February 18, 2018 
McGovern Dole Mozambique 
Impact Evaluation and Literacy Assessment
Baseline Report
Prepared for 
U.S. Department of Agriculture 
DOI, Interior Business Center, AQD 
Division 2/Branch 2 
381 Elden St 
Suite 4000 
Herndon, VA 20170 
Prepared by 
RTI International 
3040 E. Cornwallis Road 
Research Triangle Park, NC 27709 
GSA PSS Schedule No. GS00F354CA Task Order No. D16PD00635 

_________________________________
RTI International is a registered trademark and a trade name of Research Triangle Institute. 
MCGOVERN-DOLE MOZAMBIQUE IMPACT EVALUATION AND LITERACY ASSESSMENT: 
BASELINE REPORT 
RTI International 
GSA PSS Schedule No. GS00F354CA Task Order No. D16PD00635 
February 2018 
The views, opinions, and/or findings contained in this report are those of the author(s) and 
should not be construed as an official Government position, policy, or decision, unless so 
designated by other documentation. 
iv USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
CONTENTS 
Section Page 
List of Acronyms x
1. Executive Summary 1-1
2. Background & Methodology 2-4
2.1 Background ......................................................................................... 2-4
2.2 Methodology ........................................................................................ 2-5
2.2.1 Theory of Change ......................................................................... 2-5
2.2.2 Impact Evaluation ........................................................................ 2-5
2.2.3 Research Questions ...................................................................... 2-7
2.2.4 Indicators ................................................................................... 2-7
2.2.5 Instruments ................................................................................ 2-9
2.2.6 Programming ............................................................................. 2-10
2.2.7 Field Work ................................................................................. 2-10
2.2.8 Sampling .................................................................................. 2-11
2.2.9 Data Management ...................................................................... 2-16
2.3 Analysis Methodology ......................................................................... 2-17
2.3.1 Basic Quantitative Analysis .......................................................... 2-17
2.3.2 Advanced Quantitative Analysis ................................................... 2-17
2.3.3 Qualitative Analysis .................................................................... 2-17
3. Analysis 3-1
3.1 Baseline Indicators ............................................................................... 3-1
3.1.1 Literacy ...................................................................................... 3-1
3.1.2 Student Attendance ...................................................................... 3-7
3.1.3 Secondary Indicators .................................................................... 3-8
3.2 Baseline Descriptive Statistics .............................................................. 3-10
3.2.1 Students ................................................................................... 3-10
3.2.2 Teachers ................................................................................... 3-20
3.2.3 Head Teachers ........................................................................... 3-33
3.2.4 Warehouse Managers and Cooks .................................................. 3-44
3.2.5 Parents ..................................................................................... 3-47
4. Summary 4-1
5. Limitations & Recommendations for Impact Evaluation 5-6
Limitations .......................................................................................... 5-6
Recommendations ................................................................................ 5-6
6. References 6-1
7. Annex A: Within-Province Literacy Analysis 1
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report v 
8. Annex B. Precision Estimates 3
9. Annex C. Impact Evaluation recommendations (table) 5
 
vi USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
LIST OF TABLES AND FIGURES 
Figure 1. McGovern-Dole Program Coverage in Mozambique 2-4
Figure 2. Theory of Change 2-5
Table 1. Research Questions 2-7
Table 2. Literacy, Nutrition and Health, and School Management Indicators 2-8
Table 3. Instruments for Data Collection at Each School 2-9
Table 4. Description of Baseline EGRA Subtasks 2-9
Figure 3. Intervention and Comparison Districts 2-12
Table 5. Target Population across Intervention and Comparison Districts 2-12
Table 6. Summary of the Sampling Methodology and Target Sample Size 2-13
Table 7. Final Sample Counts 2-15
Table 8. Listening Comprehension Subtask Scores 3-2
Table 9. Letter Identification Subtask Scores 3-3
Table 10. Syllable Reading Subtask Scores 3-4
Table 11. Oral Reading Fluency Subtask Scores 3-5
Table 12. Reading Comprehension Subtask Scores 3-6
Table 13. Word Dictation Subtask Scores 3-6
Figure 4. Percent of Students Absent At least One Day Last Week, by Grade and 
Study Arm 3-7
Figure 5. Average Percent of Students Absent in Classroom 3-8
Figure 6. Percent of Students Achieving Nutrition Indicators, by Study Arm 3-8
Figure 7. Student Nutrition and Hygiene and Sanitation Knowledge Scores 3-9
Figure 8. Percent of Teachers Absent At least Once Last Week 3-9
Table 14. Student Responses on Household Demographic Characteristics, by 
Grade and Study Arm 3-10
Table 15. Student-reported Absenteeism, by Grade and Study Arm 3-11
Table 16. Teacher-reported Student Attendance, by Study Arm 3-12
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report vii 
Materials, by Grade and Study Arm 3-13
Arm 3-14
d, by Grade and 
Study Arm 3-15
 3-16
Table 21. Student Responses on School Meals, Current Hunger, and Household 
Hunger, by Grade and Study Arm 3-18
Table 22. Student Nutrition Practices and Knowledge, by Grade and Study Arm 3-
19
Table 23. Student Hygiene and Sanitation Practices and Knowledge, by Grade and 
Study Arm 3-20
Table 24. Teacher Profiles, by Study Arm 3-21
3-22
Table 26. Teacher-reported Absenteeism, by Study Arm 3-23
Table 27. Teacher Access to Supplies and Learning Materials, by Study Arm 3-23
Table 28. Teacher Use of School Supplies and Learning Materials, by Study Arm 3-
24
 3-25
Table 30. Teacher Perspectives on Supervision and Support, by Study Arm 3-26
Table 31. Teacher Knowledge and Skills, by Study Arm 3-27
Table 32. Teacher Perceptions of Student Attentiveness, by Study Arm 3-29
Table 33. Teacher-reported School Conditions, by Study Arm 3-31
Table 34. Teacher-reported Nutrition and Hygiene Teaching, by Study Arm 3-32
Table 35. Number of Teachers and Classes per Grade, by Study Arm 3-33
Table 36. -related Amenities and Activities at 
Schools, by Study Arm 3-34
Table 37. HT Training, by Study Arm 3-35
3-36
3-37
viii USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Arm 3-38
3-38
Study Arm 3-39
3-40
3-41
udy Arm 3-42
and Hygiene Lessons, by Study Arm 3-43
3-44
Table 47. School Feeding Infrastructure, by Intervention Arm 3-45
Table 48. Cook Training and Knowledge, by Intervention Arm 3-46
Table 49. Use of School Meals and Cook Working Days, by Intervention Arm 3-47
Table 50. Number of FGDs and Parents by District and Province Involved in 
Qualitative Data Collection 3-48
Table 51: Grade 2 and 4 EGRA Results for Maputo Province, by Study Arm 1
Table 52: Grade 2 and 4 EGRA Results for Nampula Province, by Study Arm 2
Table 53. Grade 2 Reading Fluency Means and Precision Estimates of the Mean 
ORF 3
Table 54. Grade 4 Reading Fluency Means and Precision Estimates of the Mean 
ORF 3
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report ix 
[This page intentionally left blank.] 
x USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
LIST OF ACRONYMS 
ADPP Ajuda de Desenvolvemento de Povo para Povo (Development Aid from People 
to People) 
ANOVA analysis of variance 
CI confidence interval 
clpm correct letters per minute 
csspm correct syllable sounds per minute 
cwpm correct words per minute 
DEFF Design Effect 
EGRA Early Grade Reading Assessment 
EMIS Education Management Information System 
FAS Foreign Agricultural Service 
FGD focus group discussion 
GSA U.S. General Services Administration 
HT head teacher 
ICC intra-class correlation 
IFP Instituto de Formação de Professores (Teacher Training Institute) 
ORF oral reading fluency 
PAI Planet Aid Incorporated 
PSS Professional Services Schedule 
QC quality control 
SD standard deviation 
TTC teacher training college 
UP Universidad Pedagógica de Moçambique (Mozambique Pedagogical 
University) 
USAID United States Agency for International Development 
USDA U.S. Department of Agriculture 
WHO World Health Organization 
wpm words per minute 
WV World Vision 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report xi 
[This page intentionally left blank.] 

_________________________________
RTI International is a registered trademark and a trade name of Research Triangle Institute. 
1. EXECUTIVE SUMMARY 
Background: The McGovern-Dole Program in Mozambique
In 2012, the U.S. Department of Agriculture (USDA) Foreign Agricultural Service (FAS) 
selected World Vision (WV) and Planet Aid Incorporated (PAI) to implement McGovern-Dole 
activities in Mozambique. The key objective of the McGovern-Dole program is to reduce 
hunger and improve literacy and primary education. The program provides school meals, as 
well as teacher training, health, sanitation, and hygiene interventions and other activities.
Since 2012, WV has been implementing program activities in Nampula Province, and Planet 
Aid has been working in Maputo Province. In 2015, both partners received a second cycle of 
funding to continue the programs and enhance the literacy interventions. However, some 
schools in Maputo continued to receive only school meals without the literacy intervention. 
The McGovern-Dole program theory of change posits that improved literacy, health, and 
hygiene instruction paired with regular school meals will lead to improved literacy, health, 
and dietary practices of students via various program impact pathways, including increased 
student and teacher attendance, improved student attentiveness, and increased knowledge 
of health and hygiene. The intended beneficiaries of the McGovern-Dole program are 
primary school students in targeted districts in Maputo and Nampula Provinces, their 
families, and communities. 
Baseline Literacy Assessment and Impact Evaluation
In 2016, USDA contracted RTI International to design an independent impact evaluation of 
the McGovern-Dole Mozambique program and conduct a baseline assessment. RTI has no 
role in the implementation of the McGovern-Dole program and no relationship with either 
implementing partner. The evaluation considers three groups of schools: 1) those receiving 
school feeding and literacy interventions, 2) those receiving only the school feeding 
intervention, and 3) comparison schools not receiving any McGovern-Dole interventions. 
The impact evaluation is designed to compare the following key indicators across the three 
groups (also referred to as study arms): reading ability, student attendance, student 
nutrition and hygiene knowledge and practices, and teacher attendance. 
The baseline survey was designed to support a quasi-experimental impact evaluation to be 
completed in 2021, which will compare changes in key indicators across the different study 
arms. The baseline report provides a snapshot of the situation in 2017 and does not 
constitute an evaluation of the program or its impact until they can be analyzed in 
conjunction with midline and endline data. The results are intended to support decision 
makers with information about the situation in Nampula and Maputo and share with 
implementers a wider vision of the overall status in these provinces. 
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 and, in the future, adequate exposure to the intervention. 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 intervention and comparison schools at baseline. To assess literacy, 
1-2 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
enumerators were trained to administer the Early Grade Reading Assessment (EGRA) tool 
developed by RTI,1
 which is an individually administered, 15-minute oral assessment that 
measures the basic skills a child must have to eventually be able to read fluently and with 
comprehension. The student survey also included questions on student attendance, the 
school feeding program, and nutrition and hygiene knowledge and practices. In addition, 
teachers, head teachers (HTs), and school cooks were interviewed to obtain information on 
teacher attendance, school amenities, and the school feeding program, among other 
factors. The survey was paired with qualitative data collection, which provides supplemental 
insight from parents about the McGovern-Dole program. 
Baseline Findings
Most of the baseline indicators ranged across study arms, and in some instances, 
comparison schools outperformed the intervention schools. This was expected given the 
non-random selection of intervention districts and the limited options for comparison 
districts within intervention provinces.
Student Reading 
Ability
Literacy ability is generally very low; at least 40% of Grade 2 students 
in McGovern-Dole schools received zero scores on all EGRA subtasks. 
Grade 4 students scored higher than Grade 2 students across all 
components, as expected. Students in School Feeding Only schools 
generally scored higher than those in the other two groups; however, 
it is important to note that these schools are in Maputo only. 
A higher percentage of students scored zero in the listening 
comprehension subtask in both intervention groups than in 
Comparison schools, despite the level of the text read aloud being 
representative of texts used for instruction. This is an important 
reading performance. In general, student data also show that at home, 
students speak languages other than Portuguese, which contributes an 
added challenge to literacy acquisition. 
Student 
Attendance
At the time of the baseline survey, intervention schools had been 
providing school meals for nearly a full school year (and in most cases, 
four years.) The McGovern-Dole theory of change hypothesizes that 
providing school meals will increase student attendance, but no 
significant difference in attendance was detected between 
intervention and comparison schools at baseline. There was also no 
significant difference in absence between male and female students. 
The percentage of students reported absent at least one day in the 
past week was in the range of 25 30% across study arms. 
 
1 https://globalreadingnetwork.net/resources/early-grade-reading-assessment-egra-toolkit-second-edition
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 1-3 
We also compared the number of students present in a classroom on 
the day of the survey to the official class roster. School Feeding & 
Literacy schools had the highest proportion absent (28 29%), which 
differed statistically significantly from the Comparison schools2
 (20
22%.) 
Among students who had missed school during the last week, the most 
commonly cited reason was student illness (55 60%), which helps to 
explain why we do not see higher attendance in McGovern-Dole 
schools despite the provision of meals.
Student 
Nutrition 
Practices
 inadequate, with less than half of them 
reporting a minimum acceptable diet and approximately half achieving 
minimum meal frequency. Diet diversity scores were higher, with 74% 
of students in Literacy & School Feeding schools, 54% in School 
Feeding Only schools, and 69% in Comparison schools achieving 
minimum dietary diversity. Contrary to the expectation that school 
meals would add to the home diet, most students noted that they eat 
less at home on days when they receive school meals; this was 
supported by findings from the focus group discussions (FGDs) with 
parents. 
Student 
Nutrition and 
Health 
Knowledge
hygiene/sanitation knowledge was moderate. Students in the 
Comparison schools had significantly higher knowledge scores than 
those in McGovern-Dole intervention schools. 
Teacher & Head 
Teacher 
Attendance
Teacher absence was common. Across study arms, 21 28% of teachers 
were absent during the last week, and approximately 22 30% of head 
teachers were absent. This was reported in the survey with teachers 
and head teachers and brought up by parents in FGDs as an important 
issue. 
Recommendations for the Impact Evaluation
The findings from the baseline study highlight the need to 1) continue tracking activities 
related to the program in both intervention and comparison schools within sampled 
districts; 2) invest additional effort in matching intervention and comparison schools, 
students, and teachers (e.g., propensity score matching) in future analyses; 3) conduct a 
meta-analysis to account for variance across implementers and provinces; and 4) use the 
proposed difference-in-differences 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; the intervention schools would be expected to have higher rates of 
improvement once the intervention activities take place.
 
2
 A small fraction of Comparison schools was receiving some type of school meal program (not McGovern￾Dole), and approximately half of Comparison schools were found to have a school garden. 
2-4 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
2. BACKGROUND & METHODOLOGY 
2.1 Background 
In 2012, the U.S. Department of Agriculture (USDA)/Foreign Agricultural Service (FAS) 
selected World Vision (WV) and Planet Aid Incorporated (PAI) to implement McGovern-Dole 
school feeding and some literacy activities in Mozambique. In 2015, both organizations 
received an additional cycle of funding to continue their work in providing school meals; 
further the quality of literacy activities; teacher training; school gardens; water, sanitation, 
instruction. With the 2015 funding cycle, both implementing partners began rolling out 
enhanced literacy programs, including curriculum development and teacher training. 
The coverage zones for each organization are shown in Figure 1. Schools covered in 2015 
were the same schools where the program has been active since 2012, except for a small 
number of new schools that opened because of enrollment overflow in pre-existing schools. 
Although the goals of the intervention are the 
same, the two implementers take slightly 
different approaches. What different intervention 
schools receive to achieve their targets also varies 
within each implementer. For example, for PAI in 
Maputo Province, not all McGovern-Dole schools 
receive literacy interventions: some receive 
school feeding only, while other (larger) schools 
receive both school feeding and literacy. In 
addition, some schools receive the literacy 
program in Portuguese only, whereas other 
schools receive both the mother tongue and 
Portuguese (bilingual) programs. The relatively 
new bilingual programs were incorporated by 
McGovern-Dole taking into consideration of 
evolving educational priorities in Mozambique. In 
contrast, WV delivers both school feeding and 
literacy across all targeted schools and 
implements the program in Portuguese only. 
(GSA) Professional Services Schedule (PSS) No. GS00F354CA, the USDA/FAS awarded to RTI 
-Dole Mozambique Impact 
Evaluation and Liter
evaluation and conduct the baseline data collection for its McGovern-Dole Food for 
Education projects in Mozambique. RTI has no role in the implementation of the McGovern￾Dole program and no relationship with either implementing partner. The baseline 
assessment was conducted independently from the implementing partners. 
Figure 1. McGovern-Dole Program 
Coverage in Mozambique 
Nampula/World Vision 
Maputo/Planet Aid 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-5 
2.2 Methodology 
The main objective of the impact evaluation design and the baseline data collection is to set 
the stage to answer th
McGovern￾2.2.1 Theory of Change 
To guide the evaluation, RTI developed a theory of change that links the literacy, health, 
hygiene, nutrition, and dietary components of the McGovern-Dole program and is based on 
the McGovern-Dole learning agenda. The theory of change serves as a framework for the 
design of the impact evaluation, the indicators, and the instruments used to collect baseline 
data. At the highest level, 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. 
The linkages between the key elements of the McGovern-Dole program and the other 
factors that may contribute to improved outcomes are shown in the theory of change 
diagram (Figure 2). 
Figure 2. Theory of Change 
2.2.2 Impact Evaluation 
The impact evaluation uses a quasi-experimental design in which comparison schools serve 
as the counterfactual (i.e., our best estimate of what would have happened in the absence 
McGovern-Dole SO1
Improved Literacy 
Outcomes in School￾Aged Children
McGovern-Dole 1.1
Improved Quality of 
Literacy Instruction
McGovern-Dole 1.1.2
Better Access to School 
Supplies and Materials
McGovern-Dole 1.1.4
Increased Skills and 
Knowledge of Teachers
McGovern-Dole 1.1.5
Increased Skills and Knowledge 
of School Administrators 
McGovern-Dole 1.1.3
Improved Literacy 
Instructional Materials
McGovern-Dole 1.1.1
More Consistent 
Teacher Attendance Overlap 
between 
WV and PAI
Specific to 
WV
LEGEND
McGovern-Dole 1.2
Improved 
Attentiveness
McGovern-Dole 1.2.1
Reduced Short￾term Hunger
McGovern-Dole 1.3.1.1 
Increased Access to Food/
Improved Dietary Intake
McGovern-Dole 1.3 
Improved Student 
Attendance
McGovern-Dole 1.3.5 
Increased 
Community 
Understanding 
of Benefits of 
Education
McGovern-Dole 1.3.1 
Increased 
Economic and 
Cultural Incentives
McGovern-Dole 1.3.4 
Increased 
Student 
Enrollment
McGovern-Dole 1.3.3 
Improved School 
Infrastructure
McGovern-Dole 1.3.2 
Reduced Health 
Related Absences
McGovern-Dole SO2 
Improved Health and 
Dietary Practices of
Students
McGovern-Dole 2.1
Improved 
Knowledge of 
Health and 
Hygiene 
Practices
McGovern-Dole 2.3 
Increased Knowledge 
of Nutrition
McGovern-Dole 2.4 
Increased Access 
to Clean Water 
and Sanitation 
Services
McGovern-Dole 2.5 
Increased Access 
to Preventative 
Health 
Interventions
McGovern-Dole 2.6 
Increased 
Access to 
Requisite Food 
Prep and 
Storage Tools 
and Equipment
McGovern-Dole 2.2 
Increased Knowledge 
of Safe Food Prep 
and Storage Practices
McGovern-Dole 1.2.1.1
School Feeding
Timing of the 
School Meal
Availability of 
Bednets at 
Home
Availability of 
Food at Home
 Access to Improved 
Water and Sanitation at 
 Home
Parental Knowledge and Use of 
Improved Health and Dietary 
Practices
2-6 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
of the McGovern-Dole program). An experimental design with randomized assignment of 
schools would have more accurately estimated the causal effect, but this was not possible 
because intervention districts and schools were pre-determined since phase one in 2012, 
long before the impact evaluation was planned. 
Because of the variations in interventions described above and the goal of measuring the 
impact of phase two of the interventions, the impact evaluation design was structured to 
compare outcomes across three groups of schools: 
-Dole 
interventions, including both the literacy and school feeding interventions; 
School Feeding Only: Schools receiving McGovern-Dole school feeding interventions 
without any McGovern-Dole literacy intervention (Maputo only); and 
Comparison: Schools not receiving any McGovern-Dole intervention. 
It is important to note the following: 
The aim is not to compare outcomes between the two implementing partners nor 
between Maputo and Nampula but rather to focus on the impact of the McGovern￾Dole program across both intervention provinces in Mozambique. 
At the impact evaluation design stage, no schools were identified to receive the 
literacy intervention only (i.e., without school feeding); thus, we are not pursuing a 
full factorial design. 
Schools in the School Feeding Only group are exclusively in Maputo Province, 
creating some level of imbalance that should be considered during analysis and 
when interpreting the results. 
The McGovern-Dole program in Mozambique targets Grades 1 3 for the literacy 
component and Grades 1 6 for the school feeding component. We assessed Grade 2 
and 4 students for two main reasons. First, we expect to see more variability in 
Grades 2 and 4 reading ability, which would allow us to detect significant differences. 
Second, Grade 4 students will have more years of cumulative exposure to the 
intervention, which should yield a better understanding of changes over time 
associated with the intervention activities. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-7 
2.2.3 Research Questions 
We have subdivided the overarching evaluation question into a set of primary and 
secondary questions (Table 1). 
Table 1. Research Questions 
Primary Research Questions 
Literacy 
How does reading ability among Grade 2 and 4 students who attend schools with the McGovern-Dole 
Literacy & School Feeding program compare with that of students who attend 1) schools with the 
McGovern-Dole School Feeding Only program and 2) schools without any McGovern-Dole program 
(Comparison)? 
Is there a larger skills from Grade 2 to Grade 4 in schools with the 
McGovern-Dole Literacy & School Feeding program than in 1) schools with the McGovern-Dole School 
Feeding Only program and 2) schools without any McGovern-Dole school feeding/literacy program 
(Comparison)? 
Student 
Attendance 
Is there higher student attendance at schools with the McGovern-Dole Literacy & School Feeding 
program than in 1) schools with the McGovern-Dole School Feeding Only program and 2) schools without 
the McGovern-Dole school feeding/literacy program (Comparison)? 
Secondary Research Questions 
Health & 
Nutrition 
Do a greater proportion of Grade 2 and 4 students attending schools with the McGovern-Dole Literacy & 
School Feeding program have knowledge of recommended health and nutrition practices than 
students at 1) schools with the McGovern-Dole School Feeding Only program or 2) schools without any 
McGovern-Dole school feeding/literacy program (Comparison)? 
Do a greater proportion of Grade 2 and 4 students attending schools with the McGovern-Dole Literacy & 
School Feeding program report recommended health and nutrition practices, including dietary 
diversity and meal frequency, than students in 1) schools with the McGovern-Dole School Feeding Only 
program or 2) schools without any McGovern-Dole school feeding/literacy program (Comparison)? 
Teacher 
Attendance 
Is there higher teacher attendance at McGovern-Dole Literacy & School Feeding schools than at 1) 
schools with the McGovern-Dole School Feeding Only program or 2) schools without any McGovern￾Dole school feeding/literacy program (Comparison)? 
Do McGovern-Dole Literacy & School Feeding schools have a higher percentage of teachers present
on any given day compared to 1) schools with the McGovern-Dole School Feeding Only program or 2) 
schools without any McGovern-Dole school feeding/literacy program (Comparison)? 
2.2.4 Indicators 
To answer the above research questions, a set of indicators linked to the theory of change 
was developed and used as the basis for questionnaire design, data management, and 
analysis. Table 2 highlights some of the main indicators. 
2-8 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 2. Literacy, Nutrition and Health, and School Management Indicators 
Literacy Indicators Data Level 
Reported 
% of schools with access to teaching and learning materials School 
% of teachers who are present at the school
Student performance on EGRA subtasks (Portuguese) 
% of students who report being absent or late the past week Student 
% of students who report being absent because of sickness
% of teachers who report attending the teacher training designed to prepare them to use the 
teaching and learning materials supported by the reading program 
Teacher/
Classroom 
% of schools where teachers are using the teaching and learning materials needed to teach 
reading and promoted by the program/Ministry of Education 
% of teachers with access to materials for reading instruction promoted by the program (i.e., 
teacher guides, student books, and supplementary readers) 
% of teachers who report that students are attentive before/after a meal 
Nutrition and Health Indicators 
% of students fed a school meal
School 
% of schools where meals were served on time
% of schools where cooks have correct knowledge about safe food preparation and handling
% of schools with improved sanitary facilities
% of schools where students receive deworming medication
% of students who consumed minimum meal frequency on the previous day
Student 
% of students who consumed the minimum number of food groups on the previous day
% of students receiving a minimum acceptable diet
% of students with correct knowledge of nutrition practices taught by the program
% of students with correct knowledge of hygiene and sanitation practices
% of students reporting hand washing before meals
% of trained teachers with copies of nutrition education and hygiene education materials Teacher/Classroom
% of students who report that their school meal replaces a meal at home Household
School Management Indicators 
% of schools that have enrollment and (updated) attendance registers for students, teachers, and 
HTs (i.e., attendance rate, dropout rate, transfer rate, and lateness rate by gender and grade) 
School 
% of schools that have and maintain an updated inventory of food/cooking supplies
% of schools that have a place to properly store food/cooking supplies
Note: EGRA = Early Grade Reading Assessment; HT = head teacher; USDA = U.S. Department of Agriculture. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-9 
2.2.5 Instruments 
Based on these indicators, RTI developed several data collection instruments, which are 
listed in Table 3. 
Table 3. Instruments for Data Collection at Each School 
Instrument Source Data Level Notes: 
HT Questionnaire and School Inventory HT School 
Cook Questionnaire and Kitchen Inventory Cook School Schools with 1 or more 
cooks 
Teacher Questionnaire Teacher Classroom 1 Grade 2 & 1 Grade 4 
Classroom Inventory Assessor w/ 
Teacher's help Classroom 1 Grade 2 & 1 Grade 4 
Student EGRA Student Student 8 Grade 2 & 8 Grade 4 
Student Questionnaire Student Student 8 Grade 2 & 8 Grade 4 
Parent Focus Group Discussion Guide 
Parents Parents 
In selected Literacy & 
School Feeding and School 
Feeding Only schools 
Note: EGRA = Early Grade Reading Assessment; HT = head teacher. 
All instruments and instructional dialogues (for explaining instructions to the respondents 
and asking for consent) were developed in English and circulated for peer review internally, 
with USDA, and with literacy consultants in Mozambique (for the Early Grade Reading 
Assessment [EGRA] portion). The instruments were never shared with the implementing 
partners. Once these instruments were finalized, they were adapted and translated into 
Portuguese and reviewed for accuracy. 
The only exception was the EGRA, which is an individually administered, 15-minute oral 
assessment that measures the basic skills a child must have to eventually be able to read 
fluently and with comprehension the goal of reading (Tobia & Bonifaci, 2015). The EGRA 
instrument used at baseline consisted of six subtasks and was developed directly in 
Portuguese and administered to Grade 2 and Grade 4 students in Portuguese3
 (Table 4). 
Table 4. Description of Baseline EGRA Subtasks 
# Subtask Students must: Reading Skill
1 Letter Name 
Knowledge (clpm)
Provide the name of 100 upper- and
lowercase letters presented in random
order; timed at 1 minute.
The ability to read the letters of the
alphabet naturally and without hesitation.
2 Syllable Reading 
(csspm)
Provide the sound represented by 50 
upper and lowercase letters presented 
in random order; timed at 1 minute.
The ability to recognize graphemic features 
of a letter and accurately map it to its 
corresponding sound.
 
3
 Note that EGRAs in local languages was not developed at baseline since curriculum in local language was still 
under development. EGRAs in local languages are recommended as part of the efforts for midline and endline 
instrument updates and data collection. 
2-10 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
# Subtask Students must: Reading Skill
3a ORF Read a narrative text of 57 words. Timed
at 1 minute.
The ability to read connected text with
accuracy, little effort, and at a sufficient
rate of speed.
3b Reading
Comprehension 
Respond to 5 questions (4 literal and 1
inferential) about the entire text or parts
they have read; 15 seconds to start to
answer each question.
The ability to make meaning from
(understand) what they have read.
4 Listening
Comprehension 
Listen to a connected text of 
approximately 30 words and respond 
correctly to 5 questions (4 literal and 1 
inferential); 15 seconds to start to answer 
each question. 
The ability to make sense of oral language 
and vocabulary (considered a necessary 
skill for reading comprehension).
5 Dictation Write and correctly spell five words; 20 
seconds to scribe each word. 
The ability to listen to individual words and 
spell (encode) them correctly. 
Note: clpm = correct letters per minute; csspm= correct syllable sounds per minute; ORF = oral reading fluency
The EGRA adaptation and development process included a review of relevant curriculum 
materials. For this adaptation, RTI worked with a Mozambican literacy specialist, developed 
short reading passages using appropriate words that can be used across Grades 2 and 4, and 
developed appropriate reading comprehension questions to complement the short 
passages. The adaptation process covered language analysis, the selection of appropriate 
subtasks for the language and grade levels (Grades 2 and 4), the verification of instructions, 
and the translation of directions. 
2.2.6 Programming 
All assessments and questionnaires4 mentio
an open-source suite of software developed by RTI for education and health surveys. Data 
collectors could connect to the internet. RTI programmed and tested tablet-based scripts for 
the questionnaires before beginning data collector training. During and after data collector 
training, RTI updated the Tangerine instruments to accommodate any adjustments in 
terminology or survey question items resulting from issues identified in pilot testing. 
2.2.7 Field Work 
Data collection tools and procedures were piloted and adjusted prior to the full-scale 
assessor training conducted in each province before data collection. Detailed reports of the 
piloting and adaptation process are available separately. The full-scale training for survey 
data collection included practice visits to two schools in each province. Data collection 
occurred in the third trimester of the academic year (September and October 2017) to allow 
the students to be exposed to a full school year. Quantitative assessment teams consisted of 
four members: one supervisor, one teacher/classroom assessor, and two student assessors. 
One assessment team visited one school per day. 
 
4
 Excluding the qualitative FGD guide, which cannot be administered via tablet. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-11 
Qualitative data collection teams worked separately from the quantitative teams. 
Qualitative teams consisted of two members: a moderator and a note-taker. They contacted 
selected schools in advance (the school selection process is described in the Sampling 
section below) and asked the head teacher (HT) to help them invite parents to participate in 
the focus group discussion (FGD). At each of the selected schools, one FGD was held with 
parents of students. Lead observers were hired to periodically attend FGDs to provide 
technical oversight and ensure proper implementation of the data collection methods. 
During the FGDs, hand-written notes and audio recordings were used to record the data. 
2.2.8 Sampling 
Province- and District-level Sampling Frame 
At the time of this baseline assessment, PAI was implementing the McGovern-Dole program 
in 273 schools across four districts in Maputo, and WV was implementing in 150 schools 
across two districts in Nampula. Although PAI in Maputo Provide covers schools with both 
Portuguese and local language literacy programs, and WV in Nampula Province covers 
Portuguese literacy programs only, the baseline literacy assessment was performed in 
Portuguese because the local language curriculum was still under development. At the time 
of the baseline, only Grade 1 students would have begun to receive the local language 
curriculum, so the Grade 2 and 4 students being assessed would not have been exposed. 
Selection of Comparison Districts 
We used the following criteria for selecting comparison districts: 
located within the same province as the intervention districts (Nampula and 
Maputo) 
does not have any other known literacy interventions taking place or planned for the 
foreseeable future (5 years) 
geographically close to districts with intervention schools 
has similar urban/rural makeup to the intervention districts 
has similar numbers of students and schools (thus, similar numbers of students in 
each school) 
has similar makeup of languages spoken at home 
Based on these criteria, schools in the remaining districts in Maputo, except for the City of 
Maputo, were automatically selected as comparison schools. The City of Maputo was not 
selected because of its all-urban make up. Despite this deliberate choice, Maputo 
comparison districts still tended to be more urban than the Maputo intervention districts. 
Maptuo comparison districts were also more likely to speak Portuguese at home than any of 
the intervention districts. 
The two Nampula comparison districts were selected by eliminating districts in Nampula 
that have or are planning to have some type of (non-USDA) literacy intervention or are 
primarily urban (City of Nampula) or coastal. This choice was made because such districts 
tend to have different demographic profiles than the intervention districts. 
2-12 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
The selected comparison districts are as follows: Meconta and Mogincual in Nampula 
Province; and Boane, Marracuene, and Namaacha in Maputo Province (Figure 3). 
Target Population within Districts
The population of interest within the targeted districts included the HTs, teachers, cooks, 
and Grade 2 and 4 students attending government schools in intervention and comparison 
districts. Table 5 lists the total number of schools and the expected numbers of Grade 2 and 
4 students in each province and study arm. School lists were compiled from the list frames 
provided by the Ministry of Education, PAI, and WV. 
Table 5. Target Population across Intervention and Comparison Districts 
Province Group Districts & Primary Language* No. of 
Schools 
 Students Sample 
Frame Grade 2 Grade 4 
Maputo 
Literacy & 
School 
Feeding 
Magude Changana 
139 11,475 8,544 
School and 
enrollment 
figures 
provided by 
PAI/ADPP. 
Manhiça Changana 
Matutuíne Changana/Rhonga 
Moamba Changana/Portuguese 
School 
Feeding 
Only 
Magude Changana 
134 4,745 3,323 
School and 
Enrollment 
figures 
provided by 
PAI/ADPP. 
Manhiça Changana 
Matutuíne Changana/Rhonga 
Moamba Changana/Portuguese 
Comparison 
Boane Portuguese/Changana 
133 19,382 12,722 2016 EMIS 
data Marracuene Portuguese/Changana 
Namaacha Changana/Portuguese 
Figure 3. 
Intervention and 
Comparison 
Districts 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-13 
Nampula 
Literacy & 
School 
Feeding 
Muecate
Nacarôa Macua 152 16,300 10,985 
WV 2016 
Status 
Report 
Comparison Meconta
Mogincual Macua 150 15,832 11,123 2016 EMIS 
data 
Note: ADPP = Ajuda de Desenvolvemento de Povo para Povo (Development Aid from People to People); EMIS = Education 
Management Information System; PAI = Planet Aid Incorporated; WV = World Vision. Number of students is based on 
available data at the time of design in 2016. *Primary language based on student reports of language spoken at home. See 
Table 20 for further detail. 
Sampling Design 
We selected separate samples for the quantitative and qualitative components of the 
baseline study. The quantitative sampling design was a multi-stage stratified cluster 
probability based random design aiming to represent each of the targeted populations 
(schools, students, teachers, and cooks). However, for the qualitative component, the 
sampling was purposive, aimed at giving good representation of parents with different 
characteristics in a subset of the schools from the quantitative sample. The strategies for 
selecting both samples are described in below in more detail. 
Quantitative Survey Sampling Design 
For the quantitative sample, we used a three-stage stratified cluster probability-based 
design to select a representative random sample of schools, classrooms, teachers, head 
teachers, cooks, and students. Schools were stratified by province, district, and intervention 
arm and then sampled with equal allocation. The use of equal allocation rather than 
probability proportional to size was because enrollment information was missing for many 
schools in the frame, and even when this information was available, it was determined to be 
inaccurate based on comparison with other sources. 
Classrooms and students were sampled once the data collection team reached the school. 
One Grade 2 classroom and one Grade 4 classroom were randomly sampled with equal 
probability within the sampled and verified schools. Eight students within each sampled 
Grade 2 and eight students within each sampled Grade 4 classroom were then selected with 
equal probability. Table 6 provides the details of the final targeted sample. 
Table 6. Summary of the Sampling Methodology and Target Sample Size 
Stage # Item Sample
(Sample Size) 
Stratification
(Strata Description) Sample Allocation Probability 
of Selection 
1 Schools (450) Province x Intervention 
(5) Equal (90 per stratum) Equal 
2 Classrooms (900) Grade 
(2 & 4) 
Equal (one Grade 2 + one Grade 4 per 
school) Equal 
3 Students (7,200) None 
Equal (8 students in selected Grade 2 
classroom + 8 students in selected 
Grade 4 classroom) 
Equal 
2-14 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Verification of Schools 
Once the sample was drawn, and prior to data collection, district officials were visited to 
verify that each sampled school met the following inclusion criteria. 
Intervention schools must 
be an active part of the McGovern-Dole School Feeding Only or Literacy & School 
Feeding intervention; 
be located in the specified district; 
have Grade 2 and Grade 4 enrollment; 
instruct (or intend to instruct) Grade 2 in either Changana, Ronga, or Portuguese (if 
in Maputo); and 
instruct (or intend to instruct) Grade 2 in Portuguese only (if in Nampula). 
Comparison schools must 
not be a part of any intervention and have no plan to be a part of an intervention for 
the foreseeable future (~5 years); 
be located in the specified district; and 
have Grade 2 and Grade 4 enrollment. 
Additional information collected/verified at the verification processes includes 
the number of classroom shifts in the school; 
the start and end times of shift(s); 
the total number of Grade 2 classes and Grade 4 classes; 
the total Grade 2 enrollment and Grade 4 enrollment; 
the location; and 
the HT name. 
Replacement Schools 
The verification process also helped the team compare the information provided in the list 
frames with what was on the ground. If a school did not have the required attributes for 
inclusion, it was replaced with the school that was most similar based on geographic 
location. A total of 32 schools were replaced. Seventeen comparison schools in Nampula 
were replaced because they were involved in a literacy intervention. Eight schools were 
replaced because they did not have Grade 4 enrollment. Another three schools had no 
Grade 4 enrollment but were not replaced because no additional replacement schools were 
available. The field team was instructed to proceed with the data collection in these schools. 
Additionally, seven schools were replaced because they were used in the pilot or practice 
sessions. 
Final Quantitative Sample Counts and Estimates 
A total of 448 schools were visited during September 4 October 6, 2017. The final sample 
included 429 HTs, 214 cooks (intervention schools only), 239 warehouse managers 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-15 
(intervention schools only), 412 Grade 2 teachers, 394 Grade 4 teachers, 3,420 Grade 2 
students, and 3,289 Grade 4 students (Table 7). 
Overall, the response rate was high and reflects a good representation of the targeted 
population with most of the response rates exceeding 85%, except for the population of 
cooks. Letters from the Ministry and districts helped with school collaboration and reducing 
non-response. Non-respondents were individuals who did not consent to be interviewed or 
were not present during data collection. As will be highlighted in the results sections, 
teachers and HTs report an average 25% rate of absenteeism, which is one of the main 
reasons for non-response. In addition, it is important to note that cooks are volunteer 
parents who do not necessarily go to schools daily and are not obligated to be in the school 
during the full school day, which resulted in lower response rates for this population 
specifically. However, compared to industry standards, the response rate of 79.3% of cooks 
is considered good for a population of volunteer workers. 
Table 7. Final Sample Counts 
Provinces Study Arm Schools HT Cooks Warehous
e 
Managers 
Teachers Students 
Grade 
2 
Grade 
4 
Grade 
2 
Grade 
4 
Maputo Literacy & 
School 
Feeding 
90 84 62 82 83 78 661 638 
School 
Feeding Only 
90 85 70 79 82 72 704 681 
Comparison 89 86 n/a n/a 86 83 666 616 
Nampula Literacy & 
School 
Feeding 
89 85 82 78 79 80 699 678 
Comparison 89 89 n/a n/a 82 81 690 676 
Actual Total 448 429 214 239 412 394 3,420 3,289 
Expected Total 450 450 270 270 450 450 3,600 3,600 
Overall Response Rate 99.6% 95.3% 79.3% 88.5% 91.6% 87.6% 95.0% 91.4% 
Note: HT = head teacher 
Qualitative Sample 
For the qualitative sample, we stratified schools that had completed survey data collection 
before qualitative data collection started by number of Grade 2 classes (as a proxy for 
school size), HTs
randomly selected schools to represent those categories across Maputo and Nampula 
Provinces. FGDs were conducted at 18 schools in Maputo and 12 schools in Nampula. In 
Maputo, FGDs were held in Magude (four FGDs), Manhiça (six FGDs), Matutuíne (three 
FGDs), and Moamba (five FGDs) districts. In Nampula, FGDs were held in Muecate (six FGDs) 
and Nacarôa (six FGDs). A larger number of schools was selected from Maputo to achieve 
2-16 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
representation in both the School Feeding Only and the Literacy & School Feeding study 
arms. 
2.2.9 Data Management 
Data Monitoring and Processing 
Data monitoring and quality control (QC) were conducted daily throughout data collection. 
Data were downloaded from the Tangerine site daily for each of the six instruments and 
checked by statisticians. Some checks were performed at the individual assessor level to 
check whether assessors were making common data collection errors. Regular 
communication between the RTI office and field teams ensured that any errors were 
addressed immediately. 
Preliminary analysis of variable distribution and logic QC were also conducted, such as 
checking that enrollment was larger than or equal to the number of students present and 
Weights Calculation 
Final sample weights were calculated for each stage of sampling (i.e., at the school, 
classroom, and student levels). The final weights consisted of the sampling weights adjusted 
for non-response. Sampling weights were calculated as the inverse of the probability of 
selection at each stage. School weights were calibrated to the known populations of schools 
by province and study arm. To account for appropriate inference to the target population 
and for the clustering effect in the sample, weights were calculated at each stage of the 
sampling. The complex sampling plan provides the stratification, clustering, sampling units, 
finite population correction, and final sampling weights for each stage. Unless explicitly 
mentioned otherwise, all estimates provided in the report were analyzed using the 
appropriate weights reflecting the complex sample design and adjusting for non-response. 
Sample Size Calculation 
The sample size calculation was derived using 95% confidence intervals (CIs) with a standard 
deviation (SD) of 14.5 and the average number of words read per minute in the EGRA 
5
. Some of the other 
assumptions driving the sample size calculation included aiming for a balanced sample 
design while accounting for an intra-class correlation (ICC) coefficient of 0.15 and a design 
effect of 3.25. The sample size and the sampling methodology were developed together to 
obtain the most cost-effective and statistically rigorous sample possible. The sample size 
and sampling methodology were aimed to provide a 95% CI of ±2.5 words per minute (wpm) 
for students in each province and study arm6
. Annex B provides full mean and precision 
estimates. 
 
5
 This EGRA was conducted with Grade 2 and 3 students in Nampula and Zambezia provinces at the beginning 
of the 2013 school year. The assessment was conducted by International Business & Technical Consultants Inc. 
and funded by the United States Agency for International Development (USAID). 
6
 USAID typically uses sample sizes and methodologies that provide 95% CIs of ±3.5 wpm. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 2-17 
2.3 Analysis Methodology 
The baseline analysis consists of calculating descriptive statistics and testing hypotheses to 
compare the core indicators across the Literacy & School Feeding, School Feeding Only, and 
Comparison schools. As a baseline, these results provide a snapshot of indicators at a 
moment in time and should not be interpreted as reflecting the impact of the program. 
Impact will be assessed using difference-in-differences analysis at midline and endline. 
2.3.1 Basic Quantitative Analysis 
The quantitative analysis consisted of running both un-weighted and weighted frequencies 
and means across all variables to explore the data overall. The weighted frequencies and 
descriptive statistics were followed by crosstabulations by relevant groups (e.g., grade 
level). This report highlights the results of the weighted analysis. 
For core indicators, two types of testing were conducted: 
1) For continuous variables: Analysis of variance (ANOVA) F-tests were conducted to 
investigate whether a significant difference exists between the two intervention 
groups and the comparison group, followed by multiple comparisons to detect the 
source of this significant difference. 
2) For categorical variables: Chi-square tests were conducted to compare the two 
intervention groups and the comparison group across response categories, followed 
by Kruskal-Wallis tests to investigate two-by-two comparisons to detect the source 
of the significant difference. 
2.3.2 Advanced Quantitative Analysis 
Schools across districts and provinces vary in their performance because of many exogenous 
factors, such as urban versus rural setting, the resources available to these schools, 
accessibility, and socio-economic level. When the comparison group outperformed the 
intervention groups across some indicators, it was not clear whether this difference was 
related to the purposeful selection of the intervention districts and schools because of their 
need or to other unknown reasons. For a more robust comparison across schools, RTI 
attempted to run propensity score matching at the school level to account for the variability 
among schools across the three study arms, provinces, and districts. A multinomial logistic 
regression was conducted, as were several logistic regression models. School enrollment 
was the only significant variable that could match the schools and was not linked to the 
main indicators of interest. With school enrollment, the model explained only 8% of the 
variability and was, therefore, not considered to be a robust model for matching schools. 
Although school-level information was not sufficient to achieve robust matching between 
schools across the study arms, we believe that matching at the student and teacher levels 
within schools should be conducted as part of the impact evaluation analysis effort. 
2.3.3 Qualitative Analysis 
Qualitative research methods were used to provide greater contextual understanding of the 
quantitative baseline results and account for parents perspectives and experiences with the 
program. These methods consisted of conducting FGDs among parents with children 
enrolled in School Feeding Only program and Literacy & School Feeding schools. Our goal 
2-18 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
 the impact of 
the McGovern-Dole program on their children; therefore, we did not conduct parent FGDs 
in Comparison schools
Recordings of the FGDs were transcribed and translated from the local languages into 
English for analysis. Once translated, the transcripts were coded according to 10 themes 
using the qualitative data analysis program Dedoose. Coded themes included parental 
ption of how 
gardens (when applicable). Reports compiling all FGD content relating to each of the themes 
were then generated. Researchers analyzed these reports to identify common topics 
corresponding to each of the themes in the original transcripts. 
_________________________________
RTI International is a registered trademark and a trade name of Research Triangle Institute. 
3. ANALYSIS 
This part of the report is divided into two subsections. The first subsection focuses on the 
indicators linked to the primary research questions and an assessment of differences 
between the two intervention groups and the comparison group. The second section 
provides descriptive statistics for all other indicators, reflecting a wider snapshot of the 
situation at baseline, including qualitative analysis of the focus group data. 
3.1 Baseline Indicators 
In this section, we present key indicators related to the primary and secondary research 
questions for the baseline and document statistical differences in the indicators across the 
study arms at baseline. The primary research questions focus on literacy ability and student 
attendance. The seco
practices and teacher attendance. 
3.1.1 Literacy 
The key indicator addressing the primary research questions on the literacy ability of 
students in Grades 2 and 4 is student performance on the EGRA. EGRA results are presented 
by subtask in Tables 8 13 in the sections below. The tables include mean scores for timed 
subtasks (i.e., letter identification, syllable reading, and oral reading fluency [ORF]) and 
mean percent scores for all non-timed subtasks (i.e., reading comprehension, listening 
comprehension, and dictation). Zero scores for each subtask are representative of students 
who could not, for example, identify a single letter or read a single word of connected text. 
When comparing across the study arms, it is important to consider that the School Feeding 
Only study arm exists exclusively in Maputo, whereas the Literacy & School Feeding and 
Comparison study arms represent averages from across Maputo and Nampula Provinces. 
For EGRA scores, we provide p-values to show the statistical significance of the difference 
between the score of each treatment group and that of the comparison group. Smaller p￾values (p<0.05) are indicative of strong statistical significance. 
At baseline, Grade 2 students in the Comparison group had fewer percent zero scores across 
all subtasks, indicating this group had stronger basic literacy skills compared to Grade 2 
students in both intervention groups. Grade 4 students in the Comparison group had fewer 
percent zero scores than Grade 4 students in both intervention groups on the listening 
comprehension, ORF, and reading comprehension subtasks. In contrast, Grade 4 students in 
the School Feeding Only group had fewer percent zero scores on the letter identification, 
syllable reading, and word dictation subtasks than Grade 4 students in the Comparison and 
Literacy & School Feeding groups. The section below presents student performance on the 
EGRA by subtask, grade level, and study arm. 
Listening Comprehension Subtask: 
Although listening comprehension is not a reading ability assessment, it is a prerequisite 
that students be able to understand the target language of instruction. The percentage of 
3-2 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
students with zero scores in this subtask indicate that these students were not able to 
correctly respond to a single listening comprehension question. 
In Grade 2, the percent of students with zero scores on listening comprehension varied: 
33.8% (Comparison), 50.8% (School Feeding Only), and 62.7% (Literacy & School Feeding). 
This difference was statistically significant, with a p-value of less than 0.001 in both 
intervention groups. Grade 2 students in the Comparison schools performed better on the 
Listening comprehension subtask than Grade 2 students in the intervention schools. The 
higher percentage of zero scores among Grade 2 students in both intervention groups 
suggests that these students did not understand the meaning of the words read aloud in the 
passage. 
Table 8. Listening Comprehension Subtask Scores 
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 33.8 50.8 62.7 21.2 25.6 42.4 
Mean Percent Scores 33.0 21.2 15.0 45.0 39.0 28.6 
p-values <0.001* <0.001* 0.043* <0.001* 
The p-values shown represent the statistical significance of differences in mean percent scores relative to the Comparison group. 
In Grade 4, students in Comparison schools performed better than students in both 
intervention schools. This difference was statistically significant, with p-values of 0.043 and 
less than 0.001, respectively. Grade 4 students in the Literacy & School Feeding intervention 
showed the highest percentage of zero scores in the listening comprehension subtask 
compared to Grade 4 students in the other two groups. 
A higher percentage of Grade 2 and 4 students scored zero in the listening comprehension 
subtask in both intervention groups compared to Grade 2 and 4 students in Comparison 
schools. This finding suggests these students struggled to listen with understanding to a 
story read aloud in the target language of instruction, although the level of the text was 
representative of texts used for literacy instruction in Grades 2 and 4. This is an important 
finding given that researc
performance (Scarborough, 2001). 
In general, student data also show that at home, students speak languages other than the 
target language of instruction. This finding may pose an added challenge to reading 
acquisition given that there is a mismatch between the language of instruction and the 
language spoken at home language (Heugh et al., 2007). 
Letter Identification Subtask:
In the letter identification subtask, zero scores indicate the percentage of students not able 
to correctly name a single lower- or upper-case letter. Grade 2 students in the Literacy & 
School Feeding intervention had a lower percentage of students with zero scores (42.9%) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-3 
than Grade 2 students in the School Feeding Only group (49.4%). This difference was 
statistically significant in the Literacy & School Feeding group, with a p-value of 0.003. 
Mean scores on the letter identification task (Table 9) show that on average, Grade 2 
students in the Comparison group and the School Feeding Only group could name 
approximately 11 correct letters per minute (clpm), whereas Grade 2 students in the 
Literacy & School Feeding group correctly name approximately 8 clpm. Given that letter 
knowledge is a predictor of reading acquisition (Hammill, 2004), the finding that 34 49% of 
Grade 2 students across study arms could not correctly name a single lower or upper-case 
letter helps explain the high percentage of zero scores on the syllable reading task in Grade 
2 across all study arms. 
Table 9. Letter Identification Subtask Scores 
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 34.0 49.4 42.9 10.2 8.6 13.9 
Mean Percent Scores 11.4 11.0 8.6 33.0 37.7 24.4 
p-values 0.708 0.003* 0.008* <0.001* 
The p-values shown represent the statistical significance of differences in mean percent scores relative to the Comparison group. 
The percentage of Grade 4 students in the School Feeding Only group with zero scores 
(8.6%) was lower than the percentage of Grade 4 students in the Comparison group (10.2%) 
and the Literacy & School Feeding (13.9%) groups. The differences were statistically 
significant, with a p-value of 0.008 in the School Feeding Only group and a p-value of less 
than 0.001 in the School Feeding & Literacy group. 
Mean scores on the letter identification subtask (Table 9) show that on average, Grade 4 
students in the School Feeding Only intervention could correctly name more letters per 
minute (approximately 38 clpm) than students in the Comparison group (approximately 33 
clpm) and School Feeding & Literacy group (approximately 24 clpm). In general, Grade 4 
students performed better than Grade 2 students on the letter identification subtask. 
Syllable Reading Subtask: 
In the syllable reading subtask, zero scores indicate the percentage of students not able to 
correctly read a single syllable. In Grade 2, 68 75% of students across study arms scored 
zero on the syllable reading subtask. 
Mean scores show on the syllable reading subtask (Table 10) show that on average, Grade 2 
students in the Comparison group and the School Feeding Only group correctly read 
approximately 6 syllables per minute, whereas Grade 2 students in the School Feeding & 
Literacy group correctly read approximately 4 syllables per minute. 
3-4 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
These results help explain low performance on the oral reading fluency subtask among 
Grade 2 students in both intervention groups. Research shows that when students struggle 
to read basic syllables, they often struggle with reading words made up of two or more 
syllables (Ficksman & Adelizzi, 2010). 
Table 10. Syllable Reading Subtask Scores 
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 68.2 69.7 74.7 29.8 22.1 41.0 
Mean Scores 6.5 6.0 4.0 25.3 29.4 16.8 
p-values 0.539 0.001* 0.019* <0.001* 
The p-values shown represent the statistical significance of differences in mean scores relative to the Comparison group. 
In the syllable reading subtask, Grade 4 students in the School Feeding group had fewer zero 
scores (22.1%) than Grade 4 students in the Comparison group (29.8%) and the Literacy & 
School Feeding group (41.0%). These findings were statistically significant, with p-values of 
0.019 in the School Feeding Only group and less than 0.001 in the Literacy & School Feeding 
intervention. The percentage of zero scores across study arms (22 41%) indicate that Grade 
4 students still struggle to read basic syllables, with students in the Literacy & School 
Feeding intervention performing the lowest on this foundational skill. 
Mean scores on the syllable subtask (Table 10) show that on average, Grade 4 students in 
the School Feeding Only Group could correctly read more syllables (approximately 29 
syllables per minute) than Grade 4 students in the School Feeding & Literacy Group 
(approximately 16 syllables per minute) and the Comparison group (approximately 25 
syllables per minute). 
In general, Grade 4 students performed better than Grade 2 students. Across study arms, 
Grade 2 and 4 student performance on the syllable reading subtask is in line with Grade 2 
and 4 student performance on the letter identification subtask. These findings align with 
research showing that when primary grade students struggle to correctly identify letter 
names, they often have difficulty learning about letter sounds and letter sound 
relationships, all of which are necessary steps in learning to correctly read syllables and 
simple words (Wood & McLemore, 2001) in the target language of instruction. 
Oral Reading Fluency (ORF) Subtask: 
In the oral reading fluency (ORF) subtask, zero scores indicate that students were not able to 
correctly read a single word in the passage. In Grade 2, 56% of students in the School 
Feeding Only schools and 52% of students in the Literacy & School Feeding schools could not 
read a single word, compared to 38% of students in Comparison schools (Table 11). Grade 2 
students in Comparison schools could read slightly better than students in the intervention 
groups. The difference in performance on the ORF subtask is statistically significant, with p-
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-5 
values of 0.013 for School Feeding Only schools and less than 0.001 for the School Feeding & 
Literacy schools. 
Mean scores on the ORF subtask show that on average (Table 11), Grade 2 students in the 
Comparison schools could correctly read more words per minute (approximately 5 cwpm) 
than Grade 2 students in the School Feeding Only (approximately 4 cwpm) and School 
Feeding & Literacy (approximately 3 cwpm) schools. The average number of words read 
correctly per minute across study arms indicates that Grade 2 students struggled with 
reading connected text. 
Table 11. Oral Reading Fluency Subtask Scores 
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 38.5 56.5 52.1 23.0 24.1 33.1 
Mean Scores 5.0 3.6 2.9 23.4 24.4 14.3 
p-values 0.013* <0.001* 0.595 <0.001* 
The p-values shown represent the statistical significance of differences in mean scores relative to the Comparison group. 
In Grade 4, the percentage of students with zero scores on the ORF subtask was 23 33% 
across study arms. Grade 4 students in the Literacy & School Feeding schools had the 
highest percentage of zero scores compared to Grade 4 students in the other two groups. 
This finding was statistically significant for the Literacy & School Feeding schools, with a p￾value of less than 0.001. 
Mean scores on the ORF subtask show that on average (Table 11), Grade 4 students in the 
Comparison and School Feeding Only study arms correctly read approximately 23 24 wpm, 
respectively. In contrast, students in the School Feeding & Literacy schools correctly read 
approximately 14 wpm. The average number of cwpm across study arms suggests that 
Grade 4 students struggled with reading the entire passage. Abadzi (2012) states that while 
reading rates necessary for reading comprehension may vary across languages, in general, 
students need to correctly read 45 60 wpm to understand what they read. In general, 
Grade 4 students performed better than Grade 2 students on the ORF subtask. 
Reading Comprehension Subtask: 
In the reading comprehension subtask, zero scores indicate the percentage of students not 
able to correctly respond to a single reading comprehension question (Table 12). Across 
study arms, the difference in the percentage of Grade 2 students with zero scores in the 
reading comprehension subtask is statistically significant: 87.8% (Comparison), 93.1% 
(School Feeding Only), and 92.6% (Literacy & School Feeding), with p<.001. In general, this 
finding suggests that the majority of Grade 2 students in both intervention groups were not 
able to understand what they read. This aligns with their inability to read most of the words 
in the passage. 
3-6 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 12. Reading Comprehension Subtask Scores 
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 87.8 93.1 92.6 52.2 57.6 70.9 
Mean Percent Scores 5.1 2.0 2.2 26.8 23.2 14.2 
p-values <0.001* <0.001* 0.189 <0.001* 
The p-values shown represent the statistical significance of differences in mean percent scores relative to the Comparison group. 
In Grade 4, 52.2 70.9% of students scored zero on the reading comprehension subtask. 
Grade 4 students in the Literacy & School Feeding schools (70.9%) had the highest 
percentage of zero scores in the reading comprehension subtask, compared to Grade 4 
students in the Comparison schools and School Feeding Only schools. This finding was 
statistically significant in the Literacy & School Feeding schools, with a p-value of less than 
0.001. 
When students read too slowly, as shown by the low range of ORF rates across study arms 
among Grade 4 students (14 24 cwpm), students are primarily focusing their time and 
attention on trying to decode the words, and in the process, they compromise their 
understanding of the meaning that the words convey (Samuels & Flor, 2007). 
Word Dictation Subtask: 
In the word dictation subtask, zero scores indicate that students were not able to correctly 
write a single word. In Grade 2, the percentage of students who scored zero was 74.1
81.5% across study arms (Table 13), indicating that students struggled to apply their 
knowledge of the relationship between letter names and sounds in Portuguese to write 
words found in grade-level text. Grade 2 students in the Literacy & School Feeding 
intervention had the highest percentage of zero scores compared to the other two groups. 
This finding was statistically significant, with a p-value of 0.002 in the Literacy & School 
Feeding intervention. The findings align with research that shows a strong correlation 
bet , given that across 
study arms, 68.2 74.7% of Grade 2 students could not read a single syllable (Table 10). 
Table 13. Word Dictation Subtask Scores
GRADE 2 GRADE 4 
Comparison School 
Feeding Only
Literacy & 
School Feeding
Comparison School 
Feeding Only
Literacy & 
School Feeding
Percent Zero Scores 74.7 74.1 81.5 38.1 25.2 52.1 
Mean Percent Scores 13.4 13.0 8.8 41.7 55.1 29.8 
p-values 0.828 0.002* <0.001* <0.001* 
The p-values shown represent the statistical significance of differences in mean percent scores relative to the Comparison group. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-7 
In Grade 4, the percentage of students with zero scores was 25.2 52.1%, with students in 
the Literacy & School Feeding intervention showing a higher percentage of zero scores. This 
suggests that more Grade 4 students in the Literacy & School Feeding intervention struggle 
with writing basic words than in the other two groups. The difference on the word dictation 
subtask is also statistically significant in both intervention groups (p<0.001). 
In the word dictation subtask, Grade 4 students performed better than Grade 2 students. In 
general, this finding is in line with better Grade 4 student performance than Grade 2 student 
performance on subtasks that measure foundational skills, such as letter identification, 
which are needed to develop the ability to write basic words. 
3.1.2 Student Attendance 
At baseline, no significant difference was detected in student attendance by study arm in 
either grade. The percentage of absent students was high across all groups with no 
significant differences across gender. Overall, 28% of Grade 2 students and 27% of Grade 4 
students were absent during at least one day in the last week (Figure 4). 
Figure 4. Percent of Students Absent At least One Day Last Week, by Grade and Study 
Arm 
As an additional measure of student attendance, we compared class enrollment rosters with 
the number of students present in the classroom on the day of the survey (Figure 5). School 
Feeding & Literacy schools had higher absence rates than Comparison schools (p<0.001). 
School Feeding Only schools had lower absentee rates than Comparison schools, but the 
difference was only significant in Grade 2. Student illness was cited as the primary reason 
for absences across gender. This is discussed further in Section 3.2.1. 
29.4 30.3 25.6 28.0 27.3 24.8
Comparison School
Feeding Only
Literacy &
School
Feeding
Comparison School
Feeding Only
Literacy &
School
Feeding
Grade 2 Grade 4
3-8 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Figure 5. Average Percent of Students Absent in Classroom
3.1.3 Secondary Indicators 
At baseline, we found several differences in secondary indicators by study arm. In general, 
diversity, and minimum acceptable diet (World Health Organization [WHO], 2010; Figure 6). 
Minimum meal frequency is defined as three or more meals per day. Minimum dietary 
diversity is defined as 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). Minimum acceptable diet comprises 
individuals who have both minimum meal frequency and minimum dietary diversity. Less 
than half of the students surveyed had a minimum acceptable diet. Fewer students achieved 
minimum meal frequency than minimum dietary diversity. The percentage of students with 
minimum dietary diversity and minimum acceptable diet ranged from lowest in School 
Feeding Only schools to highest in Literacy & School Feeding schools, with Comparison 
schools falling in between (both p<0.001). The percentage of students with minimum meal 
frequency was lowest in the Comparison arm and highest in the Literacy & School Feeding 
arm, with School Feeding Only arm falling in between (p<0.001). 
Figure 6. Percent of Students Achieving Nutrition Indicators, by Study Arm 
20.7 16.8
28.7 22.3 19.7
29.5
Comparison School
Feeding Only
School
Feeding &
Literacy
Comparison School
Feeding Only
School
Feeding &
Literacy
69.4
48.6
38.7
54.4 56.2
32.8
73.6
59.6
47.5
Minimum Dietary Diversity
(MDD)
Minimum Meal Frequency
(MMF)
Minimum Acceptable Diet
(MDD + MMF)
Comparison School Feeding Only Literacy & School Feeding
Grade 2 Grade 4 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-9 
We asked students eight nutrition and seven hygiene and sanitation questions based on the 
curriculum provided by the program implementers. The nutrition questions focused on food 
groups, nutrients, and consumption of a balanced diet. The hygiene and sanitation 
qu
nutrition knowledge was very low and that their hygiene/sanitation knowledge was 
moderate (Figure 7
the Literacy & School Feeding schools had the lowest nutrition and hygiene/sanitation 
knowledge scores, whereas those in Comparison schools had the highest scores; School 
Feeding Only schools fell in between (both p<0.001). 
Figure 7. Student Nutrition and Hygiene and Sanitation Knowledge Scores 
Teacher absenteeism was common. Across the study arms, 23% of teachers were absent at 
least once during the last week, based on self-report data (Figure 8). We found no significant 
differences in the percentage of teachers absent by study arm. However, the mean number 
of teachers absent differed significantly by study arm (p<0.001), based on data reported by 
HTs. School Feeding Only schools had the fewest teachers absent, and Literacy & School 
Feeding schools had the most, with Comparison schools falling in between. 
Figure 8. Percent of Teachers Absent At least Once Last Week 
3.2
1.0
3.0
0.7
2.7
0.5
 Hygiene & Sanitation Knowledge Score (max 8) Nutrition Knowledge Score (max 7)
Comparison School Feeding Only School Feeding & Literacy
20.7
27.8 24.3
Comparison School Feeding Only Literacy & School
Feeding
3-10 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
3.2 Baseline Descriptive Statistics 
In this section, we provide descriptive results for all survey variables included in the theory 
of change for the evaluation by type of respondent: students, teachers, HTs, and cooks. The 
qualitative findings from parents are also presented and triangulated with the survey 
results. 
3.2.1 Students 
Student Demographic Characteristics 
Students provided informati
characteristics. On average, students in Grade 2 were 8 years old, and students in Grade 4 
were 10 11 years old (Table 14). Across study arms, approximately half of the students 
interviewed were female. A higher percentage of students in Comparison schools had piped 
water or a tap in their homes (approximately 32%) compared to those in School Feeding 
Only (approximately 18%) and Literacy & School Feeding (approximately 15%) schools. For 
students in School Feeding Only and Literacy & School Feeding schools, wells or boreholes 
were the most common household water source, followed by communal taps. Similar 
percentages of students in Comparison schools had access to water from wells or boreholes 
and piped water. 
moderate, ranging from three to four items out of nine. On average, students in the Literacy 
& School Feeding study arm had the lowest number of household assets. 
Table 14. Student Responses on Household Demographic Characteristics, by Grade and 
Study Arm 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,386 n=666 n=1,368 n=1,349 n=616 n=1,323 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Mean ± 95% 
CI band 
Student age 8.5 [±.2] 8.1 [±.2] 8.7 [±.2] 10.9 [±.2] 10.4 [±.2] 11.3 [±.2] 
% (n) % (n) % (n) % (n) % (n) % (n) 
Female students 50.3 (702) 50.0 (344) 49.5 (680) 51.2 (685) 50.4 (310) 50.3 (670) 
Household 
water 
source 
River, stream, 
or lake 
12.7 (273) 15.5 (132) 20.0 (321) 12.7 (266) 19.3 (152) 17.5 (297) 
Well or 
borehole 
29.9 (460) 40.3 (250) 41.0 (549) 28.6 (425) 35.8 (216) 43.6 (569) 
Communal tap 21.4 (311) 24.8 (158) 23.2 (297) 20.8 (287) 23.3 (146) 20.8 (260) 
Water pipe/tap 
in home 
31.7 (268) 17.3 (73) 14.0 (130) 33.7 (300) 18.8 (64) 16.1 (145) 
Water truck or 
tank 
2.1 (24) 1.9 (17) 1.1 (21) 2.4 (29) 2.7 (12) 0.7 (18) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-11 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Other 2.2 (22) 0.2 (1) 0.7 (6) 1.8 (19) 0.1 (2) 1.3 (9) 
Household 
asset index 
Radio 57.9 (748) 61.2 (398) 58.9 (780) 56.8 (731) 57 (369) 59.8 (800) 
Mobile phone 73.3 (928) 86.8 (575) 66.7 (930) 76.1 (962) 90.5 (544) 72.4 (960) 
Television 51.8 (511) 55.9 (304) 34.4 (415) 53.8 (542) 52.9 (290) 32.9 (410) 
Computer 20.7 (216) 24.6 (106) 12.5 (139) 18.7 (190) 12.7 (74) 10.2 (107) 
Refrigerator 41.7 (379) 35.6 (164) 22.0 (229) 39.1 (368) 31.5 (161) 21.4 (225) 
Bicycle 41.0 (593) 44.3 (304) 4.09 (648) 43.2 (615) 50.4 (315) 51.0 (710) 
Motorbike 18.0 (224) 16.2 (100) 19.1 (237) 17.8 (224) 13.9 (87) 20.5 (233) 
Car/truck 24.6 (248) 32.9 (185) 14.9 (184) 22.2 (227) 29.9 (157) 14.6 (177) 
Electricity 57.9 (590) 47.2 (251) 35.6 (415) 56.0 (599) 44.7 (224) 30.4 (353) 
Mean 
± 95% CI 
Band 
Mean ± 
95% CI 
Band 
Mean ± 
95% CI 
Band 
Mean ± 
95% CI 
Band 
Mean ± 
95% CI 
Band 
Mean ± 95% 
CI Band 
Asset index (9 items) 3.8 [±.2] 4.0 [±.3] 3.1 [±.2] 3.8 [±.2] 3.8 [±.3] 3.1 [±.2] 
Note: CI = confidence interval. 
Student Absenteeism and Attendance 
Approximately 25 30% of students across the study arms were absent from school on any 
day last week (Table 15). Illness was the main reason for students being absent, with 56
66% of students who were absent saying they missed school because they were ill. Other 
top reasons for missing school were other work at home (8 18%) and unavailability of a 
school uniform (7 10%). 
Table 15. Student-reported Absenteeism, by Grade and Study Arm 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,394 n=666 n=1,368 n=1,357 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Absent from school any day last 
week 
29.3 (346) 30.3 (182) 25.6 (343) 28.0 (377) 27.3 (146) 24.8 (335) 
Reason 
for 
being 
absent 
Illness 59.2 (177) 66.4 (104) 55.8 (173) 58.3 (211) 57.5 (85) 57.5 (192) 
Had other work at 
home 
7.6 (34) 13.9 (25) 11.7 (46) 11.2 (42) 11.6 (18) 17.8 (52) 
Had to care for family 
member 
3.2 (10) 1.3 (4) 3.8 (11) 6.2 (17) 6.4 (4) 5.7 (19) 
No transportation 1.9 (7) 0.0 (0) 0.4 (4) 0.8 (3) 0.2 (1) 0.5 (1) 
3-12 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Bad weather 2.5 (8) 1 (3) 0.6 (1) 0.8 (3) 1.5 (4) 0.3 (3) 
Had an emergency 2.9 (9) 2.5 (5) 1.8 (8) 1.9 (10) 1.1 (2) 0.9 (5) 
Getting to school was 
too dangerous 
0.3 (1) 0 (0) 1.3 (2) .7 (2) 0.0 (0) 0.0 (1) 
Woke up late 1.3 (6) 0 (0) 2.6 (5) 1.3 (1) 0.9 (2) 0.9 (2) 
Could not find uniform 
or uniform was not 
ready 
6.7 (28) 6.7 (4) 7.5 (25) 9.6 (35) 8.5 (9) 5.9 (20) 
Treated poorly by 
teachers 
0.7 (2) 0.7 (1) 0.3 (3) 0.5 (2) 0 (0) 0.9 (4) 
Treated poorly by 
other students 
1.4 (3) 0 (0) 2.0 (3) 1.2 (6) 0 (0) 0.2 (1) 
Other 12.4 (38) 7.5 (15) 12.0 (47) 7.5 (34) 12.3 (8) 9.5 (31) 
Note: CI = confidence interval. 
We also asked teachers about the total number of students formally enrolled in the class. In 
general, enrollment for boys and girls was lower in School Feeding Only schools compared 
to Literacy & School Feeding and Comparison schools (Table 16). Across the study arms, the 
mean number of students present was lower than the mean number enrolled, and the 
mean number of girls enrolled was lower than the mean number of boys enrolled. 
Enrollment numbers were compared to the number of students present on the day of 
observation to generate the attendance rates discussed in Section 3.1.2. Sixteen to nineteen 
percent of teachers across the study arms did not have a student attendance register 
available. Most of those with student attendance registers completed them daily. 
Table 16. Teacher-reported Student Attendance, by Study Arm 
Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=331 n=163 n=312 
Mean ± 95% CI 
band 
Mean ± 95% CI 
band 
Mean ± 95% CI 
band 
Number of boys enrolled in the class 27.2 [±.8] 14.0 [±1.3] 23.9 [±.0] 
Number of boys present 18.3 [±1] 11.0 [±1.4] 14.9 [±1.3] 
Number of girls enrolled in the class 25.1 [±.8] 13.7 [±1.5] 22.3 [±.9] 
Number of girls present 17.7 [±1.2] 10.7 [±1.7] 14.3 [±1.2] 
Number of students absent on a typical day 7.3 [±.5] 3.3 [±.4] 7.8 [±.5] 
Number of students who are late on a typical day 7.3 [±.5] 4.6 [±.7] 6.2 [±.5] 
% (n) % (n) % (n) 
Status of class attendance register 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-13 
Register was not available 19.1 (67) 16.2 (27) 15.6 (51) 
Attendance records completed daily 65.4 (213) 69.7 (112) 70.2 (214) 
Attendance records completed weekly 5.9 (21) 7.6 (12) 5.6 (20) 
Attendance records completed biweekly 0.5 (3) 0.0 (0) 1.3 (2) 
Attendance records completed monthly 7.4 (21) 1.1 (2) 3.2 (8) 
Other 1.8 (6) 5.4 (10) 4.0 (16) 
Note: CI = confidence interval. 
ing on Access to School Supplies and Learning Materials 
The percentage of Grade 2 students who had a supplementary reader available on the day 
of the interview was 50 60% across study arms (Table 17). The percentage of Grade 2 
students who had the Portuguese textbook available on the day of the interview was 79
82% across study arms. The percentage of Grade 2 students who had the exercise notebook 
available on the day of the interview was 83 85% across study arms. When asked if they 
read books daily in the classroom or school library, across study arms, the rate of affirmative 
responses among Grade 2 students was 42 58% 
The percentage of Grade 4 students who had a supplementary reader available on the day 
of the interview was 48 53% across study arms. The percentage of Grade 4 students who 
had the Portuguese textbook available on the day of the interview was 71 82% across study 
arms. The percentage of Grade 4 students who had the exercise notebook available on the 
day of the interview was 92 94% across study arms. When asked if they read books daily in 
the classroom or school library, across study arms, the rate of affirmative responses among 
Grade 4 students was 52 62%. 
Table 17. 
Grade and Study Arm 
Grade 2 Grade 4
 Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,394 n=666 n=1,368 n=1,357 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Student has 
supplementary reader 
available today 
52.0 (655) 59.8 (405) 50.8 (690) 48.1 (603) 53.8 (341) 48.9 (658) 
Student has Portuguese 
textbook available today 
82.1 (1027) 84.0 (542) 79.2 (1016) 74.2 (1011) 82.8 (509) 71.7 (975) 
Student has exercise 
notebook available 
today 
84.4 (1169) 83.5 (538) 85.7 (1143) 94.9 (1265) 93.1 (581) 92.6 (1219) 
Student reads books 
daily in the classroom or 
school library 
42.7 (524) 58.3 (357) 47.5 (587) 52.8 (681) 62.4 (386) 55.4 (689) 
3-14 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
The percentage of Grade 2 students who reported attending an afterschool learning club 
was 16 18% across the study arms (Table 18). Reading clubs, homework clubs, and math 
clubs were the most common afterschool learning activities reported by students across the 
study arms. When asked if they had received a certificate or award in school to celebrate 
progress in reading, 8% of Grade 2 students responded affirmatively. 
The percentage of Grade 4 students who reported attending an afterschool learning club 
was 18 25% across the study arms. Reading clubs, homework clubs, and math clubs were 
the most common afterschool learning activities reported by students across the study 
arms. Among the Grade 4 students who reported attending an afterschool learning club, 
approximately 29 34% said they attended a reading club, 19 32% said they attended a 
homework club, and 17 32% said they attended a math club. When asked if they had 
received a certificate or award in school to celebrate progress in reading, 10 12% of Grade 4 
students responded affirmatively. 
Table 18. 
Grade 2 Grade 4
Comparison 
School 
Feeding 
Only 
Literacy 
+& School 
Feeding 
Comparison 
School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,394 n=666 n=1,368 n=1,357 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Attends afterschool learning club 17.2 (183) 16.2 (107) 18.6 (228) 20.3 (254) 18.7 (130) 25.2 (297) 
Type of 
afterschool 
learning club 
student 
attends 
Math club 19.0 (35) 21.5 (23) 18.4 (38) 17.5 (40) 32.1 (41) 19.5 (52) 
Science club 2.0 (2) 1.0 (1) 1.5 (2) 12.3 (21) 14.1 (16) 10.1 (21) 
Gardening club 0.9 (1) 1.3 (1) 0.0 (0) 1.3 (2) 0.4 (1) 0.4 (1) 
Reading club 23.3 (46) 35.7 (33) 35.9 (65) 29.4 (74) 33.7 (43) 34.1 (85) 
Homework club 35.4 (52) 15.8 (20) 9.7 (22) 32.7 (71) 27 (34) 19.1 (56) 
Other 1.6 (6) 10.3 (14) 13.4 (30) 5.1 (14) 13.3 (20) 13.7 (43) 
Student received a certificate or 
award in school to celebrate 
progress in reading 
8.4 (89) 8.8 (57) 8.7 (105) 10.6 (115) 16.2 (104) 12 (158) 
The percentage of Grade 2 students who reported that the teacher uses Portuguese as the 
language of instruction was 83 88% across the study arms (Table 19). The percentage of 
students in each study arm who reported that the teacher uses another language of 
instruction (Macua or Changana) was 31 41%. 
The percentage of Grade 4 students who reported that the teacher uses Portuguese as the 
language of instruction was 91 96% across the study arms. The percentage of students in 
each study arm who reported that the teacher uses another language of instruction (Macua 
or Changana) was 27 39%. Given the geographical spread of local languages, Table 19 also 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-15 
shows that 0% of students in the School Feeding Only study arm reported that the teacher 
uses Macua to teach reading since students in this study arm are located in a region of 
Mozambique where Portuguese and other mother tongues are spoken. 
When asked what the teacher does when the student is unable to answer a question, the 
most common responses among both Grade 2 and Grade 4 students across study arms were 
Table 19. onal Support Received, by Grade and Study 
Arm 
Grade 2 Grade 4
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
n=1,394 n=666 n=1,368 n=1,357 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Student￾reported 
language 
teacher 
uses to 
teach 
reading in 
class 
Portuguese 88.7 (1229) 87.3 
(561) 
83.9 (1148) 96.2 (1297) 96.6 (591) 91.2 (1222) 
Rhonga 0.1 (6) 0.7 (4) 0.5 (25) 0.2 (9) 0.3 (2) 0.5 (17) 
Changana 5.9 (101) 37.6 
(309) 
16.2 (266) 3.0 (60) 32.5 (227) 13.2 (196) 
Macua 31.7 (515) 0.0 (0) 41.5 (497) 27.5 (439) 0.0 (0) 39.1 (495) 
Other 3.3 (27) 0.1 (2) 0.3 (3) 0.6 (9) 0.0 (0) 0.8 (6) 
According 
to 
student, 
what 
teacher 
does 
when 
child is 
unable to 
answer a 
question: 
Teacher rephrases/ 
explains question 
14.7 (207) 13.8 (72) 11.4 (210) 26.2 (327) 19.1 (105) 19 (272) 
Teacher provides 
example 
14.0 (139) 7.5 (45) 5.9 (100) 14.7 (153) 8.6 (61) 8.0 (134) 
Teacher 
encourages child to 
try again 
6.6 (84) 2.7 (16) 5.9 (91) 8.5 (104) 6.1 (37) 7.0 (100) 
Teacher asks 
another student 
6.9 (95) 6.1 (31) 6.6 (87) 8 (92) 4.3 (20) 8.9 (101) 
Teacher asks again 7.5 (87) 5.0 (33) 6.3 (94) 8.4 (101) 9.4 (54) 8 (105) 
Teacher corrects 
student 
8.1 (107) 10.3 (70) 7.2 (92) 11 (134) 18.4 (110) 11 (162) 
Teacher assigns a 
student helper 
5.1 (69) 3.2 (10) 4.2 (52) 4.8 (54) 1.8 (11) 4.9 (59) 
Teacher hits the 
student 
30.0 (441) 40.5 
(294) 
35.1 (505) 23.8 (379) 37.6 (250) 34.0 (474) 
Other 4.6 (54) 6.9 (35) 5.5 (74) 5.3 (68) 5.6 (33) 8.8 (94) 
Home Literacy Practices 
study arms (Table 20). Slightly more than half of Grade 2 students reported that someone at 
home helps with homework when needed. The percentage of Grade 2 students who said 
they bring home reading books from the classroom was 64 72% across the study arms. 
Among the Grade 2 students who reported that they bring home reading books from the 
3-16 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
classroom, approximately 92 96% reported that the books are in Portuguese. When asked if 
they bring home books from the school library or book bank, the rate of affirmative 
responses was 12 14%. Among the Grade 2 students who reported that they bring home 
books from the school library or book bank, roughly 80 98% reported that the books are in 
Portuguese. Grade 2 students in all study arms reported that a language other than 
Portuguese is the most widely used at home. 
When asked about how often they read at home, a higher percentage of Grade 2 students in 
the Comparison study arm (46%) and Literacy & School Feeding study arm (49%) responded 
per 49% across the study arms. 
comparison and both intervention study arms. Slightly more than half of Grade 4 students 
reported that someone at home helps with homework when needed. The percentage of 
students who said they bring home reading books from the classroom was approximately 
68 76% across the study arms. Among the students who said they bring home reading 
books from the classroom, across the study arms, roughly 96 98% also reported that the 
books are in Portuguese. When asked if they bring home books from the school library or 
book bank, the rate of affirmative responses across the study arms was 10 18%. Among the 
students who reported that they bring home books from the school library or book bank, 
approximately 95 100% reported that the books are in Portuguese. Grade 4 students 
reported that a language other than Portuguese is the most widely used at home. 
Table 20. 
Grade 2 Grade 4 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
n= 1,394 n= 666 n=1,368 n=1,357 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Someone at home helps with 
homework when needed 
62.0 (741) 58.6 (328) 52.4 (683) 61.2 (747) 59.5 (335) 57.7 (756) 
Student brings home reading 
books from the classroom 
64.9 (816) 65.2 (424) 72 (918) 68.4 (908) 74.8 (465) 76.8 (1000) 
Language of 
classroom 
books 
brought home 
Portuguese 96.4 (705) 96.6 (384) 92.5 (785) 98.1 (853) 98.7 (456) 96.2 (938) 
Rhonga 0.3 (3) 0.0 (0) 0.4 (6) 0.1 (3) 0.0 (0) 0.0 (0) 
Changana 0.6 (9) 3.1 (16) 2.1 (17) 0.1 (1) 1.2 (6) 0.0 (2) 
Other 
language 
0.3 (3) 0.1 (1) 0.2 (2) 0.6 (4) 0.1 (1) 0.3 (2) 
Brings home books from the 
school library/book bank 
12.3 (137) 14.8 (91) 12.5 (128) 10.5 (111) 18.0 (88) 13.6 (135) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-17 
Grade 2 Grade 4 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
Language of 
school 
library/book 
bank books 
brought home 
Portuguese 98.6 (129) 98.4 (85) 80.7 (104) 100 (110) 97.6 (84) 95.4 (126) 
Rhonga 0.8 (1) 0.0 (0) 0.3 (2) 0.0 (0) 0.0 (0) 0.4 (1) 
Changana 0.3 (1) 1.6 (3) 8.1 (7) 0.0 (0) 1.6 (2) 0.7 (3) 
Macua 0.3 (1) 0.0 (0) 9.2 (11) 0.0 (0) 0.0 (0) 3.5 (6) 
Other 0.0 (0) 0.0 (0) 1.7 (1) 0.0 (0) 0.7 (1) 0.0 (0) 
Language 
child speaks 
at home 
Portuguese 38.1 (357) 21.7 (81) 12.1 (146) 40.2 (365) 23.7 (83) 10.7 (123) 
Rhonga 0.7 (20) 1.2 (10) 0.9 (43) 0.9 (22) 0.6 (5) 1.1 (51) 
Changana 17.7 (321) 74.8 (552) 31.8 (477) 16.2 (304) 75.6 (522) 34.4 (471) 
Macua 43.1 (652) 0.0 (0) 54.8 (658) 42.0 (636) 0.0 (0) 53.1 (652) 
Other 0.5 (7) 2.2 (13) -.5 (18) 0.7 (14) 0.1 (1) 0.7 (21) 
Frequency 
student reads 
on his/her 
own at home 
Never 46.3 (662) 38.4 (273) 49.8 (675) 31.9 (436) 26 (152) 29.7 (405) 
Sometimes 43.7 (547) 50.5 (283) 40.4 (519) 53.5 (676) 54.5 (339) 53.9 (705) 
Every day 9.9 (95) 11.0 (67) 9.8 (85) 14.6 (193) 19.5 (105) 16.5 (165) 
Frequency 
someone tells 
the student a 
story at home 
Never 49.3 (683) 46.4 (339) 48.1 (634) 44.3 (594) 42.2 (278) 43.6 (538) 
Sometimes 41.9 (547) 44.4 (265) 46.3 (599) 50.4 (644) 51.7 (289) 49.2 (662) 
Every day 8.8 (79) 9.2 (40) 5.6 (64) 5.3 (61) 6.2 (37) 7.3 (83) 
When asked how often they read on their own at home, the percentage of Grade 4 students 
54%. When asked how often 
someone at home tells them a story, the percentage of Grade 4 students who responded 
Nearly all students in schools with School Feeding Only and Literacy & School Feeding 
reported that their school serves school meals, whereas approximately 7% of students in 
Comparison schools said that their school serves school meals (Table 21). Among the 
students reporting school meals, approximately three quarters of those in schools with 
School Feeding Only and Literacy & School Feeding interventions said their school served a 
meal yesterday or on the previous day they were in school. Porridge was the most common 
meal in schools with School Feeding Only or Literacy & School Feeding, whereas Comparison 
schools with a feeding program mainly served food other than porridge. Of the students 
who had school meals served at their schools, more than half in some cases, more than 
two thirds said they ate fewer meals at home when they received a school meal. 
Approximately 80 90% of students said they felt less hungry on days when they received a 
school meal. 
We asked students if they felt hungry at the time of the interview, and affirmative responses 
ranged from 31% to 56% across study arms and grades. The percentage of students who 
3-18 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
said their household had experienced hunger since the beginning of the school year was 8
21% across all study arms. Among those reporting household hunger, roughly 12 34% said it 
occurred every month. 
Table 21. Student Responses on School Meals, Current Hunger, and Household Hunger, 
by Grade and Study Arm 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy 
& 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy 
& 
School 
Feeding 
n=1,386 n=666 n=1,368 n=1,349 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
School serves school meals 6.9 (114) 98.5 
(645) 
96.7 
(1317) 
7.6 (111) 98.8 (605) 97.7 
(1292) 
School meal served yesterdaya
 58.9 (79) 79.7 
(496) 
74.6 
(998) 
66.9 (83) 74.8 (457) 71.0 
(985) 
Food 
served at 
school 
meal 
yesterdaya
Porridge 28.5 (19) 99.4 
(496) 
96.9 
(980) 
38.6 (17) 98.9 (443) 97.1 
(941) 
Porridge & other 
foods 
2.2 (2) 0.2 (1) 1.4 (7) 4.0 (4) 0.4 (2) 1.0 (6) 
Other foods only 69.3 (56) 0.4 (2) 1.7 (28) 57.4 (62) 0.7 (2) 1.9 (28) 
Eats fewer meals at 
home when 
receives a school 
meal 
54.1 (60) 69.6 
(438) 
46.0 
(677) 
63.0 (68) 69.2 (415) 52.7 
(688) 
Feels less hungry 
on days when 
receives a school 
meal 
89.0 (104) 90.9 
(600) 
83.1 
(1107) 
89.4 (102) 91.1 (556) 84.9 
(1096) 
Feels hungry nowb
 55.9 (755) 40.3 
(231) 
51.9 
(725) 
46.1 (660) 30.8 (191) 39.9 
(551) 
Household hunger 21 (273) 14.7 (94) 19 (243) 20.1 (243) 7.5 (48) 17.1 
(216) 
Household 
hunger 
frequency 
Every month 28.7 (50) 15.3 (9) 16.5 (31) 26.7 (57) 12.2 (4) 33.5 (45) 
During the rainy 
season 
20.3 (38) 12.2 (8) 15.4 (31) 13.2 (35) 15.2 (4) 14.4 (31) 
Other 50.9 (92) 72.6 (53) 68.1 (99) 60.0 (94) 72.6 (27) 52.0 (87) 
aThis is question was asked only if students reported that their school serves school meals. 
bWe used Bennett & Blissett (2014) to develop this question. 
Student Nutrition Knowledge and Practices 
-report, varied across study arms (Table 
22). Higher percentages of Literacy & School Feeding (approximately 74%) and Comparison 
(approximately 69%) schools than School Feeding Only (approximately 54%) schools had 
achieved minimum dietary diversity on the previous day. The food groups consumed by the 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-19 
fewest students were dairy products (18.5%) and eggs (23.5%), whereas grains, roots, and 
tubers were consumed by nearly all students (98.5%). Approximately 40 60% of students 
across the study arms and grades achieved minimum meal frequency on the previous day. 
Achievement of minimum acceptable diet was low: 31 50% across the study arms. On 
average, students had eaten four food groups and two to three meals on the previous day. 
ow. On 
average, students correctly answered one out of eight nutrition knowledge questions. 
Table 22. Student Nutrition Practices and Knowledge, by Grade and Study Arm 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,386 n=666 n=1,368 n=1,349 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Minimum dietary 
diversitya
67.7 (905) 53.5 (356) 71.8 (899) 71.6 (930) 55.7 (348) 76.1 (937) 
Minimum meal 
frequencyb
42.8 (545) 55.3 (332) 57.4 (703) 56.9 (688) 57.5 (331) 62.4 (795) 
Minimum acceptable 
dietc
34.1 (419) 30.5 (186) 46 (516) 45.2 (537) 36.0 (211) 49.6 (596) 
Mean ± 95% 
CI band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Mean ± 95% 
CI band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Dietary diversity 4.2 [±.1] 3.7 [±.2] 4.2 [±.1] 4.3 [±.1] 3.8 [±.2] 4.4 [±.1] 
Meal frequency 2.4 [±.1] 2.6 [±.1] 2.7 [±.1] 2.6 [±.1] 2.7 [±.1] 2.8 [±.1] 
Nutrition knowledge 
score (8 items) 
0.7 [±.1] 0.4 [±.1] 0.4 [±.1] 1.3 [±.1] 1.0 [±.2] 0.7 [±.1] 
Note: CI = confidence interval 
aMinimum dietary diversity is 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). 
bMinimum meal frequency is three or more meals per day. 
cMinimum acceptable diet comprises individuals who have both minimum dietary diversity and minimum meal frequency. 
Student Hygiene and Sanitation Knowledge and Practices 
Among students in schools that provide school meals, less than half the students across all 
study arms said that they always wash their hands before eating their school meal (Table 
23
to 3.5 items out of seven, on average. Grade 2 students had lower hygiene and sanitation 
knowledge than Grade 4 students. 
3-20 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 23. Student Hygiene and Sanitation Practices and Knowledge, by Grade and Study 
Arm 
GRADE 2 GRADE 4 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=1,386 n=666 n=1,368 n=1,349 n=616 n=1,323 
% (n) % (n) % (n) % (n) % (n) % (n) 
Frequency 
of washing 
hands with 
soap or 
ash before 
school 
meal 
Never 7.4 (7) 3.1 (28) 15.4 (158) 6.9 (5) 4.4 (27) 11.2 (126) 
Sometimes 44.6 (47) 54.8 (340) 49.1 (559) 45.8 (51) 47.0 (294) 45.4 (539) 
Always 48.1 (45) 42.1 (205) 35.5 (411) 47.3 (48) 48.6 (235) 43.4 (501) 
Mean ± 95% 
CI Band 
Mean ± 
95% CI 
Band 
Mean ± 
95% CI 
Band 
Mean ± 95% 
CI Band 
Mean ± 
95% CI 
Band 
Mean ± 
95% CI 
Band 
Hygiene and sanitation knowledge 
score (7 items) 
2.9 [±.1] 2.6 [±.2] 2.5 [±.1] 3.5 [±.1] 3.5 [±.2] 3.1 [±.1] 
Note: CI = confidence interval. 
3.2.2 Teachers 
Teacher Profile 
Sixty percent of teachers interviewed in the Comparison study arm were female (Table 24), 
whereas less than half of teachers in the School Feeding Only and the Literacy & School 
Feeding study arms were female. The percentage of teachers who reported that they taught 
reading in Portuguese was 98 99% across study arms. The percentage of teachers who said 
they had attended and finished studies at a teacher training college (TTC) was 95 96% 
across the study arms. The percentage of teachers who reported they had received 
preservice training on the teaching of reading and writing to students in the primary grades 
was 58 61% across the study arms. 
The percentage of teachers who reported attaining certification was 95 97% across groups. 
Less than 20% of teachers across the study arms attended an Ajuda de Desenvolvemento de 
Povo para Povo (Development Aid from People to People [ADPP])-supported TTC. The 
percentage of teachers who reported the certificate level as the highest level of academic 
education attained was 64 71% across the study arms. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-21 
Table 24. Teacher Profiles, by Study Arm 
 Comparison School Feeding 
Only 
Literacy & 
School Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Female teacher 60.4 (153) 36.8 (53) 47.8 (124) 
Languages in 
which reading 
is taught 
Portuguese 99.7 (300) 98.2 (144) 98.8 (288) 
Rhonga 9.1 (19) 2.4 (4) 3.6 (16) 
Changana 15.6 (47) 60.4 (93) 29.1 (74) 
Macua 28.2 (112) 0.0 (0) 39.9 (122) 
Other 0.0 (0) 0.0 (0) 1.6 (7) 
Attendance 
of TTC 
Did not attend TTC 1.8 (7) 1.8 (3) 1.4 (3) 
Still attends TTC 1.9 (3) 3 (3) 3.0 (6) 
Attended and finished TTC 96.3 (292) 95.3 (141) 95.5 (283) 
TTC teacher attended 
IFP 77.5 (232) 87.9 (127) 78.5 (224) 
ADPP 13.3 (39) 9.7 (12) 16.3 (47) 
UP 9.2 (23) 2.4 (4) 5.3 (11) 
Attainment of certification 97.4 (291) 97 (142) 95.7 (282) 
Teacher is a currently trainees expected to graduate from a TTC 7.7 (20) 4.7 (8) 6.1 (17) 
TTC teacher 
is expected 
to graduate 
from 
IFP 21.5 (5) 62.5 (5) 76.6 (14) 
ADPP 2.5 (1) 0.0 (0) 0.0 (0) 
UP 76.1 (12) 37.5 (3) 23.4 (5) 
Highest level 
of academic 
education 
Form 6 (Basic) 18.8 (78) 23.1 (37) 25.0 (76) 
Certificate 64.4 (192) 71.6 (105) 69.3 (199) 
Diploma 1.4 (2) 0.0 (0) 2.1 (6) 
Advanced diploma 15.4 (30) 5.3 (5) 3.4 (10) 
Other 0.0 (0) 0.0 (0) 0.2 (1) 
During pre-service training, the teacher received specific training 
on how to teach reading and writing to students in the primary 
grades 
61.4 (190) 60.9 (93) 58.9 (175) 
Note: IFP = Instituto de Formação de Professores (Teacher Training Institute); TTC = teacher training college; UP = Universidade 
Pedagógica de Moçambique 
Teacher Training 
The percentage of teachers who reported that they attended an in-service training on how 
to teach reading in the primary grades varied across the three study arms (Table 25). Fifty￾two percent of teachers in the Literacy & School Feeding arm, 38% of teachers in the School 
Feeding Only arm, and 26% of teachers in the Comparison arm reported that they attended 
an in-service training on how to teach reading in the primary grades. The percentages of 
teachers who attended a WV or PAI training were 70% in the Literacy & School Feeding arm, 
3-22 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
63% in the School Feeding Only arm, and 10% in the Comparison arm. When asked if they 
received training on the fundamentals of reading instruction, the percentage of affirmative 
responses was 57 64% across study arms. The most common activities in which teachers 
participate during in-service training were analyzing the student exercise book, learning how 
to plan and prepare for a lesson, and learning how to assess students during the lesson. 
Table 25
 Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
This year, the teacher attended any in-service 
training on how to teach reading in the primary 
grades 
26.2 (87) 38.1 (54) 52.0 (157) 
Teacher attended WV/PAI training 10.9 (29) 63.1 (90) 70.3 (212) 
Teacher received training on the fundamentals of 
reading instruction 
57.5 (174) 58.0 (91) 64.0 (190) 
Types of 
activities 
teacher 
participates 
in during in￾service 
training 
Analysis of curriculum and syllabus 35.1 (109) 47.3 (68) 25.9 (88) 
Analysis of teacher guide 22.9 (71) 29 (42) 19.5 (63) 
Practice teaching lessons from the 
teacher guide 
23.4 (63) 22.5 (33) 19.5 (52) 
Analysis of the student exercise book 35.9 (114) 61.5 (89) 42.3 (131) 
How to plan and prepare for a lesson 64.6 (208) 81.1 (121) 67.1 (204) 
Improvisation and use of teaching and 
learning aids 
35.2 (126) 55.6 (80) 42.0 (134) 
How to assess students during the 
lesson 
41.7 (147) 65.1 (100) 54.8 (169) 
How to use assessments to adjust 
instruction 
23.1 (80) 40.8 (63) 30.7 (100) 
How to give students feedback 22.6 (67) 42.6 (66) 28.8 (98) 
How to use supplementary 
readers/decodable texts 
25.2 (87) 29.6 (48) 22.0 (72) 
How to apply new strategies for 
reading instruction 
26.8 (80) 33.7 (52) 28.2 (88) 
Teacher Absenteeism 
Across study arms, 21 28% of teachers were absent from school during the last week, and 
13 17% of teachers were late to school during the last week (Table 26). 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-23 
Table 26. Teacher-reported Absenteeism, by Study Arm 
Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n)
Teacher absent from school any day last week 20.7 (57) 27.8 (43) 24.3 (73) 
Teacher late to school any day last week 15.7 (46) 17.2 (26) 13.2 (46) 
Note: CI = confidence interval. 
Teacher Access to School Supplies and Learning Materials 
The percentage of teachers who reported that they have access to a teacher guide was 51
58% across study arms (Table 27). A higher percentage of teachers in the School Feeding 
Only study arm (72%) reported that students have exercise books that can be used during 
reading instruction than in the Comparison (58%) and Literacy & School Feeding (65%) study 
arms. Higher percentages of teachers in the School Feeding Only arm (95%) and the Literacy 
& School Feeding arm (95%) reported having access to school supplies than in the 
Comparison arm (87%). Among the teachers who reported having access to school supplies, 
most reported access to supplies such as pens/pencils, notebooks, and chalk. Teachers 
across the study arms reported low access to school supplies such as paint, markers, and 
chart paper. The percentage of teachers across the study arms who reported access to 
supplementary readers was 72 88%. 
Table 27. Teacher Access to Supplies and Learning Materials, by Study Arm 
Comparison School 
Feeding 
Only 
Literacy 
& School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Teacher has a teacher guide for reading instruction 51.8 (156) 51.5 (76) 58.2 (162)
Students have exercise books that can be used during reading instruction 58.2 (164) 72.0 (107) 65.7 (192)
Teacher has access to school supplies 87.6 (270) 95.2 (138) 95.7 (273)
Supplies 
the 
teacher 
can 
access 
Pens/pencils 91.9 (243) 93.1 (129) 88.3 (251)
Notebooks 88.7 (239) 88.7 (123) 83.8 (229)
Chalk 83.8 (225) 75.6 (104) 84.3 (234)
Cardboard 73.6 (196) 78.1 (106) 66.7 (188)
Paint 7.1 (26) 31.9 (48) 13.9 (43) 
Markers 9.7 (31) 25.6 (38) 21.9 (66) 
Chart paper 27.1 (80) 45.6 (67) 40.2 (118)
Other 12.5 (42) 10.0 (15) 16.5 (54) 
Supplementary readers available for use in this classroom 72.7 (51) 88.7 (50) 78.0 (105)
3-24 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Teacher Use of School Supplies and Learning Materials 
The percentage of teachers who said they used the teacher guide for reading instruction 
varied across the study arms (Table 28). Fifty-two percent of teachers in the Comparison 
arm, 48% of teachers in the School Feeding Only arm, and 60% of teachers in the Literacy & 
School Feeding arm reported that they used the teacher guide for reading instruction. The 
percentage of teachers who reported that students use the exercise books during reading 
instruction was 75 86% across the study arms. The percentage of teachers who reported 
using school supplies during reading instruction was 91 97% across the study arms. Of the 
teachers who said that they used school supplies during reading instruction, most reported 
using supplies such as pens/pencils, notebooks, chalk, and cardboard. Teachers across the 
study arms reported low use of school supplies such as paint, markers, and chart paper. 
Among the teachers who reported using school supplies during reading instruction, most 
indicated that they used supplementary readers once a week. 
Table 28. Teacher Use of School Supplies and Learning Materials, by Study Arm 
Comparison School 
Feeding 
Only 
Literacy & 
School Feeding 
Count n=331 n=163 n=312 
Estimates % (n) % (n) % (n) 
Teacher uses the teacher guide for reading instruction 52.9 (163) 48.2 (74) 60.8 (168) 
Student use the exercise books during reading instruction 75.3 (218) 86.9 (126) 83.2 (239) 
Teacher uses school supplies during reading instruction 91.2 (283) 95.2 (138) 97.1 (280) 
Teacher uses the 
following school 
supplies 
Pens/pencils 69.9 (202) 68.1 (91) 74.2 (205) 
Notebooks 71.0 (202) 65.0 (87) 65.5 (185) 
Chalk 77.6 (219) 80.0 (111) 80.2 (229) 
Cardboard 79.9 (231) 83.1 (115) 70.7 (203) 
Paint 12.3 (35) 20.6 (31) 11.7 (36) 
Markers 10.1 (28) 15.0 (22) 17.0 (49) 
Chart paper 28.3 (78) 49.4 (72) 39.8 (113) 
Other 11.6 (40) 11.9 (19) 18.1 (61) 
Frequency students 
use supplementary 
readers in this 
classroom 
Never 2.2 (3) 4.8 (3) 6.6 (11) 
Once a week 72.7 (51) 88.7 (50) 78.0 (105) 
Every day 25.1 (21) 6.5 (4) 15.4 (26) 
Extracurricular Activities 
The percentage of teachers who reported that special events are held at the school to 
49% across the study arms (Table 29). Sixty￾two percent of teachers in the Comparison arm, 79% of teachers in the School feeding Only 
arm, and 48% of teachers in the Literacy & School Feeding arm reported that students 
received a certificate or award in school to celebrate their progress in reading. Nineteen 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-25 
percent of teachers in the Comparison arm, 19% of teachers in the School Feeding Only arm, 
and 30% of the teachers in the Literacy & School Feeding arm reported that special events 
those, lower 
percentages of teachers in Literacy & School Feeding (45%) and School Feeding Only (67%) 
groups reported that students in their class had received a certificate or award in school to 
celebrate improvement in attendance than in the Comparison arm (82%). 
The percentage of teachers who reported the existence of an afterschool club in the school 
varied across the three study arms. Seventeen percent of teachers in Comparison schools, 
40% of teachers in School Feeding Only schools, and 60% of teachers in Literacy & School 
Feeding schools reported that an afterschool club exists in the school. Among the teachers 
who reported an afterschool literacy club in the school, most teachers across the study arms 
said that half the students in their class attend an afterschool literacy club. 
Table 29. on Extracurricular Activities, by Study Arm 
 Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Special events held at this school to celebrate 
students' progress in reading 
33.9 (95) 48.5 (71) 49.3 (131) 
Students in this class received a certificate or 
award in school to celebrate progress in reading 
62.9 (65) 79.5 (56) 48.6 (71) 
Special events held at this school to celebrate 
students' school attendance 
19.3 (59) 19.1 (28) 30.5 (82) 
Students in this class received a certificate or 
award in school to celebrate improvement in 
attendance 
81.7 (43) 67.7 (19) 45.5 (40) 
Afterschool literacy club in this school 17.4 (37) 40.7 (60) 60.5 (167) 
Number of students 
who attend an 
afterschool literacy club 
in this class 
None 3.3 (1) 5.5 (2) 3.1 (3) 
Half 61.2 (7) 60.0 (27) 73.0 (98) 
More than half 29.2 (4) 20 (10) 12.6 (19) 
All 6.3 (2) 14.5 (8) 11.2 (16) 
Teacher Perspectives on Supervision and Support 
The percentage of teachers who reported that the HT or assistant HT checked lesson plans 
was 97 98% across the study arms (Table 30). Among those teachers, higher percentages of 
teachers in the School Feeding Only and Literacy & School Feeding arms responded that the 
HT checks lesson plans once every week, whereas a higher percentage of teachers in the 
Comparison arm responded that the HT checks lesson plans daily. When asked whom they 
3-26 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
consult when help is needed, in each study arm, most teachers reported that they discuss 
issues casually with other teachers. 
When asked about the frequency with which classroom observations have been conducted 
by the assistant HT or HT since the beginning of the school year, in all study arms, most 
teachers responded that classroom observations occurred once per trimester. The 
percentage of teachers who reported that they received feedback from the HT or assistant 
HT was 92 96% across the study arms. In all study arms, higher percentages of teachers 
since 
the beginning of the school year. 
Table 30. Teacher Perspectives on Supervision and Support, by Study Arm 
Comparis
on 
School Feeding 
Only 
Literacy & School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Does the HT or assistant HT check lesson plans 98.8 (294) 98.8 (144) 97.2 (278) 
Frequency at which 
HT checks lesson 
plans 
Once every 2 3 months 3.8 (11) 6.3 (6) 10.5 (25) 
Once every month 10.3 (29) 11.3 (17) 11.6 (36) 
Once every 2 weeks 14.0 (42) 15.7 (23) 12.4 (34) 
Once every week 30.3 (82) 39.0 (56) 39.7 (107) 
Daily 41.6 (126) 27.7 (40) 25.8 (73) 
Person the teacher 
consults when help is 
needed 
Never need help 0.6 (2) 0.6 (1) 0.4 (2) 
There is no one to ask for 
help. 
1.2 (6) 0.6 (1) 0.0 (0) 
Discuss casually with other 
teachers 
78.0 (226) 78.7 (112) 72.8 (211) 
Discuss at meetings with 
other teachers 
37.6 (108) 34.3 (52) 33.8 (96) 
Reading coach 1.1 (4) 1.8 (2) 1.2 (4) 
Assistant HT 31.6 (97) 30.2 (45) 37.5 (114) 
HT 8.8 (36) 26.6 (41) 13.3 (48) 
Since the beginning of 
the school year, 
frequency of 
classroom teaching 
observations by the 
assistant HT or HT 
Never 16.1 (37) 13.7 (15) 6.2 (21) 
Once per trimester 50.4 (163) 44 (64) 56 (154) 
Monthly 20.7 (51) 25.6 (41) 28 (78) 
Weekly 12.8 (46) 16.7 (26) 9.8 (35) 
Provision of feedback from HT or assistant HT 92.6 (236) 94.4 (123) 96.6 (254) 
Since the beginning of 
the school year, 
frequency of reading 
coach visits 
Never 82.9 (225) 72.0 (101) 61.1 (169) 
Once per trimester 11.3 (33) 14.0 (19) 15.9 (46) 
Monthly 5.1 (16) 7.0 (7) 12.6 (29) 
Weekly 0.8 (4) 7.0 (10) 10.4 (32) 
Note: HT = head teacher. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-27 
Teacher Knowledge and Skills 
When asked if they could name the core reading skills, no teacher in any study arm reported 
being able to do so (Table 31). However, the percentage of teachers who named letter 
knowledge as a core reading skill was 58 66% across study arms. Less than half of HTs 
across the study arms named phonemic awareness, fluency, or comprehension as core 
reading skills. 
ading and writing progress, more than 
half of the teachers across the study arms reported using written tests and oral evaluations. 
When asked about how they check for student understanding during lessons, more than 
half of the teachers across the study arms reported asking questions of individual students. 
When asked about the techniques used to teach a new vocabulary word, teachers across 
the study arms gave varied responses. More teachers in the Comparison arm (54%) reported 
providing a verbal explanation, whereas a higher percentage of teachers of teachers in the 
School Feeding Only arm (49%) reported using real objects. Finally, a higher percentage of 
teachers in the Literacy & School Feeding arm (55%) reported using pictures. 
Table 31. Teacher Knowledge and Skills, by Study Arm 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Teacher can name the core reading skills 0.0 (0) 0.0 (0) 0.0 (0) 
The reading skills the teacher can 
name 
Letter knowledge 64.8 (195) 58.6 (89) 66.6 (195) 
Phonemic awareness 34.5 (98) 25.4 (38) 34.1 (112) 
Fluency 23.1 (65) 15.4 (23) 21.3 (72) 
Vocabulary 35.1 (114) 37.3 (57) 51.5 (146) 
Comprehension 43.1 (128) 37.3 (57) 35.3 (101) 
Grade level at which the teacher 
expects children to read text in the 
target language fluently 
Grade 1 7.1 (15) 5.9 (9) 4.5 (13) 
Grade 2 37.6 (102) 32.0 (45) 22.0 (57) 
Grade 3 25.1 (70) 27.2 (44) 42.4 (115) 
Grade 4 or higher 30.1 (111) 34.9 (49) 31.2 (94) 
Grade level at which the teacher 
expects students to write a 
coherent and comprehensible 
short story correctly 
Grade 1 4.3 (6) 3.6 (5) 2.5 (8) 
Grade 2 23.1 (62) 20.8 (31) 17.1 (45) 
Grade 3 28.4 (82) 29.2 (47) 35.2 (87) 
Grade 4 or higher 44.3 (144) 46.4 (63) 45.2 (142) 
Ways the teacher measures 
students' reading and writing 
progress 
Written tests 64.1 (198) 60.9 (93) 70.0 (199) 
Oral evaluations 66.7 (211) 66.9 (95) 62.1 (181) 
Observation 40.6 (122) 34.3 (51) 40.6 (126) 
Portfolios and other projects 4.1 (16) 5.9 (10) 4.4 (18) 
Homework 36.8 (111) 31.4 (45) 34.9 (95) 
3-28 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Classwork 31.1 (99) 30.2 (45) 32.4 (91) 
End-of-unit evaluation 6.0 (17) 5.3 (9) 5.5 (17) 
End-of-term evaluation 2.6 (11) 3.0 (4) 5.8 (16) 
Ways the teacher checks for 
student understanding during the 
lesson 
Ask questions of individual students 57.7 (187) 56.8 (81) 56 (174) 
Ask questions of whole class 46.4 (141) 40.8 (57) 38.9 (107) 
Ask questions of students in groups 25.3 (76) 25.4 (35) 25.6 (80) 
Give students a task and correct the 
written responses before the end of the 
lesson 
32.6 (118) 43.8 (65) 36.3 (102) 
Give students a task and correct the 
written responses after the end of the 
lesson 
21.2 (62) 35.5 (47) 25.6 (71) 
Other 7.7 (23) 7.1 (12) 8.8 (29) 
Ways the teacher uses the results 
of 
assessments in instruction 
Grade students 54.1 (171) 47.3 (74) 46.9 (135) 
subject matter 
59.8 (178) 56.8 (81) 53.5 (159) 
Plan teaching and learning activities 23.3 (76) 27.2 (37) 29.1 (82) 
needs 
21.3 (66) 36.7 (50) 21.4 (73) 
Arrange students in ability groups 4.7 (16) 3.6 (6) 7.0 (19) 
Other 12.0 (34) 3.6 (6) 5.1 (18) 
Techniques teacher uses to teach 
a new vocabulary word 
Pictures 40.6 (140) 43.2 (65) 55 (156) 
Realia 40.1 (131) 49.7 (72) 47.1 (141) 
Gestures 26.1 (77) 23.7 (38) 20.4 (75) 
Verbal explanation 54.9 (158) 49.1 (74) 51.9 (161) 
Other 15.7 (58) 13.6 (21) 12.9 (42) 
Frequency of teacher switching 
between official language of 
instruction and a vernacular 
language during the teaching and 
learning process 
Never 15.0 (29) 6.6 (8) 3.7 (8) 
Occasionally 58.3 (185) 63.5 (94) 67.8 (189) 
Often 21.1 (66) 22.2 (32) 18.8 (65) 
All of the time 5.6 (20) 7.8 (12) 9.7 (28) 
reading skills in the target 
language of instruction 
Weak 12.8 (47) 13.7 (18) 11.8 (39) 
Average 75.2 (223) 75.6 (112) 81.4 (231) 
Strong 8.9 (25) 9.5 (14) 6.8 (21) 
Not applicable does not teach in the 
target language 
3.1 (5) 1.2 (2) 0.0 (0) 
writing skills in the target language 
of instruction: 
Weak 11.9 (39) 12.4 (20) 9.2 (30) 
Average 71 (224) 72.8 (106) 73.5 (215) 
Not applicable does not teach in the 
target language 
13.9 (35) 13.6 (19) 17.1 (46) 
The percentage of teachers across the study arms who reported that they occasionally 
switch between the official language of instruction and a local language was 58 67%. Across 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-29 
the study arms, 75 average. Similarly, 
average was 71- 73%. Interestingly, while students were generally not competent in basic 
 performance is average. When asked at 
what grade level the teacher expects children to read text in the target language fluently, 
 varied. A higher percentage of teachers in the Comparison arm (37%) 
reported that by Grade 2, students should be able to read text in the target language 
fluently, whereas relatively more teachers in the School Feeding Only arm (34%) reported 
that students should be capable of this skill by Grade 4. In contrast, a higher percentage of 
teachers in the Literacy & School Feeding arm (42%) reported that students should be able 
to read text in the target language fluently by Grade 3. 
The percentage of teachers across the study arms who reported that students should be 
able to write a coherent and comprehensible short story correctly by Grade 4 or higher was 
44 46% across study arms. 
Across the study arms, w . In 
addition, we asked teachers in the School Feeding Only and Literacy & School Feeding 
The percentage of 
teachers who reported that more than half of students are generally attentive in class was 
62 66% across the study arms (Table 32). Teachers most often determined that students 
are attentive in class when students are seated facing and making eye contact with the 
teacher. In School Feeding Only and Literacy & School Feeding schools, less than 40% of the 
teachers said that the school meal makes their students more attentive, approximately 40% 
said the meal does not change student attentiveness, and approximately 20% said it makes 
their students less attentive. 
Table 32. Teacher Perceptions of Student Attentiveness, by Study Arm 
 Comparison School 
Feeding 
Only 
Literacy & 
School Feeding 
n=331 n=163 n=312 
% (n) % (n) % (n) 
Students are attentive in class 91.8 (282) 97.0 (143) 96.1 (278) 
Approximate number of 
students who are generally 
attentive in this class 
Less than half 7.7 (23) 4.3 (7) 9.1 (24) 
Half 20.4 (64) 5.5 (8) 12.0 (37) 
More than half 63.4 (170) 66.9 (93) 62.6 (173) 
All 8.5 (22) 23.3 (34) 16.3 (43) 
Ways teacher gauges if 
students are attentive in 
class 
Students are seated facing and 
making eye contact with the 
teacher. 
55.8 (162) 59.8 (86) 47.9 (141) 
Students actively listen and act 
upon teacher directions. 
38.1 (137) 41.4 (59) 37.4 (122) 
3-30 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
 Comparison School 
Feeding 
Only 
Literacy & 
School Feeding 
Students are working 
independently, in pairs, or in small 
groups on the current learning 
activity. 
14.8 (46) 14.8 (24) 13.1 (39) 
Students respond to questions 
posed by the teacher. 
43.1 (139) 42.6 (67) 43.1 (132) 
Students ask the teacher or 
another student for help related to 
the current learning activity. 
14.6 (49) 10.1 (15) 18.7 (53) 
Students ask the teacher 
questions related to the current 
learning activity. 
16.8 (51) 13.6 (21) 13.0 (41) 
Students are reading. 2.6 (11) 8.3 (10) 6.3 (18) 
Students are writing. 4.7 (20) 4.1 (7) 4.2 (15) 
Students are listening to a story. 2.1 (8) 0.6 (1) 3.5 (11) 
Other 9.0 (27) 3.0 (5) 6.1 (19) 
Student attentiveness in 
relation to the school meal 
Students are more attentive after 
the school meal. 
- 37.2 (55) 36.6 (108) 
Students are less attentive after 
the school meal. 
- 21.3 (28) 22.7 (65) 
Students show the same 
attentiveness before and after 
lunch 
- 41.5 (59) 40.7 (104) 
School Conditions 
and relationships both between the school leadership and the community and among 
students. The most common issues identified with school conditions were the structure of 
the school buildings and lack of clean drinking water (Table 33). Issues with school 
structures affected approximately 60% of schools, across study arms. In terms of the lack of 
clean drinking water, School Feeding Only schools were most affected (41%), followed by 
Comparison schools (33%) and Literacy & School Feeding schools (26%). Other issues, such 
as issues related to latrines and relationships with parents, affected less than one third of 
the schools, with Comparison schools being the most affected, followed by School Feeding 
Only schools and Literacy & School Feeding schools. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-31 
Table 33. Teacher-reported School Conditions, by Study Arm 
Comparison School Feeding 
Only 
Literacy & 
School 
Feeding 
n=329 n=163 n=312 
 % (n) % (n) % (n) 
Issues with structure of school buildings 60.0 (196) 62.1 (96) 63.3 (177) 
Issues with the surrounding area 21.5 (73) 15.4 (25) 9.6 (32) 
Lack of clean drinking water 32.9 (122) 41.4 (67) 25.7 (93) 
Issues with latrines for boys 31.3 (105) 15.4 (25) 9.4 (30) 
Issues with latrines for girls 26.8 (90) 14.2 (22) 10.0 (33) 
Issues with latrines for teachers 29.4 (96) 19.5 (31) 12.5 (36) 
Poor relationships with parents/community 16.6 (56) 10.7 (18) 11.2 (34) 
Bullying at school 3.6 (10) 3.0 (4) 3.8 (10) 
Other 19.9 (76) 24.3 (36) 29.7 (94) 
Nutrition, Health, and Hygiene Teaching 
The percentage of teachers trained to teach about nutrition varied widely across the three 
study arms (Table 34). Seventy-two percent of teachers in School Feeding Only schools, 52% 
of teachers in Literacy & School Feeding schools, and 12% of teachers in Comparison schools 
said they were trained to teach nutrition. The percentage of teachers who had lesson plans 
or other materials for teaching nutrition lessons was 27 36% across the study arms. Of the 
teachers who said they had nutrition lesson plans, across the study arms, 13 29% were 
unable to show any such lesson plans. Among the teachers who said they had lesson plans, 
70% in the Literacy & School Feeding and 80% in the School Feeding Only schools had other 
materials, such as picture cards or posters for teaching nutrition. A higher percentage of 
teachers in School Feeding Only schools taught nutrition at least once per week (70%) 
compared to Literacy & School Feeding schools (58%) and Comparison schools (49%). A 
knowledge of nutrition as average compared to Literacy & School Feeding and Comparison 
schools (approximately 51%). 
Overall, fewer teachers reported that they were trained to teach about hygiene and 
sanitation than about nutrition. A higher percentage of teachers in School Feeding Only 
schools (39%) said they were trained to teach about hygiene and sanitation than in Literacy 
& School Feeding (30%) or Comparison (23%) schools. The percentage of teachers who had 
lesson plans for teaching about hygiene and sanitation was low: 24 29% across the study 
arms. Of the teachers who said they had hygiene and sanitation lesson plans, across the 
study arms, 13 33% did not have lesson plans that they could show to the interviewer. 
Approximately three quarters of the teachers in the School Feeding Only and Literacy & 
School Feeding study arms had other materials, such as posters or picture cards, for 
teaching about hygiene and sanitation. In contrast, just 41% of teachers in Comparison 
3-32 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
schools had these materials. Sixty-seven percent of teachers in the School Feeding Only 
group teach about hygiene and sanitation at least once per week compared to 
approximately 55% of teachers in the Literacy & School Feeding and Comparison study 
arms. A higher percentage of teachers in School Feeding Only schools (65%) rated their 
ene and sanitation as average compared to those in the Literacy 
& School Feeding and Comparison schools (roughly 56%). 
Table 34. Teacher-reported Nutrition and Hygiene Teaching, by Study Arm 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=329 n=163 n=312 
 % (n) % (n) % (n) 
Trained on teaching about nutrition 12.4 (35) 72.2 (105) 52.1 (153) 
Has lesson plans or other materials for teaching nutrition 
lessons 
26.6 (66) 32.1 (49) 36.4 (109) 
Able to show 
All nutrition lesson plans 34.8 (23) 28.6 (13) 24.0 (24) 
Some nutrition lesson plans 45.9 (27) 58.9 (31) 46.9 (51) 
No nutrition lesson plans 19.3 (16) 12.5 (6) 29.1 (35) 
Has other materials for teaching nutrition, such as picture 
cards or posters 
35.8 (25) 80.4 (40) 69.9 (75) 
Frequency of nutrition lessons 
Daily 4.1 (8) 12.8 (19) 9.8 (21) 
A few times per week 32.3 (71) 27.4 (43) 25.1 (67) 
Once per week 12.3 (34) 29.3 (40) 22.9 (64) 
A few times per month 17.9 (46) 17.7 (21) 16.2 (44) 
Once per month 5.3 (12) 2.4 (4) 4.9 (14) 
Less than once per month 7.4 (22) 3.0 (5) 1.8 (8) 
Weak 38.3 (120) 18.4 (25) 41.3 (112) 
Average 52.2 (137) 72.8 (100) 51.2 (146) 
Strong 9.6 (23) 8.9 (13) 7.5 (22) 
Trained on teaching about hygiene and sanitation 23.1 (56) 38.5 (57) 29.9 (96) 
Has lesson plans or other materials for teaching hygiene and 
sanitation lessons 
29.0 (72) 23.7 (35) 27.7 (88) 
Able to show 
All hygiene and sanitation lesson plans 30.2 (24) 32.5 (10) 22.9 (22) 
Some hygiene and sanitation lesson plans 37.0 (32) 55.0 (20) 56.2 (49) 
No hygiene and sanitation less plans 32.8 (19) 12.5 (5) 20.9 (19) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-33 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Has other materials for teaching hygiene and sanitation, such 
as picture cards or posters 
40.8 (33) 72.5 (24) 77.6 (66) 
Frequency of hygiene and sanitation lessons 
Daily 20.9 (49) 27.2 (36) 18.1 (49) 
A few times per week 19.2 (53) 19.5 (31) 18.9 (56) 
Once per week 15.4 (50) 20.7 (30) 17.4 (56) 
A few times per month 21.1 (66) 19.5 (30) 18.4 (50) 
Once per month 2.6 (9) 3.0 (5) 4.6 (14) 
Less than once per month 1.4 (5) 1.2 (2) 1.9 (8) 
Weak 22.0 (79) 15.0 (24) 28.6 (79) 
Average 56.6 (169) 65.0 (89) 56.2 (164) 
Strong 21.4 (43) 20.0 (26) 15.2 (37) 
3.2.3 Head Teachers 
School Profile 
We asked our sample of Grade 2 and 4 teachers for information about their individual 
characteristics, and we also asked HTs to provide data on all teachers in their schools and 
school amenities. The average number of teachers per school varied across the study arms 
(Table 35). 
Table 35. Number of Teachers and Classes per Grade, by Study Arm 
Comparison School Feeding 
Only 
Literacy & 
School Feeding 
n=175 n=85 n=169 
Mean ± 95% CI Mean ± 95% CI Mean ± 95% CI 
Total teachers 12.7 [±1.1] 6.7 [±1.0] 10.1 [±.7] 
 Male 6.1 [±.4] 4.4 [±.6] 5.8 [±.3] 
 Female 6.7 [±.8] 2.4 [±.5] 5.1 [±.6] 
Number of teachers currently teaching Grade 2 2.3 [±.2] 1.2 [±.1] 1.9 [±.1] 
Number of teachers currently teaching Grade 4 1.8 [±.1] 1.1 [±.0] 1.5 [±.2] 
Number of Grade 2 classes (streams) 2.3 [±.2] 1.4 [±.2] 1.9 [±.1] 
Number of Grade 4 classes (streams) 1.7 [±.1] 1.3 [±.2] 1.4 [±.1] 
Note: CI = confidence interval. 
School Feeding Only schools had the fewest teachers (approximately 7), whereas 
Comparison schools had the most (approximately 13). The numbers of male and female 
3-34 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
teachers were evenly split in Literacy & School Feeding and Comparison schools, but there 
were fewer female than male teachers in School Feeding Only schools. Schools had slightly 
more Grade 2 than Grade 4 teachers, but the number of teachers for each grade was one
two across arms. The number of classes per grade (streams) mirrored the number of 
teachers. 
We asked HTs to report about health-related amenities and activities at their schools, 
including school gardens, deworming, latrines, and tippy taps. The percentage of schools 
with gardens varied across the study arms (Table 36). More Comparison schools had 
gardens (50%) than Literacy & School Feeding (41%) or School Feeding Only (29%) schools. 
Comparison and Literacy & School Feeding schools tended to have gardens during the rainy 
season only (91% and 65%, respectively), whereas School Feeding Only schools more often 
had year-round gardens (76%). The majority of HTs in schools with gardens said they use the 
produce for school meals only when garden products are available. 
Most schools coordinate with local health facilities to provide deworming medication to 
students. Higher percentages of School Feeding Only (89%) and Literacy & School Feeding 
(87%) schools ensured that students were dewormed than Comparison schools (75%). 
Similarly, more School Feeding Only (88%) and Literacy & School Feeding (87%) schools than 
Comparison schools (64%) facilitated the deworming of students during this school year. 
Higher percentages of School Feeding Only (82%) and Literacy & School Feeding (87%) 
schools had improved latrines than Comparison schools (39%). Among schools with 
improved latrines, the average numbers of latrines were 3.1 in School Feeding Only, 4.3 in 
Literacy & Scho
observations, the percentage of schools with clean latrines was low overall but higher in the 
intervention study arms (44% in School Feeding Only and 42% in Literacy Plus School 
Feeding) than in the Comparison study arm (27%). Of the schools with improved latrines, 
the numbers of schools with tippy taps or hand-washing stations near the latrines were low 
(41% and 45% in intervention schools and 32% in Comparison schools), and few of these 
schools had ash or soap at all the tippy taps or hand-washing stations (9 28% across study 
arms). 
Table 36. Report of Health-related Amenities and Activities at Schools, 
by Study Arm 
Comparison School 
Feeding Only 
Literacy & 
School Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
School has a garden 50.0 (83) 29.4 (25) 41.4 (69) 
Period 
garden 
operates 
In rainy season only 91.4 (75) 24.0 (6) 65.4 (44) 
Year round 8.6 (8) 76.0 (19) 34.6 (25) 
Use of 
garden to 
Every day 0.0 (0) 0.0 (0) 1.4 (1) 
3 4 times per week 2.3 (2) 4.0 (1) 1.4 (1) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-35 
supply food 
for school 
mealsa
1 2 times per week 0.0 (0) 8.0 (2) 5.7 (4) 
Never 22.1 (18) 8.0 (2) 13.1 (9) 
During the rainy season 20.9 (16) 8.0 (2) 16.2 (11) 
Only when school garden products are available 47.4 (37) 72.0 (18) 58.0 (39) 
Students are given deworming medication 74.7 (130) 89.4 (76) 87.0 (144) 
Students received deworming medication at least once during 
this school year 
64.3 (113) 88.2 (75) 81.2 (137) 
School has improved latrines 38.9 (71) 82.4 (70) 86.5 (146) 
Cleanliness 
of latrines 
All latrines are dirty with feces on the slabs 17.1 (12) 11.4 (8) 13 (19) 
Some latrines are dirty 56.1 (40) 44.3 (31) 45.4 (66) 
All latrines are clean with no feces on the slabs 26.9 (19) 44.3 (31) 41.5 (61) 
School has tippy taps or hand-washing stations 
near improved latrinesb
32.4 (23) 41.3 (29) 45.2 (65) 
All tippy taps or hand-washing stations have ash 
or soapc
8.5 (2) 27.6 (8) 9.0 (6) 
Note: HT = head teacher. 
aUse of garden to supply food for school meals includes schools with a school garden. Some schools without school feeding but with 
gardens provide students with food from the garden when it is available. 
bOf the schools with improved latrines. 
cOf the schools with improved latrines and tippy taps. 
Head Teacher Training 
In the Literacy & School Feeding arm, approximately 50% of the HTs attended WV or PAI 
training (Table 37). Among the HTs who reported attending WV or PAI training, 47% 
reported attending the training during this academic year. In the School Feeding Only arm, 
none of the HTs reported attending WV or PAI training this academic year. 
HTs reported receiving training on the fundamentals of reading instruction (37 53%), school 
management (69 82%), and the importance of access to and use of school supplies and 
materials (25 38%) across the arms. 
Table 37. HT Training, by Study Arm 
Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=175 n=85 n=169 
 % (n) % (n) % (n) 
% of HTs who attended WV or PAI training 1.1 (2) 0.0 (0) 50.9 (83) 
% of HTs who attended training this academic year 0.0 (0) 0.0 (0) 47.9 (78) 
% of HTs who received training on the fundamentals of 
reading instruction 
37.3 (63) 34.9 (29) 53.4 (89) 
% of HTs who report receiving training on school 
management 
68.1 (119) 82.1 (69) 69.3 (117) 
% of HTs who report receiving training about the 
importance of access to and use of school supplies 
and materials 
26.9 (47) 24.7 (21) 38.7 (65) 
3-36 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Note: HT = head teacher; PAI = Planet Aid Incorporated; WV = World Vision. 
Head Teach Knowledge and Skills 
The percentage of HTs who reported they could name the core reading skills was 76 93% 
across the study arms (Table 38). However, when asked to name the core reading skills, few 
HTs across the study arms could do so. HTs in the Literacy & School Feeding arm named 
more of the core reading skills than HTs in the School Feeding Only and Comparison arms. 
Table 38
 Comparison School 
Feeding Only 
 Literacy & 
School Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
HTs able to name the core reading skills 93.7 (57) 76.9 (20) 90.7 (75) 
Core reading skills 
named by HTs 
Letter knowledge 5.9 (11) 4.7 (4) 21.4 (35) 
Phonemic awareness 10.2 (18) 8.2 (7) 30.9 (51) 
Fluency 5.9 (11) 9.4 (8) 21.8 (36) 
Vocabulary 15.1 (27) 17.6 (15) 30.9 (51) 
Comprehension 17.1 (30) 15.3 (13) 22.2 (37) 
Other 11 (19) 9.4 (8) 7.1 (12) 
Ways HTs identify 
academic progress 
Classroom observation 58.3 (102) 48.2 (41) 62.5 (105) 
Monitor student results on tests 
given by teachers 
55.9 (98) 50.6 (43) 48.1 (82) 
Evaluate students orally myself 32.3 (55) 29.4 (25) 41.5 (69) 
homework 
32.7 (56) 32.9 (28) 24.8 (42) 
Teachers provide progress reports 35.8 (64) 36.5 (31) 26.3 (45) 
End-of-term evaluations 37.9 (66) 28.2 (24) 32.6 (55) 
Feedback from parents 7.0 (12) 0.0 (0) 1.2 (2) 
Feedback from school counselors 5.3 (9) 3.5 (3) 0.0 (0) 
Feedback from school committees 1.1 (2) 1.2 (1) 0.6 (1) 
Other 14.8 (27) 16.5 (14) 12.4 (21) 
Note: HT = head teacher. 
Across the study arms, the most common approaches that HTs reported using to identify 
academic progress were conducting classroom observations and monitoring student results 
on tests given by teachers. 
Head Reporting on the Training of Teachers 
The percentage of HTs who reported that teachers in their schools attended a literacy 
training this academic year was 7 14% across the study arms (Table 34). Of the HTs who 
reported that teachers attended a literacy training this academic year, 100% of HTs in the 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-37 
School Feeding Only schools and 79% of the HTs in the Literacy & School Feeding schools 
reported that teachers in their school attended WV or PAI training. Of the 13% of HTs in the 
Comparison arm who reported that teachers in the school attended the literacy training this 
academic year, approximately 9% also reported that teachers in the school had attended 
WV or PAI training. Overall, the mean number of Grade 2 teachers who attended literacy 
training this academic year in the Comparison arm was greater than the mean numbers of 
Grade 2 teachers in the School Feeding Only and Literacy & School Feeding arm who 
attended training. The mean number of Grade 4 teachers who attended literacy training this 
academic year in the Comparison arm was greater than the mean numbers of Grade 4 
teachers in the School Feeding Only and the Literacy & School Feeding arms who did so. 
Table 39. 
Comparison School 
Feeding Only 
Literacy & School 
Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
HTs who report that teachers in the school 
attended literacy training this academic year 
13.3 (22) 7.2 (6) 14.1 (23) 
HTs who report the teachers in the school 
attended WV or PAI training (of the HTs who 
reported teachers attended literacy training) 
9.1 (2) 100 (5) 79.2 (72) 
Mean [95% CI] Mean [95% CI] Mean [95% CI] 
Mean number of female teachers in the school 
who attended literacy training this academic year 
1.3 [±1.6] 0.4 [±.5] 3.6 [±3.6] 
Mean number of male teachers in the school who 
attended literacy training this academic year 
1.9 [±.8] 1.0 [±.6] 2.6 [±.6] 
Mean number of Grade 2 teachers who attended 
literacy training this academic year 
2.5 [±1.6] 0.7 [±.4] 1.4 [±.7] 
Mean number of female Grade 2 teachers who 
attended literacy training this academic year 
2.0 [±1.3] 0.0 [±0] 1.5 [±1.5] 
Mean number of male Grade 2 teachers who 
attended literacy training this academic year 
0.7 [±.5] 0.7 [±.4] 0.8 [±.2] 
Mean number of Grade 4 teachers who attended 
literacy training this academic year 
1.6 [±.9] 0.7 [±.4] 0.9 [±.4] 
Mean number of female Grade 4 teachers who 
attended literacy training this academic year 
1.1 [±.8] 0.2 [±.3] 0.1 [±.2] 
Mean number of male Grade 4 teachers who 
attended literacy training this academic year 
0.5 [±.3] 0.3 [±.4] 0.7 [±.1] 
Note: CI = confidence interval; HT = head teacher; PAI = Planet Aid Incorporated; WV = World Vision. 
 Reporting on the number of Teacher Training Sessions 
When asked about the expected number of WV or PAI teacher training sessions, the mean 
numbers reported by HTs were higher in the School Feeding Only (2.9) and Literacy & School 
3-38 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Feeding (6.5) groups than in the Comparison (0) group (Table 39). When asked about the 
number of WV teacher training sessions held to date, the mean number reported by HTs in 
the Literacy & School Feeding study was 2.6. When asked about the number of PAI teacher 
training sessions held to date, the mean number reported by HTs in the School Feeding Only 
arm was 3, whereas that in the Literacy & School Feeding arm was 2.4. 
Table 39
Comparison School 
Feeding Only 
Literacy & 
School Feeding 
n=175 n=85 n=169 
Mean [95% 
CI band] 
Mean [95% CI 
band] 
 Mean [95% CI 
band] 
Expected number of WV or PAI teacher training sessions 0.0 [±0] 2.9 [±.6] 6.5 [±.4] 
Number of WV teacher training sessions held to date 0.1 [±.1] 0.1 [±.1] 2.6 [±.2] 
Number of PAI teacher training sessions held to date 0.0 [±.0] 3.0 [±.2] 2.4 [±.3] 
Note: CI = confidence interval; HT = head teacher; PAI = Planet Aid Incorporated; WV = World Vision. 
 Reporting on the Number of Reading Coaches 
Higher percentages of HTs in the School Feeding Only (47%) and Literacy & School Feeding 
(59%) arms reported that reading coaches train teachers at the school than in the 
Comparison arm (24%) (Table 40). Among the HTs who reported that reading coaches train 
teachers at the school, the mean number of trainings provided by reading coaches this 
academic year ranged from approximately one to three across the study arms. 
Table 40
Comparison School Feeding Only Literacy & School Feeding
n=175 n=85 n=169 
% (n) % (n) % (n) 
% of HTs who report that reading coaches 
train teachers at this school 
24.6 (43) 47 (39) 59.7 (97) 
% of HTs who report that reading coaches 
supervise teachers at this school 
48.5 (20) 59.5 (22) 82.0 (78) 
Mean [95% CI] Mean [95% CI] Mean [95% CI]
Mean number of trainings provided by 
reading coaches this academic year 
2.0 [±.4] 1.9 [±.3] 3.2 [±.2] 
Frequency at which 
reading coaches 
supervise teachers 
at this school 
Never 45.1 (18) 42.9 (15) 11.2 (10) 
Once a year 15.3 (6) 11.4 (4) 3.4 (3) 
Once per term 21.5 (8) 8.6 (3) 27.4 (24) 
Once a month 12.8 (5) 11.4 (4) 30.5 (27) 
More than once a 
month 
5.2 (2) 25.7 (9) 27.5 (25) 
Note: CI = confidence interval; HT = head teacher. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-39 
Higher percentages of HTs in the Literacy & School Feeding (82%) and School Feeding Only 
(59%) arms reported that reading coaches supervise teachers at the school than in the 
Comparison arm (48%). Among the HTs who reported that reading coaches supervise 
teachers at the school, higher percentages of HTs in the Comparison (45%) and School 
Feeding 
contrast, a higher percentage of HTs in the Literacy & School Feeding arm (30%) reported 
that reading coaches supervise teachers once a month. 
Reporting on Access to School Supplies and Learning Materials 
Higher percentages of HTs in the School Feeding (29%) and Literacy & School Feeding (52%) 
study arms reported that the school has a library/book bank than in the Comparison arm 
(17%) (Table 41). Among the HTs who reported that the school has a library/book bank, 
100% also reported that the print material is available in Portuguese. In contrast, the 
percentage of HTs across the study arms who reported that the print material in the 
library/book bank is available in the mother tongue was 8 12%. Among the HTs in the 
Comparison arm, the most common types of print material reported as being available in 
the library/book bank were primers/student manuals and leveled books. Among the HTs in 
the School Feeding Only and Literacy & School Feeding arms, the most common types of 
print material reported as being available in the library/book bank were storybooks and 
primers/manuals. 
Table 41
Arm 
Comparison School 
Feeding Only 
Literacy & 
School Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
HTs who report that the school has a library/book bank 17.1 (30) 29.4 (25) 52.2 (86) 
HTs who report 
specific types of 
print material 
available 
Flipbooks 19.1 (6) 36.0 (9) 18.6 (16) 
Storybooks 61.8 (19) 84 (21) 84.1 (72) 
Primers/student manuals 70.2 (21) 84 (21) 57.9 (50) 
Fables 39.6 (12) 40 (10) 50.3 (43) 
Leveled books 63.6 (19) 52 (13) 53.5 (46) 
Supplementary readers 50.2 (15) 40 (10) 40.8 (35) 
Guidelines for parents 6.2 (2) 0.0 (0) 1.2 (1) 
Reading posters 29.8 (9) 28 (7) 20.8 (18) 
Other 2.3 (4) 8.2 (7) 4.3 (7) 
HTs who report that the print material in the library/book bank 
is available in Portuguese 
100.0 (30) 100.0 (24) 100.0 (86) 
HTs who report that the print material in the library/book bank 
is available in the mother tongue 
9.8 (3) 8.3 (2) 12.7 (11) 
Note: HT = head teacher. 
3-40 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Reporting on Extracurricular Activities 
A higher percentage of HTs in Literacy & School Feeding schools (33%) reported that the 
school has a literacy afterschool program than HTs in the Comparison (12%) and School 
Feeding Only (15%) schools (Table 42). Higher percentages of HTs in the School Feeding Only 
(72%) and Literacy & School Feeding (73%) arm reported that the school has other types of 
afterschool programs than HTs in the Comparison arm (14%). 
Among the HTs in the School Feeding Only arm who reported that the school has other 
types of afterschool programs, the most common types were the Math Learning Club and 
the School Gardening Club. Among the HTs in the Literacy & School Feeding arm who 
reported that the school has other types of afterschool programs, the most common types 
were the Help with HW club, Math Learning Club, and Science Club. Among the HTs in the 
Comparison arm who reported that the school has other types of afterschool programs, the 
most common types were the Help with HW Club and the Math Learning Club. 
Higher percentages of HTs in the School Feeding Only (55%) and Literacy & School Feeding 
(55%) arms reported that 
progress in reading than in the Comparison arm (38%). Higher percentages of HTs in the 
Literacy & School Feeding (48%) and School Feeding Only (35%) arms report that special 
events are held at the school to recognize effective reading instruction than in the 
Comparison arm (29%). Higher percentages of HTs in the Comparison (47%) and Literacy & 
School Feeding (55%) arms reported that special events are held at the school to recognize 
teacher attendance than in the School Feeding Only arm (27%). 
Table 42
Comparison School Feeding 
Only 
Literacy & School 
Feeding 
n=175 n=85 n=169 
% (n % (n) % (n) 
HTs who report that the school has a literacy afterschool 
program 
12.9 (22) 15.3 (13) 33.5 (55) 
HTs who report that their school has other types of 
afterschool programs 
14.3 (25) 72.9 (62) 73.2 (124) 
Other types 
of afterschool 
programs 
reported by 
HTs 
Help with HW Club 5.7 (10) 8.2 (7) 23.3 (39) 
Math Learning Club 4.0 (7) 36.5 (31) 30.3 (52) 
Science Learning Club 2.4 (4) 10.6 (9) 23.3 (39) 
School Garden Learning Club 1.2 (2) 23.5 (20) 12.6 (22) 
Other 7.6 (14) 55.3 (47) 42.7 (73) 
HTs who report that special events are held at the school to 38.7 (68) 55.3 (47) 55.7 (93) 
HTs who report that special events are held at the school to 
recognize effective reading instruction 
29.0 (50) 35.3 (30) 48.0 (79) 
HTs who report that special events are held at the school to 
recognize teacher attendance 
47.1 (81) 27.1 (23) 55.9 (92) 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-41 
Note: HT = head teacher. 
The percentage of HTs who reported satisfaction with parental involvement in their 
and in the school was less than half across all three study arms (Table 43). 
When asked about the ways that parents show evidence of support for the school, the most 
common answers given by HTs across the study arms were volunteering time and resources. 
The percentage of HTs who reported that parents volunteer in the school feeding program 
was higher in the School Feeding Only arm (45%) than in the Literacy & School Feeding 
(28%) and Comparison arms (4%). 
When asked about the frequency of parental involvement according to teachers, HTs in the 
Comparison (47%) and Literacy & School Feeding (43%) arms reported that parents were 
involved once a quarter. In contrast, HTs in the School Feeding Only arm (64%) reported that 
according to teachers, the frequency of parental involvement was more than once a month. 
Table 43
Comparison School 
Feeding Only 
Literacy & 
School Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
% of HTs who report satisfaction with parental involvement in 31.3 (56) 19.0 (16) 36.7 (62)
school
30.2 (53) 19.0 (16) 35.0 (59) 
Ways that parents 
show evidence of 
support for the school 
None 19.1 (33) 7.1 (6) 9.9 (17) 
Visiting the school 18.9 (34) 16.5 (14) 19.2 (32)
Actively seeking information about 
school matters 
26.2 (45) 23.5 (20) 29.2 (49) 
Following up on academic progress 24 (42) 27.1 (23) 26.6 (45) 
Volunteering time 37.7 (66) 47.1 (40) 40.5 (68) 
Volunteering in school feeding 4.5 (8) 45.9 (39) 27.8 (48) 
Volunteering resources 38.7 (66) 40.0 (34) 33.4 (56)
Participating in school 
meetings/committees 
32.0 (57) 32.9 (28) 36.2 (61)
Other 8.8 (15) 7.1 (6) 6.7 (11) 
Frequency of parental 
involvement reported 
by teachers 
Once a year 18.0 (24) 0.0 (0) 9.2 (13) 
Once a quarter 47.7 (63) 26.0 (20) 43.3 (62)
Once a month 15.2 (20) 9.1 (7) 12.7 (18) 
More than once a month 19.1 (25) 64.9 (50) 34.9 (51) 
Note: HT = head teacher. 
Teacher Absenteeism 
Across study arms, most schools (88 96%) had teacher attendance registers (Table 44). HTs 
reported that on average, one two teachers were absent yesterday or the last day school 
 absence were 
3-42 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
allocating the class to another teacher or combining students together into one class. HTs 
themselves were frequently absent (22 30%) or late to school (10 19%). 
Table 44
Comparison School 
Feeding Only 
Literacy & 
School 
Feeding 
n=175 n=85 n=169 
Mean ± 95% 
CI band 
Mean ± 95% 
CI band 
Mean ± 95% 
CI band 
Number of teachers absent yesterday or last day school was in 
session 
1.4 [±.2] 1.0 [±.3] 1.7 [±.2] 
% (n) % (n) % (n) 
Teacher attendance register available 93.5 (164) 88.2 (75) 96.0 (162) 
How 
class is 
handled 
if teacher 
is absent 
Let class proceed without teacher 2.7 (5) 1.2 (1) 0.6 (1) 
Allocate class to another teacher 55.0 (95) 62.4 (53) 71.8 (119) 
Join all students in one class 26.9 (46) 15.3 (13) 15.0 (25) 
Dismiss students for the day 0.5 (1) 1.2 (1) 1.2 (2) 
Distribute students among other classes 6.6 (12) 11.8 (10) 7.5 (13) 
Other 8.3 (15) 8.2 (7) 4.0 (7) 
HT absent from school any day last week 21.8 (38) 29.4 (25) 30.1 (51) 
HT late to school any day last week 9.6 (16) 10.6 (9) 18.8 (31) 
Note: CI = confidence interval; HT = head teacher. 
Nutrition and Hygiene Training and Teaching 
We asked HTs to report on teacher training on nutrition and hygiene/sanitation to collect 
information about all teachers in each school and triangulate the results with those 
hygiene/sanitation mirrored those of Grade 2 and 4 teachers (Table 45). In the School 
Feeding Only and Literacy & School Feeding schools, more teachers were trained on 
nutrition than hygiene/sanitation, whereas the reverse was true in Comparison schools. 
y of teaching 
nutrition and hygiene/sanitation did not align with those of teachers. HTs described less 
availability of teaching materials for nutrition and hygiene/sanitation than teachers. They 
also reported less frequent teaching of nutrition, although their estimates of the frequency 
of hygiene/sanitation teaching were relatively close to those of teachers in the School 
Feeding Only and Literacy & School Feeding arms. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-43 
Table 45
Hygiene Lessons, by Study Arm 
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=175 n=85 n=169 
Mean ± 95% 
CI band 
Mean ± 
95% CI 
band 
Mean ± 
95% CI 
band 
Teachers trained on teaching nutrition 0.5 [±.4] 3.5 [±.4] 4.5 [±.6] 
Teachers trained on teaching hygiene and sanitation 1.1 [±.4] 1.5 [±.4] 2.3 [±.5] 
% (n) % (n) % (n) 
Teachers have materials for teaching nutrition 12.4 (22) 59.0 (49) 66.0 (109) 
Frequency 
of teaching 
nutrition 
lessons to 
each class 
Once per week 15.2 (22) 64.6 (51) 60.9 (91) 
Once every two weeks 6.1 (9) 7.6 (6) 8.5 (13) 
Once per month 7.7 (11) 12.7 (10) 14.3 (21) 
Once per term 38.6 (56) 8.9 (7) 6.1 (9) 
Never 32.4 (45) 6.3 (5) 10.2 (15) 
Teachers have materials for teaching hygiene and sanitation 
22.9 (39) 25.0 (21) 55.7 (90) 
Frequency 
of teaching 
hygiene 
and 
sanitation 
lessons to 
each class 
Once per week 35.0 (53) 70.0 (56) 66.9 (104) 
Once every two weeks 8.5 (13) 6.2 (5) 7.8 (12) 
Once per month 6.9 (10) 5.0 (4) 7.3 (11) 
Once per term 26.3 (39) 6.2 (5) 7.2 (11) 
Never 23.4 (34) 12.5 (10) 10.8 (17) 
Note: CI = confidence interval; HT = head teacher. 
School Feeding Program 
Nearly all School Feeding Only schools and all Literacy & School Feeding schools had a 
school feeding program, whereas only a few Comparison schools did (Table 46). In School 
Feeding Only schools, the program was supported exclusively by PAI/ADPP; in Literacy & 
School Feeding schools, the program was supported nearly equally by PAI/ADPP and WV; in 
Comparison schools, the program was supported by other organizations. Approximately 
92% of School Feeding Only and Literacy & School Feeding schools served school meals 
daily, whereas all Comparison schools with a feeding program served meals daily. The main 
issue with school feeding programs was insufficient number of pots, utensils, or dishes, as 
mentioned by 21
school feeding programs, lack of water, lack of cooks/unavailability of cooks, lack of 
remuneration for cooks, lack of soap, and lack of firewood were mentioned most frequently. 
3-44 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 46
Comparison School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=175 n=85 n=169 
% (n) % (n) % (n) 
School has a school feeding program 5.4 (10) 98.8 (84) 100.0 (168) 
Organization 
supporting school 
feedinga
PAI/ADPP 0.0 (0) 100.0 (84) 46.4 (81) 
WV 0.0 (0) 0.0 (0) 50.7 (82) 
Community support 0.0 (0) 0.0 (0) 0.6 (1) 
Other 100.0 (10) 0.0 (0) 4.1 (7) 
Frequency of school 
mealsa
Daily 100.0 (10) 92.8 (77) 92.2 (155) 
A few times per week 0.0 (0) 4.8 (4) 2.9 (5) 
Whenever food is available 0.0 (0) 2.4 (2) 4.9 (8) 
Issues with school 
feeding programa,b
No issues 10.0 (1) 29.4 (25) 17.6 (30) 
Food is not always available. 10.0 (1) 2.4 (2) 4.1 (7) 
Kitchen/warehouse not constructed or in 
disrepair 
30.0 (3) 1.2 (1) 4.2 (7) 
Not enough pots, utensils, or dishes 40.0 (4) 21.2 (18) 31.7 (53) 
Meals not prepared on time 0.0 (0) 7.1 (6) 6.0 (10) 
Other 70.0 (7) 51.8 (44) 68.7 (115) 
Note: ADPP = Ajuda de Desenvolvemento de Povo para Povo (Development Aid from People to People); HT = head 
teacher; PAI = Planet Aid Incorporated; WV = World Vision. 
aOf schools with a school feeding program. 
b
3.2.4 Warehouse Managers and Cooks 
We originally planned to collect information about school feeding infrastructure and training 
from cooks, but during the pilot phase, we found that warehouse managers handle many of 
the logistical issues related to the school feeding program at each school. Therefore, we 
divided the questions between the manager and the head cook. We also determined that 
warehouse managers and cooks were not always present to answer questions. 
School Feeding Infrastructure 
More schools in School Feeding Only schools (92%) compared to Literacy & School Feeding 
schools (79%) had a school feeding log book where they tracked the quantities of food 
received and used (Table 47). In approximately 90% of schools with a log book, the log was 
filled out and up to date. Nearly all schools in both intervention arms had a kitchen and 
storeroom for the school feeding program. Furthermore, in both intervention arms, the 
percentages of schools in which the kitchen and storeroom were in very good condition 
were low (23% in School Feeding Only and 28% in Literacy & School Feeding schools). 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-45 
We found some gaps in the availability of materials for school feeding programs. Half of the 
Literacy & School Feeding schools and more than one third of School Feeding Only schools 
had ever had a gap in their supply of corn-soy blend. Most schools had cooking pots and 
utensils, but 14% of Literacy & School Feeding and 8% of School Feeding Only schools did 
not. More School Feeding Only schools had sufficient numbers of cooking pots and utensils 
than Literacy & School Feeding schools. Approximately 80% of the schools had bowls and 
spoons for students, but only 50 60% had them in sufficient quantities. Clean water was 
available at 68% of Literacy & School Feeding and 60% of School Feeding Only schools. Many 
schools did not have soap for cleaning cooking materials and bowls. School Feeding Only 
schools more often had soap available to clean pots and bowls (61%) than Literacy & School 
Feeding schools (36%). A higher percentage of School Feeding Only than Literacy & School 
Feeding schools had a place for cooks and students to wash their hands and soap for hand 
washing. The availability of soap was the largest gap, especially in Literacy & School Feeding 
schools. 
School meals are most often served between 9 and 10 a.m. in both intervention study arms, 
which puts the meal in the middle of the morning school session. Approximately half the 
cooks reported that school meals are always served on time. 
Table 47. School Feeding Infrastructure, by Intervention Arm 
School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=79 n=168 
% (n) % (n) 
School has a school feeding log book. 92.4 (73) 78.5 (125) 
Log book is filled out and up to date. 89.0 (65) 89.0 (111) 
School has a kitchen and storeroom for school feeding program. 98.7 (78) 98.2 (156) 
Condition of 
the kitchen 
and 
storeroom 
Not adequate (holes in roof or walls, no door, and/or no lock) 16.7 (13) 19.8 (32) 
Moderate (complete building but needs repairs) 60.3 (47) 52 (83) 
Very good (complete building with no repairs needed) 23.1 (18) 28.1 (44) 
School has ever had a gap in the supply of corn-soy blend. 36.7 (29) 50.3 (80) 
Cooking pots and utensils are available. 92.4 (73) 86.1 (137) 
Number of pots and cooking utensils is sufficient. 75.9 (60) 56.9 (91) 
School has bowls and spoons for students. 78.6 (55) 80.1 (114) 
Number of bowls and spoons is sufficient. 61.4 (43) 51.8 (74) 
School has clean water available. 60.0 (42) 67.7 (97) 
Soap is available to clean bowls, spoons, pots, and cooking utensils. 61.4 (43) 35.5 (52) 
School has a place for cooks to wash their hands. 74.3 (52) 58.7 (84) 
Soap is available for cooks to wash their hands. 60.0 (42) 28.5 (42) 
School has a place for children to wash their hands before eating. 72.9 (51) 65.2 (93) 
Soap is available for children to wash their hands before eating. 45.7 (32) 19.0 (28) 
3-46 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
School 
Feeding 
Only 
Literacy & 
School 
Feeding 
Time of day 
when meals 
are served 
clock hour 0.0 (0) 1.9 (3) 
60.8 (48) 56.7 (90) 
10.1 (8) 14.1 (22) 
12.7 (10) 12.7 (20) 
13.9 (11) 10.9 (18) 
After noon 2.5 (2) 3.7 (6) 
School 
meals are 
served on 
time 
Never 5.1 (4) 3.9 (6) 
Sometimes 46.8 (37) 46.9 (74) 
Always 48.1 (38) 49.2 (79) 
Cook Training and Knowledge 
Less than half of the school cooks interviewed had received training (Table 48). Among 
those trained, the timing of the last training varied between the two intervention study 
arms, with 45% of School Feeding Only cooks and 27% of Literacy & School Feeding cooks 
receiving training less than 6 months ago. Cooks had low to moderate knowledge of food 
safety (three four out of eight items correct) and hygiene (two three out of five items 
correct). 
Table 48. Cook Training and Knowledge, by Intervention Arm 
School 
feeding only
Literacy + 
School 
feeding 
n=70 n=144 
% (n) % (n) 
School cook training received 41.4 (29) 47.3 (67) 
Timing of 
last school 
cook 
training 
Less than 6 months ago 44.8 (13) 27.2 (19) 
6- 11 months ago 17.2 (5) 17.9 (12) 
12 or more months ago 37.9 (11) 54.8 (37) 
Mean ± 95% 
CI Band 
Mean ± 
95% CI 
Band 
Knowledge of food safety (8 items) 3.4 [±.3] 3.8 [±.2] 
Knowledge of hygiene (5 items) 2.4 [±.2] 2.5 [±.2] 
Note: CI = confidence interval. 
Use of School Meals 
Nearly 90% of cooks in both intervention study arms said that they usually eat the school 
meal when they are at the school cooking (Table 49). Fifty-three percent of School Feeding 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-47 
Only schools and 40% of Literacy & School Feeding schools did not usually have leftovers 
from school meals. When there were leftovers, cooks most often gave more porridge to 
students or ate it themselves. On average, interviewers found two three cooks present at 
each school, and cooks reported that they prepare three four school meals per week. 
Table 49. Use of School Meals and Cook Working Days, by Intervention Arm 
School 
Feeding 
Only 
Literacy & 
School 
Feeding 
n=70 n=144 
% (n) % (n) 
Cooks usually eat the school meal 87.1 (61) 89.7 (128) 
Use of 
leftover 
food at 
end of 
school 
meala
Usually there are no leftovers 52.9 (37) 39.5 (57) 
Give more to students 27.1 (19) 37.3 (53) 
Give food to siblings/other children at school who are not students 7.1 (5) 4.9 (7) 
Give it to teachers 7.1 (5) 7.8 (11) 
Cooks eat it 15.7 (11) 16.1 (23) 
Cooks take it home 4.3 (3) 4.7 (7) 
Other 7.1 (5) 14.8 (21) 
Mean ± 95% 
CI Band 
Mean ± 95% 
CI Band 
Number of cooks present on day of survey 2.3 [±.3] 2.6 [±.2] 
Number of days per week cooks prepare meals at school 3.6 [±.3] 3.7 [±.2] 
Note: CI = confidence interval. 
a
3.2.5 Parents 
Overall, a total of 30 FGDs with parents were conducted. Fourteen and 16 FGDs were 
completed in rural and urban schools, respectively. Twenty-three of the schools where FGDs 
were held received the Literacy & School Feeding intervention, whereas seven received 
School Feeding Only (Table 50). 
A total of 285 parents participated in the FGDs. We included mothers and fathers in the 
same groups because there were no sensitive questions, and we did not expect different 
old, with half between 29 and 43 years old. Fifty-seven percent of the participants were 
female. On average, the participants had 3.4 children in total, and 47%, 34%, and 19% had a 
child in Grade 2, Grade 4, and both grades, respectively. Fifty-six percent reported speaking 
only a local language at home, 30% reported speaking both a local language and 
Portuguese, and 14% reported speaking Portuguese only. 
3-48 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 50. Number of FGDs and Parents by District and Province Involved in Qualitative 
Data Collection 
Literacy & School Feeding School Feeding Only 
# of 
FGDs
Mean # 
of 
Parents 
% of 
Parents 
with 
Grade 
2 Child 
% of 
Parents 
with 
Grade 
4 Child 
# of 
FGDs
Mean # 
of 
Parents 
% of 
Parents 
with 
Grade 
2 Child 
% of 
Parents 
with 
Grade 
4 Child 
Maputo 
Magude 1 7.0 57% 43% 3 11.3 71% 53% 
Manhiça 4 7.3 79% 59% 2 8.0 75% 44% 
Matutuíne* 3 7.7 - - 
Moamba* 3 6.7 33% 56% 2 11.5 61% 39% 
Nampula 
Muecate 6 10.5 75% 46% - - - - 
Nacarôa 6 11.8 51% 63% - - - - 
Total 23 9.3 63% 55% 7 10.4 68% 47% 
Note: FGD = focus group discussion.
* Demographic information was missing for three FGDs in Matutuíne and two in Moamba. Average and percent values are 
based on non-missing data only. 
in order to be so
Parent, Namina Rio School, Muecate, Nampula 
Parents value education for their children. They consistently stated that they expected their 
children to gain knowledge, including reading and writing; find employment after finishing 
school; support th
hope he continues to study to acquire knowledge and one day becomes a doctor, a teacher 
uture 
employment opportunities but also teaches children good behavior and helps them be 
e in schools 
were mixed, with some parents being very pleased, while others expressed frustration that 
teachers were not more patient and persistent in helping each child succeed. 
Most parents seemed very satisfied with the level of education their child received at 
school, citing the fact that children are gaining skills in reading and writing as proof that the 
school is working to provide an education. However, some parents expressed deep concerns 
about and extreme disappointment in their school because children were progressing 
through grade levels without advancement in their reading and writing skills. For example, 
one parent stated that her son was advanced to Grade 5 without the ability to read or write. 
Three consistent problems regarding the quality of education were identified by parents: 1) 
teacher absenteeism, 2) inadequate education materials (particularly school books), and 3) 
the lack of a water source at the school. Without water, school meals cannot be prepared. 
Parents often stated that their children must wake up extremely early (e.g., 12:00 a.m.) to 
fetch water before school and that some schools require children bring water to attend 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-49 
classes. Other requests made by parents related to adding grade levels in the schools to 
offer complete elementary and secondary education, incorporating English into the 
curriculum, providing security for the school to prevent thievery and vandalism, providing 
assistance with purchasing school uniforms, addressing the need for additional classrooms 
and latrines, and fulfilling the need for additional shelters (i.e., roofs) where children can 
read and study during the rainy season. 
he 
overall literacy education in schools were mixed. While many parents indicated that they 
observed their children successfully reading and writing at home, parents in two of the 
School Feeding Only schools stated that the literacy education was not working because the 
children are advancing grade levels despite lacking the ability to read and write. A few 
themselves were not literate. 
Parents pointed to poor teaching quality and teacher absenteeism as problems leading to 
poor literacy program performance. The main suggestion for improvement was to increase 
have are incomplete, and not everyone has them which makes it difficult to study because 
on a lack of trust in those running the program when supplies are not consistently available. 
Children were promised materials, including books, notebooks, and snacks, and when these 
materials did not arrive, children stopped attending school. However, other schools have 
reported receiving materials, including notebooks and pencils, and parents viewed these 
parents were pleased with the literacy program, they greatly valued the program and 
everal 
parents proposed increasing the number of teachers implementing the school literacy 
program. 
Parents reported that nearly all the reading and writing activities that children engage in at 
home revolve around homework assigned by school teachers. Parents often mentioned that 
teachers should continue to assign homework as a means of ensuring that their children 
continue to engage in reading and writing activities. According to the parents, parental 
involvement in such activities at home mainly consists of encouraging their children to 
complete their homework. Many parents mentioned that children ask them questions about 
homework and that they try to answer such questions as best they can. However, the 
parents indicated that they were not always able to fully assist their children. Lack of fluency 
in Portuguese and illiteracy were consistently identified as the main barriers that prevent 
parents from engaging in reading and writing activities with their children at home. Some 
parents mentioned that older siblings in higher grade levels often help their younger 
3-50 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
and writing activities are conducted solely at home; however, sometimes, a lack of materials 
parents noted that children engage in reading and writing activities daily, either once or 
twice a day. 
Nearly all parents mentioned that school workbooks were the main source of reading 
reading material included non-school books purchased by parents, newspapers, television 
ads, phone text messages, and the bible. Some parents stated that their children have 
pencils and notebooks in which they can write and draw pictures; however, parents more 
frequently indicated a lack of writing materials as a problem. Thus, parents reported that 
many children use sticks and wire to write and draw pictures in the dirt. 
Parents identified the following as factors that motivate their children to engage in reading 
and writing activities at home: assigned homework, access to adequate materials, parental 
guidance or supervision, support from older siblings, and peer support. Barriers included a 
lack of materials, a lack of electricity to provide light and allow children to read and write in 
the evening, a lack of time because of the need to complete household chores, insufficient 
teachers to provide additional support for low-achieving students, and the desire for 
children to play rather than complete homework. Parents also noted that children need to 
indicated 
available to complete homework. 
he Language Used in School and at Home 
Parents reported that local languages (i.e., Changana and Macua) are the most common 
languages used in the home setting, whereas Portuguese is used in schools. Very few 
parents indicated that they had the ability to read and write in Portuguese. Furthermore, 
they reported that the inability to read and write in Portuguese makes helping their children 
with homework challenging. When parents reported that they did engage in reading and 
writing activities at home with their children, they mainly described doing so in the local 
language rather than Portuguese. Some parents suggested that the sole use of Portuguese 
at school when children use Changana at home might hinder their ability to learn at school. 
One parent stated, 
is important and suggested using both Portuguese and the local language in school. 
erceptions of School Meals 
-Parent, Maguaza School, Moamba District, Maputo 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 3-51 
All parents expressed that they were pleased with the presence of the school feeding 
program. Indeed, the parents frequently mentioned that school meals help motivate their 
children to attend school more regularly and stay at school longer and prevent children 
from dropping out o
children because it guarantees that at least they have something to eat, and it encourages 
result of the school meals. Others noted that their children appeared to be stronger and 
happier when a school feeding program was implemented. All parents requested that the 
-being and success 
in school. 
Parents identified the main problem with the school feeding program as the lack of a water 
source for cooking meals at the school. One parent noted that installing a water source near 
the school would also improve school attendance, the consistent provision of school meals, 
and the hygienic preparation of school meals. Many parents suggested providing more 
variety in the types of foods offered through the school feeding program. Suggested foods 
included beans, meat, rice, and pasta. One parent proposed adding peanuts to the school 
porridge to increase the calories consumed by the children. Several parents also 
recommended improving the flavor of the porridge by adding salt or sugar. Other 
suggestions for improving the school feeding program included increasing the number of 
cooks, ensuring that no gaps occur in the supply of school meals, and constructing more 
buildings with roofs where children can eat during the rainy season. 
Changes in Diets at Home due to School Meals 
Although a few parents stated that the amount of food consumed at home by their children 
did not change because of the school meals, most indicated that their school-going children 
the school meals, children no longer complain about being hungry when they arrive home 
tea, prior to leaving for school. However, with school meals, they no longer do so and 
instead eat only at school. Others stated that their children either skip dinner entirely or eat 
very little for dinner because when school meals are served, they are not hungry in the 
evening. 
Some parents said that their children never missed school because of illness, whereas 
others stated that fever, cough, nausea, malaria, and/or diarrhea had caused their children 
to miss some school. Parents explained that at school, children mainly learned about 
personal hygiene, including bathing, washing their face, and washing hands, and about 
washing dishes. Most parents reported that their children applied these messages in their 
behavior at home, although some parents indicated that their children did not bathe 
regularly. No parents discussed improved sanitation, such as increased or proper use of 
3-52 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
latrines; instead, they only mentioned that their children learned to wash their hands after 
going to the bathroom. 
Nine of the schools where FGDs were held had school gardens. Parents noted that school 
gardens were beneficial because they allow students to have a variety of foods; however, 
many barriers prevent gardens from being successful. At most schools, parents are 
responsible for managing the school garden, and at some schools, children are also involved 
in the farming practices. Parents frequently indicated that a lack of seeds and other tools is 
the most serious challenge to having a successful school garden. 
_________________________________
RTI International is a registered trademark and a trade name of Research Triangle Institute. 
4. SUMMARY 
In this section, we return to our theory of change and describe the baseline findings related 
to the elements of the theory. We also triangulate results across types of participants and 
data collection methods, wherever possible. At baseline, we found several differences 
between groups in the key indicators for the primary and secondary evaluation questions 
and in the factors that may contribute to those key indicators. 
Student literacy: Baseline EGRA scores varied across treatment arms. In general, 
performance on the EGRA was better among Grade 4 students than Grade 2 students, as 
expected. In almost all cases, and for both grades, as the difficulty of the subtasks increased, 
Students in McGovern-Dole intervention schools struggle 
more with listening comprehension than students in Comparison schools. 
Students in School Feeding & Literacy schools can read fewer words per minute than in 
School Feeding Only and Comparison schools. Oral reading fluency (ORF rates across all 
groups were lower than the minimum rate necessary for students to read the entire 
passage. Percent zero scores were generally higher for the syllable sound task than for the 
ORF task, indicating that more students could read at least one word of a short passage than 
were able to read a syllable. This pattern was similar across all arms and in both grades. This 
finding suggests that students have not fully acquired the decoding skills necessary to read 
fluently and with comprehension but may have recognized or memorized familiar words 
that they frequently encounter in grade-level text. 
that their 
children read and write well, whether others saw their children promoted to the next grade 
despite struggling with reading and writing. 
Student attentiveness: Most teachers said that they can tell whether their students are 
attentive when they face the teacher and make eye contact or when they respond to 
questions posed by the teacher. Less than 40% of the teachers said that the school meal 
makes their students more attentive, approximately 40% said the meal does not change 
student attentiveness, and roughly 20% said it makes their students less attentive. 
Access to and use of school supplies: S is 
instruction. However, data from parent FGDs indicated that children do not have consistent 
access to school materials. Only slightly more than half of teachers said they had access to a 
teacher guide, and their reported use of the teacher guide was generally aligned with their 
reported access to the teacher guide. Although more teachers in intervention schools than 
in comparison schools reported having access to basic school supplies, most teachers across 
groups reported low access to materials such as markers and chart paper. 
Literacy instructional materials: Students had lower access to supplementary readers than 
to Portuguese textbooks and exercise notebooks. This point was supported by the data from 
teachers, who reported using supplementary readers approximately once a week. More 
than 50% of schools in the Literacy & School Feeding intervention have a library/book bank 
4-2 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
with print materials, such as storybooks and primers/manuals in Portuguese. Less than 50% 
of schools in the School Feeding Only intervention have a library/book bank. Most 
intervention schools have few print materials in the mother tongue available. This finding 
aligns with parent FGD data, which revealed that parents would like to see an increase in 
the quantity of reading materials available to their children. 
Extracurricular activities: Less than 50% of the HTs in the intervention and comparison 
schools reported the presence of an afterschool literacy program. Among the teachers who 
reported the existence of a literacy club at the school, across intervention and comparison 
schools, most indicated that half the students in their class attend an afterschool literacy 
club. Consistent with the teacher data, the HT data showed that more than 50% of the 
schools in the School Feeding Only 
Less than 50% of all sampled schools 1) hold special events at the school to celebrate 
 or 2) hold special events at the school to recognize effective 
reading instruction. More than 50% of schools in the Literacy & School Feeding study arm 
hold special events at the school to recognize teacher attendance. In contrast, this trend is 
not observed on the same magnitude in the School Feeding Only study arm. 
Skills and knowledge of teachers: Most teachers reported attending and finishing studies at 
a TTC. Additionally, most reported that they had attained certification. Slightly more than 
50% of teachers received specific training on how to teach reading and writing to students 
in the primary grades as part of their preservice training. More than half of the teachers 
earned a certificate as their highest level of education. According to HTs, more than 50% of 
teachers in the Literacy & School Feeding intervention receive training from reading 
coaches, whereas less than half of teachers in the School Feeding Only intervention receive 
training from reading coaches. More than 50% of the teachers in both interventions are 
supervised by reading coaches. However, these data are not consistent with teacher data 
indicating that more than 50% of teachers reported that reading coaches had not visited 
since the beginning 
support, most teachers in intervention schools reported that HTs or assistant HTs check 
lesson plans once a week and conduct classroom observations once per trimester. Nearly all 
the teachers in both comparison and intervention schools receive feedback from HTs or 
assistant HTs. When they need help, teachers in both comparison and intervention schools 
most often have casual discussions with other teachers. More than 50% of the teachers in 
comparison and intervention schools could not name all the core reading skills. More than 
50% of the teachers in the comparison and intervention schools use written tests and oral 
 and writing 
progress. To check student understanding during lessons, more than 50% of teachers in 
both comparison and intervention schools question students individually. Data on teacher 
knowledge and skills contradict on the instructional support 
received. For example, when asked what the teacher does when a student is unable to 
answer a question, the most common response from students in both comparison and 
intervention schools was that the teacher hit the student. 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 4-3 
Skills and knowledge of Head Teachers: This year, 50% of the HTs in the Literacy & School 
Feeding group attended a WV or PAI training. Fifty-three percent of the HTs in the Literacy & 
School Feeding group received training on the fundamentals of reading instruction, 
compared to 35% in School Feeding Only schools. More than 65% of the HTs in both the 
comparison and intervention schools received training on school management. Less than 
40% of the HTs in comparison and intervention schools received training on access to and 
the use of school supplies and materials. Few HTs in comparison and intervention schools 
could name the core reading skills. Classroom observations and monitoring of student 
results on tests given by teachers were the most common approaches used to identify 
academic progress by HTs in both comparison and intervention schools. 
: Sixty to 75% of Grade 2 and 4 students bring reading 
books home from the classroom, and nearly all the books brought home are in Portuguese. 
However, 50% of Grade 2 students in the Literacy & School Feeding intervention said that 
they never read on their own at home, and 50% of Grade 2 students in the School Feeding 
Only intervention said they sometimes read on their own at home. More than 50% of Grade 
4 students across groups said they sometimes read on their own at home. Across the study 
arms, less than a quarter of Grade 2 and 4 students bring books home from the school 
library or book bank. 
Parental involvement: Less than 50% of the HTs across the study arms are satisfied with the 
level of parental participation in terms of 
and in the school. Among the HTs who reported that some parents do participate, across 
the study arms, the most common forms of participation consisted of volunteering time and 
resources to the school. More parents in the School Feeding Only arm said they volunteered 
their time in the school feeding program than parents in the Literacy & School Feeding arm. 
Slightly more than 50% of Grade 2 and 4 students across groups reported that someone at 
home helps with homework when needed. This result is in line with parent FGD data, which 
show that parental involvement at home occurs mainly in the form of encouragement to 
complete assigned homework. In the FGDs, some parents reported that older siblings help 
their younger siblings with homework and with practicing reading and writing. Across 
groups, more than half of students in Grade 2 and 4 speak a language other than Portuguese 
at home. This finding is consistent with parent FGD data showing that parents are not fully 
fluent in Portuguese. When asked if someone reads them stories, the most common 
response among Grade 2 students was that no one at home tells them stories. In contrast, 
among Grade 4 students, the most common response was that someone at home 
sometimes tells them stories. 
Access to food/dietary intake and short-term hunger: More than 50% of the students said 
they eat fewer meals at home on days when they receive a school meal, indicating that 
school meals replace home meals for many students. Most students reported that they feel 
less hungry on days when they receive a school meal. These points were supported by the 
qualitative data obtained from parents, many of whom reported that their children skip 
meals at home on school days, in part because they are less hungry when they have a school 
meal. 
4-4 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Timing of the school meal: At most schools, meals are served at 9:30 a.m., which is in the 
middle of the morning session. This timing may have an impact on student attentiveness, if 
meals do indeed improve attentiveness, because the first half of the school day is over 
before the meal is served. 
School feeding: Nearly all of the School Feeding Only and Literacy & School Feeding schools 
serve school meals; however, it appears that meals are not served daily in all schools. 
Instead, on average, schools serve meals three to four times per week. More than half of 
the intervention schools do not serve school meals on time. Parents appreciate the school 
feeding program because it encourages their children to go to school. They recommended 
enhancing or changing the foods included in school meals. A small number of Comparison 
schools had formal school feeding programs, but half of the Comparison schools also 
offered some type of school food during the rainy season when food from their school 
gardens was available. The presence of school feeding activities in some of the Comparison 
schools should be considered in the final evaluation analysis. 
Access to requisite food preparation and storage tools and equipment: The conditions of 
the kitchen and storeroom were very good in only approximately one quarter of the 
intervention schools. Some schools had experienced a gap in the supply of corn-soy blend. 
Most schools had cooking pots and utensils available, but at least one quarter noted that 
the number of these utensils was not sufficient. Roughly 20% of schools lacked bowls and 
spoons for students. At least one third of the schools did not have clean water available. 
Soap for cooks and students to wash their hands was lacking, especially in Literacy & School 
Feeding schools. 
Knowledge of safe food preparation and storage practices: Less than 50% of the school 
cooks had received training. Their knowledge of food safety and hygiene was low to 
moderate. 
Knowledge of health and hygiene practices: Students had very low knowledge of nutrition 
and moderate knowledge of hygiene and sanitation. In intervention schools, one quarter to 
one half of the teachers were not trained on teaching about nutrition, and only 
approximately one third had nutrition lesson plans. Fewer teachers were trained to teach 
about hygiene and sanitation than about nutrition, and even fewer had lesson plans for 
hygiene and sanitation than for nutrition. 
Health and dietary practices of students: We found that many students across study arms 
had inadequate diets, with inadequate meal frequency being more common than low 
dietary diversity. Less than half of the students reported that they always wash their hands 
with soap or ash before eating the school meal. 
School infrastructure and access to water and sanitation services: Less than 50% of the 
intervention schools had a garden. Most intervention schools had improved latrines, but 
less than half of those with improved latrines had tippy taps or hand-washing stations. 
According to teachers, the most serious issues with the school environment were the 
dilapidated state of the school buildings and the lack of clean drinking water. Parents who 
participated in the FGDs also mentioned these as major problems. The lack of water impacts 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 4-5 
the students in some intervention schools because they are required to get up early to fetch 
water to bring to school to make the porridge. 
Access to preventive health interventions: Most intervention schools offered deworming to 
students. 
Student attendance and health absences: Student attendance was difficult to capture 
accurately. We used student recall and compared class enrollment lists to the number of 
students present on the day of data collection. Across the study arms, more than one 
quarter of students were absent during the last week with no significant differences 
between the groups and no significant differences by gender. Students were most often 
absent because they were ill. 
Community understanding of the benefits of education: Parents who participated in the 
FGDs understood that education and reading are important because their children need 
literacy skills to be gainfully employed later in life and to be able to support themselves and 
their family members. Parents mentioned insufficient quantities of books and workbooks for 
students in some schools, which led students to drop out. 
Teacher attendance: Approximately one quarter of teachers and HTs were absent during 
the last week. 
5-6 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
5. LIMITATIONS & RECOMMENDATIONS FOR IMPACT 
EVALUATION 
In this section, we describe the limitations of the study and offer recommendations relevant 
to future data collection efforts (midline and endline) and impact evaluation analysis. 
Limitations and corresponding recommendations from a methodological perspective are 
included as a table in Annex C. 
Limitations 
The School Feeding Only study arm is only available in Maputo. As a result, the design is 
unbalanced, which will impact the distribution of degrees of freedom and the analytical 
approach taken during the midterm analysis and final impact evaluation. 
Intervention districts were matched with comparison districts manually based on the 
limited information available on these districts and current information on future donor 
activities. 
At baseline, we found that some comparison schools served school meals. 
Most intervention schools have participated in school feeding activities since 2012, so 
the school-level 
Because this study was not a randomized controlled trial, intervention schools may be 
inherently different from comparison schools. The purposive selection of intervention 
schools for inclusion in the program could be related to their past poor performance, 
which could be one of the factors explaining why comparison schools seem to 
outperform intervention schools on some indicators. 
Recommendations 
Difference-in-difference analysis should be used at midline and endline to account for 
the lack of a true baseline and the inherent differences between the three groups at 
baseline. 
Other donor and governmental interventions in target districts should be tracked over 
the course of the program so they can be accounted for or considered in subsequent 
parts of the evaluation. 
Matching at the school, teacher, and student levels should be considered in future 
analyses to control the variance. No matching below the district level was attempted in 
the baseline assessment. 
Based on findings from the baseline quantitative analysis, we recommend adding 
additional qualitative data collection from teachers and cooks/warehouse managers to 
facilitate a more comprehensive mixed methods approach at midline and endline. This 
would produce more information to explain and support the quantitative analysis 
findings. For example, to reduce changes to the questionnaire and shed light on data 
regarding access to and use of the teacher guide, mother tongue instruction, and 
materials such as those from the school library or book bank, we recommend that 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 5-7 
qualitative in-depth interviews with teachers be conducted at midline. Furthermore, 
adding comparison schools to the qualitative interviews would help elucidate issues at 
these schools compared to intervention schools. 
Given that the literacy intervention will include mother tongue instruction and materials 
in some areas, we recommend designing an EGRA in the new target language of 
instruction, in addition to the Portuguese EGRA at midline. We also recommend keeping 
the question that asks about the percentage of teachers who report access to and use of 
mother tongue materials. Additionally, we recommend retaining the question on 
We recommend the inclusion of a meta-analysis of the impacts in Maputo and Nampula 
to overcome the confounding factors and variance issues resulting from the differences 
in program implementation strategies across implementers, baseline levels, and 
contextual factors in the two provinces. 

_________________________________
RTI International is a registered trademark and a trade name of Research Triangle Institute. 
6. REFERENCES 
Abadzi, H. (2012). Developing cross-language metrics for reading fluency measurement: 
Some issues and options. Global Partnership for Education working paper series on learning. 
No. 6. Washington DC: World Bank. 
Bennett, C., & Blissett, J. (2014). Measuring hunger and satiety in primary school 
children. Validation of a new picture rating scale. Appetite, 78, 40 48. 
Hammill, D. D. (2004). What we know about correlates of reading. Exceptional Children, 
70(4), 453 468. 
Heugh, K., Benson, C., Bogale, B., Yohannes, M. A. G. (2007). Final report study on 
medium of instruction in primary schools in Ethiopia. Commissioned by the Ministry of 
Education. 
Samuels, S. J., & Flor, R. F. (1997). The importance of automaticity for developing 
expertise in reading. Reading & Writing Quarterly, 13, 107 121 
Scarborough, H. (2001). Connecting early language and literacy to later reading 
(dis)abilities: Evidence, theory and practice. In S. Neuman & D. Dickison (Eds.), Handbook of 
family literacy (pp. 97 110). New York: Guilford Press. 
Tobia, V., & Bonifacci, P. (2015). The simple view of reading in a transparent 
orthography: The stronger role of oral comprehension. Reading and Writing, 28(7), 939
957. 
Weiser, B., & Mathes, P. (2011). Using encoding instruction to improve the reading and 
spelling performances of elementary students at risk for literacy difficulties: A best-evidence 
synthesis. Review of Educational Research, 81(2), 170 200. 
WHO. (2010). Indicators for assessing infant and young child feeding practices Part 2 
Measurement. Geneva, Switzerland: WHO. 
Wood, J., & McLemore, B. (2001). Critical components in early literacy Knowledge of 
the letters of the alphabet and phonics instruction. Florida Reading Quarterly, 38(2), 1 8. 
6-2 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report A-1 
7. ANNEX A. WITHIN-PROVINCE LITERACY ANALYSIS 
Table 51 presents the Early Grade Reading Assessment (EGRA) results for Maputo Province 
by study arm. In Maputo Province, intervention arms include Comparison schools, School 
Feeding Only schools, and Literacy & School Feeding schools. The table provides the percent 
zero scores for all subtasks, mean scores for timed subtasks, and percent mean scores for 
untimed subtasks. 
Table 51: Grade 2 and 4 EGRA Results for Maputo Province, by Study Arm 
Maputo Province Grade 2 Grade 4 
Subtask Study Arm Mean Scores
Percent Zero 
Scores Mean Scores
Percent 
Zero Scores
Letter Identification 
(clpm) 
Comparison 16.4 29.4% 45.5 3.5% 
School Feeding Only 11 49.4% 37.7 8.6% 
Literacy & School Feeding 11.7 42.3% 37 5.6% 
Syllable Reading 
(csspm) 
Comparison 9.9 57.3% 36.6 14%
School Feeding Only 6 69.7% 29.4 22.1%
Literacy & School Feeding 5.9 66.8% 28.2 21.2% 
ORF 
 
 
Comparison 7.6 26.9% 35.6 10.7% 
School Feeding Only 3.6 56.5% 24.4 24.1%
Literacy & School Feeding 3.9 49.9% 24.5 20.7% 
Percent 
Mean Scores
Percent Zero 
Scores
Percent 
Mean Scores
Percent 
Zero Scores
Reading 
Comprehension 
Comparison 8.7% 80.3% 43.7% 31.7%
School Feeding Only 2% 93.1% 23.2% 57.6% 
Literacy & School Feeding 3.7% 87.9% 26.7% 50% 
Listening 
Comprehension 
Comparison 49.2% 12.7% 63.9% 4.4% 
School Feeding Only 21.2% 50.8% 39% 25.6%
Literacy & School Feeding 24.6% 44.9% 44.1% 20.3% 
Word Dictation 
Comparison 22.4% 61.4% 65.4% 19.1% 
School Feeding Only 13% 74.1% 55.1% 25.2% 
Literacy & School Feeding 15.6% 71.1% 55% 27.1%
Note: clpm = correct letters per minute; csspm = correct syllable sounds per minute; EGRA = Early Grade Reading 
Assessment; ORF = oral reading fluency. 
Sample sizes for Grade 2: Comparison, n=704; School Feeding Only, n=666; Literacy & School Feeding, n=661 
Sample sizes for Grade 4: Comparison, n=681; School Feeding Only, n=616; Literacy & School Feeding, n=638 
Table 52 presents the EGRA results for students in Nampula Province in Grades 2 and 4 by 
study arm. The study arms in Nampula include Comparison schools and schools that will 
receive the Literacy & School Feeding intervention. The table provides the percent zero 
scores for all subtasks, mean scores for timed subtasks, and percent mean scores for 
untimed subtasks. 
A-2 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Table 52: Grade 2 and 4 EGRA Results for Nampula Province, by Study Arm 
Nampula Province Grade 2 Grade 4 
Subtask Study Arm Mean Scores
Percent Zero 
Scores Mean Scores
Percent 
Zero Scores
Letter Identification 
(clpm) 
Comparison 5.4 39.6% 18.6 17.8%
Literacy & School Feeding 6.4 43.4% 14.8 20.2% 
Syllable Reading 
(csspm) 
Comparison 2.3 81.6% 12.4 47.7% 
Literacy & School Feeding 2.7 80.2% 8.1 56.3% 
ORF
 
 
Comparison 1.8 52.6% 9.4 37.1%
Literacy & School Feeding 2.2 53.5% 6.3 42.6% 
Percent 
Mean Scores
Percent Zero 
Scores
Percent 
Mean Scores
Percent 
Zero Scores
Reading 
Comprehension 
Comparison .7% 96.9% 7.6% 75.7% 
Literacy & School Feeding 1.2% 96% 4.5% 87.1% 
Listening 
Comprehension 
Comparison 13.1% 59.6% 23.3% 40.3%
Literacy & School Feeding 8.3% 75.2% 16.6% 59.4%
Word Dictation 
Comparison 2.4% 91% 14.7% 59.9% 
Literacy & School Feeding 4% 88.8% 10.4% 71.4% 
Nampula Province Grade 2 Grade 4 
Subtask Study Arm Mean Scores
Percent Zero 
Scores Mean Scores
Percent 
Zero Scores
Letter Identification 
(clpm) 
Comparison 5.4 39.6% 18.6 17.8% 
Literacy & School Feeding 6.4 43.4% 14.8 20.2%
Syllable Reading 
(csspm) 
Comparison 2.3 81.6% 12.4 47.7%
Literacy & School Feeding 2.7 80.2% 8.1 56.3% 
ORF 
 
 
Comparison 1.8 52.6% 9.4 37.1% 
Literacy & School Feeding 2.2 53.5% 6.3 42.6% 
Percent 
Mean Scores
Percent Zero 
Scores
Percent 
Mean Scores
Percent 
Zero Scores
Reading 
Comprehension 
Comparison .7% 96.9% 7.6% 75.7%
Literacy & School Feeding 1.2% 96% 4.5% 87.1% 
Listening 
Comprehension 
Comparison 13.1% 59.6% 23.3% 40.3% 
Literacy & School Feeding 8.3% 75.2% 16.6% 59.4% 
Word Dictation 
Comparison 2.4% 91% 14.7% 59.9%
Literacy & School Feeding 4% 88.8% 10.4% 71.4%
Note: clpm = correct letters per minute; csspm = correct syllable sounds per minute; EGRA = Early Grade Reading 
Assessment; ORF = oral reading fluency. 
Sample sizes for Grade 2: Comparison, n=690; Literacy & School Feeding, n=707 
Sample sizes for Grade 4: Comparison, n=676; Literacy & School Feeding, n=685 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report A-3 
8. ANNEX B. PRECISION ESTIMATES 
Table 53 shows the weighted oral reading fluency (ORF) mean estimate and the confidence 
in the mean estimates by province and study arm. All mean estimates for Grade 2 fall well 
within the ±2.5 wpm 95% confidence interval (CI) band. All groups fall under ±1.0 wpm, 
except for the Maputo Comparison group, which has a band of ±1.3 wpm. Table 54 provides 
the same information for the Grade 4 students. The mean ORF for Grade 4 students in 
Nampula stayed within the expected 95% confidence bands: ±1.5 wpm for the intervention 
group and ±1.7 wpm for the Comparison group. However, the mean ORF for Grade 4 
students in Maputo exceeded the expected precision. The precision values in the 
Comparison group and the School Feeding Only group were similar: ±2.8 and ±3.2 wpm, 
respectively. The Literacy & School Feeding group had a precision of ±4.0 wpm because of 
the combined high inter-class correlation (ICC) and standard deviation (SD), which were not 
expected. Overall, the estimates were sufficiently precise. 
Table 53. Grade 2 Reading Fluency Means and Precision Estimates of the Mean ORF 
 Provinces Study Arm n Range 
(wpm) 
Mean 
 (wpm) 
95% CI 
Band 
 (wpm) 
SD 
 (wpm) 
ICC DEFF 
Maputo Literacy & School 
Feeding 
661 [0,46] 3.9 ±0.9 7.6 0.134 2.8 
School Feeding 
Only 
666 [0,55] 3.6 ±0.7 8 -0.018 0.7 
Comparison 703 [0,89] 7.6 ±1.3 13.4 0.183 3 
Nampula Literacy & School 
Feeding 
707 [0,69] 2.2 ±0.9 5.6 0.427 7.4 
Comparison 690 [0,64] 1.8 ±0.7 6.2 0.108 2.7 
Note: CI = confidence interval; DEFF = Design effect; ICC = inter-class correlation; SD = standard deviation; wpm = words 
per minute. 
Table 54. Grade 4 Reading Fluency Means and Precision Estimates of the Mean ORF 
Provinces Study Arm n 
Range 
 (wpm) 
Mean 
 (wpm) 
95% CI 
Band 
 (wpm) 
SD 
 (wpm) ICC DEFF 
Maputo Literacy & School 
Feeding 
638 [0,112] 24.5 ±4.0 22.1 0.252 4.5 
School Feeding 
Only 
616 [0,115] 24.4 ±3.2 24.1 -0.005 0.9 
Comparison 680 [0,122] 35.6 ±2.8 26.9 0.06 2.3 
Nampula Literacy & School 
Feeding 
685 [0,92] 6.3 ±1.5 12.2 0.115 2.6 
Comparison 676 [0,89] 9.4 ±1.7 14.7 0.109 2.5 
Note: CI = confidence interval; DEFF = Design effect; ICC: inter-class correlation; SD = standard deviation; wpm = words 
per minute. 
A-4 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report A-5 
9. ANNEX C. IMPACT EVALUATION RECOMMENDATIONS (TABLE) 
Source Limitation Consequences Recommendations 
Exogenous 
Factors & 
Program 
Design 
Different conditions, 
environment, and access 
to resources between 
Maputo and Nampula 
High variance 
expected which might 
mask some of the 
significant differences 
Meta-analysis for 
midterm and final impact 
evaluation 
Exogenous 
Factors & 
Implementer 
Strategy 
Different implementation 
activities by Planet Aid 
and World Vision 
Implementer and 
province variances are 
confounded resulting 
in increased variance 
which might mask 
some of the 
significance 
differences 
Meta-analysis for 
midterm and final impact 
evaluation analysis 
Implementer 
Strategy 
Program phased out for 
developing curriculum 
Different levels of 
exposures for students 
across different 
cohorts in midline and 
endline 
Special care in the 
analysis on the exposure 
of each cohort relative to 
implementer activities 
and inclusion of 
indicators for intensity of 
the treatment 
Implementer 
Strategy 
No schools receiving a 
intervention 
The quasi￾experimental design is 
not a full factorial 
design 
Special care in midterm 
and final impact 
evaluation analysis 
Implementer 
Strategy 
School Feeding Only study 
arm is only available in 
Maputo 
Unbalanced design, 
which will impact the 
distribution of degrees 
of freedom 
Special care to account 
for this difference during 
the midterm and final 
impact evaluation 
analysis 
Exogenous 
Factor & 
Implementer 
Strategy 
Intervention changes in 
response to endogenous 
and exogenous factors 
(e.g., schools with Literacy 
Only in Maputo because 
of other school feeding 
programs overlapping) 
Implications on 
sampling and data 
collection efforts 
Keep track of activities by 
implementers over time 
and update school lists 
and relevant activities to 
account for these 
changes in sampling and 
analysis 
A-6 USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report 
Exogenous 
Factor 
Limited information on 
current and future 
activities on literacy and 
nutrition at the district 
and individual school level 
Manual matching of 
intervention and 
comparison districts 
Need additional model 
based matching effort 
(e.g., propensity score 
matching) and continue 
collecting information 
about relevant school 
level activities 
Exogenous 
Factors 
Some comparison 
schools serve school 
meals or have 
overlapping activities 
Confounding factors 
due to contamination 
of the comparison 
schools 
Continue tracking other 
donor and governmental 
institution activities 
throughout the 
intervention period to 
account for them in the 
impact evaluation 
analysis 
Exogenous 
Factors 
Wide range of differences 
across schools within 
districts with limited 
information available 
during sampling 
Large variance which 
could impact the 
ability to detect 
significant differences 
Conduct matching at 
school, teacher, and 
student level prior to 
impact evaluation 
analysis 
Program 
Design 
Intervention schools 
participated in nutrition 
activities since 2012 
Not a true baseline 
across all indicators 
Use of difference-in￾difference approach for 
midterm and final impact 
evaluation analysis 
relative to 2017 
Program 
Design 
Not a randomized control 
trial and purposive 
selection of districts and 
school for inclusion in the 
program because of past 
poor performance 
Comparison schools 
might be better and 
different relative to 
intervention schools 
because of inherent 
selection bias 
Use propensity score 
matching and difference￾in-difference approach 
for midterm and final 
impact evaluation 
Baseline 
Results 
Unexpected trends in 
quantitative findings 
Need to explain 
reasons for the 
observed trends 
Use a mixed methods 
approach by adding more
qualitative data 
collection from teachers, 
cooks, and warehouse 
managers in midline and 
endline data collection 
Include comparison 
schools in qualitative 
USDA McGovern Dole Mozambique Impact Evaluation and Literacy Assessment: Final Baseline Report A-7 
data collection and 
analysis 
Avoid too many changes 
to quantitative 
instruments for 
consistency and 
comparability 
Program 
Design and 
Implementer 
Strategy 
Integration of mother 
tongue literacy program 
No measurements at 
baseline of mother 
tongue reading 
assessment 
Develop Early Grade 
Reading Assessment 
(EGRA) in mother tongue 
based on curriculum 
developed once available 
for midline and endline 
Assume zero scores at 
baseline