Survey
Did you attend the teacher training 
held this academic year? Yes = 1; No = 0 1 = 100% responded yes (1); 0= less than 
100% said yes (1) 
Did teachers at the school attend the 
literacy training this year? Yes = 1; No = 0 1 = 100% responded yes (1); 0= less than 
100% said yes (1) 
Do reading coaches supervise 
teachers at the school? Yes = 1; No = 0 1 = 100% responded yes (1); 0= less than 
100% said yes (1) 
Have you received training on the 
fundamentals of reading instruction? 
Yes = 1; No = 0 1 = 100% responded yes (1); 0= less than 
100% said yes (1) 
Have you received training on school 
management? Yes = 1; No = 0 1 = 100% responded yes (1); 0= less than 
100% said yes (1) 
Teacher 
Survey
During your pre-service training, did 
you receive any specific training on 
how to teach reading and writing to 
students in the primary grades? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
This year, have you attended any in￾service training on how to teach 
reading in the primary grades? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Have you attended the WV/PAI 
training? Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Have you received training on the 
fundamentals of reading instruction? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Does the Head Teacher or Assistant 
Head Teacher check your lesson 
plans? 
Yes = 1; No = 0 1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Since the beginning of the school 
year, how frequently does the 
Assistant Head Teacher or Head 
Teacher observe your teaching? 
(1) Quarterly, (2) Monthly, 
(3) weekly = 1; else=0 
1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Since the beginning of the school 
year, how frequently does the 
reading coach visit you? 
(1) Quarterly, (2) Monthly, 
(3) weekly = 1; else=0 
1 = 50% or more responded yes (1); 0= less 
than 50% said yes (1) 
Head teacher and teacher training Fidelity Score Range per school 0-12
 
Abt Associates June 2022 ▌78
SAMPLE SIZE BY OUTCOME INDICATOR
Exhibit A7. Sample Size, Data Source, and Level of Analysis by Outcome Indicator
Outcome Data 
Source
Level of 
Analysis
Literacy and 
School 
Feeding
School 
Feeding Only Comparison
Main Impacts
Student reading performance1 EGRA 
assessment Student 5,504 2,576 5,474
Student attendance Student 
survey 
Student 5,491 2,574 5,473
Student attentiveness Teacher 
survey 
Teacher 634 311 631
Secondary Impacts
Teacher-reported receipt of 
training on the fundamentals 
of reading instruction
Teacher 
survey 
Teacher
634 311 631
Teacher-reported absence of 
at least one day last week
Teacher 
survey 
Teacher 634 311 632
Teacher has access to school 
supplies
Teacher 
survey 
Teacher 631 310 631
Student reports feeling hungry 
now
Student 
survey 
Student 5,504 2,576 5,476
Student achieved minimum 
meal frequency
Student 
survey 
Student 5,405 2,543 5,421
Student achieved minimum 
dietary diversity
Student 
survey 
Student 5,502 2,576 5,476
Student achieved minimum 
acceptable diet
Student 
survey 
Student 5,427 2,545 5,429
Student’s nutrition knowledge 
score 
Student 
survey 
Student 5,503 2,576 5,471
Student’s hygiene and 
sanitation knowledge score
Student 
survey 
Student 5,503 2,576 5,471
Student reports absence last 
week as a result of illness
Student 
survey 
Student 5,491 2,574 5,473
Head teacher reports school 
has improved latrines
Head teacher 
survey 
Head teacher 346 174 352
Head teacher reports school 
has tippy taps or hand￾washing stations
Head teacher 
survey 
Head teacher
346 174 352
Head teacher reports all tippy 
taps or hand-washing stations 
have ash or soap
Head teacher 
survey 
Head teacher
346 174 352
1 Student reading performance includes average reading comprehension scores and percent zero reading comprehension scores
Abt Associates June 2022 ▌79
SAMPLE POPULATION CHARACTERISTICS
Exhibit A8. Student Responses on Household Demographic Characteristics
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Early-Grade Students
Student age (years) 8.5 9.2 0.7 8.1 8.6 0.5 8.7 9.4 0.7
Female students (%) 50.3 49.7 -0.6 50.0 46.9 -3.1 49.5 50.3 0.8
Household source of water
River, stream, or lake (%) 12.7 10.3 -2.4 15.4 9.6 -5.7 20.0 14.6 -5.4
Well or borehole (%) 29.8 21.1 -8.7 40.3 23.5 -16.8 40.9 20.8 -20.2
Communal tap (%) 21.5 27.6 6.1 24.9 36.8 11.9 23.4 39.3 15.9
Water pipe / tap in home (%) 31.6 39.0 7.4 17.4 26.2 8.8 13.9 24.0 10.1
Water truck or tank (%) 2.1 0.1 -2.0 1.9 2.4 0.5 1.1 0.3 -0.8
Other (%) 2.2 1.6 -0.6 0.2 0.9 0.7 0.7 1.0 0.3
Household asset index
Radio (%) 57.9 50.0 -7.9 61.2 53.0 -8.2 58.9 45.3 -13.6
Mobile phone (%) 73.3 74.8 1.5 86.8 95.5 8.7 66.7 72.3 5.6
Television (%) 51.7 51.6 -0.1 56.1 46.8 -9.2 34.4 30.5 -3.9
Computer (%) 20.6 18.0 -2.5 24.7 15.5 -9.2 12.5 10.2 -2.3
Refrigerator (%) 41.6 36.7 -4.9 35.8 26.2 -9.5 21.9 20.9 -1.0
Bicycle (%) 40.8 30.5 -10.3 44.4 37.3 -7.1 48.9 36.9 -12.0
Motorbike (%) 18.1 17.1 -1.0 16.2 14.5 -1.7 19.1 16.9 -2.2
Car/truck (%) 24.5 15.2 -9.3 33.1 22.6 -10.5 14.9 11.2 -3.7
Electricity (%) 57.9 56.4 -1.5 47.4 45.3 -2.1 35.6 32.2 -3.4
Asset index (9 items) 3.8 3.5 -0.3 4.0 3.6 -0.4 3.1 2.7 -0.4
Mid-Grade Students
Student age (years) 10.9 10.9 -0.1 10.4 10.8 0.4 11.3 11.4 0.1
Female students (%) 51.2 50.4 -0.7 50.2 52.2 2.0 50.3 50.9 0.6
Household source of water
River, stream, or lake (%) 12.7 9.8 -2.9 19.4 13.4 -6.0 17.5 14.1 -3.4
Well or borehole (%) 28.6 15.5 -13.2 35.6 24.7 -10.9 43.4 22.3 -21.0
Communal tap (%) 20.8 24.4 3.6 23.3 28.3 5.0 21.2 36.0 14.8
Water pipe / tap in home (%) 33.5 48.5 15.0 18.9 28.5 9.6 16.0 26.0 10.0
Water truck or tank (%) 2.5 0.6 -1.9 2.7 1.9 -0.8 0.7 0.7 0.1
Other (%) 1.8 1.2 -0.6 0.1 2.6 2.4 1.3 0.9 -0.4
Household asset index
Radio (%) 57.0 50.7 -6.2 57.0 60.7 3.7 59.7 50.0 -9.6
Mobile phone (%) 76.1 85.5 9.4 90.6 94.4 3.8 72.3 76.7 4.4
Television (%) 53.6 62.6 9.0 53.0 46.8 -6.2 32.9 36.9 4.1
Computer (%) 18.5 21.6 3.1 12.7 14.4 1.6 10.1 12.6 2.5
Refrigerator (%) 38.9 47.3 8.4 31.6 30.4 -1.2 21.3 27.5 6.2
Bicycle (%) 43.3 30.2 -13.1 50.7 43.0 -7.7 50.9 36.9 -14.0
Motorbike (%) 17.9 14.2 -3.7 14.0 12.7 -1.2 20.5 19.8 -0.7
Car/truck (%) 22.1 23.3 1.2 30.1 22.6 -7.5 14.5 12.6 -1.9
Electricity (%) 56.1 67.2 11.1 44.9 43.9 -1.0 30.5 37.8 7.3
Asset index (9 items) 3.8 4.0 0.2 3.8 3.7 -0.2 3.1 3.1 0.0
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌80
Exhibit A9. Language of Literacy Instruction and Difference with Language Spoken at Home
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Early-Grade Students
Teacher-reported language of instruction
Portuguese (%) 99.6 100.0 0.4 98.5 99.0 0.6 99.1 99.1 0.0
Rhonga (%) 10.6 3.4 -7.2 1.8 0.0 -1.8 1.4 1.4 0.0
Changana (%) 13.9 12.2 -1.8 58.6 40.4 -18.2 25.3 21.7 -3.7
Macua (%) 37.5 36.0 -1.5 0.0 0.0 0.0 50.5 45.1 -5.4
Other (%) 0.0 0.3 0.3 0.0 12.4 12.4 0.7 10.1 9.4
Language student speaks at home 
Portuguese (%) 38.0 52.5 14.5 21.8 40.0 18.2 12.1 22.4 10.4
Rhonga (%) 0.7 0.1 -0.6 1.2 0.5 -0.7 0.9 1.6 0.7
Changana (%) 17.7 7.4 -10.2 74.7 59.1 -15.6 31.8 20.8 -11.0
Macua (%) 43.3 38.6 -4.7 0.0 0.0 0.0 54.8 54.8 -0.0
Other (%) 0.4 1.4 0.9 2.3 0.3 -1.9 0.5 0.4 -0.1
Students whose language spoken 
at home differs from language of 
instruction (%)
18.5 14.5 -3.9 26.6 28.2 1.7 17.8 21.3 3.6
Mid-Grade Students
Teacher-reported language of instruction
Portuguese (%) 99.6 100.0 0.4 99.5 98.8 -0.7 99.4 100.0 0.6
Rhonga (%) 4.5 7.1 2.6 2.4 0.8 -1.6 0.8 3.1 2.2
Changana (%) 10.4 12.9 2.4 61.4 35.9 -25.5 21.1 17.5 -3.6
Macua (%) 35.5 23.4 -12.1 0.0 0.0 0.0 48.2 43.5 -4.7
Other (%) 0.1 0.2 0.1 0.0 0.1 0.1 1.7 0.7 -1.0
Language student speaks at home
Portuguese (%) 40.1 65.2 25.1 23.8 41.2 17.3 10.7 28.2 17.5
Rhonga (%) 0.9 0.1 -0.7 0.6 0.6 -0.0 1.1 0.7 -0.4
Changana (%) 16.3 5.8 -10.5 75.5 58.2 -17.3 34.4 26.2 -8.3
Macua (%) 42.1 27.8 -14.4 0.0 0.0 0.0 53.1 44.5 -8.6
Other (%) 0.7 1.1 0.4 0.1 0.1 -0.1 0.7 0.4 -0.3
Students whose language spoken 
at home differs from language of 
instruction (%)
20.3 13.2 -7.1 27.3 31.0 3.7 23.8 23.3 -0.6
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌81
Exhibit A10. Teachers Characteristics
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Female teachers (%) 60.1 62.5 2.4 37.2 41.9 4.8 47.8 44.8 -3.0
Teach in multi-grade class (%) 16.0 9.8 -6.2 41.0 31.8 -9.2 13.2 9.4 -3.8
Attendance at Teacher Training College (TTC)
Did not attend TTC (%) 1.8 0.4 -1.4 1.8 0.0 -1.8 1.4 0.0 -1.4
Still attends TTC (%) 1.9 0.5 -1.3 3.0 0.0 -3.0 3.0 0.9 -2.1
Attended and finished TTC (%) 96.3 99.1 2.7 95.2 100.0 4.8 95.5 99.1 3.5
TTC teacher attended
Primary Teacher Training Institute
(%) 75.8 80.9 5.1 85.6 83.3 -2.3 75.4 80.3 4.9
ADPP (%) 13.1 9.0 -4.1 9.6 10.6 1.0 15.6 11.5 -4.1
Pedagogical University (%) 9.1 8.0 -1.0 2.4 1.0 -1.4 5.1 4.7 -0.4
Attainment of certification (%) 99.5 98.5 -1.0 98.8 99.5 0.7 97.1 99.8 2.7
Highest level of education (baseline)
Form 6 (Basic) (%) 19.0 23.4 25.0
Certificate (%) 64.6 71.3 69.3
Diploma (%) 1.4 0.0 2.1
Advanced Diploma (%) 15.1 5.4 3.4
Other (%) 0.0 0.0 0.2
Highest level of education (endline)
Secondary Education (%) 15.8 12.2 16.8
Middle School (%) 61.3 72.6 69.2
Bachelor’s Degree or Higher (%) 22.9 15.2 13.9
During pre-service training, the 
teacher received specific training 
on how to teach reading and 
writing to students in the primary 
grades (%)
61.4 66.5 5.1 60.5 58.1 -2.4 58.9 70.5 11.6
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌82
SUPPLEMENTAL FINDINGS RELATED TO PRIMARY IMPACTS
Exhibit A11. Reading Comprehension Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -7.3 -17.9 -10.6 6.0 0.08 5,474 2,576
Average mean scores (percentage 
correct) 6.5 7.7 1.2 2.9 0.69 5,474 2,576
Mid-Grade Students
Percent zero scores -8.5 -15.0 -6.5 6.6 0.32 5,474 2,576
Average mean scores (percentage 
correct) 10.1 8.4 -1.8 4.6 0.70 5,474 2,576
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A12. Reading Comprehension Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -7.3 -17.4 -10.0** 3.6 0.00 5,474 5,504
Average mean scores (percentage 
correct) 6.5 11.3 4.8** 1.9 0.01 5,474 5,504
Mid-Grade Students
Percent zero scores -8.5 -23.0 -14.4** 4.4 0.00 5,474 5,504
Average mean scores (percentage 
correct) 10.1 17.7 7.6* 3.3 0.02 5,474 5,504
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A13. Listening Comprehension Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -10.8 -38.5 -27.6** 4.3 0.00 5,476 2,575
Average mean scores (percentage 
correct) 14.3 30.7 16.4** 3.5 0.00 5,476 2,575
Mid-Grade Students
Percent zero scores -9.5 -24.7 -15.2** 3.6 0.00 5,476 2,575
Average mean scores (percentage 
correct) 17.5 25.8 8.3** 3.0 0.01 5,476 2,575
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌83
Exhibit A14. Listening Comprehension Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -10.8 -38.3 -27.4** 4.7 0.00 5,476 5,503
Average mean scores (percentage 
correct) 14.3 37.0 22.8** 3.3 0.00 5,476 5,503
Mid-Grade Students
Percent zero scores -9.5 -22.8 -13.3** 3.0 0.00 5,476 5,503
Average mean scores (percentage 
correct) 17.5 26.8 9.3** 3.2 0.00 5,476 5,503
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A15. Letter Identification Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 5.4 -20.6 -26.0** 5.2 0.00 5,476 2,576
Average mean scores (percentage 
correct) 13.3 14.6 1.3 3.9 0.74 5,476 2,576
Average mean scores per minute 17.5 17.2 -0.3 4.4 0.95 5,456 2,575
Mid-Grade Students
Percent zero scores 1.6 1.9 0.2 2.3 0.92 5,476 2,576
Average mean scores (percentage 
correct) 19.1 17.1 -2.0 3.1 0.52 5,476 2,576
Average mean scores per minute 28.8 21.3 -7.5 4.4 0.09 5,456 2,575
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A16. Letter Identification Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 5.4 -9.4 -14.8** 4.5 0.00 5,476 5,504
Average mean scores (percentage 
correct) 13.3 16.3 2.9 2.6 0.26 5,476 5,504
Average mean scores per minute 17.5 18.6 1.2 3.3 0.73 5,456 5,488
Mid-Grade Students
Percent zero scores 1.6 0.1 -1.5 2.2 0.50 5,476 5,504
Average mean scores (percentage 
correct) 19.1 25.9 6.8* 2.8 0.02 5,476 5,504
Average mean scores per minute 28.8 29.4 0.6 4.5 0.89 5,456 5,488
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌84
Exhibit A17. Syllable Identification Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -30.9 -45.0 -14.2 7.2 0.05 5,474 2,573
Average mean scores (percentage 
correct) 9.1 12.8 3.7 4.9 0.45 5,474 2,573
Average mean scores per minute 6.6 8.0 1.4 2.9 0.63 5,445 2,556
Mid-Grade Students
Percent zero scores 48.0 37.8 -10.2* 4.6 0.03 5,474 2,573
Average mean scores (percentage 
correct) 11.5 10.7 -0.8 3.9 0.85 5,474 2,573
Average mean scores per minute 11.6 9.1 -2.6 3.4 0.45 5,445 2,556
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A18. Syllable Identification Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores -30.9 -21.0 9.9 5.3 0.06 5,474 5,502
Average mean scores (percentage 
correct) 9.1 15.5 6.4* 2.8 0.02 5,474 5,502
Average mean scores per minute 6.6 9.5 2.9 1.9 0.13 5,445 5,464
Mid-Grade Students
Percent zero scores 48.0 38.5 -9.5 4.8 0.05 5,474 5,502
Average mean scores (percentage 
correct) 11.5 22.8 11.3** 3.6 0.00 5,474 5,502
Average mean scores per minute 11.6 14.9 3.3 4.2 0.43 5,445 5,464
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A19. Oral Reading Fluency Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 23.4 -7.0 -30.4** 6.5 0.00 5,474 2,576
Average mean scores (percentage 
correct) 6.8 10.5 3.7 3.6 0.30 5,474 2,576
Mid-Grade Students
Percent zero scores 3.5 -1.6 -5.1 4.1 0.22 5,474 2,576
Average mean scores (percentage 
correct) 13.3 12.3 -1.0 4.5 0.83 5,474 2,576
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌85
Exhibit A20. Oral Reading Fluency Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 23.4 6.0 -17.4** 4.8 0.00 5,474 5,504
Average mean scores (percentage 
correct) 6.8 11.9 5.1* 2.1 0.02 5,474 5,504
Mid-Grade Students
Percent zero scores 3.5 -1.0 -4.5 4.4 0.30 5,474 5,504
Average mean scores (percentage 
correct) 13.3 23.3 10.0** 3.3 0.00 5,474 5,504
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A21. Word Dictation Estimates for School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 20.3 6.8 -13.5** 4.8 0.01 5,476 2,576
Average mean scores (percentage 
correct) 2.7 10.9 8.2 5.9 0.16 5,476 2,576
Mid-Grade Students
Percent zero scores 7.3 8.1 0.8 2.8 0.76 5,476 2,576
Average mean scores (percentage 
correct) 7.4 5.6 -1.7 3.8 0.64 5,476 2,576
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A22. Word Dictation Estimates for School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Percent zero scores 20.3 9.5 -10.8* 4.2 0.01 5,476 5,504
Average mean scores (percentage 
correct) 2.7 11.5 8.8** 3.3 0.01 5,476 5,504
Mid-Grade Students
Percent zero scores 7.3 6.4 -0.9 3.6 0.80 5,476 5,504
Average mean scores (percentage 
correct) 7.4 14.2 6.9* 3.1 0.03 5,476 5,504
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌86
Exhibit A23. Impact Estimates on Student Attendance (Student Self-Reported Absence At least One Day Last 
Week) in School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Grade 3 -1.8 -6.4 -4.6 5.8 0.43 5,473 2,574
Grade 5 -5.9 -1.1 4.8 4.5 0.28 5,473 2,574
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A24. Impact Estimates on Student Attendance (Student Self-Reported Absence At least One Day Last 
Week) in School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Grade 3 -1.8 -7.5 -5.7 3.6 0.11 5,473 5,491
Grade 5 -5.9 -2.0 3.9 3.8 0.31 5,473 5,491
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A25. Impact Estimates on Student Attentiveness as Perceived by Teachers in School Feeding Only 
Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Grade 3 5.4 1.9 -3.5 4.1 0.40 632 311
Grade 5 6.1 -1.1 -7.2* 3.3 0.03 632 311
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A26. Impact Estimates on Student Attentiveness as Perceived by Teachers in School Feeding and 
Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Grade 3 5.4 1.3 -4.1 3.5 0.24 632 634
Grade 5 6.1 3.0 -3.0 2.9 0.29 632 634
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌87
SUPPLEMENTAL FINDINGS RELATED TO SECONDARY IMPACTS
Exhibit A27. Teachers' Perception on Student Attentiveness
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Approximate number of students who are generally attentive in this class
Less than half 7.1 8.8 1.7 4.2 2.1 -2.1 8.7 8.6 -0.1
Half 18.7 17.2 -1.5 5.4 10.8 5.4 11.5 16.9 5.4
More than half 57.6 57.0 -0.6 64.7 65.5 0.8 60.0 62.0 2.0
All 7.9 17.0 9.2 22.2 21.6 -0.5 15.6 12.5 -3.1
Ways teacher gauges if students are attentive in class
Students are seated facing and 
making eye contact with the 
teacher
55.7 54.8 -0.9 59.3 53.5 -5.7 47.9 59.5 11.6
Students actively listen and act 
upon teacher directions 38.4 47.6 9.2 41.3 46.0 4.6 37.4 60.1 22.7
Students are working 
independently, in pairs, or in small 
groups on the current learning 
activity
14.4 7.6 -6.9 15.0 5.6 -9.4 13.1 17.5 4.4
Students respond to questions 
posed by the teacher 43.4 52.7 9.3 43.1 53.5 10.4 43.1 60.2 17.1
Students ask the teacher or 
another student for help related to 
the current learning activity
14.4 7.5 -6.9 10.2 7.1 -3.1 18.7 16.6 -2.1
Students ask the teacher 
questions related to the current 
learning activity
16.7 18.9 2.2 13.8 13.1 -0.6 13.0 17.3 4.3
Students are reading 2.3 13.8 11.5 8.4 2.0 -6.4 6.3 13.1 6.9
Students are writing 4.4 13.1 8.7 4.2 1.5 -2.7 4.2 15.1 10.9
Students are listening to a story 2.2 8.4 6.2 0.6 0.0 -0.6 3.5 5.4 1.9
Other 9.1 3.7 -5.4 1.8 8.1 6.3 6.1 3.8 -2.4
Student attentiveness in relation to the school meal
Students are more attentive after 
the school meal 20.8 44.7 23.9 36.5 72.6 36.1 34.4 54.8 20.4
Students are less attentive after 
the school meal 10.3 17.3 7.0 19.8 12.3 -7.4 21.3 13.2 -8.1
Students show the same 
attentiveness before and after 
lunch
17.6 26.0 8.4 40.7 15.1 -25.7 38.2 27.5 -10.7
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Exhibit A28. Impact on Teacher-Reported Instructional Input in School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
0.3 14.5 14.3 9.3 0.13 625 310
Teacher-reported absence of at 
least one day last week -0.6 -12.2 -11.6 6.7 0.08 632 311
Teacher has access to school 
supplies 11.5 0.2 -11.3** 3.3 0.00 631 310
Mid-Grade Students
Abt Associates June 2022 ▌88
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
-8.9 -5.0 3.9 9.1 0.67 625 310
Teacher-reported absence of at 
least one day last week -1.2 -3.4 -2.2 6.7 0.74 632 311
Teacher has access to school 
supplies 8.7 3.3 -5.5 4.1 0.18 631 310
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A29. Impact on Teacher-Reported Instructional Input in School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
0.3 15.5 15.3* 7.0 0.03 625 632
Teacher-reported absence of at 
least one day last week -0.6 -1.7 -1.1 6.6 0.87 632 634
Teacher has access to school 
supplies 11.5 1.5 -10.0** 3.7 0.01 631 631
Mid-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
-8.9 19.0 27.9** 7.5 0.00 625 632
Teacher-reported absence of at 
least one day last week -1.2 1.4 2.6 5.8 0.66 632 634
Teacher has access to school 
supplies 8.7 0.7 -8.1* 3.9 0.04 631 631
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌89
Exhibit A30. Impact on Feeding Outcomes in School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Student reports feeling hungry now -8.2 0.9 9.1 6.2 0.14 5,476 2,576
Student achieved minimum meal 
frequency 7.0 -27.2 -34.2** 6.2 0.00 5,421 2,543
Student achieved minimum dietary 
diversity -28.8 -26.5 2.3 6.2 0.71 5,476 2,576
Student achieved minimum 
acceptable diet -7.0 -20.8 -13.8* 5.8 0.02 5,429 2,545
Student's nutrition knowledge 
score -0.4 0.0 0.5** 0.1 0.00 5,471 2,576
Mid-Grade Students
Student reports feeling hungry now -2.1 12.2 14.3* 6.7 0.04 5,476 2,576
Student achieved minimum meal 
frequency 1.4 -15.4 -16.8** 4.9 0.00 5,421 2,543
Student achieved minimum dietary 
diversity -24.6 -31.2 -6.7 7.3 0.36 5,476 2,576
Student achieved minimum 
acceptable diet -9.2 -23.1 -13.9* 6.7 0.04 5,429 2,545
Student's nutrition knowledge 
score -0.5 -0.1 0.4 0.2 0.06 5,471 2,576
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A31. Impact on Feeding Outcomes in School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Student reports feeling hungry now -8.2 -15.2 -6.9 4.9 0.16 5,476 5,504
Student achieved minimum meal 
frequency 7.0 -7.7 -14.7** 5.1 0.00 5,421 5,405
Student achieved minimum dietary 
diversity -28.8 -28.0 0.8 4.7 0.86 5,476 5,502
Student achieved minimum 
acceptable diet -7.0 -20.8 -13.8** 4.4 0.00 5,429 5,427
Student's nutrition knowledge 
score -0.4 0.4 0.9** 0.1 0.00 5,471 5,503
Mid-Grade Students
Student reports feeling hungry now -2.1 -2.0 0.1 4.6 0.98 5,476 5,504
Student achieved minimum meal 
frequency 1.4 -2.5 -3.9 4.2 0.35 5,421 5,405
Student achieved minimum dietary 
diversity -24.6 -34.3 -9.8* 4.6 0.03 5,476 5,502
Student achieved minimum 
acceptable diet -9.2 -22.2 -13.1** 4.2 0.00 5,429 5,427
Student's nutrition knowledge 
score -0.5 0.3 0.8** 0.2 0.00 5,471 5,503
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌90
Exhibit A32. Impact on Student Knowledge and Health Practices in School Feeding Only Schools
........................................................ School Feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Student's hygiene and sanitation 
knowledge score 0.8 1.1 0.3 0.3 0.25 5,471 2,576
Student reports absence last week 
as a result of illness 0.3 -17.8 -18.1 10.2 0.08 1,325 566
Mid-Grade Students
Student's hygiene and sanitation 
knowledge score 1.0 0.9 -0.0 0.2 0.84 5,471 2,576
Student reports absence last week 
as a result of illness 1.7 -2.1 -3.8 12.1 0.76 1,325 566
All Students
Head teacher reports school has 
improved latrines 11.4 11.8 0.3 7.5 0.97 352 174
Head teacher reports school has 
tippy taps or hand-washing 
stations
29.4 1.7 -27.7** 9.0 0.00 352 174
Head teacher reports all tippy taps 
or hand-washing stations have ash 
or soap
26.7 5.7 -21.0** 7.0 0.00 352 174
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Exhibit A33. Impact on Student Knowledge and Health Practices in School Feeding and Literacy Schools
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Student's hygiene and sanitation 
knowledge score 0.8 1.6 0.9** 0.2 0.00 5,471 5,503
Student reports absence last week 
as a result of illness 0.3 13.2 13.0 8.2 0.11 1,325 1,219
Mid-Grade Students
Student's hygiene and sanitation 
knowledge score 1.0 1.3 0.3* 0.1 0.02 5,471 5,503
Student reports absence last week 
as a result of illness 1.7 10.6 9.0 9.1 0.33 1,325 1,219
All Students
Head teacher reports school has 
improved latrines 11.4 8.6 -2.8 5.0 0.57 352 346
Head teacher reports school has 
tippy taps or hand-washing 
stations
29.4 19.6 -9.8 6.1 0.11 352 346
Head teacher reports all tippy taps 
or hand-washing stations have ash 
or soap
26.7 21.0 -5.7 4.2 0.18 352 346
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌91
GENDER DISAGGREGATED IMPACT OF READING COMPREHENSION
Exhibit A34. Impact on Students’ Reading Comprehension in School Feeding Only Schools, by Grade and 
Gender
........................................................ School feeding Only
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Boys
Student reading performance 
(Average Reading Comprehension 
Scores) 
5.9 8.8 2.9 3.1 0.35 2,745 1,238
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-7.2 -20.1 -12.8 6.6 0.05 2,745 1,238
Student attendance 1.2 -8.1 -9.3 6.8 0.16 2,743 1,236
Girls
Student reading performance 
(Average Reading Comprehension 
Scores) 
8.0 6.7 -1.3 3.5 0.72 2,729 1,338
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-8.6 -15.5 -6.9 6.2 0.27 2,729 1,338
Student attendance -2.6 -4.0 -1.4 8.6 0.87 2,730 1,338
Mid-Grade Students
Boys
Student reading performance 
(Average Reading Comprehension 
Scores) 
10.6 9.9 -0.7 5.2 0.89 2,745 1,238
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-10.1 -13.4 -3.3 8.0 0.68 2,745 1,238
Student attendance -5.3 -1.6 3.7 6.0 0.54 2,743 1,236
Girls
Student reading performance 
(Average Reading Comprehension 
Scores) 
8.3 6.1 -2.2 6.1 0.72 2,729 1,338
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-4.8 -14.1 -9.3 7.7 0.23 2,729 1,338
Student attendance -6.4 0.6 7.0 5.5 0.21 2,730 1,338
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌92
Exhibit A35. Impact on Students’ Reading Comprehension in School Feeding and Literacy Schools, by Grade 
and Gender
........................................................ School Feeding and Literacy
C Diff T Diff Impact SE p-value C N T N
Early-Grade Students
Boys
Student reading performance 
(Average Reading Comprehension 
Scores) 
5.9 11.0 5.1* 2.2 0.02 2,745 2,737
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-7.2 -16.8 -9.5* 4.4 0.03 2,745 2,737
Student attendance 1.2 -13.6 -14.8** 4.9 0.00 2,743 2,731
Girls
Student reading performance 
(Average Reading Comprehension 
Scores) 
8.0 12.1 4.1 2.3 0.08 2,729 2,767
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-8.6 -19.1 -10.5** 3.9 0.01 2,729 2,767
Student attendance -2.6 -2.0 0.6 4.7 0.90 2,730 2,760
Mid-Grade Students
Boys
Student reading performance 
(Average Reading Comprehension 
Scores) 
10.6 21.7 11.1* 4.3 0.01 2,745 2,737
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-10.1 -28.0 -17.9** 5.6 0.00 2,745 2,737
Student attendance -5.3 -7.2 -1.9 4.7 0.69 2,743 2,731
Girls
Student reading performance 
(Average Reading Comprehension 
Scores) 
8.3 13.8 5.5 4.1 0.19 2,729 2,767
Student reading performance 
(Percent Zero Reading 
Comprehension Scores)
-4.8 -17.7 -12.9* 5.5 0.02 2,729 2,767
Student attendance -6.4 3.6 9.9 5.1 0.05 2,730 2,760
Note: “C Diff” denotes difference in the comparison group. “T Diff” denotes difference in the treatment group. “Impact” denotes the difference in 
differences. “SE” denotes standard error. “C N” denotes sample size in the comparison group. “T N” denotes sample size in the treatment 
group.
Abt Associates June 2022 ▌93
SUPPLEMENTAL FINDINGS RELATED TO MEDIATING IMPACTS OF 
IMPLEMENTATION
Exhibit A36. Impact of Adequate Implementing Programs on Student Reading Comprehension
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores -7.2 -21.2 -14.0 -17.0 -9.8*
(10.6) (4.5)
Average mean scores (percentage 
correct) 6.3 8.9 2.6 11.7 5.4*
(4.0) (2.3)
Mid-Grade Students
Percent zero scores -8.3 -15.9 -7.5 -27.2 -18.9**
(7.4) (5.3)
Average mean scores (percentage 
correct) 9.8 7.4 -2.4 20.7 10.9**
(5.2) (3.8)
Note: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 1,054 School Feeding Only; 2,959 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference in differences. 
Exhibit A37. Impact of Adequate Implementing Programs on Oral Comprehension
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores -11.4 -38.5 -27.1** -41.8 -30.4**
(6.6) (6.4)
Average mean scores (percentage 
correct) 14.5 29.3 14.9** 42.3 27.8**
(3.8) (4.3)
Mid-Grade Students
Percent zero scores -9.9 -23.7 -13.8** -23.7 -13.8**
(4.6) (3.7)
Average mean scores (percentage 
correct) 17.5 25.9 8.4* 30.5 13.0**
(3.8) (3.9)
Note: Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 1,053 School feeding only; 2,959 Literacy and school feeding. 
Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences estimates. **
p<0.01; * p<0.05. 
Abt Associates June 2022 ▌94
Exhibit A38. Impact of Adequate Implementing Programs on Letter Identification
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores 5.9 -17.9 -23.8** -8.9 -14.8**
(7.1) (5.3)
Average mean scores (percentage 
correct) 13.5 15.0 1.5 18.1 4.6
(5.2) (2.9)
Average mean scores per minute 17.8 17.0 -0.8 20.7 2.9
(5.5) (3.8)
Mid-Grade Students
Percent zero scores 2.1 3.6 1.4 -0.3 -2.4
(3.0) (2.7)
Average mean scores (percentage 
correct) 19.3 14.7 -4.6 30.6 11.3**
(4.2) (3.4)
Average mean scores per minute 29.1 18.6 -10.4 34.3 5.2
(5.4) (5.2)
Note: Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 1,054 School feeding only; 2,959 Literacy and school feeding.
Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences estimates. **
p<0.01; * p<0.05. 
Abt Associates June 2022 ▌95
Exhibit A39. Impact of Adequate Implementing Programs on Syllable Identification
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores -33.0 -42.9 -9.9 -21.9 11.1
(9.4) (6.1)
Average mean scores (percentage 
correct) 9.1 15.7 6.6 16.2 7.1*
(6.9) (3.0)
Average mean scores per minute 6.6 9.3 2.7 9.7 3.1
(3.7) (2.1)
Mid-Grade Students
Percent zero scores 45.6 37.2 -8.3 40.9 -4.7
(6.2) (6.2)
Average mean scores (percentage 
correct) 11.4 5.2 -6.2 28.4 17.0**
(4.3) (4.2)
Average mean scores per minute 11.7 5.7 -6.0 18.3 6.6
(3.6) (4.9)
Note: Standard errors appear in parentheses. Sample sizes: 5,474 Comparison; 1,053 School feeding only; 2,958 Literacy and school feeding. 
Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences estimates. **
p<0.01; * p<0.05.
Exhibit A40. Impact of Adequate Implementing Programs on Oral Reading Fluency
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores 23.7 -11.1 -34.8** 4.5 -19.2**
(10.2) (5.7)
Average mean scores (percentage 
correct) 6.7 12.8 6.1 11.2 4.5
(5.2) (2.6)
Mid-Grade Students
Percent zero scores 3.9 -0.8 -4.6 -6.3 -10.2*
(4.6) (5.1)
Average mean scores (percentage 
correct) 13.2 10.0 -3.3 26.2 13.0**
(5.0) (4.0)
Note: Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 1,054 School feeding only; 2,959 Literacy and school feeding.
Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences estimates. **
p<0.01; * p<0.05.
Abt Associates June 2022 ▌96
Exhibit A41. Impact of Adequate Implementing Programs on Word Dictation
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Percent zero scores 20.6 12.0 -8.6 7.9 -12.7**
(6.7) (4.6)
Average mean scores (percentage 
correct) 2.6 14.3 11.6 12.0 9.4**
(7.9) (3.5)
Mid-Grade Students
Percent zero scores 7.7 8.7 1.0 4.1 -3.6
(3.4) (4.4)
Average mean scores (percentage 
correct) 7.3 3.5 -3.7 16.5 9.3*
(3.6) (4.0)
Note: Standard errors appear in parentheses. Sample sizes: 5,476 Comparison; 1,054 School feeding only; 2,959 Literacy and school feeding.
Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences estimates. **
p<0.01; * p<0.05. 
Exhibit A42. Impact of Adequate Implementing Programs on Student Attendance
Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students -1.5 6.9 8.4 -7.6 -6.1
(6.1) (3.7)
Mid-Grade Students -5.5 -4.0 1.5 -5.3 0.1
(7.1) (4.1)
Note: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 1,052 School feeding only; 2,956 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference in differences 
estimates. ** p<0.01; * p<0.05. 
Exhibit A43. Impact of Adequate Implementing Programs on Student Attentiveness
Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students 5.2 3.7 -1.4 -2.5 -7.7
(5.4) (4.0)
Mid-Grade Students 5.9 3.0 -2.9 3.9 -1.9
(4.7) (3.0)
Note: “Early-Grade Students” were in grade 2 at baseline and grade 3 at endline. “Mid-Grade Students” were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 632 Comparison; 125 School feeding only; 336 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05.
Abt Associates June 2022 ▌97
SUPPLEMENTAL FINDINGS RELATED TO SECONDARY IMPACTS
Exhibit A44. Impact on Student Reported Access to School Supplies and Study Materials
........................................................ Comparison School Feeding Only School Feeding and Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Student has supplementary reader 
available today -23.0 -10.1 12.9 -8.6 14.5**
(7.5) (5.6)
Student has Portuguese/Local 
Language textbook available today -38.9 -47.2 -8.3 -51.0 -12.1*
(6.8) (4.9)
Student has exercise notebook 
available today 16.5 15.6 -0.9 9.4 -7.1*
(3.4) (2.8)
Student reads books daily in the 
classroom or school library -9.0 -11.0 -2.0 -8.1 0.9
(6.7) (5.1)
Mid-Grade Students
Student has supplementary reader 
available today -23.7 -11.8 11.9 -18.6 5.1
(7.2) (5.7)
Student has Portuguese/Local 
Language textbook available today -37.8 -48.0 -10.2 -54.0 -16.3**
(6.7) (5.1)
Student has exercise notebook 
available today 4.7 7.3 2.6 4.7 0.0
(2.5) (1.8)
Student reads books daily in the 
classroom or school library -12.8 -11.2 1.6 -2.5 10.3*
(6.5) (5.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 2,574 School feeding only; 5,491 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌98
Exhibit A45. Impact on Teacher Access to Supplies and Learning Materials
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Teacher has a teacher guide for 
reading instruction 5.4 14.4 9.0 14.7 9.3
(8.0) (7.2)
Students have exercise books that 
can be used during reading 
instruction
16.1 -6.7 -22.8** 3.6 -12.6*
(7.0) (6.0)
Teacher has access to school 
supplies 11.5 0.2 -11.3** 1.5 -10.0**
(3.3) (3.7)
Mid-Grade Students
Teacher has a teacher guide for 
reading instruction 3.3 4.5 1.3 -4.1 -7.3
(8.4) (7.4)
Students have exercise books that 
can be used during reading 
instruction
39.8 17.0 -22.8** 31.8 -8.0
(8.0) (6.1)
Teacher has access to school 
supplies 8.7 3.3 -5.5 0.7 -8.1*
(4.1) (3.9)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 628 Comparison; 309 School feeding only; 623 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A46. Percentage of Teachers Reporting Access to School Supply Items
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Pens/pencils 92.2 95.8 3.6 93.0 88.7 -4.3 88.3 95.6 7.3
Notebooks 89.4 95.0 5.6 88.6 86.7 -1.9 83.8 92.8 9.0
Chalk 84.5 95.9 11.4 75.3 90.8 15.5 84.3 86.6 2.3
Cardboard 73.3 49.7 -23.7 77.8 58.5 -19.4 66.7 60.8 -5.9
Paint 7.2 11.3 4.1 31.0 27.2 -3.8 13.9 18.3 4.4
Markers 9.5 20.9 11.4 24.7 45.6 21.0 21.9 35.9 14.1
Chart paper 26.8 24.4 -2.4 44.9 52.3 7.4 40.2 38.4 -1.7
Other 12.6 12.4 -0.2 10.1 16.4 6.3 16.5 13.6 -2.9
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌99
Exhibit A47. Impact on Head Teacher Reporting Access to Library/Book Banks and Materials
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
HTs who report that the school has 
a library/book bank 5.7 -15.6 -21.3** 7.0 1.2
(7.9) (5.2)
HTs who report that the print 
material in the library/book bank is 
available in Portuguese
5.7 -14.1 -19.8* 5.3 -0.4
(7.8) (5.2)
HTs who report that the print 
material in the library/book bank is 
available in the mother tongue
0.6 2.9 2.2 12.8 12.1**
(4.0) (3.7)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 352 Comparison; 174 School feeding only; 346 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A48. Percentage of Head Teachers Reporting Availability of Specific Types of Print Materials
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Flipbooks 19.1 59.0 39.9 36.0 31.3 -4.7 18.6 63.8 45.2
Storybooks 61.8 70.5 8.7 84.0 68.8 -
15.2 84.1 79.4 -4.6
Primers/student manuals 70.2 27.8 -
42.4 84.0 31.3 -
52.8 57.9 53.0 -5.0
Fables 39.6 36.5 -3.1 40.0 31.3 -8.8 50.3 52.9 2.6
Leveled books 63.5 30.9 -
32.7 52.0 18.8 -
33.3 53.5 34.9 -18.6
Supplementary readers 50.2 33.0 -
17.2 40.0 62.5 22.5 40.8 56.2 15.4
Guidelines for parents 6.2 2.4 -3.8 0.0 6.3 6.3 1.2 21.0 19.8
Reading posters 29.8 46.2 16.4 28.0 12.5 -
15.5 20.8 44.4 23.6
Other 2.3 14.9 12.7 8.3 18.8 10.4 4.3 9.9 5.7
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌100
Exhibit A49. Impact on Teacher Reporting on Teacher Training
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
This year, the teacher attended 
any in-service training on how to 
teach reading in the primary 
grades
-8.5 -12.0 -3.5 -12.2 -3.7
(7.6) (5.8)
Teacher received training on the 
fundamentals of reading instruction -3.3 14.0 17.3 15.7 19.0**
(9.4) (7.1)
Teacher can name the core 
reading skills 16.6 18.8 2.2 3.8 -12.8*
(8.0) (5.3)
Mid-Grade Students
This year, the teacher attended 
any in-service training on how to 
teach reading in the primary 
grades
12.5 -9.7 -22.2** 6.0 -6.5
(8.4) (6.4)
Teacher received training on the 
fundamentals of reading instruction -7.1 -5.4 1.7 19.9 26.9**
(9.0) (7.5)
Teacher can name the core 
reading skills 0.5 10.3 9.8 14.8 14.3*
(7.6) (6.2)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 631 Comparison; 311 School feeding only; 634 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌101
Exhibit A50. Teachers Reporting on Teacher Training
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Percentage of teacher attended 
WV/PAI training 10.9 . . 63.5 64.3 0.8 70.0 71.3 1.3
Frequency of the reading coach visiting the teacher (this academic school year)
Never 76.0 65.5 -10.5 66.5 56.1 -10.4 59.0 35.5 -23.5
Quarterly 10.4 9.5 -1.0 13.2 13.1 -0.0 15.4 29.0 13.7
Monthly 4.7 2.1 -2.6 6.6 7.1 0.5 12.2 9.0 -3.1
Weekly 0.7 1.2 0.5 6.6 8.1 1.5 10.0 14.2 4.2
Types of activities teacher participates in during in-service training
Analysis of curriculum and syllabus 34.9 46.3 11.4 46.7 55.9 9.2 25.9 46.8 20.9
Analysis of teacher guide 23.0 12.1 -11.0 28.1 18.6 -9.5 19.5 34.3 14.8
Practice teaching lessons from the 
teacher guide 23.6 20.0 -3.5 21.6 28.8 7.3 19.5 38.6 19.1
Analysis of the student exercise 
book 36.1 26.7 -9.4 61.1 28.8 -32.3 42.3 36.2 -6.1
How to plan and prepare for a 
lesson 64.8 45.4 -19.4 80.8 55.9 -24.9 67.1 56.8 -10.4
Improvisation and use of teaching 
and learning aids 35.4 32.6 -2.8 55.1 28.8 -26.3 42.0 44.5 2.6
How to assess students during the 
lesson 42.1 23.8 -18.2 64.7 35.6 -29.1 54.8 45.4 -9.4
How to use assessments to adjust 
instruction 23.3 11.9 -11.4 40.1 18.6 -21.5 30.7 21.6 -9.1
How to give students feedback 22.8 17.6 -5.2 43.1 13.6 -29.6 28.8 22.5 -6.4
How to use supplementary 
readers/decodable texts 25.4 1.5 -23.9 28.7 15.3 -13.5 22.0 20.1 -1.9
How to apply new strategies for 
reading instruction 27.0 35.1 8.1 32.9 44.1 11.1 28.2 46.2 18.0
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌102
Exhibit A51. Teacher Knowledge and Skills
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
The reading skills the teacher can name
Letter knowledge 65.0 94.0 29.0 58.1 87.3 29.3 66.6 94.9 28.3
Phonemic awareness 34.5 39.7 5.2 24.6 56.6 32.1 34.1 71.1 37.0
Fluency 22.9 17.3 -5.6 14.4 41.0 26.6 21.3 57.1 35.8
Vocabulary 35.4 58.9 23.5 36.5 57.2 20.7 51.5 68.9 17.4
Comprehension 43.5 57.0 13.5 36.5 60.2 23.7 35.3 71.3 36.0
Grade level at which the teacher expects children to read text in the target language fluently
Grade 1 7.1 1.5 -5.5 6.0 3.5 -2.5 4.4 0.4 -4.0
Grade 2 36.9 16.4 -20.5 31.1 11.1 -20.0 21.2 11.3 -9.9
Grade 3 25.0 45.9 20.9 27.5 33.8 6.3 40.9 40.3 -0.6
Grade 4 or higher 29.6 35.6 6.0 35.3 50.0 14.7 30.1 47.2 17.1
Grade level at which the teacher expects students to write a coherent and comprehensible short story correctly
Grade 1 4.2 2.7 -1.5 3.6 0.0 -3.6 2.5 0.5 -1.9
Grade 2 22.2 13.2 -8.9 21.0 4.0 -16.9 16.6 7.9 -8.8
Grade 3 27.8 36.6 8.8 29.3 29.3 -0.0 34.4 32.2 -2.2
Grade 4 or higher 43.1 47.1 4.0 45.5 64.6 19.1 44.1 57.7 13.5
Ways the teacher measures students' reading and writing progress
Written tests 63.8 77.2 13.4 61.1 65.2 4.1 70.0 81.2 11.2
Oral evaluations 66.9 74.4 7.5 66.5 81.3 14.8 62.1 78.6 16.5
Observation 40.9 29.0 -11.9 34.7 25.3 -9.5 40.6 44.4 3.8
Portfolios and other projects 4.1 4.0 -0.1 6.0 1.5 -4.5 4.4 6.7 2.3
Homework 36.6 38.7 2.0 31.7 32.8 1.1 34.9 44.5 9.5
Classwork 30.9 42.8 11.9 30.5 49.0 18.5 32.4 42.0 9.6
End-of-unit evaluation 6.0 17.9 11.9 5.4 15.2 9.8 5.5 24.3 18.8
End-of-term evaluation 2.6 17.4 14.8 3.0 9.6 6.6 5.8 26.7 20.9
Ways the teacher checks for student understanding during the lesson
Ask questions of individual 
students 57.4 72.3 14.9 56.3 62.1 5.8 56.0 82.7 26.7
Ask questions of whole class 46.5 49.7 3.3 40.7 48.0 7.3 38.9 58.1 19.2
Ask questions of students in 
groups
25.5 17.4 -8.1 25.1 21.2 -3.9 25.6 32.7 7.2
Give students a task and correct 
the written responses before the 
end of the lesson
32.8 45.1 12.3 44.3 51.5 7.2 36.3 53.8 17.5
Give students a task and correct 
the written responses after the end 
of the lesson
20.9 35.3 14.4 35.9 25.8 -10.2 25.6 29.3 3.7
Ask students to put thumbs up or 
down 2.8 2.2 -0.6 3.6 1.5 -2.1 5.2 6.4 1.2
Other 12.0 4.8 -7.2 7.8 4.0 -3.7 14.5 9.5 -5.0
Ways the teacher uses the results of students' oral and written assessments in instruction
Grade students 53.7 59.1 5.4 47.9 64.1 16.2 46.9 69.9 22.9
Evaluate students' understanding 
of subject matter 59.5 0.0 -59.5 56.3 1.0 -55.3 53.5 2.2 -51.3
Plan teaching and learning 
activities 23.5 39.8 16.3 27.5 25.3 -2.3 29.1 45.2 16.1
Adapt teaching to better suit 
students' needs 21.5 0.0 -21.5 37.1 0.0 -37.1 21.4 0.0 -21.4
Other 16.3 11.4 -4.9 7.2 10.6 3.4 11.7 18.4 6.7
Techniques teacher uses to teach a new vocabulary word
Pictures 40.9 56.9 16.0 42.5 57.6 15.1 55.0 68.5 13.5
Abt Associates June 2022 ▌103
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Realia 40.5 50.1 9.7 49.1 49.5 0.4 47.1 49.9 2.8
Gestures 26.3 26.4 0.1 24.0 22.2 -1.7 20.4 35.4 15.0
Verbal explanation 55.3 76.8 21.5 49.7 73.2 23.5 51.9 74.7 22.8
Other 15.8 3.0 -12.8 13.8 8.1 -5.7 12.9 6.3 -6.6
Frequency of teacher switching between official language of instruction and a vernacular language during the 
teaching and learning process
Never 14.3 19.6 5.3 6.6 10.6 4.0 3.7 5.0 1.3
Occasionally 58.6 59.7 1.2 62.3 55.6 -6.7 67.5 56.6 -10.9
Often 21.2 15.0 -6.2 22.2 20.2 -2.0 18.8 30.3 11.5
All of the time 5.6 5.5 -0.2 7.8 13.6 5.9 9.7 7.6 -2.1
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Exhibit A52. Impact on Head Teacher Reported Teacher Training
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
% of HTs who attended WV or PAI 
training -6.4 -10.5 -4.1 -8.0 -1.6
(4.7) (2.8)
% of HTs who received training on 
the fundamentals of reading 
instruction
2.3 27.8 25.5** 28.1 25.7**
(9.4) (6.9)
% of HTs who report receiving 
training on school management 4.5 -2.4 -6.9 17.0 12.5*
(7.7) (6.2)
% of HTs who report receiving 
training about the importance of 
access to and use of school 
supplies and materials
-2.9 8.0 10.9 11.2 14.0*
(8.9) (7.1)
Head teacher can name the core 
reading skills 12.3 17.0 4.7 12.7 0.4
(10.9) (7.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 352 Comparison; 174 School feeding only; 346 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌104
Exhibit A53. Head Teacher Knowledge and Skills
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
The reading skills the head teacher can name
Letter knowledge 5.9 92.3 86.3 4.8 85.7 81.0 21.4 91.0 69.6
Phonemic awareness 10.2 54.9 44.7 8.3 58.9 50.6 30.9 84.2 53.2
Fluency 5.9 35.9 30.0 9.5 44.6 35.1 21.8 71.3 49.4
Vocabulary 15.2 63.4 48.3 17.9 57.1 39.3 30.9 80.4 49.5
Comprehension 17.1 61.3 44.2 15.5 78.6 63.1 22.2 80.8 58.6
Ways the head teacher identifies academic progress
Classroom observation 58.3 61.8 3.5 48.8 61.4 12.6 62.5 75.6 13.1
Monitor student results on tests 
given by teachers 55.9 58.1 2.2 51.2 48.9 -2.3 48.1 56.8 8.7
Evaluate students orally myself 32.2 41.8 9.5 29.8 39.8 10.0 41.5 51.2 9.7
Check students' assignments or 
homework 32.7 20.4 -12.3 33.3 23.9 -9.5 24.8 30.4 5.5
Teachers provide progress reports 35.9 36.1 0.3 35.7 27.3 -8.4 26.3 41.3 15.0
End-of-term evaluations 37.9 42.0 4.1 28.6 38.6 10.1 32.6 42.5 9.9
Feedback from parents 7.0 11.3 4.3 0.0 2.3 2.3 1.2 8.2 6.9
Feedback from school counselors 5.3 9.5 4.2 3.6 4.5 1.0 0.0 11.5 11.5
Feedback from school committees 1.1 8.9 7.7 1.2 3.4 2.2 0.6 5.8 5.2
Other 14.8 9.5 -5.2 16.7 8.0 -8.7 12.4 4.5 -7.9
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Exhibit A54. Percentage of Head Teachers Reporting Teachers Attended Literacy Training and WI or PAI 
Training
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
HTs who report that teachers in 
the school attended literacy 
training this academic year
12.8 49.6 36.8 7.1 100.0 92.9 13.9 100.0 86.1
HTs who report the teachers in the 
school attended WV or PAI training 
(of the HTs who reported teachers 
attended literacy training)
1.1 0.0 -1.1 6.0 81.8 75.9 44.0 92.8 48.9
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌105
Exhibit A55. Impact on Head Teacher Reported Number of Male and Female Teachers who Attended 
Literacy Training
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Mean number of female teachers 
in the school who attended literacy 
training this academic year
1.0 -0.1 -1.2** 0.8 -0.3
(0.2) (0.3)
Mean number of male teachers in 
the school who attended literacy 
training this academic year
0.5 0.9 0.4* 1.1 0.6**
(0.2) (0.2)
Mean number of early-grade
teachers who attended literacy 
training this academic year
0.7 0.4 -0.3 1.2 0.4*
(0.2) (0.2)
Mean number of female early￾grade teachers who attended 
literacy training this academic year
0.5 0.0 -0.4* 0.5 0.1
(0.2) (0.1)
Mean number of male early-grade
teachers who attended literacy 
training this academic year
0.3 0.4 0.1 0.6 0.3**
(0.1) (0.1)
Mean number of mid-grade
teachers who attended literacy 
training this academic year
0.6 0.4 -0.2 0.9 0.3
(0.2) (0.1)
Mean number of female mid-grade
teachers who attended literacy 
training this academic year
0.4 -0.0 -0.4** 0.4 -0.0
(0.1) (0.1)
Mean number of male mid-grade
teachers who attended literacy 
training this academic year
0.3 0.5 0.2* 0.6 0.3**
(0.1) (0.1)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 352 Comparison; 174 School feeding only; 346 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. Note that at baseline teachers were asked if they attended literacy training this academic year; at endline 
teachers were asked if they attended literacy training between 2017 and 2019.
Abt Associates June 2022 ▌106
Exhibit A56. Impact on Head Teachers’ Reported Reading Coaches Training and Supervising Teachers, and 
Number of Trainings Reading Coaches Provided
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
% of HTs who report that reading 
coaches train teachers at this 
school
11.9 17.5 5.6 33.0 21.1**
(9.6) (6.2)
% of HTs who report that reading 
coaches supervise teachers at this 
school
11.9 17.5 5.6 33.0 21.1**
(9.6) (6.2)
Mean number of trainings provided 
by reading coaches this academic 
year if there is training by reading 
coaches
19.7 26.0 6.3 40.3 20.6**
(8.5) (5.8)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 352 Comparison; 174 School feeding only; 346 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A57. Percentage of Head Teachers Reporting on Frequency Reading Coaches Supervise Teachers
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Never 45.2 27.4 -17.7 41.2 32.7 -8.5 11.2 13.6 2.3
Once a year 15.3 54.1 38.8 11.8 24.5 12.7 3.4 45.5 42.1
Once per term 21.5 7.6 -13.9 8.8 14.3 5.5 27.4 17.2 -10.2
Once a month 12.8 7.1 -5.7 11.8 26.5 14.8 30.5 23.1 -7.5
More than once a month 5.2 3.8 -1.4 26.5 2.0 -24.4 27.5 0.6 -26.8
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌107
Exhibit A58. Impact of Head Teachers’ Reported Teacher Attendance and Head Teacher Attendance
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Average number of teachers 
absent yesterday or last day 
school was in session
0.5 -0.0 -0.5 -0.3 -0.8
(0.7) (0.6)
Teacher attendance register 
available -39.2 -45.9 -6.8 -60.3 -21.1**
(8.1) (5.6)
HT absent from school any day 
last week 12.5 13.8 1.3 2.9 -9.6
(8.9) (6.8)
HT late to school any day last 
week 15.5 27.1 11.6 4.2 -11.4
(7.7) (6.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 231 Comparison; 110 School feeding only; 182 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A59. Head Teacher Reporting on How Class is Handled when Teacher is Absent
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Let class proceed without teacher 2.7 5.3 2.6 1.2 3.4 2.2 0.6 2.9 2.3
Allocate class to another teacher 54.7 76.6 21.9 61.9 71.6 9.7 70.9 80.8 9.9
Join all students in one class 26.7 18.8 -7.9 15.5 8.0 -7.5 14.8 20.3 5.5
Dismiss students for the day 0.5 5.5 4.9 1.2 0.0 -1.2 1.1 6.0 4.9
Distribute students among other 
classes 6.6 5.8 -0.8 11.9 3.4 -8.5 7.4 4.0 -3.4
Other 8.3 14.9 6.6 8.3 21.6 13.3 4.0 10.1 6.1
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌108
Exhibit A60. Impact on Students Reporting School Serving Schools Meals, Current Hunger, and Household 
Hunger
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
School serves school meals -0.1 -28.9 -28.7** -26.4 -26.3**
(7.8) (3.7)
Eats fewer meals at home when 
receives a school meal -1.8 -45.5 -43.7** -27.5 -25.6**
(7.8) (4.0)
Feels less hungry on days when 
receives a school meal -0.3 -22.9 -22.6** -17.6 -17.3**
(8.7) (4.5)
Feels hungry now -8.2 0.9 9.1 -15.2 -6.9
(6.2) (4.9)
Any household hunger during 
school year 10.3 13.2 2.9 -3.1 -13.4**
(4.8) (3.8)
Mid-Grade Students
School serves school meals 2.4 -39.8 -42.2** -29.0 -31.4**
(6.5) (4.0)
Eats fewer meals at home when 
receives a school meal -2.5 -45.5 -42.9** -37.5 -35.0**
(6.4) (3.5)
Feels less hungry on days when 
receives a school meal 2.6 -33.9 -36.5** -18.7 -21.3**
(6.3) (4.4)
Feels hungry now -2.1 12.2 14.3* -2.0 0.1
(6.7) (4.6)
Any household hunger during 
school year 6.6 12.4 5.8 1.1 -5.5
(4.5) (4.5)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 2,574 School feeding only; 5,488 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌109
Exhibit A61. Percentage of Students Reporting that School Served School Meals Yesterday, and Frequency of 
Household Hunger
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
At schools that serve meals: 
School meal served yesterday 4.2 3.6 -0.6 77.4 1.1 -76.3 71.3 41.8 -29.5
Every month 28.6 61.7 33.1 15.0 27.6 12.6 16.5 24.5 8.1
During the rainy season 20.3 10.6 -9.7 12.3 7.5 -4.8 15.7 23.7 8.0
Other 51.1 27.8 -
23.3 72.8 64.9 -7.9 67.8 51.7 -16.1
At schools that serve meals: 
School meal served yesterday 5.4 5.0 -0.3 72.6 1.0 -71.6 68.9 35.1 -33.7
Every month 26.5 53.7 27.1 12.2 11.8 -0.4 33.3 23.3 -10.0
During the rainy season 13.1 3.7 -9.5 15.2 11.5 -3.8 14.3 19.8 5.4
Other 60.3 42.7 -
17.7 72.6 76.7 4.2 52.3 56.9 4.6
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Exhibit A62. Head teacher Reporting on School Feeding Program
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
School has a school feeding 
program
5.3 14.3 8.9 98.8 37.5 -61.3 100.0 64.7 -35.3
Organization supporting school feeding (if program is present)
PAI/ADPP 0.0 0.0 0.0 100.0 100.0 0.0 46.1 18.7 -27.4
WV 0.0 0.0 0.0 0.0 0.0 0.0 51.0 79.6 28.7
Community support 0.0 59.3 59.3 0.0 0.0 0.0 0.0 0.0 0.0
Other 100.0 40.7 -59.3 0.0 0.0 0.0 4.7 1.7 -3.0
Frequency of school meals (if program is present)
Daily 100.0 0.0 -100.0 91.6 0.0 -91.6 92.3 0.0 -92.3
A few times per week 0.0 78.3 78.3 4.8 100.0 95.2 2.9 98.1 95.2
Whenever food is available 0.0 21.7 21.7 2.4 0.0 -2.4 4.8 1.9 -3.0
Issues with school feeding program (if program is present)
No issues 10.0 33.3 23.3 30.1 45.5 15.3 18.1 26.7 8.5
Food is not always available 10.0 18.5 8.5 2.4 12.1 9.7 4.1 17.8 13.7
Kitchen/warehouse not 
constructed or in disrepair 30.0 18.5 -11.5 1.2 0.0 -1.2 4.2 9.8 5.6
Not enough pots, utensils, or 
dishes 40.0 3.7 -36.3 20.5 6.1 -14.4 31.5 24.2 -7.3
Not enough volunteer cooks 0.0 14.8 14.8 7.2 15.2 7.9 5.9 14.9 9.0
Other 70.0 25.9 -44.1 53.0 42.4 -10.6 68.3 47.2 -21.1
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌110
Exhibit A63. Impact on Teacher-Reported Nutrition and Hygiene Teaching
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Early-Grade Students
Teacher had training on teaching 
nutrition -3.8 -19.8 -16.0* -25.9 -22.1**
(7.1) (4.7)
Teacher has nutrition lesson plans -17.8 -3.5 14.3* -19.4 -1.7
(6.5) (5.8)
Teacher had training on teaching 
hygiene and sanitation -3.7 -0.2 3.5 -2.1 1.6
(7.3) (6.9)
Teacher has hygiene and 
sanitation lesson plans -18.9 -3.0 15.8* -2.5 16.4**
(6.4) (4.9)
Mid-Grade Students
Teacher had training on teaching 
nutrition -8.2 0.6 8.8 -11.6 -3.4
(7.0) (4.9)
Teacher has nutrition lesson plans -25.2 4.6 29.8** -15.3 9.9
(6.5) (5.3)
Teacher had training on teaching 
hygiene and sanitation 20.4 3.8 -16.6* 11.3 -9.1
(7.6) (6.4)
Teacher has hygiene and 
sanitation lesson plans -13.3 0.3 13.6 -9.6 3.7
(7.6) (6.5)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 632 Comparison; 311 School feeding only; 634 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A64. Teachers' Perception of Student Knowledge of Nutrition, Hygiene, and Sanitation
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Teacher rating of student knowledge of nutrition
Weak 38.5 29.7 -8.9 18.6 18.9 0.3 41.2 22.1 -19.1
Average 51.8 49.5 -2.3 72.4 72.1 -0.4 51.3 63.6 12.3
Strong 9.7 20.9 11.2 9.0 9.0 0.0 7.5 14.3 6.8
Teacher rating of student knowledge of hygiene and sanitation
Weak 22.1 9.2 -13.0 15.2 4.0 -11.2 28.6 6.7 -21.8
Average 56.3 55.7 -0.5 64.6 77.3 12.8 56.1 69.5 13.4
Strong 21.6 35.1 13.5 20.3 18.7 -1.6 15.3 23.8 8.4
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌111
Exhibit A65. Impact on Head Teacher Reported Teacher Training on Nutrition, Hygiene, and Health
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Teachers trained on teaching 
nutrition -58.2 -40.0 18.2 -115.1 -56.9
(64.0) (45.1)
Teachers trained on teaching 
hygiene and sanitation 10.9 38.3 27.4 -3.4 -14.3
(70.7) (52.0)
Teachers have materials for 
teaching nutrition 5.4 17.5 12.2 1.5 -3.9
(8.2) (6.1)
Teachers have materials for 
teaching hygiene and sanitation 22.7 21.6 -1.1 19.2 -3.5
(9.0) (6.8)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 279 Comparison; 171 School feeding only; 321 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A66. Head Teacher Reported Frequency of Nutrition, Hygiene, and Sanitation Lessons
........................................................ Comparison School Feeding 
Only
School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
Frequency of teaching nutrition lessons
Once per week 15.3 12.9 -2.4 64.1 62.2 -1.9 60.5 43.0 -17.5
Once every two weeks 6.1 5.4 -0.7 7.7 11.0 3.3 9.1 8.5 -0.7
Once per month 7.7 8.1 0.4 12.8 9.8 -3.1 14.2 13.1 -1.0
Once per term 38.6 0.0 -38.6 9.0 0.0 -9.0 6.1 0.0 -6.1
Never 32.3 52.4 20.1 6.4 7.3 0.9 10.1 21.7 11.6
Frequency of teaching hygiene and sanitation lessons
Once per week 35.0 29.3 -5.8 69.6 59.3 -10.3 66.5 56.4 -10.1
Once every two weeks 8.6 4.5 -4.1 6.3 8.1 1.8 8.5 8.7 0.2
Once per month 6.8 13.0 6.2 5.1 8.1 3.1 7.2 11.4 4.2
Once per term 26.3 0.0 -26.3 6.3 0.0 -6.3 7.1 0.0 -7.1
Never 23.3 29.3 6.0 12.7 8.1 -4.5 10.7 7.0 -3.7
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌112
Exhibit A67. Impact on Students Receiving Deworming Medication
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
Difference Difference Difference in 
Differences Difference Difference in 
Differences
(a) (b) (b)-(a) (c) (c)-(a)
Students are given deworming 
medication -15.6 -1.6 14.0 0.3 15.9**
(7.3) (5.9)
Students received deworming 
medication at least once during 
this school year
-39.8 -37.1 2.7 -22.2 17.6**
(8.5) (6.2)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 352 Comparison; 174 School feeding only; 346 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A68. Head Teacher Reported Cleanliness of latrines
........................................................ Comparison School Feeding Only School Feeding and 
Literacy
BL EL Diff BL EL Diff BL EL Diff
All latrines are dirty with feces on 
the slabs 16.8 5.3 -11.5 11.6 4.5 -7.0 13.0 8.8 -4.3
Some latrines are dirty 55.2 27.1 -28.1 44.9 28.4 -16.5 45.4 37.8 -7.7
All latrines are clean with no feces 
on the slabs 28.0 45.7 17.8 43.5 65.9 22.4 41.5 52.9 11.4
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌113
Exhibit A69. Cooks Reporting on School Infrastructure
........................................................ School Feeding Only School Feeding and Literacy
BL EL Diff BL EL Diff
School has a school feeding log book 92.3 52.5 -39.8 78.6 68.1 -10.5
Log book is filled out and up to date 84.6 81.0 -3.6 71.4 83.2 11.8
School has a kitchen and storeroom for 
school feeding program 98.7 97.5 -1.2 98.2 100.0 1.8
Cooking pots and cooking utensils (such as 
stirrers or scoops) are reported to be 
available
92.3 82.5 -9.8 86.2 88.6 2.4
Cooking pots and cooking utensils (such as 
stirrers or scoops) are reported to be 
sufficient
75.6 50.0 -25.6 57.2 54.2 -3.0
School has bowls and spoons for students 78.3 89.7 11.4 80.2 89.6 9.4
Number of bowls and spoons for students 
sufficient for each meal session 60.9 69.2 8.3 52.1 68.4 16.3
School has ever had a gap in the supply of 
corn-soy blend 37.2 62.1 24.9 50.0 77.3 27.3
Condition of the kitchen and storeroom
Not adequate (holes in roof or walls, no door, 
and/or no lock) 16.7 9.0 -7.7 19.7 9.6 -10.1
Moderate (complete building but needs 
repairs) 59.0 56.4 -2.6 52.4 58.3 5.9
Very good (complete building with no repairs 
needed) 23.1 34.6 11.5 28.0 32.1 4.1
School has clean water available 59.4 72.5 13.1 67.2 64.7 -2.5
Soap is available to clean bowls, spoons, 
pots, and cooking utensils 60.9 79.3 18.4 36.0 48.1 12.1
School has a place for cooks to wash their 
hands 73.9 73.8 -0.1 59.0 76.6 17.6
Soap is available for cooks to wash their 
hands 59.4 75.0 15.6 28.3 54.4 26.1
School has a place for children to wash their 
hands before eating 72.5 72.5 0 64.7 78.5 13.8
Soap is available for children to wash their 
hands before eating 44.9 72.5 27.6 18.9 53.9 35
School meals are served on time
Never 5.1 0.0 -5.1 3.8 1.9 -1.9
Sometimes 47.4 41.4 -6.0 47.3 41.3 -6.0
Always 47.4 58.6 11.2 48.9 56.8 7.9
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌114
Exhibit A70. Cook Training and Knowledge
........................................................ School Feeding Only School Feeding and Literacy
BL EL Diff BL EL Diff
School cook training received 40.6 70.0 29.4 47.0 80.8 33.8
Timing of last school cook training
Less than 6 months ago 18.8 0.0 -18.8 13.0 46.6 33.6
6- 11 months ago 7.2 3.6 -3.6 8.6 7.4 -1.2
12 or more months ago 14.5 96.4 81.9 26.2 46.0 19.8
Cook knowledge of food safety and hygiene
Cook's knowledge of food safety (8 items) 3.4 5.2 1.8 3.8 5.5 1.7
Cook's knowledge of hygiene (5 items) 2.3 2.8 0.5 2.7 3.2 0.5
Note: ""BL" indicates baseline; "EL" indicates endline; "Diff" indicates difference.
Abt Associates June 2022 ▌115
SUPPLEMENTAL FINDINGS RELATED TO IMPLEMENTATION BY ITEMS
Exhibit A71. Percentage of Schools that Adequately Implemented Literacy, Training, and Feeding Program 
Items
......................................................................................................................
.............
Comparison
School 
Feeding 
Only
School 
Feeding and 
Literacy
Literacy Items
Do you have a school literacy program? 36.2 56.8 71.8
Does the school have a library/book bank? 22.0 18.2 57.1
Does the school have a literacy afterschool program? 14.1 39.8 50.3
Are there other types of afterschool learning clubs? 14.1 39.8 50.3
Do you have a Teacher Guide for reading instruction? 70.6 77.3 80.0
Do you use the Teacher Guide for reading instruction? 71.8 80.7 81.7
Do each of your students have exercise books that can be used during 
reading instruction? 
93.8 89.8 91.7
Do your students use the exercise books during reading instruction? 97.7 97.7 97.2
Do you have access to school supplies? 100.0 98.9 99.4
Do you use any school supplies in your classroom during reading 
instruction?
99.4 100.0 100.0
Is there an afterschool literacy club in this school? 10.2 39.8 51.7
Do your students attend an afterschool literacy club at this school? 8.5 31.8 41.1
Training Items
Did you attend the teacher training held this academic year? 0.0 85.2 90.6
Did teachers at the school attend the literacy training this year? 0.0 81.8 92.8
Do reading coaches supervise teachers at the school? 30.5 58.0 85.6
Have you received training on the fundamentals of reading instruction? 36.7 71.6 79.4
Have you received training on school management? 71.8 84.1 84.4
During your pre-service training, did you receive any specific training on 
how to teach reading and writing to students in the primary grades?
79.1 73.9 87.4
This year, have you attended any in-service training on how to teach 
reading in the primary grades? 
46.3 46.6 67.8
Have you attended the WV/PAI training? 0.0 85.2 88.0
Have you received training on the fundamentals of reading instruction? 65.5 84.1 92.3
Does the Head Teacher or Assistant Head Teacher check your lesson 
plans?
100.0 100.0 100.0
Since the beginning of the school year, how frequently does the Assistant 
Head Teacher or Head Teacher observe your teaching?
96.6 98.9 98.9
Since the beginning of the school year, how frequently does the reading 
coach visit you?
23.7 45.5 72.7
Feeding Items
Does the school have a school feeding program? 15.3 37.5 63.3
How often do students receive the school meals? 10.2 36.4 58.2
Does the school have a garden? 13.6 33.0 20.9
How often is the garden used to supply food for school meals? 1.1 3.4 2.3
Does the school have a kitchen and storeroom for the school feeding 
program?
81.5 97.5 100.0
Are cooking pots and cooking utensils (such as stirrers or scoops) 
available?
96.3 82.5 88.5
Do you supplement the corn-soy blend with other foods? 3.7 8.8 1.3
Does the school have bowls and spoons for students? 77.8 32.5 60.3
Is there soap available to clean the bowls, spoons, pots, and cooking 
utensils?
92.6 28.8 32.7
Is there a supply of clean water (from a borehole or rainwater tank) 
available at the school?
85.2 72.5 65.4
Abt Associates June 2022 ▌116
......................................................................................................................
.............
Comparison
School 
Feeding 
Only
School 
Feeding and 
Literacy
Is there a place for children to wash their hands before eating? 100.0 72.5 78.8
Is there soap available for children to wash their hands before eating? 92.6 72.5 54.5
Did you receive any training for being a school cook? 44.4 70.0 80.1
Is there a place for cooks to wash their hands? 85.2 73.8 76.9
Is there soap for cooks to wash their hands? 96.3 75.0 55.1
Exhibit A72. Impact on Reading Comprehension by Literacy Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Do you have a school literacy program? 2.1 4.4*
(3.5) (2.0)
Does the school have a library/book bank? 4.6 4.1*
(6.0) (1.9)
Does the school have a literacy afterschool program? 4.1 6.6**
(4.4) (2.4)
Are there other types of afterschool learning clubs? 4.1 6.6**
(4.4) (2.4)
Do you have a Teacher Guide for reading instruction? 2.3 4.7*
(3.1) (1.9)
Do you use the Teacher Guide for reading instruction? 0.5 4.7*
(3.2) (1.9)
Do each of your students have exercise books that can be used during reading 
instruction? 1.6 5.8**
(3.1) (1.9)
Do your students use the exercise books during reading instruction? 1.7 4.8*
(3.0) (1.9)
Do you have access to school supplies? 1.4 4.9**
(2.9) (1.9)
Do you use any school supplies in your classroom during reading instruction? 1.2 4.8**
(2.9) (1.9)
Is there an afterschool literacy club in this school? 2.8 2.5
(4.6) (2.2)
Do your students attend an afterschool literacy club at this school? -2.8 5.0
(3.0) (2.6)
Mid-Grade Students
Do you have a school literacy program? -3.7 7.1*
(4.6) (3.5)
Does the school have a library/book bank? -4.2 9.0*
(7.1) (3.5)
Does the school have a literacy afterschool program? -4.3 8.6*
(5.3) (3.9)
Are there other types of afterschool learning clubs? -4.3 8.6*
(5.3) (3.9)
Do you have a Teacher Guide for reading instruction? -1.4 9.8**
(4.3) (3.2)
Do you use the Teacher Guide for reading instruction? -4.9 9.8**
(4.8) (3.2)
Abt Associates June 2022 ▌117
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Do each of your students have exercise books that can be used during reading 
instruction? -2.5 8.5*
(4.7) (3.4)
Do your students use the exercise books during reading instruction? -2.3 7.6*
(4.6) (3.4)
Do you have access to school supplies? -1.7 7.7*
(4.6) (3.3)
Do you use any school supplies in your classroom during reading instruction? -1.8 7.6*
(4.6) (3.3)
Is there an afterschool literacy club in this school? -2.3 12.5**
(5.4) (3.3)
Do your students attend an afterschool literacy club at this school? 1.0 12.6**
(4.5) (3.5)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,474 Comparison; 1,436 School feeding only; 3,951 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A73. Impact on Reading Comprehension by Training Item Adequately Implemented by Schools in 
the Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Did you attend the teacher training held this academic year? 0.5 4.7*
(3.1) (1.9)
Did teachers at the school attend the literacy training this year? 3.7 4.9*
(3.1) (1.9)
Do reading coaches supervise teachers at the school? 1.5 4.6*
(3.1) (1.9)
Have you received training on the fundamentals of reading instruction? 2.8 3.5
(3.0) (1.8)
Have you received training on school management? 1.6 5.3**
(2.7) (2.0)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? -0.7 5.4**
(3.1) (2.0)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? -3.4 6.9**
(2.5) (2.0)
Have you attended the WV/PAI training? 0.5 5.4**
(3.2) (1.9)
Have you received training on the fundamentals of reading instruction? 1.0 5.1**
(3.2) (1.9)
Does the Head Teacher or Assistant Head Teacher check your lesson plans? 1.2 4.8**
(2.9) (1.9)
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? 1.3 4.3*
(3.0) (1.8)
Abt Associates June 2022 ▌118
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Since the beginning of the school year, how frequently does the reading coach 
visit you? 3.7 4.5*
(4.0) (1.9)
Mid-Grade Students
Did you attend the teacher training held this academic year? -1.7 7.3*
(4.9) (3.4)
Did teachers at the school attend the literacy training this year? 0.1 7.0*
(4.3) (3.3)
Do reading coaches supervise teachers at the school? 1.1 7.3*
(4.1) (3.4)
Have you received training on the fundamentals of reading instruction? 1.7 7.4*
(3.9) (3.5)
Have you received training on school management? 3.6 8.0*
(3.8) (3.5)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? -2.0 8.5*
(5.1) (3.4)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? 2.3 11.8**
(4.2) (3.2)
Have you attended the WV/PAI training? -2.0 8.7*
(4.9) (3.4)
Have you received training on the fundamentals of reading instruction? -2.1 7.3*
(4.9) (3.4)
Does the Head Teacher or Assistant Head Teacher check your lesson plans? -1.8 7.6*
(4.6) (3.3)
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? -1.6 7.4*
(4.6) (3.3)
Since the beginning of the school year, how frequently does the reading coach 
visit you? 0.1 10.2**
(5.6) (3.4)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,474 Comparison; 2,174 School feeding only; 4,919 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A74. Impact on Reading Comprehension by Feeding Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Does the school have a school feeding program? -0.9 4.9*
(4.1) (2.1)
How often do students receive the school meals? -0.7 4.5*
(4.2) (2.1)
Does the school have a garden? 2.3 6.4
(4.8) (4.4)
How often is the garden used to supply food for school meals? -5.2 21.1**
Abt Associates June 2022 ▌119
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
(3.8) (4.4)
Does the school have a kitchen and storeroom for the school feeding program? 0.7 4.7*
(3.1) (2.0)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? 0.1 5.2*
(3.4) (2.0)
Do you supplement the corn-soy blend with other foods? -6.7* 3.6
(3.0) (2.3)
Does the school have bowls and spoons for students? 2.9 5.9**
(4.9) (2.1)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? -0.6 3.8
(4.9) (2.9)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? -0.4 5.3*
(3.2) (2.3)
Is there a place for children to wash their hands before eating? 1.8 4.8*
(3.5) (2.1)
Is there soap available for children to wash their hands before eating? 0.4 3.0
(3.6) (2.3)
Did you receive any training for being a school cook? 0.2 5.1**
(3.8) (1.9)
Is there a place for cooks to wash their hands? 2.0 4.7*
(3.6) (2.0)
Is there soap for cooks to wash their hands? 0.8 3.5
(3.5) (2.3)
Mid-Grade Students
Does the school have a school feeding program? -7.2 12.1**
(6.2) (2.9)
How often do students receive the school meals? -7.0 12.4**
(6.3) (2.9)
Does the school have a garden? -4.8 7.3
(6.1) (6.9)
How often is the garden used to supply food for school meals? 14.5 10.7*
(10.0) (4.2)
Does the school have a kitchen and storeroom for the school feeding program? -1.4 7.9*
(4.8) (3.3)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? -1.0 7.6*
(5.3) (3.5)
Do you supplement the corn-soy blend with other foods? 6.3 16.6**
(7.3) (5.7)
Does the school have bowls and spoons for students? -0.3 13.5**
(6.4) (2.8)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? -7.5 14.8**
(7.1) (3.6)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? -4.1 6.2
(5.1) (3.8)
Is there a place for children to wash their hands before eating? -3.0 7.7*
(5.2) (3.6)
Is there soap available for children to wash their hands before eating? -3.1 8.9*
(5.4) (4.3)
Did you receive any training for being a school cook? -3.2 11.2**
(5.7) (3.2)
Abt Associates June 2022 ▌120
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Is there a place for cooks to wash their hands? -2.4 8.0*
(5.4) (3.6)
Is there soap for cooks to wash their hands? -2.1 7.9
(5.3) (4.4)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,474 Comparison; 950 School feeding only; 3,479 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A75. Impact on Student Attendance by Literacy Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Do you have a school literacy program? 3.0 -5.4
(5.8) (3.8)
Does the school have a library/book bank? 3.3 -4.6
(9.5) (3.9)
Does the school have a literacy afterschool program? -0.2 -6.0
(7.3) (4.3)
Are there other types of afterschool learning clubs? -0.2 -6.0
(7.3) (4.3)
Do you have a Teacher Guide for reading instruction? 1.2 -7.4*
(5.5) (3.5)
Do you use the Teacher Guide for reading instruction? -3.3 -7.3*
(6.7) (3.5)
Do each of your students have exercise books that can be used during reading 
instruction? -3.3 -5.9
(6.2) (3.7)
Do your students use the exercise books during reading instruction? -3.6 -6.3
(6.0) (3.6)
Do you have access to school supplies? -4.6 -5.7
(5.9) (3.6)
Do you use any school supplies in your classroom during reading instruction? -4.6 -5.7
(5.8) (3.6)
Is there an afterschool literacy club in this school? -1.1 -10.2**
(5.5) (3.9)
Do your students attend an afterschool literacy club at this school? -3.7 -10.6*
(6.6) (4.2)
Mid-Grade Students
Do you have a school literacy program? 3.8 2.6
(6.1) (4.2)
Does the school have a library/book bank? 12.9* 0.8
(5.6) (4.5)
Does the school have a literacy afterschool program? 7.0 1.4
(5.1) (4.7)
Are there other types of afterschool learning clubs? 7.0 1.4
(5.1) (4.7)
Do you have a Teacher Guide for reading instruction? 1.1 2.8
Abt Associates June 2022 ▌121
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
(5.3) (4.0)
Do you use the Teacher Guide for reading instruction? 2.7 3.0
(4.9) (4.0)
Do each of your students have exercise books that can be used during reading 
instruction? 4.7 2.1
(4.7) (3.7)
Do your students use the exercise books during reading instruction? 4.5 4.4
(4.6) (3.8)
Do you have access to school supplies? 4.7 3.7
(4.5) (3.8)
Do you use any school supplies in your classroom during reading instruction? 4.8 3.9
(4.5) (3.8)
Is there an afterschool literacy club in this school? 7.4 -1.4
(5.0) (4.1)
Do your students attend an afterschool literacy club at this school? 5.0 0.2
(5.3) (4.3)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 1,435 School feeding only; 3,941 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A76. Impact on Student Attendance by Training Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Did you attend the teacher training held this academic year? -7.9 -5.3
(6.0) (3.7)
Did teachers at the school attend the literacy training this year? -2.1 -5.9
(5.5) (3.6)
Do reading coaches supervise teachers at the school? -4.1 -5.5
(6.4) (3.6)
Have you received training on the fundamentals of reading instruction? -6.9 -5.9
(5.6) (3.5)
Have you received training on school management? -0.4 -5.6
(5.7) (3.8)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? -7.5 -5.9
(6.4) (3.7)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? 4.3 -3.2
(5.3) (3.5)
Have you attended the WV/PAI training? -3.7 -5.3
(6.3) (3.6)
Have you received training on the fundamentals of reading instruction? -5.4 -5.8
(6.2) (3.6)
Does the Head Teacher or Assistant Head Teacher check your lesson plans? -4.6 -5.7
(5.8) (3.6)
Abt Associates June 2022 ▌122
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? -5.0 -4.5
(5.9) (3.5)
Since the beginning of the school year, how frequently does the reading coach 
visit you? 4.3 -7.3*
(4.8) (3.5)
Mid-Grade Students
Did you attend the teacher training held this academic year? 5.8 1.8
(4.7) (3.7)
Did teachers at the school attend the literacy training this year? 6.3 3.8
(4.1) (3.9)
Do reading coaches supervise teachers at the school? 0.1 3.0
(5.7) (3.9)
Have you received training on the fundamentals of reading instruction? 5.8 1.9
(5.8) (4.1)
Have you received training on school management? 6.8 2.6
(4.1) (3.8)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? 3.2 3.5
(4.9) (3.9)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? -1.1 -0.1
(7.6) (4.1)
Have you attended the WV/PAI training? 4.5 4.3
(4.7) (3.8)
Have you received training on the fundamentals of reading instruction? 4.0 4.2
(4.8) (3.9)
Does the Head Teacher or Assistant Head Teacher check your lesson plans? 4.8 3.9
(4.5) (3.8)
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? 4.5 3.8
(4.5) (3.9)
Since the beginning of the school year, how frequently does the reading coach 
visit you? -2.8 1.7
(7.4) (4.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 2,172 School feeding only; 4,907 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A77. Impact on Student Attendance by Feeding Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Does the school have a school feeding program? -3.8 -6.8*
(8.9) (3.4)
How often do students receive the school meals? -3.1 -7.1*
(9.0) (3.5)
Abt Associates June 2022 ▌123
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Does the school have a garden? 1.9 -11.3*
(6.8) (5.5)
How often is the garden used to supply food for school meals? -6.8 -34.5**
(17.9) (6.0)
Does the school have a kitchen and storeroom for the school feeding program? -3.1 -5.9
(6.2) (3.7)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? -4.3 -5.8
(6.5) (3.7)
Do you supplement the corn-soy blend with other foods? -12.4 -21.9
(9.9) (21.0)
Does the school have bowls and spoons for students? 2.9 -8.1*
(6.8) (3.5)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? -6.5 -6.4
(9.0) (3.9)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? -6.0 -4.4
(6.8) (4.1)
Is there a place for children to wash their hands before eating? -3.1 -5.3
(7.0) (3.7)
Is there soap available for children to wash their hands before eating? -4.6 -4.0
(6.9) (3.9)
Did you receive any training for being a school cook? -3.7 -7.4*
(7.3) (3.6)
Is there a place for cooks to wash their hands? -1.3 -5.0
(7.1) (3.6)
Is there soap for cooks to wash their hands? -2.9 -4.0
(6.9) (3.9)
Mid-Grade Students
Does the school have a school feeding program? 6.0 -1.9
(6.4) (4.0)
How often do students receive the school meals? 6.4 -2.0
(6.4) (4.1)
Does the school have a garden? 11.7* 6.9
(5.1) (7.0)
How often is the garden used to supply food for school meals? 33.4** 2.7
(8.5) (5.7)
Does the school have a kitchen and storeroom for the school feeding program? 4.1 2.3
(4.6) (3.7)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? 4.1 1.3
(5.0) (3.8)
Do you supplement the corn-soy blend with other foods? 13.5* 18.2
(5.8) (19.2)
Does the school have bowls and spoons for students? 2.5 -5.1
(8.7) (3.8)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? 3.4 -4.2
(7.3) (4.8)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? 1.9 2.4
(5.1) (4.1)
Is there a place for children to wash their hands before eating? 2.7 4.1
(5.1) (4.1)
Is there soap available for children to wash their hands before eating? 4.6 4.6
Abt Associates June 2022 ▌124
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
(5.2) (4.4)
Did you receive any training for being a school cook? 3.5 0.4
(5.5) (3.9)
Is there a place for cooks to wash their hands? 3.6 2.1
(5.1) (4.0)
Is there soap for cooks to wash their hands? 4.3 4.5
(5.0) (4.6)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 5,473 Comparison; 948 School feeding only; 3,467 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A78. Impact on Student Attentiveness by Literacy Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Do you have a school literacy program? -7.3 -5.7
(4.1) (3.6)
Does the school have a library/book bank? -7.4* -1.4
(3.6) (3.2)
Does the school have a literacy afterschool program? -7.3 -2.1
(4.7) (3.7)
Are there other types of afterschool learning clubs? -7.3 -2.1
(4.7) (3.7)
Do you have a Teacher Guide for reading instruction? -5.4 -3.7
(4.3) (3.7)
Do you use the Teacher Guide for reading instruction? -5.5 -3.7
(4.2) (3.7)
Do each of your students have exercise books that can be used during reading 
instruction? -3.1 -4.5
(4.2) (3.6)
Do your students use the exercise books during reading instruction? -3.4 -4.0
(4.2) (3.5)
Do you have access to school supplies? -3.5 -4.1
(4.1) (3.5)
Do you use any school supplies in your classroom during reading instruction? -3.5 -4.1
(4.1) (3.5)
Is there an afterschool literacy club in this school? -3.9 -6.7
(5.9) (4.3)
Do your students attend an afterschool literacy club at this school? -3.3 -3.3
(7.3) (4.2)
Mid-Grade Students
Do you have a school literacy program? -10.6** -2.6
(3.4) (3.0)
Does the school have a library/book bank? -0.9 -2.4
(6.0) (3.1)
Does the school have a literacy afterschool program? -8.1** -2.1
(2.9) (3.0)
Abt Associates June 2022 ▌125
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Are there other types of afterschool learning clubs? -8.1** -2.1
(2.9) (3.0)
Do you have a Teacher Guide for reading instruction? -5.2 -3.0
(3.4) (2.9)
Do you use the Teacher Guide for reading instruction? -7.1* -3.6
(3.6) (2.9)
Do each of your students have exercise books that can be used during reading 
instruction? -5.5 -3.2
(3.3) (2.9)
Do your students use the exercise books during reading instruction? -7.2* -2.9
(3.4) (2.9)
Do you have access to school supplies? -7.2* -3.0
(3.3) (2.9)
Do you use any school supplies in your classroom during reading instruction? -7.2* -3.0
(3.3) (2.9)
Is there an afterschool literacy club in this school? -3.0 -3.5
(4.5) (2.9)
Do your students attend an afterschool literacy club at this school? -8.1** -2.6
(3.0) (3.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 632 Comparison; 177 School feeding only; 448 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A79. Impact on Student Attentiveness by Training Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Did you attend the teacher training held this academic year? -3.1 -6.0
(4.3) (3.5)
Did teachers at the school attend the literacy training this year? -2.6 -4.0
(4.4) (3.5)
Do reading coaches supervise teachers at the school? -2.6 -4.4
(4.5) (3.6)
Have you received training on the fundamentals of reading instruction? -3.8 -2.3
(4.6) (3.3)
Have you received training on school management? -4.5 -5.6
(4.3) (3.5)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? -2.4 -3.6
(4.3) (3.6)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? -1.8 -3.0
(5.2) (3.4)
Have you attended the WV/PAI training? -3.2 -3.3
(4.3) (3.6)
Have you received training on the fundamentals of reading instruction? -5.7 -3.9
(4.2) (3.5)
Abt Associates June 2022 ▌126
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Does the Head Teacher or Assistant Head Teacher check your lesson plans? -3.5 -4.1
(4.1) (3.5)
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? -3.5 -3.9
(4.1) (3.5)
Since the beginning of the school year, how frequently does the reading coach 
visit you? -4.3 -4.4
(5.4) (3.7)
Mid-Grade Students
Did you attend the teacher training held this academic year? -5.7 -3.7
(3.3) (2.9)
Did teachers at the school attend the literacy training this year? -7.0 -3.4
(3.6) (2.9)
Do reading coaches supervise teachers at the school? -7.4* -3.2
(2.9) (2.8)
Have you received training on the fundamentals of reading instruction? -7.3 -2.6
(3.8) (3.0)
Have you received training on school management? -7.5* -3.9
(3.5) (2.9)
During your pre-service training, did you receive any specific training on how to 
teach reading and writing to students in the primary grades? -9.5** -3.2
(3.2) (2.9)
This year, have you attended any in-service training on how to teach reading in 
the primary grades? -3.4 -2.5
(4.4) (2.9)
Have you attended the WV/PAI training? -5.7 -3.1
(3.3) (2.9)
Have you received training on the fundamentals of reading instruction? -7.2* -3.4
(3.5) (2.9)
Does the Head Teacher or Assistant Head Teacher check your lesson plans? -7.2* -3.0
(3.3) (2.9)
Since the beginning of the school year, how frequently does the Assistant Head 
Teacher or Head Teacher observe your teaching? -5.8 -2.9
(3.2) (2.9)
Since the beginning of the school year, how frequently does the reading coach 
visit you? -3.9 -4.2
(4.0) (2.9)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 632 Comparison; 265 School feeding only; 562 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Exhibit A80. Impact on Student Attentiveness by Feeding Item Adequately Implemented by Schools in the 
Treatment Arms
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
Early-Grade Students
Does the school have a school feeding program? -6.9 -4.4
(4.5) (4.0)
Abt Associates June 2022 ▌127
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
How often do students receive the school meals? -7.1 -6.7
(4.5) (3.9)
Does the school have a garden? -1.0 -8.4
(5.0) (6.0)
How often is the garden used to supply food for school meals? -7.8* -6.9
(3.9) (3.6)
Does the school have a kitchen and storeroom for the school feeding program? -4.3 -4.1
(4.2) (3.6)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? -3.9 -3.4
(4.5) (3.7)
Do you supplement the corn-soy blend with other foods? -7.9* -1.0
(3.9) (2.4)
Does the school have bowls and spoons for students? -7.2 -3.2
(3.8) (4.7)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? -6.9 -2.2
(3.8) (6.4)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? -3.9 -8.0*
(4.7) (3.7)
Is there a place for children to wash their hands before eating? -2.0 -6.0
(4.5) (3.7)
Is there soap available for children to wash their hands before eating? -2.1 -4.8
(4.5) (4.2)
Did you receive any training for being a school cook? -3.9 -1.1
(4.8) (3.4)
Is there a place for cooks to wash their hands? -2.0 -6.0
(4.5) (3.7)
Is there soap for cooks to wash their hands? -2.2 -4.9
(4.5) (4.1)
Mid-Grade Students
Does the school have a school feeding program? -3.5 -2.9
(4.2) (3.1)
How often do students receive the school meals? -3.6 -2.7
(4.3) (3.1)
Does the school have a garden? -8.3** -5.3
(3.1) (3.2)
How often is the garden used to supply food for school meals? -8.6** -6.9*
(3.1) (2.9)
Does the school have a kitchen and storeroom for the school feeding program? -7.1* -3.1
(3.5) (2.9)
Are cooking pots and cooking utensils (such as stirrers or scoops) available? -7.1 -3.3
(3.7) (2.9)
Do you supplement the corn-soy blend with other foods? -8.6** -1.7
(3.1) (3.4)
Does the school have bowls and spoons for students? -1.9 -1.7
(5.3) (3.4)
Is there soap available to clean the bowls, spoons, pots, and cooking utensils? -1.5 -1.3
(5.4) (3.3)
Is there a supply of clean water (from a borehole or rainwater tank) available at 
the school? -8.1** -4.2
(2.9) (2.9)
Is there a place for children to wash their hands before eating? -5.1 -3.2
Abt Associates June 2022 ▌128
...................................................................................................................................
School Feeding 
Only
School Feeding 
and Literacy
Difference in 
Differences
Difference in 
Differences
(3.6) (2.9)
Is there soap available for children to wash their hands before eating? -5.2 -2.6
(3.6) (3.1)
Did you receive any training for being a school cook? -7.6 -1.7
(3.9) (3.0)
Is there a place for cooks to wash their hands? -5.2 -3.3
(3.6) (2.9)
Is there soap for cooks to wash their hands? -5.2 -3.0
(3.6) (3.0)
Note: "Early-Grade Students" were in grade 2 at baseline and grade 3 at endline. "Mid-Grade Students" were in grade 4 at baseline and grade 
5 at endline. Standard errors appear in parentheses. Sample sizes: 632 Comparison; 120 School feeding only; 391 Literacy and school 
feeding. Sample sizes vary for certain outcomes due to missing data. Asterisks indicate statistical significance of difference-in-differences
estimates. ** p<0.01; * p<0.05. 
Abt Associates June 2022 ▌129
OUTCOME LEVELS AT BASELINE AND ENDLINE BY TREATMENT ARM
Exhibit A811. Outcome Levels at Baseline and Endline by Treatment Arm for Main Outcomes
........................................................ Baseline Endline
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Early-Grade Students
Student reading performance 
(Average Reading Comprehension 
Scores) 
4.9 2.0 2.2 10.8 12.2 13.1
Student reading performance 
(Percent Zero Comprehension 
Scores) 
88.0 93.2 92.6 81.6 73.4 75.0
Student attendance 28.7 30.1 25.2 27.1 20.8 18.5
Student attentiveness 92.2 95.6 95.4 96.9 99.0 96.7
Mid-Grade Students
Student reading performance 
(Average Reading Comprehension 
Scores) 
26.7 23.1 14.2 41.6 35.1 33.3
Student reading performance 
(Percent Zero Comprehension 
Scores) 
52.3 57.8 70.9 38.4 40.6 45.1
Student attendance 27.9 27.1 24.7 22.1 24.5 21.7
Student attentiveness 91.2 98.7 96.9 96.2 96.8 98.1
Note: Table reports weighted average outcome by treatment arm at baseline and endline. Difference-in-differences calculated by simple 
subtraction will differ from impact estimates because impacts are based on the impact analysis models described in Section 3.4. 
Exhibit A822. Outcome Levels at Baseline and Endline by Treatment Arm for Instructional Inputs
........................................................ Baseline Endline
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Early-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
57.6 53.8 67.2 53.4 71.8 84.0
Teacher-reported absence of at 
least one day last week 22.6 27.5 25.0 23.8 13.6 24.6
Teacher has access to school 
supplies 86.5 97.8 96.4 98.0 99.0 97.4
Mid-Grade Students
Teacher-reported receipt of 
training on the fundamentals of 
reading instruction
55.2 61.8 58.7 51.4 60.0 78.4
Teacher-reported absence of at 
least one day last week 17.4 27.6 23.2 16.5 26.3 24.4
Teacher has access to school 
supplies 89.0 92.0 94.6 97.3 97.9 95.2
Note: Table reports weighted average outcome by treatment arm at baseline and endline. Difference-in-differences calculated by simple 
subtraction will differ from impact estimates because impacts are based on the impact analysis models described in Section 3.4. 
Abt Associates June 2022 ▌130
Exhibit A833. Outcome Levels at Baseline and Endline by Treatment Arm for Feeding Outcomes and 
Nutrition Knowledge
........................................................ Baseline Endline
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Early-Grade Students
Student reports feeling hungry now 55.0 39.7 51.4 47.8 38.1 37.9
Student achieved minimum meal 
frequency 42.7 55.3 57.6 49.9 41.9 48.0
Student achieved minimum dietary 
diversity 67.8 53.5 71.8 38.2 34.6 43.2
Student achieved minimum 
acceptable diet 34.1 30.5 46.0 26.5 20.1 23.8
Student's nutrition knowledge 
score 
0.7 0.4 0.4 0.3 0.6 0.8
Mid-Grade Students
Student reports feeling hungry now 46.0 30.7 39.8 43.5 41.8 37.8
Student achieved minimum meal 
frequency 56.9 57.4 62.4 62.0 54.7 59.0
Student achieved minimum dietary 
diversity 71.8 56.0 76.3 50.0 29.9 40.7
Student achieved minimum 
acceptable diet 45.2 36.3 49.6 39.1 21.2 26.5
Student's nutrition knowledge 
score 
1.3 1.0 0.7 0.8 1.1 1.0
Note: Table reports weighted average outcome by treatment arm at baseline and endline. Difference-in-differences calculated by simple 
subtraction will differ from impact estimates because impacts are based on the impact analysis models described in Section 3.4. 
Exhibit A844. Outcome Levels at Baseline and Endline by Treatment Arm for Student Knowledge and Use of 
Recommended Health Practices
........................................................ Baseline Endline
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Comparison
School 
Feeding 
Only
School 
Feeding 
and 
Literacy
Early-Grade Students
Student's hygiene and sanitation 
knowledge score 2.9 2.6 2.5 3.6 3.5 4.2
Student reports absence last week 
as a result of illness 59.3 66.4 55.8 55.7 51.7 70.2
Mid-Grade Students
Student's hygiene and sanitation 
knowledge score 3.5 3.5 3.1 4.6 4.3 4.5
Student reports absence last week 
as a result of illness 58.3 57.3 57.5 64.5 55.7 67.3
All Students
Head teacher reports school has 
improved latrines 39.5 82.1 86.0 51.9 90.9 95.4
Head teacher reports school has 
tippy taps or hand-washing 
stations
12.6 33.3 38.8 36.9 55.7 54.2
Abt Associates June 2022 ▌131
Head teacher reports all tippy taps 
or hand-washing stations have ash 1.1 8.3 3.5 24.0 29.5 21.4
or soap
Note: Table reports weighted average outcome by treatment arm at baseline and endline. Difference-in-differences calculated by simple 
subtraction will differ from impact estimates because impacts are based on the impact analysis models described in Section 3.4.
APPENDIX B: DATA COLLECTION INSTRUMENTS 
Data Collection Instruments 
1. Student Survey (English)
2. Head Teacher Questionnaire (English)
3. Teacher Questionnaire (English)
4. Cook/Warehouse Manager Questionnaire (English)
5. Ministry of Education Interview (English)
6. Parent Focus Groups & Discussion Guide (English)
Abt Associates June 2022 ▌132
1. Student Questionnaire (English)
1. Is the student a girl?
2. How old are you?
3. What class are you in right now?
a. 3rd Grade
b. 5th Grade
4. What class were you in last year?
a. 1st Grade
b. 2nd Grade
c. 3rd Grade
d. 4th Grade
e. 5th Grade
f. Don’t Know/Refuse to Respond
5. Did you go to pre-primary?
6. I would like to see what schoolbooks you have with you today. Please show me your:
a. School Books
b. Supplementary reader
c. textbook
d. exercise book
7. During the teaching of reading, what language does your teacher use to teach you and your
classmates?
a. Português
b. Ronga
c. Changana
d. Macua
e. Other
8. What does the teacher normally do when you are unable to answer a question or you answer a
question incorrectly?
a. Teacher rephrases/explains the question
b. Teacher provides an example
c. Teacher encourages the student to try again
d. Teacher asks another student
e. Teacher asks again
f. Teacher corrects the student
Abt Associates June 2022 ▌133
g. Teacher assigns a student helper
h. Other
i. Don’t Know/Refuse to Respond
9. Do you have someone at home who helps you with TPC?
a. No
b. Yes
c. I don’t have TPC
d. Don’t Know/Refuse to Respond
10. Were you absent from school any day last week? Why were you absent?
a. No, was not absent from school last week
b. Yes, illness
c. Yes, because there was other work at home
d. Yes, because I had to take care of a family member
e. Yes, no transportation
f. Yes, because of bad weather
g. Yes, because of an emergency
h. Yes, because getting to school was too dangerous
i. Yes, because I woke up late
j. Yes, because I could not find my uniform, or because my uniform was not ready on time
in the morning
k. Yes, because I am treated poorly by teachers at school
l. Yes, because I am treated poorly by other students at school
m. Other
n. Don’t Know/Refuse to Respond
11. Do you read books in your classroom or in your school library every day?
12. Do you attend an afterschool learning club?
13. What type (s) of afterschool learning club do you attend?
a. Afterschool Reading/Writing Club
b. Help with Homework Club
c. Math Learning Club
d. Science Learning Club
e. School Garden Learning Club
f. Other
14. Do you bring home reading books from your classroom?
Abt Associates June 2022 ▌134
15. If you bring home reading books from your classroom, what language are they in?
a. Português
b. Ronga
c. Changana
d. Macua
e. Other
16. Do you bring home books from the school library/book bank?
17. If you bring home reading books from the school library/book bank, what language are they in?
a. Português
b. Ronga
c. Changana
d. Macua
e. Other
18. What language do you normally speak with your family at home?
a. Português
b. Ronga
c. Changana
d. Macua
e. Other
19. In addition to school books, do you also read newspapers or other materials at home?
20. How often do you read on your own at home? Never, sometimes, or every day?
a. Never
b. Sometimes
c. Every day
d. Don’t Know/Refuse to Respond
21. How many times do you read aloud for everyone to hear when you're at home? Never, sometimes
or every day?
a. Never
b. Sometimes
c. Every day
d. Don’t Know/Refuse to Respond
22. How many times do you have someone read to you at home? Never, sometimes or every day?
a. Never
b. Sometimes
Abt Associates June 2022 ▌135
c. Every day
d. Don’t Know/Refuse to Respond
23. When you're at home, how often do you tell stories to others? Never, sometimes or
24. every day?
a. Never
b. Sometimes
c. Every day
d. Don’t Know/Refuse to Respond
25. When you're at home, how often do they tell you stories? Never, sometimes or every day?
a. Never
b. Sometimes
c. Every day
d. Don’t Know/Refuse to Respond
26. Do you have drawing or writing material at home?
27. Do you usually draw or write when you're at home?
28. How often do you draw or write? Never, sometimes or every day?
a. Never
b. Sometimes
c. Every day
d. Don’t Know/Refuse to Respond
29. Do you tend to be very attentive in classes?
30. What are you doing in the classroom to show that you are always paying attention?
a. Sit facing the teacher and looking them in the eye when they’re speaking.
b. Listen attentively and follow the teacher’s explanation
c. Work alone, in pairs, or in small groups during learning activities.
d. Answer all the questions that the teacher asks
e. I ask the teacher or aide when I don’t understand what was explained.
f. I ask questions about the explanation given
g. Read
h. Write
i. Listen to the story being told.
j. Other
31. Which phrase best matches your level of attention in the classroom?
a. My attention is greater before school lunch
Abt Associates June 2022 ▌136
b. My attention is much higher after school lunch
c. My level of attention before and after school lunch is almost always the same
d. Don’t Know/Refuse to Respond
32. Is there a certificate or award in school to celebrate progress in reading?
33. Have you ever received a distinction or diploma to celebrate your success in reading?
34. Did you eat any food before you arrived at school today?
a. No
b. Yes, at home
c. Yes, on the way to school
d. Other, specify
e. Don’t Know/Refuse to Respond
35. How hungry are you right now
a. Very hungry
b. Somewhat hungry
c. Just right
d. Somewhat full
e. Very full
36. Does your school serve school meals?
37. Was a school meal served yesterday or the last day you were at school?
38. Did you eat the school meal yesterday or the last day you were at school??
a. No
b. Yes, breakfast
c. Yes, lunch
d. Don’t Know/Refuse to Respond
39. What type of food was served during the school meal yesterday or the last day you were at
school?
a. Porridge only
b. Porridge plus other foods, specify types of other foods
c. Other foods only, specify types of other foods
d. Don’t Know/Refuse to Respond
40. On the days when you eat a meal at school, do you usually get:
a. Same number of meals at home
b. Fewer number of meals at home
41. On the days when you eat a meal at school, do you feel:
Abt Associates June 2022 ▌137
a. More hungry than usual
b. Less hungry than usual
c. Same amount hungry as usual
42. Please describe everything you ate yesterday during the day or night, whether at home or outside
the home.
43. Starting from the morning, what was the first thing you ate after waking up? Please tell me
everything you ate at that time.
44. PROBE IF MIXED DISH: What ingredients were in the dish? Anything else?
45. Yesterday during the day or at night, did you eat or drink:
a. Any foods made from grains, like rice, maize, spaghetti, sorghum, bread, porridge, or
nsima?
b. Any vegetables or roots that are orange-colored inside, like pumpkin, carrots, or sweet
potatoes that are orange inside?
c. Any white roots and tubers or plantains, such as white potatoes (batatas reno),
manioc/cassava, plantains, or other roots that are white inside?
d. Any dark green leafy vegetables, such as chinese cabbage, cassava leaves, mustard
greens, rape, or dark green lettuce
e. Any fruits that are dark yellow or orange inside, like ripe mango or ripe papaya?
f. Any other fruits like banana, maca, lemon, orange, tangerine, guava, grapes, or apple?
g. Any other vegetables like tomato, onion, cauliflower, eggplant, mushroom, cucumber, or
cabbage?
h. Any meat made from animal organs, such as gizzards, liver, kidney, heart or other organ
meats
i. Any other types of meat or poultry, like beef, pork, lamb, goat, chicken, or duck?
j. Any eggs from chicken, duck, or other birds?
k. Any fish or seafood, whether fresh or dried fish, prawns, shrimp, crab, clams, or other
seafood?
l. Any beans or peas, such as beans, lentils, or peas
m. Any nuts or seeds, like groundnuts, almonds, pumpkin seeds, or sesame seeds
n. Any milk or milk products, such as fresh milk, powdered milk, yogurt, or cheese
o. Insects, such as termites, snails, or spiders?
p. Oils and fats, such as vegetable oil, butter, or margarine
q. Fortified foods, such as corn-soy porridge eaten at school?
46. Yesterday during the day and at night, how many times did you eat, including meals and snacks?
Abt Associates June 2022 ▌138
47. How many food groups do you know in Mozambique?
48. What are the names of the Mozambican food groups?
a. Staple Foods (base foods)
b. Foods to grow (protein rich foods)
c. Foods to protect (protein-rich foods)
d. Energy-rich foods (high in calories)
49. Name one type of food that has a lot of protein (or is a food to grow)
a. Fruit
b. Vegetables
c. Maize
d. Bread
e. Meat
f. Fish
g. Eggs
h. Other (specify)_______________________
i. Don’t Know
50. What are nutrients? (Mark the first response given by the student)
a. The organs of animals
b. The four food groups
c. The part of food that provides nutrition for the body
d. Others (specify)
e. Don’t Know
51. What is a balanced diet? (Mark the first response given by the student)
a. Eating from the four food groups
b. Eating a variety of different types of foods
c. Others (specify)
d. Don’t Know
52. Why is it important to eat a balanced diet? (Mark the first response given by the student)
a. To make meals more interesting
b. To stay healthy and grow well
c. To have enough energy to study or play
d. To get all the nutrientes your body needs
e. Other (specify)
f. Don’t Know
Abt Associates June 2022 ▌139
53. What should you wash your hands with? (Mark the first response given by the student)
a. Water
b. Water and soap or ash
c. Other (specify)
d. Don’t Know
54. How long should you take to wash your hands? (Mark the first response given by the student)
a. At least 20 seconds
b. At least 2 minutes
c. At least 5 minutes
d. Other (specify)
e. Don’t Know
55. When should you wash your hands?
a. Before preparing meals
b. Before eating
c. After using the toilet
56. How often do you wash your hands with soap or ash before eating your school meal?
a. Never
b. Sometimes
c. Always
d. Don’t Know
57. Why is it important to wash your hands? (Mark the first response given by the student)
a. To prevent germs from spreading
b. To prevent diarrhea
c. To stay healthy
d. Because my parents or teachers told me to
e. Other (specify)
f. Don’t know
58. What can you do to prevent diarrhea? (Mark the first response given by the student)
a. Wash your hands before preparing food, before eating, and after using the toilet
b. Drink clean water from a protected well
c. Use a latrine
d. Make sure food is well cooked (hot) before you eat it
e. Cover food to keep flies off it.
f. Other (specify)
Abt Associates June 2022 ▌140
g. Don’t Know
59. What should you do to prevent malaria?
a. Sleep in the mosquito net
b. Avoid mosquito bites(use long pants and long sleeves at night and sunset)
c. Other(specify)
d. Don’t Know
60. Since the beginning of the school year, were there times when you and your family did not have
enough food to eat?
61. How often did this happen?
a. Every Month
b. During Rainy season
c. Other (specify)
62. Does your house have electricity?
63. Where do you normally get your water from at home?
a. River, stream or lake
b. Well or borehole
c. Communal tap
d. Water pipe / tap in your home
e. Water truck or tank
f. Other
g. Don’t Know/Refuse to Respond
64. What do you cook food with in your house?
a. Firewood
b. Coal stove
c. Oil stove
d. Gas stove
e. Electric stove
f. Other
g. Don’t Know/Refuse to Respond
65. What type of toilet does your family use at your home?
a. No toilet
b. Pit toilet (including shared and communal)
c. Flush/eastern toilet outside your home
d. Flush/eastern toilet inside your home
Abt Associates June 2022 ▌141
e. Other
f. Don’t Know/Refuse to Respond
66. Where does your family get water that you use at home?
a. Stream, river, pond, lake
b. Unprotected well
c. Protected (borehole) well)
d. Piped water to yard
e. Piped water into home
67. Does your family have the following items in your home?
68. Point to appropriate pictograms.
a. Radio
b. Mobile phone
c. Television
d. Computer
e. Refrigerator
f. Bicycle
g. Motorbike
h. Car/truck
69. Do you sleep in the mosquito net?
70. What language do you usually speak at home?
a. Português
b. Rhonga
c. Changana
d. Macua
e. Other
Abt Associates June 2022 ▌142
2. Head Teacher Questionnaire (English)
1. What is your position at the school?
a. Head Teacher?
b. Assistant Head Teacher?
c. Other?
d. Don’t know / Refuse.
2. In what year did you start your position at this school?
3. Is the Head Teacher / Assistant Head Teacher female?
4. What date did classes start this year?
5. Between 2017 and 2019, did you attend any WV trainings?
6. Between 2017 and 2020, did you attend any PAI trainings?
7. Between 2017 and 2020, did you attend any teacher trainings?
8. Between 2017 and 2019, did you modify any aspects of the WV meals or literacy programs?
9. Between 2017 and 2019, did you modify any aspects of the PAI meals or literacy programs?
10. Do you believe these adaptations were successful at meeting the needs of your teachers and
students?
11. How many teachers are currently employed at this school?
a. Number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
12. Do you have a school literacy program?
13. If yes, in which language?
a. Portuguese?
b. Rhonga?
c. Changana?
d. Macua?
e. Other (specify)?
14. Besides the ministry of education, what other organization supports the school’s literacy
program?
a. PAI/ADPP?
b. World Vision?
c. USAID?
d. Other (specify)?
Abt Associates June 2022 ▌143
e. Don’t know/Refuse
15. Did teachers at the school attend the literacy training this year?
16. Have they attended the WV training?
17. Have they attended the PAI training?
18. Of these teachers, how many are currently teaching Standard 2?
a. Total number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
19. Of these teachers, how many are currently teaching Standard 3?
a. Total number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
20. Of these teachers, how many are currently teaching Standard 4?
a. Total number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
21. Of these teachers, how many are currently teaching Standard 5?
a. Total number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
22. How many teacher training sessions are expected to be held this academic year?
23. How many teacher training sessions have been held this year up until today?
24. How many streams each are there for Standard 3 and 5?
a. Number of Standard 3 streams at the school?
b. Number of Standard 5 streams at the school?
25. How many teachers are currently trainees expected to graduate from ADPP-supported Teacher
Training Colleges?
a. Total number of teachers?
b. Male?
c. Female?
Abt Associates June 2022 ▌144
d. Don’t know / Refuse
26. How many teachers have already graduated from ADPP-supported Teacher Training Colleges?
a. Total number of teachers?
b. Male?
c. Female?
d. Don’t know / Refuse
27. Could I please see your teacher attendance register?
a. How many teachers were absent yesterday (or on the last day school was in session)
across the school?
b. Record unavailable?
c. Don’t know / Refuse
28. What do you do with a class whose teacher is absent?
a. Let the class proceed without a teacher?
b. Allocate that class to another teacher?
c. Join all students in one class?
d. Bring in a teacher from outside?
e. Dismiss students for the day?
f. Send students to the playground?
g. Distribute students among other classrooms?
h. Other?
i. Don’t know / Refuse
29. Is it easy to find and retain good teachers at this school?
30. Between 2017 and 2019, did reading coaches train teachers at your school?
31. Between 2017 and 2019, how many trainings did reading coaches provide?
32. Between 2017 and 2019, did reading coaches supervise teachers at the school?
33. Between 2017 and 2019, how often did reading coaches supervise teachers at the school?
a. Once in the year?
b. Once per term?
c. Once a month?
d. More than once a month?
e. Don’t know / Refuse
34. Have you received training on the fundamentals of reading instruction?
35. Can you name the core reading skills?
36. Please name as many as you can.
Abt Associates June 2022 ▌145
a. Letter knowledge.
b. Phonemic awareness.
c. Fluency.
d. Vocabulary.
e. Comprehension.
37. Have you received training on school management?
38. How do you know whether students are progressing academically?
a. Classroom observation
b. Monitor pupil results on tests given by teachers
c. Evaluate students orally myself
d. Check students’ assignments or homework
e. Teachers provide progress reports
f. End-of-term evaluations
g. Feedback from parents
h. Feedback from school counselors
i. Feedback from school committees
j. Other
k. Don’t know/Refuse
39. Are there special events held at the school to celebrate students’ progress in reading?
40. Are there special events held at the school in recognition of effective reading instruction?
41. Are you generally satisfied with parents’ involvement in the school?
a. Strongly agree
b. Agree
c. Disagree
d. Strongly Disagree
42. What ways have parents evidenced support for this school?
a. None
b. Visiting the school
c. Actively seeking information about school matters
d. Following up on academic progress
e. Volunteering time
f. Volunteering in school feeding program
g. Volunteering resources
h. Participating in school committees/meetings
Abt Associates June 2022 ▌146
i. Other
j. Don’t know/Refuse
43. How frequently are these types of supports received from parents?
a. Once in the year
b. Once a quarter
c. Once a month
d. More than once a month
e. Don’t know/Refuse
44. Between 2017 and 2019, did you receive training to learn about the importance of access and use
of school supplies and materials?
45. Do teachers have access to a Teacher's Handbook for teaching reading?
46. Do teachers have access to school supplies?
a. No
b. Yes
c. Don’t know/Refuse
d. Pens/pencils
e. Notebooks
f. Chalk
g. Cardboard
h. Paint
i. Markers
j. Chart paper
k. Other (Specify) __________________
47. Do students receive materials that are used in teaching and learning reading?
a. No
b. Yes
c. Don’t know/Refuse
d. Student exercise book
e. Student notebook
f. Supplementary reading books / decodeable texts
g. Notebook
h. Pens / Pencils
i. Other (Specify) __________________
48. Does the school have a library/book bank?
Abt Associates June 2022 ▌147
a. No
b. Yes
c. Don’t know/Refuse
d. Flipbooks
e. Storybooks
f. Primers
g. Fables
h. Leveled books
i. Supplementary readers/decodable texts
j. Guidelines for parents
k. Reading posters
l. Other
49. Is the print material in the library/book bank available in Portuguese?
50. Is the print material in the library/book bank available in the mother tongue?
51. This year, how many books or other printed materials have students ordered from the
library/book bank?
a. Printed material consulted by 3rd grade:
b. Printed material consulted by 5th grade:
52. Are there other types of afterschool learning clubs?
a. No
b. Yes
c. Don’t know/Refuse
d. Help with Homework Club
e. Math Learning Club
f. Science Learning Club
g. School Garden Learning Club
h. Other
53. Were you absent from school any day last week?
54. Were you late to school any day last week?
55. How many teachers are currently teaching Standard 2?
56. How many teachers are currently teaching Standard 4?
57. Between 2017 and 2019, How many teachers at this school received training on teaching
nutrition?
58. Does/do the teacher/teachers have materials for teaching nutrition?
Abt Associates June 2022 ▌148
59. How often do they teach lessons on nutrition to each class?
a. Never
b. Once per week
c. Once every two weeks
d. Once per month
e. Once per term
60. How many teachers at this school received training on teaching hygiene and sanitation?
61. Does/do the teacher/teachers have materials for teaching hygiene and sanitation?
62. How often do they teach lessons on hygiene and sanitation to each class?
a. Once per week
b. Once every two weeks
c. Once per month
d. Once per term
e. Never
63. Does the school have a school feeding program?
64. What organization supports the school feeding program?
a. PAI/ADPP
b. World Vision
c. Government of Mozambique
d. Other (specify)
e. Don’t know
65. Are students on a three days per week schedule, or other schedule less than five days per week?
a. No
b. Yes
c. Depends on which month this academic year
66. Do students receive school meals only on days they attend school?
67. How often do students receive the school meals?
a. Daily
b. A few times per week
c. Whenever food is available
68. Please describe any issues this school has faced related to the school feeding program.
a. No issues
b. Food is not always available
c. Kitchen/warehouse not constructed or in disrepair
Abt Associates June 2022 ▌149
d. Not enough pots, utensils, or dishes
e. Not enough volunteer cooks
f. Meals not prepared on time
g. Other (specify) ______________________
69. Does the school have a garden? (Answered based on surveyor observations)
70. Does the garden operate year-round or during the rainy season only?
a. Rainy season only
b. Year round
71. How often is the garden used to supply food for school meals?
a. No school feeding program
b. Every day
c. 3-4 times per week
d. 1-2 times per week
e. Never
f. During the rainy season
g. Only when school garden products are available
72. Are the students in the school provided with deworming medication (albendazole or
mebendazole)?
73. If yes, have students in the school received deworming medication at least once during this school
year?
74. Does the school have improved latrines? (Answered based on surveyor observations)
75. How many improved latrines does the school have? (Answered based on surveyor observations)
76. How clean are the latrines? (Answered based on surveyor observations)
a. All latrines are dirty with feces on the slabs
b. Some latrines are dirty
c. All latrines are clean with no feces on the slabs
77. Are there tippy-taps or another type of hand washing station near the latrines? (Answered based
on surveyor observations)
78. Do the tippy-taps or hand washing stations have ash or soap? (Answered based on surveyor
observations)
a. None have ash or soap
b. Some have ash or soap
c. All have ash or soap
Abt Associates June 2022 ▌150
79. Due to COVID-19, in many schools students have missed a year or more of literacy. Are you
implementing the curriculum from the prior standard to your students? (For example, are
Standard 2 students studying the Standard 1 curriculum?)
a. No
b. Yes
c. Depends on the grade
d. Don’t know/Refuse
80. Did your school provide literacy learning during COVID-19? (For example, radio activities?)
81. While schools were shut down during COVID-19 [from March 2020 to March 2021], did
students still receive deworming?
82. While schools were shut down during COVID-19 [from March 2020 to March 2021], did
students receive take-home rations?
83. How frequently were take-home rations given to students between March 2020 and March 2021?
a. One time, early in the shutdown
b. Weekly
c. Monthly
d. Other (specify) ____________________
84. Will students be able to catch up to their grade level by the end of the school year? How much of
the year will be spent on catching up to this years’ curriculum? Please tell us your thoughts.
85. Comparing the school between 2017 to now, what has changed in terms of literacy activities,
sanitation, and health and nutrition?
86. How much of these changes are because of help from World Vision?
87. Comparing the school between 2017 to now, what has changed in terms of literacy activities,
sanitation, and health and nutrition?
88. How much of these changes are because of help from PAI?
89. Which of these changes do you think will stay over time? What do you need to maintain changes?
90. Our survey firm, Ernst and Young, may call some people back if there is anything we would like
to clarify. Would you mind if we call you back if needed? If yes, what is your phone number?
Abt Associates June 2022 ▌151 
3. Teacher Questionnaire (English)
1. Identify the shift.
a. All day
b. Morning shift
c. Afternoon shift
2. Was the teacher present in the classroom at the beginning of the shift?
3. Is the teacher female?
4. In what languages do you teach reading?
a. Portuguese
b. Rhonga
c. Changana
d. Macua
5. What is the primary (target) language you teach children to read and write in?
a. Portuguese
b. Rhonga
c. Changana
d. Macua
6. In what year did you start your position at this school?
7. What teacher training college did you attend?
a. Did not attend a teacher training college
b. Still attends a teacher training college
c. Institute of Primary Teaching (IMAP)
d. Primary Teaching Training Institute
e. ADPP Teacher Training Schools for the Future
f. Pedagogical University – UP
g. Other (Specify)
8. Did you attain certification?
9. Are you currently a trainee expected to graduate from a Teacher Training College?
10. What teacher training college are you expected to graduate from?
a. Institute of Primary Teaching (IMAP)
b. Primary Teaching Training Institute
c. ADPP Teacher Training Schools for the Future
d. Pedagogical University – UP
e. Other (Specify)
Abt Associates June 2022 ▌152
11. Identify the grade in which you teach reading.
a. Grade 2
b. Grade 4
12. What subjects do you teach in?
a. Portuguese Reading
b. Portuguese Writing
c. Reading (Other language)
d. Writing (Other language
e. Health and Nutrition Education
f. Games, Sport and Visual Arts
g. Other (Specify)
h. Don’t know/Refuse
13. What is your highest level of academic education?
a. Secondary Education
b. Middle School
c. Bachelor
d. Licensure
e. Masters
f. Other (Specify)
14. During your pre-service training, did you receive any specific training on how to teach reading
and writing to students in the primary grades?
15. This year, have you attended any in-service training on how to teach reading in the primary
grades?
16. What kinds of activities did you participate in during in-service training?
a. Analysis of the curriculum and syllabus
b. Analysis of the teacher guide
c. Practice teaching lessons from the teacher guide
d. Analysis of the student exercise book
e. How to plan and prepare for a lesson
f. Improvisation and use of teaching and learning aids
g. How to assess students during the lesson
h. How to use assessments to adjust instruction
i. How to give students feedback
j. How to use supplementary readers/decodable texts
Abt Associates June 2022 ▌153
k. How to apply new strategies for reading instruction
l. Don’t know/Refuse
17. Between 2017 and 2019, did you attend any teacher trainings?
18. Between 2017 and 2019, did you attend any WV trainings?
19. Between 2017 and 2020, did you attend any PAI trainings?
20. Have you received training on the fundamentals of reading instruction?
21. Can you name the core reading skills?
22. Please name as many core reading skills as you can. (Tick all that apply)
a. Letter knowledge
b. Phonemic awareness
c. Fluency
d. Vocabulary
e. Comprehension
23. Is the class that you teach a multi-grade class?
24. Could I please see your student attendance register?
a. Register was not available to be examined
b. Attendance records were completed daily
c. Attendance records were completed weekly
d. Attendance records were completed biweekly
e. Attendance records were completed monthly
f. Other (Specify)
g. Unclear how often the attendance records were completed
25. Record the date of the most recent attendance record entry.
26. How many boys are enrolled in this class?
27. How many girls are enrolled in this class?
28. How many boys in this class are repeaters?
29. How many girls in this class are repeaters?
30. On a typical day, how many students are absent?
31. On a typical day, how many students are late? We define “late” to be arriving at least 15 minutes
after the start of the first class.
32. Does the Head Teacher or Assistant Head Teacher check your lesson plans?
33. How often during this school year have your lesson plans been checked? (Read the responses)
a. Once every 2-3 months
b. Once every month
Abt Associates June 2022 ▌154
c. Once every two weeks
d. Once every week
e. Daily
f. Don’t know/Refuse
34. When you need some help with your teaching, who do you consult? (Read the responses)
a. Never need help
b. There is no one to ask for help
c. Discuss casually with other teachers
d. Discuss at meetings with other teachers
e. Reading Coach
f. Assistant Head Teacher
g. Head Teacher
h. Don’t know/Refuse
i. Other (Specify)
35. Since the beginning of the school year, how frequently does the Assistant Head Teacher or Head
Teacher observe your teaching? (Read the responses)
a. Never
b. Quarterly
c. Monthly
d. Weekly
e. Don’t know/Refuse
36. Does the Assistant Head Teacher or Head Teacher provide feedback on your teaching?
37. Since the beginning of the school year, how frequently does the reading coach visit you? (Read
the responses)
a. Never
b. Quarterly
c. Monthly
d. Weekly
e. Don’t know/Refuse
38. Do you have a Teacher Guide for reading instruction?
39. Do you use the Teacher Guide for reading instruction?
40. Do each of your students have exercise books that can be used during reading instruction?
41. Do your students use the exercise books during reading instruction?
42. Do you have access to school supplies?
Abt Associates June 2022 ▌155
a. School supplies available:
i. Pens/pencils
ii. Notebooks
iii. Chalk
iv. Cardboard
v. Paint
vi. Markers
vii. Chart paper
viii. Other (Specify)
43. Do you use any school supplies in your classroom during reading instruction?
a. School supplies available:
i. Pens/pencils
ii. Notebooks
iii. Chalk
iv. Cardboard
v. Paint
vi. Markers
vii. Chart paper
viii. Other (Specify)
44. How many books are registered for your class?
45. How frequently do you switch languages between the official language of instruction and a
vernacular language during the teaching and learning process? (Read the responses)
a. Never
b. Occasionally
c. Often
d. All of the time
e. Don’t know/Refuse
46. How would you rate the reading skills of your students in the target language of instruction?
Weak, Average or Strong?
47. How would you rate the writing skills of your students in the target language of instruction?
Weak, Average or Strong?
48. How do you measure your students’ reading and writing progress? (Do NOT READ the options)
a. Written tests
b. Oral evaluations
Abt Associates June 2022 ▌156
c. Observation
d. Portfolios and other projects
e. Homework
f. Classwork
g. End of unit evaluation
h. End-of-term evaluation
i. Other (Specify)
j. Don’t know/Refuse
49. How do you check for student understanding during the lesson? (Do NOT READ the options)
a. Ask questions to individual students
b. Ask questions to whole class
c. Ask questions to students in groups
d. Give students a task and correct the written responses
e. before the end of the lesson
f. Give students a task and correct the responses after
g. the end of the lesson.
h. Ask students to put thumbs up or down
i. Other (Specify)
j. Don’t know/Refuse
50. How do you use the results of students’ oral and written assessments in your teaching? (Do NOT
READ the options)
a. Grade students
b. Evaluate students’ understanding of subject
c. matter
d. Plan teaching and learning activities
e. Adapt teaching to better suit students’ needs
f. Arrange students in ability groups
g. Other (Specify)
h. Don’t know/Refuse
51. What techniques do you use to teach a new vocabulary word? (Do NOT READ the options)
a. Pictures
b. Real objects
c. Gestures
d. Verbal explanation
Abt Associates June 2022 ▌157
e. Don’t know/Refuse
f. Other (Specify)
52. In your class, how many parents/guardians have reviewed the Students' TPC? None, some, most
or all? (Read the options, mark ONLY one answer)
a. None
b. Some
c. Majority
d. All
e. Don’t know/Refuse
53. Overall, are you satisfied with the involvement of parents and guardians in the school work of
their students?
54. At what grade level do you expect children to read text in the target language fluently (accurately
and to use punctuation marks correctly)?
a. Grade 1
b. Grade 2
c. Grade 3
d. Grade 4 or higher
e. Don’t know/Refuse
55. At what grade level do you expect children to write a coherent and comprehensible short story
correctly?
a. Grade 1
b. Grade 2
c. Grade 3
d. Grade 4 or higher
e. Don’t know/Refuse
56. Are students attentive in class?
57. How many students are generally attentive in class?
a. Less than half
b. Half
c. More than half
d. All
58. How do you know if students are attentive in class?
a. Students are seated facing the teacher and making eye contact with the teacher while the
teacher is talking
Abt Associates June 2022 ▌158
b. Students actively listen and act upon teacher directions
c. Students are working independently, in pairs, or in small groups on the current learning
activity
d. Students respond to questions posed by the teacher
e. Students ask the teacher or another student for help related to the current learning activity
f. Students ask the teacher questions related to the current learning activity
g. Students are reading
h. Students are writing
i. Students are listening to a story
j. Other (Specify)
59. Are there special events held at this school to celebrate students’ progress in reading?
60. Have any of your students received a certificate or award in school to celebrate progress in
reading?
61. Are there special events held at this school to celebrate students’ school attendance?
62. Have any of your students received a certificate or award in school to celebrate improvement in
attendance?
63. Is there an afterschool literacy club in this school?
64. Do your students attend an afterschool literacy club at this school?
65. If yes, specify the name of the afterschool literacy club.
66. How many students in your class attend an afterschool literacy club at this school?
a. None
b. Half
c. More than half
d. All
67. Were you absent from school any day last week?
68. If so, why were you absent?
a. No, I wasn't absent from school last week
b. Yes, I was sick
c. Yes, I went to collect my salary
d. Yes, I did other work
e. Yes, I did not receive/received too little, received
f. irregular payment
g. Yes, a lack of motivation
h. Yes, I had family responsibilities
Abt Associates June 2022 ▌159
i. Yes, I had no transportation
j. Other (Specify)
k. Don’t know/Refuse
69. Were you late from school any day last week?
70. If so, why were you late?
a. No, I wasn't absent from school last week
b. Yes, I was sick
c. Yes, I went to collect my salary
d. Yes, I did other work
e. Yes, I did not receive/received too little, received irregular payment
f. Yes, a lack of motivation
g. Yes, I had family responsibilities
h. Yes, I had no transportation
i. Other (Specify)
j. Don’t know/Refuse
71. Is there anything at this school that you identify as an issue, or is missing?
a. If so, explain.
b. Issues with the structure of the school building (walls,
c. windows, roof, etc.)
d. Issues with the surrounding area
e. Lack of clean drinking water
f. Issues with latrines for boys
g. Issues with latrines for girls
h. Issues with latrines for teachers
i. Poor relationships with parents/community
j. Bullying at school
k. Other (Specify)
l. Don’t know/Refuse
72. Did you receive training on teaching nutrition?
73. Do you have lesson plans or other materials that you can use to teach lessons on nutrition?
74. May I see the lesson plans or materials? (Do NOT READ the options)
a. Has all the nutrition lesson plans provided by the
b. implementing organization
c. Has some of the nutrition lesson plans
Abt Associates June 2022 ▌160
d. Has no nutrition lesson plans
e. Has other materials, like picture cards or posters
f. Has no other materials
75. How often do you do lessons on nutrition?
a. Daily
b. A few times per week
c. Once per week
d. A few times per month
e. Once per month
f. Less than once per month
76. What classes do you teach nutrition classes in?
a. Grade 1
b. Grade 2
c. Grade 3
d. Grade 4
e. Grade 5
77. How would you rate your students’ knowledge of nutrition?
a. Weak
b. Average
c. Strong
d. Don’t know/Refuse
78. Have you received training on teaching about hygiene and sanitation?
79. Have you received training on hygiene and sanitation practices?
80. Do you have lesson plans or other materials that you can use to teach lessons on hygiene and
sanitation?
81. May I see the lesson plans or materials? (Do NOT READ the options)
a. Has all the nutrition lesson plans provided by the
b. implementing organization
c. Has some of the nutrition lesson plans
d. Has no nutrition lesson plans
e. Has other materials, like picture cards or posters
f. Has no other materials
82. How often do you do lessons on hygiene and sanitation?
a. Daily
Abt Associates June 2022 ▌161 
b. A few times per week
c. Once per week
d. A few times per month
e. Once per month
f. Less than once per month
83. What classes do you teach hygiene and sanitation classes in?
a. Grade 1
b. Grade 2
c. Grade 3
d. Grade 4
e. Grade 5
84. How would you rate your students’ knowledge of hygiene and sanitation?
a. Weak
b. Average
c. Strong
d. Don’t know/Refuse
85. Does the school serve meals?
86. Which statement most accurately reflects the level of student attentiveness in your class?
a. The students show the same level of attentiveness after the school meal
b. The students are less attentive after school meal
c. The students show the same level of attention before and after school meals
d. Don’t know/Refuse
87. Did you eat a school meal yesterday or on the last day that you were at school?
88. Do you receive corn-soy flour from the school meal program to take home with you?
89. How many kilograms of corn-soy blend do you receive per month?
90. How do you think the school feeding program impacts your students’ learning?
a. School meals distract students
b. School meals make students more attentive
c. School meals don’t affect student learning
d. Other (Specify)
91. Did you provide literacy instruction to your students during COVID-19? (For example, radio
activities?)
92. Will students be able to catch up to their grade level by the end of the school year?
Abt Associates June 2022 ▌162
93. How much of the year will be spent on catching up to this years’ curriculum? Please tell us your
thoughts.
94. Our survey firm, Ernst and Young, may call some people back if there is anything we would like
to clarify. Would you mind if we call you back if needed? If yes, what is your phone number?
Abt Associates June 2022 ▌163
4. Cook/Warehouse Manager Questionnaire (English)
1. Is there a log book for this school year showing how many children were present and how many
meals were served?
2. May I see the log book?
a. Log book is not filled out or up to date
b. Log book is filled out and up to date
3. Check the attendance record for the previous school day (with the head teacher) and cross check
the number of school meals prepared that day.
a. Number present and number of meals cooked are the same
b. Number present is lower than number of meals cooked
c. Number present is higher than number of meals cooked
4. Does the school have a kitchen and storeroom for the school feeding program? (Answered based
on surveyor observations)
5. What is the condition of the kitchen/storeroom for school feeding? (Answered based on surveyor
observations)
a. Not adequate (e.g., holes in roof or walls, no door, no lock on storeroom, etc.)
b. Moderate (complete building, but could use some repairs)
c. Very good (complete building, no repairs needed)
6. Do you ever run out or have no supply of corn-soy blend?
a. For part of each month
b. Once every 2-3 months
c. Once in the school year
7. Are cooking pots and cooking utensils (such as stirrers or scoops) available? (Answered based on
surveyor observations)
8. Is the number of pots and cooking utensils sufficient?
9. What do you do if you are lacking materials (corn-soy blend, pots, utensils) for preparing the
school meals?
a. Nothing
b. Inform the head teacher
c. Inform the project
d. Other (specify)
10. At what time or times of day do you serve the school meal?
a. Time 1: ____________________
b. Time 2: ____________________
Abt Associates June 2022 ▌164
11. At what time do students in grade 2 eat?
a. Time:________________
12. At what time do students in grade 4 eat?
a. Time:________________
13. How often would you say that the school meals are served on time?
a. Never
b. Sometimes
c. Always
14. How many grams of corn-soy blend should a child receive during the school lunch?
15. How many cooks are present today? (Answered based on surveyor observations)
16. How many of these cooks are parents with at least one child attending the school?
17. Do you supplement the corn-soy blend with other foods?
18. How often is the meal supplemented with other foods?
a. Every day
b. 3-4 times per week
c. 1-2 times per week
d. Never
e. During the rainy season
f. Only when school garden products are
g. available
19. What types of foods do you supplement the corn-soy blend with?
a. Fresh maize
b. Vegetables
c. Fruits
d. Other/specify ______________
20. Does the school have bowls and spoons for students? (Answered based on surveyor observations)
21. Are there enough bowls and spoons for students for each meal session?
22. Are the bowls and spoons cleaned after each meal?
23. Is there soap available to clean the bowls, spoons, pots, and cooking utensils? (Answered based
on surveyor observations)
24. Is there a supply of clean water (from a borehole or
25. rainwater tank) available at the school? (Answered based on surveyor observations)
26. Is there a place for children to wash their hands before
27. eating? (Answered based on surveyor observations)
Abt Associates June 2022 ▌165
28. Is there soap available for children to wash their hands
29. before eating? (Answered based on surveyor observations)
30. How many days per week do you cook at the school?
31. In what year did you start as a cook?
32. Did you receive any training for being a school cook?
33. When was the last time you received training related to being a school cook?
a. Less than 6 months ago
b. 6-11 months ago
c. 12 or more months ago
34. How can you get rid of bacteria and other germs in food?
a. Wash it with clean water
b. Cook it
c. Keep it in a safe place away from animals
35. What do you do to make sure the food you serve to the children at school is safe?
a. Cook it for long enough
b. Wash hands with soap before preparing the food
c. Wash bowls and spoons with clean water and soap
d. Wash cooking pots and utensils with clean water and soap
e. Have children wash hands before eating
36. Is there a place for cooks to wash their hands? (Answered based on surveyor observations)
37. Is there soap for cooks to wash their hands? (Answered based on surveyor observations)
38. What should you wash your hands with?
a. Water
b. Water and soap or ash
c. Without water
39. How long should you take to wash your hands?
a. Less than 2 minutes
b. 2 minutes
c. More than 2 minutes
40. When should you wash your hands?
a. Before preparing food
b. Before eating food
c. After using the toilet
41. Do you usually eat the school meal when you are at the school cooking?
Abt Associates June 2022 ▌166
42. What do you do with leftover corn-soy blend at the end of the meal?
a. Give more to students
b. Give it to siblings/other children at
c. school who are not students
d. Give it to teachers
e. Cooks eat it
f. Cooks take it home
g. Other/specify ______________
h. Don’t know
43. What do you receive in return for cooking the food?
a. Nothing
b. Payment
c. Corn soy blend
d. Don’t know/No answer
e. Other/specify ______________
44. While schools were shut down during COVID-19 [from March 2020 to March 2021], did anyone
take care of the school gardens?
45. What changes have you seen in terms of health and nutrition since you started working as a cook,
if any? [probe: is the cooking space larger than before? More cookware? More knowledge of
nutrition? Changes due to COVID-19?]
46. What changes have you seen in terms of health and nutrition since you started working as a cook,
if any? [probe: is the cooking space larger than before? More cookware? More knowledge of
nutrition? Changes due to COVID-19?]
Abt Associates June 2022 ▌167
5. Ministry of Education Interview (English)
1. What is your involvement in the program?
2. What are your opinions about the literacy program?
a. Probe: Positive aspects, innovations on what was new
b. Challenges
i. Working with parents?
ii. Working with teachers?
iii. Working with donors?
3. How did external factors such as COVID-19 influence program implementation?
4. What goals do you think were or were not achieved?
5. Are there any aspects of the program that you intend to continue? Why or why not?
a. Probe: is there any budget set aside for this? What resources and personnel are necessary?
Why are they or are they not being provided?
6. What are your opinions about the school feeding program?
a. Probe: Positive aspects
b. Challenges
i. Working with parents?
ii. Working with teachers?
iii. Working with donors?
7. What are your opinions about the school garden program?
a. Probe: Positive aspects
b. Challenges
i. Working with parents?
ii. Working with teachers?
iii. Working with donors?
8. How did external factors such as COVID-19 influence program implementation?
9. What goals do you think were or were not achieved?
10. How did COVID-19 influence goals?
a. Probe: are most students now behind a grade level? Learning at a slower rate due to
coming to school fewer days a week?
11. Are there any aspects of the program that you intend to continue? Why or why not?
a. Probe: is there any budget set aside for this?
Abt Associates June 2022 ▌168
6. Parent Focus Groups & Discussion Guide
1. How many children do each of you have at this school?
2. What is your involvement in the school as a parent?
3. In what other ways are parents involved in the school?
4. Are you generally satisfied with the level of parental involvement at the school? Why? Why not?
5. What do you expect your children to gain from attending this school?
6. How well is the school meeting your expectations?
7. What are your opinions about the literacy program at this school, if there is one?
a. Probe: Positive aspects
b. Challenges
8. What language do you speak at home?
9. What types of reading or storytelling activities do you do at home?
10. In what language do you carry out reading or storytelling activities at home?
11. Do you review your child’s homework?
12. Does your child read at home?
13. How often does your child read at home?
14. In what language does your child read at home?
15. Does your child have materials to draw or write at home?
16. Does your child draw or write at home?
17. How often does your child draw or write at home?
18. What recommendations do you have for improving the literacy program at this school? (skip if
there is no literacy program)
19. How often do your children miss school due to illness?
a. Probe: Types of illnesses, duration of missing school
20. Are your children less likely to experience health-related absences when receiving meals at
school?
21. What types of health practices do your children learn about at school?
22. How do they apply their knowledge of improved health practices at home?
23. What are your opinions about the school feeding program at this school, if there is one?
a. Probe: Positive aspects, Challenges (skip below questions if no program)
24. What are the benefits of your child or children receiving a meal at school?
25. What are your thoughts about the type of food that is provided for the meal?
26. How does the school meal affect the amount of food or number of times per day you give your
school-going child or children at home?
Abt Associates June 2022 ▌169
27. How does the school meal affect the types of food you give your school-going child or children at
home?
28. What recommendations do you have for improving the school feeding program?
29. What are your opinions about the school garden, if there is one?
a. Probe: Positive aspects
b. Challenges
30. What recommendations do you have for improving the school garden?
31. Tell me about what happened during COVID-19?
a. What kind of education did your children get during school closures due to COVID-19, if
any?
i. Literacy?
ii. Health?
b. Did your children still receive deworming?
c. Did you children receive any take-home rations?
d. If so, what kind of rations did they receive?
e. How frequently were take-home rations given to your children between March 2020 and
March 2021?
32. How is COVID-19 affecting school now?
a. Shifted schedules to 3 days a week?
b. Literacy a grade level behind?
33. Do you think any of the knowledge from the trainings provided by World Vision (if in Maputo)
or PAI (if in Nampula) will continue to impact the school in the future? On which topics? (health,
literacy)?
34. What would it take to sustain the knowledge gained? (More trainings? More materials?)
Abt Associates June 2022 ▌170