UGANDA PERFORMANCE AND IMPACT 
EVALUATION FOR LITERACY ACHIEVEMENT 
AND RETENTION ACTIVITY (LARA)
FINAL IMPACT AND PERFORMANCE 
EVALUATION REPORT
Prepared under Contract No.: AID-617-C-16-00003 
ALICIA MENENDEZ FOR NORC
UGANDA PERFORMANCE AND IMPACT 
EVALUATION FOR LITERACY 
ACHIEVEMENT AND RETENTION 
ACTIVITY (LARA)
FINAL IMPACT AND PERFORMANCE 
EVALUATION REPORT
April 2021 
Prepared under Contract No.: AID-617-C-16-00003 
Submitted to: 
USAID/Uganda 
Submitted by: 
NORC at the University of Chicago 
4350 East West Highway, 8th Floor 
Bethesda, MD 20814 
Attention: Ritu Nayyar-Stone 
Tel: 301- 634-9489; E-mail: nayyarstone-ritu@norc.org
PHOTO CREDIT 
Literacy Hour in P2 Classroom, Ugandan Primary School; by Alicia Menendez
DISCLAIMER 
The authors’ views expressed in this publication do not necessarily reflect the views of the United States 
Agency for International Development or the United States Government.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | i
TABLE OF CONTENTS
TABLES.........................................................................................................................................III
FIGURES .......................................................................................................................................III
ACRONYMS..................................................................................................................................V
GLOSSARY................................................................................................................................. VII
EXECUTIVE SUMMARY..............................................................................................................X
EVALUATION PURPOSE..........................................................................................................................................X
LARA BACKGROUND.............................................................................................................................................XI
EVALUATION DESIGN, DATA COLLECTION, AND SAMPLES ................................................................XI
SRGBV FINDINGS....................................................................................................................................................XIII
RETENTION AND ATTENDANCE FINDINGS............................................................................................. XIV
EARLY GRADE READING FINDINGS.............................................................................................................. XIV
RECOMMENDATIONS......................................................................................................................................... XVI
SCHOOL-RELATED GENDER-BASED VIOLENCE.................................................................................. XVI
EARLY GRADE READING..............................................................................................................................XVII
POLICY BRIEFS ......................................................................................................................................................XVIII
1 INTRODUCTION ..................................................................................................................1
1.1 EVALUATION PURPOSE .............................................................................................................................1
1.2 EVALUATION COMPONENTS.................................................................................................................1
1.3 EVALUATION REPORT STRUCTURE.....................................................................................................2
2 LARA BACKGROUND ..........................................................................................................3
2.1 PROGRAM OBJECTIVES ..............................................................................................................................3
2.2 PROGRAM ACTIVITIES................................................................................................................................3
R1. INCREASED CAPACITY TO DELIVER EARLY GRADE READING ..................................................3
R2. IMPROVED RETENTION IN PRIMARY GRADES ..................................................................................5
3 EVALUATION DESIGN ........................................................................................................9
3.1 METHODOLOGY..........................................................................................................................................9
3.2 SAMPLE............................................................................................................................................................10
ANALYTIC SAMPLE SIZES..................................................................................................................................11
3.3 INSTRUMENTS .............................................................................................................................................12
ii | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
EARLY GRADE READING DATA COLLECTION TOOLS.......................................................................12
SRGBV DATA COLLECTION TOOLS ...........................................................................................................13
RETENTION AND ATTENDANCE DATA COLLECTION TOOLS .....................................................14
PERFORMANCE EVALUATION DATA COLLECTION TOOLS ............................................................14
3.4 LIMITATIONS................................................................................................................................................15
4 FINDINGS ............................................................................................................................ 16
SRGBV FINDINGS......................................................................................................................................................16
4.1 EFFECTS OF R2 ACTIVITIES ON SRGBV INTERMEDIATE OUTCOMES...................................16
GENDER ATTITUDES..........................................................................................................................................16
CAREGIVERS ..........................................................................................................................................................17
TEACHERS ..............................................................................................................................................................17
SCHOOL CLIMATE ..............................................................................................................................................18
DISCIPLINARY METHODS.................................................................................................................................21
PREVALENCE OF VIOLENCE............................................................................................................................25
REPORTING OF VIOLENCE..............................................................................................................................28
EVIDENCE OF FIDELITY OF IMPLEMENTATION......................................................................................30
4.2 RETENTION AND ATTENDANCE FINDINGS (R&A).....................................................................41
4.3 EARLY GRADE READING PROFICIENCY FINDINGS ............................................................................47
EVIDENCE OF FIDELITY OF IMPLEMENTATION OF EARLY GRADE READING ACTIVITIES....51
5 CONCLUSIONS.................................................................................................................. 70
5.1 SRGBV..............................................................................................................................................................70
5.2 EGR...................................................................................................................................................................71
6 RECOMMENDATIONS ...................................................................................................... 73
7 POLICY BRIEFS................................................................................................................... 75
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | iii
TABLES
Table 1. Training on Early Grade Reading ...................................................................................................................3
Table 2. Contents of Journeys Activity Handbooks..................................................................................................6
Table 3. SRGBV Journeys Training ................................................................................................................................7
Table 4. NORC LARA P&IE Data Collection Tasks (2017 – 2020)....................................................................10
Table 5. Number of Schools in Each Sample, by Language....................................................................................10
Table 6. Grades of learners in NORC LARA samples ...........................................................................................11
Table 7. Sample Characteristics...................................................................................................................................11
Table 8. Focus Group Discussions for EGR and SRGBV.......................................................................................12
Table 9. Inequitable Gender Attitude Index1 at Midline .........................................................................................16
Table 10. Learner1 School Climate, Percent Agreeing with Statements............................................................18
Table 11. Violent and non-violent disciplinary methods employed by caregivers ...........................................21
Table 12. Disciplinary methods employed by teachers..........................................................................................22
Table 13. Learners that suffered violence..................................................................................................................26
Table 14. Teachers’ use of a cane or stick to hit learners, as reported by learners.......................................27
Table 15. Non-verbal disclosure of sexual violence perpetrated by teachers..................................................28
Table 16. Learner Protection Resources...................................................................................................................28
Table 17: LARA AMELP Indicators..............................................................................................................................31
Table 18. Teacher training.............................................................................................................................................36
Table 19. Type of LARA training received by T2 teachers under R2 activities................................................36
Table 20. Learners’ reasons for not attending school1...........................................................................................44
Table 21. EGRA Effects on P3 Learners.....................................................................................................................48
Table 22. Percentage of learners in class that had their LARA reading book...................................................61
FIGURES
Figure 1. Selection of Treatment and Control Groups............................................................................................9
Figure 2. School Climate Indicators by Treatment Arm, Luganda Schools .......................................................20
Figure 3. School Climate Indicators by Treatment Arm, Runyankore/Rukiga Schools ..................................20
Figure 4. Percent of caregivers, teachers, and head teachers that think corporal punishment is an 
effective form of discipline.............................................................................................................................................23
Figure 5. Reports of hitting a learner with a cane or stick being effective at midline .....................................24
Figure 6. Learners’ reported number of types of violence suffered1 ..................................................................27
Figure 7. Teacher knowledge........................................................................................................................................35
Figure 8. Head teacher knowledge..............................................................................................................................35
Figure 9. LARA SRGBV activities implemented in T2 schools, teachers and head teachers.........................37
Figure 10. LARA SRGBV activities implemented in T2 schools, learners..........................................................40
Figure 11. Learners’ enrollment status in Y3T3 by treatment group and region.............................................42
Figure 12. Grade retention in Y3T3 by treatment group and region .................................................................42
Figure 13. Learners’ average attendance by treatment group and region .........................................................43
Figure 14. Learner self-reports being absent at least one school day on the week before the survey, by 
treatment and region ......................................................................................................................................................44
Figure 15. Teachers’ attendance status by grade .....................................................................................................46
Figure 16. Learners report teacher was absent at least one school day the week before the survey .......46
Figure 17. Oral Reading Fluency (cwpm) in local language at midline, Luganda region ..................................50
iv | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 18. Oral Reading Fluency (cwpm) in local language at midline, Runyankore/Rukiga region .............50
Figure 19. Head teacher participation in LARA EGR training...............................................................................52
Figure 20. P3 teacher participation in EGR training and refresher training.......................................................53
Figure 21. P3 teachers that taught in same school in 2018 ...................................................................................53
Figure 22. LARA EGR support supervision for P3 teachers in 2019 ..................................................................55
Figure 23. Opinion of P3 teachers about LARA EGR support supervision.......................................................55
Figure 24. Frequency that CCTs observed P3 teachers in 2019..........................................................................56
Figure 25. Printed P3 teacher guide for lesson planning ........................................................................................58
Figure 26. Percent of classes where P3 learners have reading books in the language of instruction, as 
reported by P3 teacher..................................................................................................................................................60
Figure 27. Percent of P3 learners observed having LARA reading books .........................................................60
Figure 28. P3 Teachers’ opinion about LARA pupil reading book.......................................................................61
Figure 29. Frequency of use of CAM forms (for those with CAM forms) ........................................................62
Figure 30. Availability of LARA supplementary reading materials in English and story cards in P3 classes, 
as reported by the teacher............................................................................................................................................64
Figure 31. Reading practices of P3 learners in the classroom ..............................................................................65
Figure 32. Head teachers’ awareness of the ULC ...................................................................................................67
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | v
ACRONYMS
C1, C2 Cluster 1, Cluster 2
CAM Continuous Assessment Monitoring
CAPI Computer-Assisted Personal Interview
CBO Community-Based Organization
CCT Coordinating Center Tutors
CDO Community Development Officer
CRO Classroom Observations
CSR Center for Social Research 
DEO District Education Officer
DPSWO District Probation and Social Welfare Officers 
CLA Collaborating, Learning, and Adapting
COP Chief of Party
EGR Early Grade Reading
EGRA Early Grade Reading Assessment
FA Field Assistants
FGD Contracting Officer Representative
GBV Gender-Based Violence
GPE Global Partnership for Education
IE Impact Evaluation
IP Implementing Partner
KII Key Informant Interviews
LARA Literacy Achievement and Retention Activity 
LC Local Councils
LOE Level of Effort
MDES Minimum Detectable Effect Size
MoES Ministry of Education and Sports 
MoGLSD Ministry of Gender, Labor, and Social Development
NORC NORC at the University of Chicago
ORF Oral Reading Fluency 
P1, P2, P3, P4, P6 Primary 1, Primary 2, Primary 3, Primary 4, Primary 6
P&IE Performance and Impact Evaluation
vi | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
PE Performance Evaluation
PTC Primary Teacher Colleges
R1, R2 Result 1, Result 2
RCT Randomized Controlled Trial
R&A Retention and Attendance 
RTI Research Triangle Institute
RTRR Reporting, Tracking, Referral, Response
RWI Research World International 
SBCC Social Behavior Change Communication
SCA School Change Agreement
SEL Social and Emotional Learning
SHRP School Health and Reading Program
SRGBV School-Related Gender-Based Violence 
STS School-to-School
T1, T2 Treatment 1, Treatment 2
TASO The AIDS Support Organization
TG Teacher Guides
ToT Training of Teachers
ULC Uganda Learning/Literacy Campaign 
UKU Uganda Kids Unite
USAID United States Agency for International Development
VACiS Violence Against Children in Schools
VCCMC Village Child Case Management Committees
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | vii
GLOSSARY
fBaseline. A survey to collect data prior to the start of the intervention. 
Beneficiary or beneficiaries. Beneficiaries are the individuals, households, firms, facilities, villages or 
similar that are exposed to an intervention with beneficial intentions. 
Cluster sample. A multi-stage sample design, in which a sample is first drawn of geographical areas 
(e.g. sub-districts or villages), and then a sample of households, firms, or facilities is drawn from within 
the selected districts. The design results in larger standard errors than would occur in a simple random 
sample, but is often used to minimize cost. 
Comparison Group. A group of individuals whose characteristics are similar to those of the 
treatment groups (or participants) but who do not receive the intervention. The comparison group is 
called a control group when the evaluator can ensure, under trial conditions, that no confounding 
factors affect it.
Confidence interval (CI). A range of values around a value measured from a sample that reveals how 
precisely the sample value reflects the population value. A larger confidence interval reflects lower 
precision. For example, if the average age of a sample is 36, then a smaller confidence interval (from 35 
to 37) suggests that the sample average age is likely a more precise estimate of the population average 
age than if the confidence interval were larger (ranging from 34 to 38, for example). 
Confidence level. The level of certainty that the true value of impact (or any other statistical estimate) 
will be included within a specified range. 
Control Group. A special case of the comparison group, in which the evaluator can control the 
environment and so limit confounding factors.
Counterfactual. The state of the world in the absence of the intervention. For most impact 
evaluations, the counterfactual is the value of the outcome for the treatment group in the absence of the 
intervention. However, studies should also pay attention to unintended outcomes, including effects on 
non-beneficiaries. 
Convenience sample. Also known as reliance on available subjects, a convenience sample is a 
nonprobability sample that relies on data collection from population members who are easy to reach (or 
conveniently available). This method does not allow for generalizations and is of limited value in social 
science research.
Effect Size. The size of the relationship between two variables (particularly between program variables 
and outcomes). See also “Minimum detectable effect .”
Experimental Design. See “Randomized Control Trial.”
Fidelity of Implementation. The degree to which an intervention or program is delivered as 
designed.
viii | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Impact. How an intervention alters the state of the world. Impact evaluations typically focus on the 
effect of the intervention on the outcome for the beneficiary population. 
Impact evaluation. A study of the attribution of changes in the outcome to the intervention. Impact 
evaluations have either an experimental or quasi-experimental design.
Intervention The project, program, or policy that is the subject of the impact evaluation. 
Intra-class correlation coefficient (ICC). This is a descriptive statistic that is used when data are 
clustered into groups. The statistic ranges from 0 and 1 and measures how closely members of a group 
resemble each other in certain observable characteristics. ICCs can also be used to gauge the 
consistency of measurement across observers. Large impact evaluation studies applying statistical means 
to construct a counterfactual require a sufficiently large sample size (n) to ensure statistical power. 
Minimum detectable effect. The smallest effect size the researcher deems necessary to detect in the 
impact evaluation. Used to perform the power calculation necessary to determine the required sample 
size. 
Mixed methods. The use of both quantitative and qualitative methods in an impact evaluation design. 
Sometimes called Q-squared or Q2. 
Outcome. A variable, or variables, that measure the impact of the intervention. 
Power. The ability of a study to detect an impact. Conducting a power calculation is a crucial step in 
impact evaluation design, 
Power calculation. A calculation of the sample required for the impact evaluation, which depends on 
the minimum effect size and the required level of confidence. 
Purposive sample. A form of non-probability sampling in which researchers rely on their own 
judgment when choosing members of the population to participate in their study
Random assignment. An intervention design in which members of the eligible population are assigned 
at random to either the treatment group or the control group. That is, whether someone is in the 
treatment or control group is solely a matter of chance, and not a function of any of their characteristics 
(either observed or unobserved). 
Randomized Controlled Trial (RCT). An impact evaluation design in which random assignment has 
been used to allocate the intervention amongst members of the eligible population. Since there should 
be no correlation between participant characteristics and the outcome, and differences in outcome 
between the treatment and control can be fully attributed to the intervention, i.e. there is no selection 
bias. However, RCTs may be subject to several types of bias and therefore need to follow strict 
protocols. Also called “Experimental design.”
Sample. A subset of the population being studied. The sample is drawn randomly from the sampling 
frame. In a simple, random sample, all elements in the frame are equally likely to be selected, but usually,
more complex sampling designs are used, requiring the use of sample weights in analysis. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | ix
Sampling frame. The complete list of the population of interest in the study. This is not necessarily 
the whole population of the country or area being studied, but rather is restricted to the eligible 
population, e.g. grade 3 teachers in public schools. For a school survey, the sampling frame would be all 
schools in the area of study. 
Secondary data. Data that has been collected for another purpose, but may be reanalyzed in a 
subsequent study. 
Standard deviation. A measure of the amount of variation or dispersion in a set of values. A low 
standard deviation indicates that the values tend to be close to the mean of the set, while a high 
standard deviation indicates that the values are spread over a wider range.
Statistical power. The probability that a study will find a treatment effect given that there is a 
treatment effect to be detected. 
Statistical significance. The likelihood that a treatment effect found in a study is not a result of 
chance. A higher statistical significance indicates a higher likelihood that the observed treatment effect is 
not the result of chance.
Survey. The collection of information using (1) a pre-defined sampling strategy, and (2) a survey 
instrument. A survey may collect data from individuals, households, or other units such as schools or 
hospitals.
Survey instrument. A pre-designed form (questionnaire) used to collect data during a survey. 
Theory of change. Laying out the underlying causal chain linking inputs, activities, outputs, and 
outcomes, and identifying the assumptions required if the intervention is to be successful. A theory of 
change is the starting point for theory-based impact evaluation. 
Treatment group. The group of people, firms, or facilities who receive the intervention. Also called 
participants or beneficiaries. 
x | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
EXECUTIVE SUMMARY 
EVALUATION PURPOSE
NORC at the University of Chicago (NORC), in partnership with subcontractor Panagora Group, 
serves as the independent evaluator for the Performance and Impact Evaluation (P&IE) of the Literacy 
Achievement and Retention Activity (LARA), implemented by RTI International (RTI), in Uganda. LARA 
is a 6-year (April 2015‒April 2021)1 USAID-funded activity designed to support the Ugandan Ministry of 
Education and Sports (MoES) in its efforts to improve early grade reading (EGR) and retention.
LARA concentrates on two results: Result 1 (R1) improved capacity to deliver early grade reading in 
three local languages2 and English; and Result 2 (R2) improved retention in early primary grades, through 
the reduction of school-related gender-based violence (SRGBV) resulting in safer school environments. 
Both results are intended to contribute to the overall objective of improving reading skills.
The Uganda LARA P&IE activity (April 2016 - April 2021) has two objectives 1) To assess the impact of 
LARA on learners’ literacy skills and retention rates; and 2) To assess the performance of LARA in
terms of project management, learning, design, implementation, results, and sustainability.
 There are two clusters of schools where LARA activities take place. Cluster 1 (C1) schools 
started receiving the intervention in 2015-2016 prior to the start of this evaluation. The focus 
of NORC’s evaluation is Cluster 2 (C2) schools that LARA only reached in 2017 after NORC 
had collected baseline data. NORC’s LARA P&IE covers C2 schools located in Luganda and 
Runyankore/Rukiga language areas.
The LARA P&IE consists of two components: (1) an Impact Evaluation (IE), and (2) a Performance 
Evaluation (PE). The IE addresses the following questions:
 What is the impact of R1 activities on reading performance and retention rates?
 What is the additional impact of R2 activities on reading performance and what is the impact 
of R2 activities on SRGBV intermediate outcomes?
 What is the total impact that R1 + R2 activities have on reading performance and retention 
rates?
The purpose of the PE is to provide insights into LARA program implementation and the IE results. 
Thus, the PE focuses on the following questions:
 Is LARA on track to achieve results by the end of the project?
 What are the key outcomes of the project?
 
1
LARA was originally conceived as a 5-year activity, but it was extended for a sixth year.
2 The three local languages are Luganda, Runyankore-Rukiga and Runyoro-Rutooro. Since there was only one district with three CCTs that 
predominately used the Runyoro-Rutooro languages, and therefore insufficient sample size, NORC’s evaluation is restricted to the 12 districts 
dominated by Luganda and Runyankore/Rukiga.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | xi
 Were activities carried out as planned? 
 Are activities sufficient and relevant to achieving Result 1 and Result 2?
 What factors accelerated or inhibited the achievement of LARA results?
LARA BACKGROUND
LARA aims to improve the reading skills of 1.3 million primary-grade learners in government schools in 
31 districts throughout Uganda. LARA has two main goals: 
R1 (Increased capacity to deliver early grade reading) focuses on strengthening the capacity of 
MoES and other educational stakeholders to deliver EGR in three local languages in Primary 1 (P1) to 
Primary 3 (P3) with a transition to English in Primary 4 (P4). The R1 main activities of LARA are:
 Teacher training in EGR methodology
 Continuous assessment monitoring (CAM) 
 Distribution of teaching and learning materials and supplemental reading materials
 Teacher support supervision 
 Uganda Learning/Literacy Campaign (ULC)
R2 (Improved retention in primary grades) endeavors to further improve children’s retention and 
active participation in early primary grades (P1–P4) through the reduction of SRGBV, resulting in the 
establishment of a positive and supportive school climate for learning. The main activities are:
 Journeys training and materials. 
 Uganda Kids Unite (UKU) 
 Community subgrants 
 SRGBV-related social behavior change campaign (SBCC) 
 National support for the implementation of Uganda’s National Child Policy (2017–2022)
EVALUATION DESIGN, DATA COLLECTION, AND SAMPLES
The LARA P&IE uses a mixed-methods approach combining a randomized controlled trial (RCT) design 
and qualitative methods. Randomization of treatment assignment was conducted at the Coordinating 
Center Tutors (CCT) level, assigning all schools under a CCT to one of three arms: treatment T1 
(receiving R1 EGR activities only), treatment T2 (receiving R1 EGR and R2 SRGBV activities), or the 
control group (receiving no activities). Within each CCT, schools were randomly selected for data 
collection at baseline (in 2017). The same schools were visited in 2019 for the midline data collection, 
with EGR and SRGBV data collected between July and October 2019 respectively. The PE qualitative 
data was collected in March 2020.
xii | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
The IE also estimates the effect of the R1 and R2 activities on learner attendance, a major predictor of 
dropout and grade repetition. Every school term, NORC checks the retention and attendance (R&A) of 
just over 3,500 learners who were selected at baseline from 71 schools when they were in P1 and P4. 
We conducted classroom observations (CROs) during unannounced visits in term 3 of each school year 
to assess the fidelity of implementation of the EGR activities and compare teaching practices of 30 
treatment and control schools. The table below lists all data collection efforts conducted by the LARA 
P&IE team thus far, between 2017 and the first school term of 2020.
NORC LARA P&IE Data Collection Tasks (2017 – 2020)
YEAR EGR SRGBV R&A CRO PE
2017 Term 1: 
Baseline
Term 1: 
Baseline
Term 1: Round 0 (Sample creation)
Term 2: Round 1
Term 3: Round 2
Term 3: 
Round 1
2018 Term 1: Round 3
Term 2: Round 4
Term 3: Round 5
Term 3: 
Round 2
2019 Term 3: 
Midline
Term 2: 
Midline
Term 1: Round 6
Term 2: Round 7
Term 3: Round 8
Term 3: 
Round 3
2020 Term 1: Round 9 Term 1: Final PE
The table below includes the number of schools sampled for data collection in each district.
Number of Schools in Each Sample, by Language 
LANGUAGE EGR SRGBV R&A CRO* PE*
Luganda 132 40 35 15 2
Runyankore/Rukiga 132 40 36 16 2
TOTAL 264 80 71 31 4
* The schools selected for the CROs and PE are not intended to represent the universe of C2 schools in the Luganda and Runyankore/Rukiga 
language regions.
The schools selected for the EGR and R&A sample are stratified by dominant language (Luganda or 
Runyankore/Rukiga) and treatment status (T1, T2, or control). For EGR surveys, 44 schools were 
selected in each arm for a total of 264 schools. At baseline (2017), 20 P1 learners were randomly 
selected in each school to be assessed. At midline (2019), we randomly selected 20 P3 learners from 
each of the same school. As part of the EGR sample, we also interviewed the teacher in charge of the 
grade and the head teacher.
For the R&A study, twelve schools were selected in each arm—except for the control group with 11 
schools—for a total of 71 schools. At baseline we created a random sample of 30 learners enrolled in 
P1 and 30 learners enrolled in P4. We followed them every school term. 
The sample for SRGBV survey-related data includes only schools in the T1 or T2 groups, given that the 
relevant comparison is between schools that received the EGR activities and schools that received the
EGR plus the SRGBV activities. The sampled schools are stratified by language and treatment status with 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | xiii
each of the four arms including 20 schools—80 schools total. However, schools in both language regions 
can be pooled together during analysis, in contrast to EGR. At baseline, we interviewed 20 learners 
enrolled in P2, 20 in P4, and 20 in P6 in each school. At midline we interviewed 20 P4 and 20 P6 
learners -and some P5 learners when the numbers were insufficient- in the same schools. In addition, we 
interviewed learners’ caregivers, teachers and head teachers, and conducted a school inventory. 
The table below indicates the learners’ grades of enrollment at the time of the data collection. 
Grades of learners in NORC LARA samples 
EVALUATION EGR R&A SRGBV
Baseline (2017) P1 P1 and P4 P2, P4, and P6
Midline (2019) P3 P3 and P6 (mostly) P4 and P6 (some P5)
Finally, in October 2017 and 2018 we observed teachers’ reading instructional practices in P1 
classrooms, and at midline (2019) in P3 classrooms.
SRGBV FINDINGS
We find a very modest effect of SRGBV activities. We find that teachers in T2 schools show more 
gender-equitable attitudes than teachers in T1 schools and that this difference is statistically significant. 
There is no statistically significant difference in learner school climate indicators between T1 and T2 
schools.3 Qualitative findings also reveal several themes related to fear of going to school, feelings of 
safety to and from school, and fear of reporting violence. 
In general, the use of physical and emotional violence remains high. For the most part, there is no 
statistically significant difference in the approach to discipline used between T1 and T2 groups. However, 
there is a statistically significant reduction of six percentage points in the percent reporting taking a 
child’s privileges away as punishment in T2 vs. T1 caregivers. This is a reduction in a non-violent 
discipline approach. 
Looking specifically at corporal punishment, we find no statistically significant differences between T1 
and T2 schools at baseline or midline when caregivers, teachers, and head teachers were asked whether 
or not they believed corporal punishment was an effective form of discipline. Few teachers (around 12 
percent at midline) say that corporal punishment is an effective form of discipline, however, around 60 
percent of them report hitting learners on the buttocks with an object such as a stick, broom, cane, or 
belt; 22 percent use public humiliation and 19 percent cursed learners as a discipline method at midline. 
Qualitative data indicate similar findings, despite teachers having received trainings on alternative 
discipline methods.
The prevalence of violence is still very high and there are no statistically significant differences in the 
percentage of learners that were victims of violence between T1 and T2 schools. The majority of 
learners report suffering emotional (98 percent) and physical violence at school (94 percent) and around 
38 percent report sexual violence (these are results for T2 schools). We do see some statistically 
significant differences in the incidence of some sub-categories of violence. For example, 59 percent of 
 
3 The 4 indicators include: (i) learners are sometime afraid to go to school for fear of punishment, (ii) learners feel safe when they are at school, 
(iii) learners feel safe on the way to and from school, and (iv) learners fear reporting when someone older touches their private parts at school. 
xiv | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
learners in T1 schools report being hit in school with an object (cane, stick, belt, or book) by a teacher. 
In T2 schools, it is 52 percent and the difference is statistically significant. The frequency of this type of 
behavior is identical in T1 and T2 schools. 
The proportion of learners that report having a trusted adult to whom they can report violence is 49 
and 58 percent in T1 and T2 schools respectively. The difference between T1 and T2 schools is 
statistically significant, suggesting that LARA activities had an effect in this domain. 
Evidence of implementation of SRGBV activities in T2 schools is far from universal.The duration of the 
LARA SBCC campaign to prevent corporal punishment and promote positive discipline methods was 
too short for teachers and caregivers to fully grasp and internalize new knowledge and cultivate new 
social expectations and disapproval that would support building new social norms promoting positive 
discipline methods. 
Indicators on implementation for teachers and head teachers show higher rates of familiarity with 
Journeys materials and experience with Uganda Kids Unite groups. Over 85% of teachers and head 
teachers reported having seen Journeys materials in their schools, launching a Uganda Kids Unite group,
and participating in a Uganda Kids Unite meeting or similar activity with learners. However, only 25 
percent of learners participated in any activity related to violence prevention and only 52 
percent have seen any materials related to Journeys. Implementation seems to need attention.
RETENTION AND ATTENDANCE FINDINGS
Using the panel of learners that we followed every term since the baseline, we find statistically significant 
differences in the learners’ enrollment status by treatment group at midline among the 
Runyankore/Rukiga language area schools, but the data does not suggest significant differences between 
treatment groups in the Luganda language area. Learners in T1 and T2 Runyankore/Rukiga area schools 
tend to stay enrolled in the original school at significantly higher rates than those from control schools. 
In the Runyankore/Rukiga dominant language regions, learners in T1 schools are 
significantly less likely to transfer to other schools and learners from T2 schools are 
significantly less likely to drop out, compared to those from control schools. However, there 
are no statistically significant differences in grade repetition by treatment group in either 
region.
We measured learners’ and teachers’ absenteeism during our unannounced visits. Absenteeism remains 
high among learners and unchanged since baseline. The average attendance rate is 81 percent in the 
Luganda speaking region and 83 percent in the Runyankore/Rukiga speaking region. We do not find 
statistically significant differences when comparing attendance rates by treatment groups in each 
language area. On average, slightly over 80 percent of classrooms visited had a teacher present during 
instruction time. In the rest of the cases, the teacher was somewhere else in the school –in another 
classroom or not– or absent. We do not find significant differences between the teachers’ attendance 
status by treatment group in either region.
EARLY GRADE READING FINDINGS
At the end of 2019, LARA shows positive effects on the reading performance of cohort 2 P3 
learners in Luganda and Runyankore/Rukiga dominant language regions. The effects of the program are 
similar in each region, giving us great confidence in the findings, and, as expected, LARA had stronger 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | xv
effects in local languages than in English. We found no difference in reading performance of P3 
learners between T1 and T2 schools which indicates that SRGBV-related program 
components brought no additional benefit to learners’ reading ability. We show below the 
number of correct words per minute that the average learner reads from a short grade 2 level 
paragraph by treatment group. In both treatment groups, the learners’ performance is significantly 
better than in the control group; however, the differences between T1 and T2 are not significantly 
different. 
An improvement in reading performance was found across groups of learners with 
different reading abilities. EGRP reduced the number of zero scores among P3 learners and also 
increased the percentage of learners that reach at least 20 cwpm in the oral reading fluency subtask. 
This is an important achievement however, reading performance remains, on average, quite low 
for the treated learners by the end of P3. Learners’ decoding skills are low; on average, learners in 
T1 and T2 schools can identify less than 14 correct letter sounds in a minute. Reading ability is low as 
well. Over a quarter of P3 learners benefitted by LARA are non-readers – they cannot read a single 
word from a short grade 2 level paragraph – and on average, oral reading fluency is around 17 words 
per minute, which is far from the levels needed to comprehend the text. 
Oral Reading Fluency in local language – P3 Learners
Nevertheless, the improvement due to LARA EGR activities is significant. The progress achieved by the 
program is a good base upon which to build. With that aim, we explored in detail the different 
components of LARA ERG activities to identify what works well and what needs to be improved. 
The strengths of the LARA program are evident when comparing instructional reading practices 
between treatment and control schools. Many, although not all, treatment classrooms implemented the 
program on the day of classroom observation. Having LARA reading books in the lessons supported a 
greater engagement by learners with text, and more opportunities to read an extended text. Learners in 
control classrooms and treatment classrooms not implementing the program read no extended text, but 
rather engaged in choral recitation of single words or short text written on the blackboard. LARA also 
xvi | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
offered opportunities to engage with literacy-specific skills and content around the phonetic, semantic,
and syntactic aspects of language. Lessons tend to focus on the letter, syllable, word, sentence, and 
extended text levels, although around half of the observed teachers completed 50 percent or less of the 
lesson plan for the day. The more challenging aspects of the program were left out. P3 teachers are 
more likely to follow aspects of the program that closely resembled the structure of P1 and P2 lessons 
than those that require additional and more challenging work on vocabulary and language structure. 
Lessons lack opportunities for learners to read extended texts, and particularly to read silently and 
independently.
The fact that these aspects were often not exploited does not diminish the importance of how the 
program explicitly instructs learners how to read. In control and treatment classrooms not 
implementing the program, instruction focused on content topics (health, culture, gender), and the 
informational content relevant to these topics was emphasized rather than literacy-specific skills and 
content. 
Neither CAM forms nor alternative methods are used by P3 teachers to conduct a 
continuous assessment of learners. Sometimes, teachers listen to individual learners read, but their 
feedback is quite restricted. One of the crucial methods of reading practice and assessment (I do, we do, 
you do) potentially contributed to a very repetitive, chorused discourse, empty of evaluative potential. 
The above observations suggest that further work is needed during teaching, training, and support 
supervision visits to instill rich, engaging, and motivating language learning in the classroom, which is the 
foundation for developing literacy. Support supervision visits are less frequent than they should
be. In particular, support supervision by CCTs and district education officers is very low. 
This seems to be a consequence of lack of time, lack of means to reach the schools, and competing 
responsibilities. Head teachers are not engaged in support supervision either. This creates an important 
challenge that was already present in USAID/SHRP – the activity that preceded LARA. If support 
supervision is not properly and credibly embedded in the education system, it will not be 
sustainable and most likely will disappear once the LARA ends. 
LARA distributed plenty of reading books among treatment schools that were very well-received. 
However, in many classes, reading books are not in the hand of learners as they should. This seems to 
be the result of schools’ policies that keep learner’s books in storage perhaps to avoid them being
damaged, undermining the full potential of the program. The lack of opportunities that learners have to 
handle books and experience individual reading is reinforced by the lack of reading materials at home. 
Supplementary reading materials and reading cards are scarce and the lending system for the few 
materials available does not work well. It is hard to overestimate the importance of having access to a 
quality, print text in the development of children’s literacy and learning and in the creation of positive 
behaviors and habits that support reading development.
RECOMMENDATIONS
SCHOOL-RELATED GENDER-BASED VIOLENCE
Improve the school climate: Making the school climate safer for learners is within the domain of head 
teachers and teachers’ responsibilities. More can be done here in two key areas. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | xvii
 School infrastructure: Many learners continue to feel unsafe in school due to a lack of boundary wall 
and latrines that lack locks on the doors. Future programs on SRGBV that include funding for a 
boundary wall may result in quick payoffs. At a minimum, it is easy to insist that school latrines have 
functioning locks to ensure program support. 
 Change head teacher and teacher behavior: Train all teachers on positive, non-physical disciplinary 
methods, SRGBV prevention and response, not victim-blaming survivors of violence, and effective, 
non-re-victimizing communication through guidance and counseling. Going beyond an SBCC 
approach, start with teacher discussion groups to reflect on gender norms, as well as school and 
community expectations for teachers’ behaviors that underpin SRGBV. Groups can work toward 
developing a shared understanding of the negative effects of SRGBV on child development and 
academic achievement as well as new, positive norms that define a safe, supportive, nurturing, and 
reliable teacher. A SBCC campaign that sends a clear message that it is acceptable to talk with 
children respectfully, use non-violent, positive disciplinary methods, and not use harsh physical or 
psychological punishment could be embedded within a norm change approach. On-going and in￾service training for teachers is required beyond the limited implementation of one program; the 
cascading model of training does not seem to be working.
Include SRGBV prevention activities as part of the school hours, rather than having SRGBV as an 
extracurricular activity. Undertake and test a pilot that changes the approach towards SRGBV. For 
example, all grades P1-P7 could have one hour of SRGBV prevention per week.
SRGBV encompasses a large and complicated range of issues to address as one of the multiple program 
components. Future programs should have SRGBV as an explicit focus, with activities that address the 
problem from several angles directly, including social and gender norm change strategies at a peer 
reference group level, and school and community levels specifically addressing norms that underpin and 
perpetuate SRGBV to evaluate how the intervention affects students’ exposure to SRGBV. Journeys 
focuses largely on socio and emotional learning with SRGBV interspersed throughout; a more targeted 
and streamlined message with more time to learn the material is critical. 
Revise the program that relies on district support supervision. District officials do not have the time and 
resources to visit the schools for support supervision. Given that the current system is not working and 
there are challenges to changing it, it would be best to empower the head teachers so they can provide 
the necessary support supervision within their schools.
Work at the national level with MoES: Ministry officials understand the importance and seem open to 
the idea of including SRGBV prevention and response in their primary teacher college curriculum that 
focuses on teacher development and management. This is important and will have a long-term impact on 
reducing SRGBV. Additionally, explore working with MoES to change the current Reporting, Tracking, 
Referral, and Response process whereby only head teachers report cases of SRGBV cases to the LC or 
district. School teachers being allowed to report incidents can help prevent head teachers from 
becoming gatekeepers and interfering with child violence survivors seeking justice. 
EARLY GRADE READING
Teacher training and support: The high level of implementation fidelity in treatment schools potentially 
offers a good base on which to build, to extend teachers beyond excessive repetition, and to encourage 
more learner speaking, oral language development, and engagement with the text’s meaning. We 
xviii | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
recommend the inclusion of more discussion and demonstration of these activities during training and 
during support visits. In particular, conducting demonstrations in the actual classroom could be very 
valuable for teachers.
Some MoES officials, based on their belief that the EGR program is a success, are pushing it to be fully 
integrated into pre-service training offered at Primary Teacher Colleges (PTCs) and Universities. This 
type of action is promising and should be carefully considered in future programs.
Improve teacher support supervision: A larger fraction of teachers in treatment schools received more 
frequent support supervision than those in control schools; however, there are still many teachers that 
do not receive supervision at all or do not receive it frequently enough to make it useful. In particular, 
support supervision by head teachers, CCTs, and district education officers is very low. Evidence 
suggests that including follow-up classroom visits and teacher support increases learning gains (see the
2018 World Development Report). We recommend exploring this challenge and focusing on how to 
effectively scale support supervision within the education system to ensure the sustainability of the 
program, given that the current approach is not working.
Continuous Assessment. LARA or future programs needs to revise its approach to training teachers in 
conducting a continuous assessment of learners. CAM forms are not being used, teachers are not 
creating alternatives, and even oral feedback to learners seems insufficient.
Putting reading books in learners’ hands. Additional work and training needs to be done with head 
teachers and teachers to persuade them that reading books and supplemental materials are only useful if 
they are in the hands of the children. This idea should be reinforced during support supervision visits. 
Reading materials to take home are insufficient and schools are reluctant to lend them to the children. In 
the future, it would be worth considering alternatives to create inexpensive products, for example,
newsprint materials, which even if not designed to last years, can be given to children to read at home 
or in school.
POLICY BRIEFS
The last section of this report includes four policy briefs that summarize the P&IE’s findings and 
recommendations regarding the following topics:
 The High Absenteeism of Teachers and Learners in Uganda
 Why Lule still can’t read at the end of P3?
 Teachers’ implementation of LARA in classrooms
 Ending Teacher-to-Learner Violence as an Entry-point to Addressing School-related Gender￾based Violence
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 1
1 INTRODUCTION
1.1 EVALUATION PURPOSE
NORC at the University of Chicago (NORC), in partnership with subcontractor Panagora Group, 
serves as the independent evaluator for the Performance and Impact Evaluation (P&IE) of the Literacy 
Achievement and Retention Activity (LARA), implemented by RTI International (RTI), in Uganda. LARA 
is a 6-year (April 2015‒April 2021)
4 USAID-funded activity designed to support the Ugandan Ministry of 
Education and Sports (MoES) in its efforts to improve early grade reading (EGR) and retention.
LARA concentrates on two results: Result 1 (R1) improved capacity to deliver early grade reading in 
three local languages5 and English; and Result 2 (R2) improved retention in early primary grades, through 
the reduction of school-related gender-based violence (SRGBV) resulting in safer school environments. 
Both of these results are intended to contribute to the overall objective of improving reading skills.
The Uganda LARA P&IE activity (April 2016 - April 2021) has two objectives:
 To assess the impact of LARA on learners’ literacy skills and retention rates
 To assess the performance of LARA in terms of project management, learning, design, 
implementation, results, and sustainability.
 There are two clusters of schools where LARA activities take place. Cluster 1 (C1) schools 
started receiving the intervention in 2015-2016 prior to the start of this evaluation. The focus 
of NORC’s evaluation is Cluster 2 (C2) schools that LARA only started reaching in 2017 after 
NORC had collected baseline information for both EGR and SRGBV in term 1 of 2017. 
NORC’s LARA P&IE covers C2 schools located in areas where two languages dominate: 
Luganda and Runyankore/Rukiga.
1.2 EVALUATION COMPONENTS
The LARA P&IE consists of two components: (1) an Impact Evaluation (IE), and (2) a Performance 
Evaluation (PE). The IE addresses the following questions:
 What is the impact of R1 activities on reading performance and retention rates?
 What is the additional impact of R2 activities on reading performance and what is the impact 
of R2 activities on SRGBV intermediate outcomes?
 What is the total impact that R1 + R2 activities have on reading performance and retention 
rates?
 The IE also estimates the effect of the R1 and R2 activities on learner attendance, a major 
predictor of dropout and grade repetition. 
 
4
LARA was originally conceived as a 5-year activity, but it was extended for a sixth year.
5 The three local languages are Luganda, Lunyankore/Rukiga and Runyoro-Rutooro. Since there was only one district with three CCTs that 
predominately used the Runyoro-Rutooro languages, and therefore insufficient sample size, NORCs evaluation is restricted to the 12 districts 
dominated by the Luganda and Runyankore/Rukiga languages.
2 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
The purpose of the PE is to provide insights into LARA program implementation and the IE results.
Thus, the PE focuses on the following questions:
 Is LARA on track to achieve results by the end of the project?
o What are the key outcomes of the project?
o Were activities carried out as planned? Adaptions by LARA that were not in their 
original design are noted in italics. 
o Are activities sufficient and relevant to achieving Result 1 and Result 2?
 What factors accelerated or inhibited the achievement of LARA results?
The LARA P&IE uses a mixed-methods approach combining a randomized controlled trial (RCT) design 
and qualitative methods. Randomization of treatment assignment was conducted at the Coordinating 
Center Tutors (CCT) level, assigning all schools under a CCT to treatment T1 (receiving R1 EGR 
activities only), treatment T2 (receiving R1 EGR and R2 SRGBV activities), or the control group 
(receiving no activities). 
This report summarizes the findings of the midterm IE and the PE. The midterm IE quantitative and 
qualitative data was gathered between July and October 2019, while the PE qualitative data was 
collected in March 2020.
1.3 EVALUATION REPORT STRUCTURE
Section 2 of this report provides the background on LARA. We present the program objectives and 
activities as originally designed and planned. Changes in implementation and adaptations are noted 
where appropriate when we describe activity outcomes and examine the fidelity of implementation. 
Section 3 presents the evaluation design including the methodology, sample, instruments, and limitations. 
In Section 4, we present the combined findings from the impact and performance evaluation in three 
sections: section 4.1 is the effect of R2 activities on SRGBV intermediate outcomes, section 4.2 presents 
the retention and attendance findings, and section 4.3 the early grade reading proficiency findings (this 
also includes findings from the classroom observations). Section 5 includes our conclusions for SRGBV 
followed by EGR. In the final Section 6, we suggest recommendations. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 3
2 LARA BACKGROUND
2.1 PROGRAM OBJECTIVES
LARA aims to improve the reading skills of 1.3 million primary-grade learners in government schools in 
31 districts throughout Uganda. LARA has two main goals: 
 R1 (Increased capacity to deliver early grade reading) focuses on strengthening the capacity of 
MoES and other educational stakeholders to deliver EGR in three local languages (Luganda, 
Runyankore/Rukiga, and Runyoro/Rutooro) in Primary 1 (P1) to Primary 3 (P3) with a 
transition to English in Primary 4 (P4).
 R2 (Improved retention in primary grades) endeavors to further improve children’s retention
and active participation in early primary grades (P1–P4) through the reduction of SRGBV,
resulting in the establishment of a positive and supportive school climate for learning.
 All LARA activities are intended to work towards Result 1 and/or Result 2. Both Results 1 and 
2 use a phased approach, starting with 16 C1 districts during Year 1 and continuing with 15 C2 
districts in Year 2. NORC P&IE focuses on C2 districts. Thus, unless specified, all program 
activities and outcomes reported are focused on C2 districts. 
2.2 PROGRAM ACTIVITIES
R1. INCREASED CAPACITY TO DELIVER EARLY GRADE READING
LARA is expanding the EGR methodology that was implemented by the USAID/Uganda School Health 
and Reading Program (SHRP)6. The R1 main activities of LARA in C2 districts are:
 Training: Training of teachers in EGR methodology to strengthen their reading instruction 
skills followed by refresher trainings. As part of the training program, teachers receive a 
teaching guide and pupil books. The C2 EGR training is detailed below in Table 1.
Table 1. Training on Early Grade Reading
TRAINING CLASS DAYS REFRESHER 
TRAINING
REFRESHER TRAINING 
CONTENT
January 2017 P1 5 Feb/March 2019; 1 
day
Components, Literature hour, 
Oral literacy
January 2017 P2 5 Feb/March 2019; 1 
day
Components, Literature hour, 
Oral literacy
January 2018 P3 5 Term 2 2019; 1 day Components, English, Literature 
2 (writing)
January 2019
May 2019
September 2019
P4 3 (English)
2 (Local Lang)
2 (English)
Term 2 2019; 1 day Components, Transition, 
Writing
 
6 The School Health and Reading Program (SHRP) had two sets of intervention: An Early Grade Reading intervention targeted at learners 
Grades 1-3, and an HIV/AIDS intervention targeted at upper primary and secondary school learners.
4 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
 LARA uses a training of trainers (ToT) approach, which starts with training lead facilitators and 
conducting training sessions for individuals who would provide training to classroom teachers. 
Trainers were identified among former teachers, CCTs7 , and district education staff. Over 
time, LARA has identified ways to improve the ToT and included more scripted training 
guides. In Year 3, LARA adapted the EGR training program by refining the content and changing the 
frequency and duration of training from five continuous days to a 3-2-2 day program to allow time to 
absorb and practice the material.
 Continuous Assessment Monitoring (CAM): CAM is a process to collect evidence about 
the learner’s acquired reading skills. It is not a test that produces a score or grade that can be 
shared with caregivers. The teacher guides describe how to use CAM forms to assess the
learners’ reading performance across multiple competencies. LARA does not deliver CAM 
forms to teachers, but they can photocopy or create their own based on the examples 
provided in the teacher guides.
 Teaching and learning materials: LARA distributed teacher guides and P1 to P4 pupil 
reading books (also known as primers). The teacher guides and pupil reading books cover the 
reading and writing hours for every week of the full academic year. 
 Originally developed by SHRP, the pupil books are written in English and local languages, 
reflect the cultural diversity of the regions in Uganda, and progress through the thematic 
curriculum. Early on, there were challenges with book distribution, but most of those were resolved 
when district offices and head teachers were given responsibility for distribution as a cost reduction 
measure by LARA. LARA’s goal, consistent with MoES’ policies, was to put a reading book into
the hands of each learner. LARA provided training for head teachers on book care and 
storage, and each school received a padlock and metal box where books are to be returned at 
the end of each school day after being inventoried. The reading books distributed in 2016 
were supposed to be replaced in 2019 by the MOES because they have a 3-year life span, but 
this has not happened yet.
 Support Supervision: LARA Field Assistants (FAs), who are former teachers, provide 
supervisory support and mentorship to primary teachers trained in EGR. They are expected to 
visit teachers in their classrooms to observe EGR lessons and provide instructional support 
once per school term (three times a year in total). The last time that FAs were expected to
provide support and observe the degree to which teachers demonstrated the competencies 
learned in the EGR training was in 2017 for P1 and P2 teachers, 2018 for P3 teachers, and 
2019 for P4 teachers. During their visits to P1 and P2 classes, FAs also assessed the reading 
skills of three randomly selected learners through “learner checks.”
 Finally, FAs also built the skills of MoEs officials, district officials, and CCTs for district-led 
supportive supervision. Initially, FAs joined district teams to conduct support supervision to 
accelerate the reach and frequency of school support in supporting teachers’ use of EGR 
materials and methodology. By Year 4, 14 out of 30 districts were leading and financing their 
 
7 CCTs are school support workers in charge of monitoring education quality. Each CCT is responsible for a certain number of schools within 
a district (one district typically has multiple CCTs).
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 5
own support supervision exercise to monitor progress, performance, and implementation of 
EGR. 
 LARA has also trained head teachers and zonal head teachers on teaching support supervision. 
 Uganda Learning/Literacy Campaign (ULC): The ULC, implemented in local languages 
between February and April 2018, used a social and behavior change communications (SBCC) 
approach to encourage caregivers to read with their children. LARA provided subgrants to 
community-based organizations (CBOs) to implement SBCC campaigns that included at least 
20 drama shows in each C2 district, as well as home visits with caregivers. LARA also trained 
field teams on interpersonal communication across four thematic areas of supporting 
children’s reading: the value of education, the role of the caregiver, creating time for reading, 
and the caregiver-child relationship. The goal of these trainings was to more effectively engage 
with caregivers during home visits. At the national level, it consisted of two television 
commercials (running daily for two weeks), one radio program (running weekly for one 
month), and four radio advertisements (running daily for one month).
 Supplemental reading materials The ULC was complemented by distributing story cards
and supplemental reading books in English to C1 and C2 schools. The story cards are one 
page, double-sided, and laminated for durability. LARA developed eight different story cards 
that are grade-level appropriate: four for P1 and P2, and four for P3 and P4. Field teams of 
trained interpersonal communicators met with caregivers to teach them how to use the story 
cards, even if the caregiver is not literate. The story cards were intended for pupils to take 
home to read to/with their caregivers. LARA trained teachers and head teachers on a lending 
protocol of supplemental reading materials.
R2. IMPROVED RETENTION IN PRIMARY GRADES
R2 is implemented in close collaboration and partnership with the MoES and the Ministry of Gender, 
Labor, and Social Development (MGLSD). According to the LARA Theory of Change, the secondary 
program final result, “Result 2,” envisioned a long-term outcome, “improved retention in primary 
grades.” This outcome will be pursued through medium-term capacity-building within the government 
school system and communities to create, “positive and supportive learning environments” (LARA 
AMELP Year Four). The program aimed to work on multiple levels, both school- and community-wide, 
utilizing a synergetic approach to training community leaders and members as well as education sector 
staff at national, district, and school levels. LARA sought to strengthen the education system and 
community capacities to cultivate safe and enabling environments for primary grade children to attend 
school, advance from one grade to the next, avoid dropping out, and thereby improve their early grade 
reading skills. The main activities of LARA’s R2 in C2 districts are:
 Journeys training: The Journeys curriculum is implemented using three activity handbooks, 
each aimed at a target population for SRGBV prevention and to strengthen the process of 
recognition, response, reporting, and referral. The three activity handbooks include the 
content shown in Table 2. Note that the Journeys for Pupils is actually targeted toward
teachers to use when working with learners in the school on SRGBV issues. Table 3. SRGBV 
Journeys Training days per target population.
6 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Table 2. Contents of Journeys Activity Handbooks
FOR TEACHERS AND 
SCHOOL STAFF
FOR PUPILS (TARGETED / 
GEARED TO TEACHERS)
FOR COMMUNITIES
Introduction; Guidelines; the 
Journeys Five-Step process for 
inspiring change; Community 
based management; Planning 
and evaluation forms.
Activity Themes:
Understanding Positive and 
Supportive Schools
Barriers to Positive and 
supportive Schools
Child Protection Basics
Violence Against Children in 
Schools
Response to Violence Against 
Children in Schools
Annexes: Information briefs 
(SRGBV definitions; Positive 
Discipline Responses: 
Alternatives to Corporal 
Punishment; Discipline versus 
Punishment; Ten Keys to Safer 
Schools Strategies for Improving 
School Climate); Uganda 
Teacher’s Code of Conduct; 
Uganda Children’s Act; 
Glossary; Bibliography; Sources 
of Activities.
Introduction: Guidelines; 
Registration Form; Planning and 
Evaluation Forms and 
Attendance Registers
Activity Themes:
Knowing Myself, My Friends, 
and My School (activities 1-13)
Building Positive Relationships, 
Understanding and Solving 
Social Challenges (activities 14-
26)
Making Decisions When Facing 
Violence; Being a Good Friend 
(activities 17-39)
Annexes: The Journey’s Five￾Step Process for Inspiring 
Change; Definitions of SRGBV; 
Positive Discipline Responses: 
Alternatives to Corporal 
Punishment; Discipline versus 
Punishment; Ten Keys to Safer 
Schools: Strategies for 
Improving School Climate; 
Government of Uganda’ 
Teachers Code of Conduct; 
Government of Uganda 
Children’s Act; Social and 
Emotional Learning 
Competencies; Matrix of 
Activities by Social and 
Emotional Learning 
Competencies; Matrix of 
Activities by Journeys Goals; 
Matrix of Activities by Thematic 
Life Skills and Values; Glossary; 
Bibliography; Sources of 
Activities 
Introduction; Guidelines on 
Using this Handbook; The 
Journeys Five-Step Process for 
Inspiring Change; Community 
based Case Management; 
Planning and Evaluation Forms
Activity Themes:
Understanding Safe and Caring 
Schools and Communities 
(activities 1.1-1.7)
Barriers to Positive and 
Supportive Schools (activities 
2.1-2.6)
Child Protection Basics 
(activities 3.1-3.2)
Violence Against Children in 
Schools (activities 4.1-4.7)
Response to Violence Against 
Children in Schools (activities 
5.1-5.9)
Annexes: Information Briefs 
(SRGBV definitions; Positive 
Discipline Responses: 
Alternatives to Corporal 
Punishment; Discipline versus 
Punishment; Ten Keys to Safer 
School Strategies for Improving 
School Climate); Uganda 
Children’s Act; Glossary; 
Bibliography
 The SRGBV training aimed at delivering Journeys in cohort 2 districts is detailed below.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 7
Table 3. SRGBV Journeys Training
 Following a ToT model, CCTs and Community Development Officers (CDOs)—supported by 
LARA FAs, trainers, and Regional Program Officers—built the capacity of school change agents 
(SCAs, including head teachers and teachers) to improve school climate, and to respond to 
and prevent SRGBV. The SCAs are in charge of training all school personnel on the knowledge 
foundations of school climate, and facilitating dialogue on gender norms and power relations 
that produce and perpetuate SRGBV. The FAs conducted once-per-term support visits to 
SCAs in Years 2 and 3, and were trained on supporting teacher patrons and SCAs to conduct 
Journeys dialogue in 2019.
 Uganda Kids Unite (UKU): Teacher patrons support the establishment and implementation 
of UKU groups, using the Journey’s for Pupils handbook. UKU groups, initiated in Year 3, 
allow primary school learners to develop an understanding of SRGBV, as well as develop their 
socio-emotional learning and individual and collective sense of agency to avoid, challenge, and 
prevent SRGBV. UKU lead facilitators were trained in December 2017, and the UKU Teacher 
Patron training was conducted in 2018. The plans included training 10 teachers per school in 
all T2 schools (which includes 2 SCAs) on Journeys for Pupils. 
 Community subgrants: Provide grants to community-based organizations (CBOs) to train 
and support community change agents (CCAs) to work together to contribute to building a 
safe and caring school climate and catalyze community SRGBV response and prevention 
initiatives. To this end, in Year 3, LARA distributed Journeys for Communities handbooks to 
subgrantees and communities. The program also trained subgrantees on grants management, 
the Journeys approach, and Reporting Tracking Referral and Response (RTRR). The CCAs 
were trained in the Journeys program for communities in 2018 and expected to cover 142 
schools across three R2 implementing school districts. In Years 3 and 4, subgrantees mapped 
referral networks in their districts, and established Village Child Case Management 
Committees (VCCMCs) to receive, refer, and track progress on SRGBV cases of violence
against pupils reported to them. Uganda’s 2014 “Reporting, Tracking, Referral, and Response” 
(RTRR) guidelines were to be operationalized through VCCMCs. Each VCCMC has around 
seven members, a CCA, a school change agent, a head teacher, a Community Development 
YEAR No. OF DAYS TARGET 
2017 5 2 School Change Agents (SCAs) (1 male; 1 female)
2017 2 Head Teacher (HTs)
2018 2 Refresher training for SCAs and HTs
2018 1 Cluster learning event for SCAs and community change agents (CCAs) 
to discuss what is working or not
2018 N/A CCAs (caregivers, para social workers, community leaders) get trained 
in Journeys for the community
2019 N/A 10 teachers were targeted for training as teacher patrons to facilitate 
Journeys for Pupils activities through UKU clubs.
8 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Officer, a representative of a Village Health Team, the Local Council One representative, and 
religious leaders.8
 SRGBV-related SBCC: Support a targeted SBCC initiative that builds widespread 
awareness of the attitudes, gender norms, and power relations that produce and maintain 
SRGBV in an effort to equalize these power dynamics. This initiative also promotes actions 
that prevent SRGBV. The SBCC pilot activities were implemented in Year 3. RTI partnered 
with a Ugandan communications organization to develop the SBCC campaign for the 
prevention of corporal punishment in communities and schools. Community-based 
organizations as LARA subgrantees implemented the SBCC campaign rollout in schools in 15 
districts between June and August 2019. 
 National support: Contribute to implementation of Uganda’s National Child Policy (2017–
2022) to make ending violence against children a country-wide priority. LARA has also worked 
to strengthen MoES knowledge, commitment, and action in SRGBV prevention. Key activities 
under this component include:
o Technical support to MoES Violence Against Children in Schools (VACiS) working group 
and quarterly meetings – started in Y2
o Engaging MoES, MoGLSD, foundational bodies to end violence against children; training 
of trainers in SRGBV and Journeys events on the national and district levels; and 
providing Journeys material to district duty bearers
o Support to MoES on child protection policies and follow-up
 
8 More details regarding VCCMCs are found in Annex C. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 9
3 EVALUATION DESIGN
3.1 METHODOLOGY
The LARA P&IE uses a mix of mutually reinforcing qualitative and quantitative methods that reflect the 
program design, research questions, and indicators. NORC combines the results of each technique to 
capture the diversity of opinions and perceptions of both beneficiaries and stakeholders. The qualitative 
data supplements and enriches quantitative findings by addressing research questions not well-suited to 
quantitative analysis. In particular, when evaluating the effect of SRGBV activities on retention and 
different forms of violence prevalence, we explore possible causal pathways through which the 
intervention is operating, and how different contextual factors may affect outcomes.
The IE combines an RCT design and qualitative methods. Since each CCT is responsible for many 
schools within a district, randomizing at the school level would have required a CCT to treat schools 
under their jurisdiction differently if some were treatment schools and others controls, leading to a high 
risk of 'contamination' between treatment and control groups. To avoid this problem, randomization 
was done at the CCT level, assigning all schools under a CCT to treatment T1 (receiving R1 EGR 
activities only), treatment T2 (receiving R1 EGR and R2 SRGBV activities), or the control group 
(receiving no activities). Within each CCT, schools were randomly selected for data collection at 
baseline (in 2017). The same schools were visited in 2019 for the midline data collection, with EGR and 
SRGBV data collected between July and October 2019, respectively. Figure 1. Selection of Treatment 
and Control Groups illustrates the selection of treatment and control groups process.
Figure 1. Selection of Treatment and Control Groups
The IE also estimates the effect of the R1 and R2 activities on learner attendance, a major predictor of 
dropout and grade repetition. Therefore, every school term – three per year – NORC checks the 
retention and attendance (R&A) of just over 3,500 learners who were selected at baseline from 71 
10 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
schools. Furthermore, we conduct classroom observations (CROs) in term 3 of each school year to 
assess the fidelity of implementation of the EGR activities and compare teaching practices of 30
treatment and control schools. This report includes the results of CROs conducted in October of 2017, 
2018, and 2019 during unannounced visits.
The goal of the PE is to provide insights into LARA program implementation and the IE results, so an 
extensive review of LARA documentation was conducted to obtain information to answer the key 
evaluation questions and to determine what additional data was required to address these questions or 
help explain the quantitative findings of the IE. Key Informant Interviews (KIIs) with stakeholders at the 
national, district, and school levels were conducted in March 2020.
Table 4 lists all data collection efforts conducted by the LARA P&IE team thus far, between 2017 and the 
first school term of 2020.
Table 4. NORC LARA P&IE Data Collection Tasks (2017 – 2020)
YEAR EGR SRGBV R&A CRO PE
2017 Term 1: 
Baseline
Term 1: 
Baseline
Term 1: Round 0 (Sample creation)
Term 2: Round 1
Term 3: Round 2
Term 3: 
Round 1
2018 Term 1: Round 3
Term 2: Round 4
Term 3: Round 5
Term 3: 
Round 2
2019 Term 3: 
Midline
Term 2: 
Midline
Term 1: Round 6
Term 2: Round 7
Term 3: Round 8
Term 3: 
Round 3
2020 Term 1: Round 9 Term 1: Final PE
3.2 SAMPLE
The random sample selected for EGR includes schools in C2 districts dominated by the Luganda and 
Runyankore/Rukiga languages in the sample frame. The school sample drawn for SRGBV data collection 
(a subset of EGR schools) spans 11 of the 12 intended districts—a fact that does not affect the 
representativeness of the sample. Table 5 includes the number of schools sampled for data collection in 
each district.
Table 5. Number of Schools in Each Sample, by Language 
LANGUAGE EGR SRGBV R&A CRO* PE*
Luganda 132 40 35 15 2
Runyankore/Rukiga 132 40 36 16 2
Total 264 80 71 31 4
* The schools selected for the CROs and PE are not intended to represent the universe of C2 schools in the Luganda and Runyankore/Rukiga 
language regions.
The NORC evaluation team selected schools for the EGR, SRGBV, and retention and attendance (R&A) 
samples using a sample frame of all government primary schools, including information on the district, 
coordinating center, language, and LARA treatment status. The schools selected for the EGR and R&A 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 11
sample are stratified by dominant language (Luganda or Runyankore/Rukiga) and treatment status (T1, 
T2, or control). For EGR surveys, 44 schools were selected in each arm for a total of 264 schools. At 
baseline (2017), 20 P1 learners were randomly selected in each school to be assessed. At midline (2019), 
we randomly selected 20 P3 learners from each of the same schools. As part of the EGR sample, we
also interviewed the teacher in charge of the grade and the head teacher.
For the R&A study, a total of 71 schools were selected – twelve schools in each arm, and eleven in the 
control. At baseline we created a random sample of 30 learners enrolled in P1 and 30 learners enrolled 
in P4. We followed them every school term. At midline most of them were enrolled in P3 and P6. 
The sample for SRGBV survey-related data includes only schools in the T1 or T2 groups, given that the 
relevant comparison is between schools that received the EGR activities and schools that received the 
EGR plus the SRGBV activities. The sampled schools are stratified by language and treatment status with 
each of the four arms including 20 schools—80 schools total. However, schools in both language regions 
can be pooled together during analysis, in contrast to EGR. At baseline, we interviewed 20 learners 
enrolled in P2, 20 in P4, and 20 in P6 in each school. At midline, 20 P4 and 20 P6 learners - and some P5 
learners when the numbers were insufficient – were interviewed in the same schools. In addition, we 
interviewed learners’ caregivers, teachers, and head teachers, and conducted a school inventory. 
Table 6 below indicates the learners’ enrollment grades at the time of data collection. 
Table 6. Grades of learners in NORC LARA samples 
TIME OF DATA 
COLLECTION
EGR R&A SRGBV
Baseline (2017) P1 P1 and P4 P2, P4 and P6
Midline (2019) P3 P3 and P6 (mostly) P4 and P6 (and some P5)*
Note: * When there were not enough P6 learners in the school we added learners from P5.
Finally, in October 2017 and 2018 we observed teachers’ reading instructional practices in P1 
classrooms, and at midline (2019) in P3 classrooms.
ANALYTIC SAMPLE SIZES
Table 7 displays the number of completed interviews for each survey type, in addition to the gender and 
local language distributions. For both data collections, approximately half the learners interviewed were 
female, and more than 50% of teachers for EGR data collection and caregivers for SRGBV data 
collection were female. Additionally, 28% of SRGBV teachers and 30% of head teachers for both EGR 
and SRGBV data collections were female. The samples for all surveys were nearly balanced between 
respondents from Runyankore/Rukiga and Luganda-dominant schools. 
Table 7. Sample Characteristics
SURVEY NUMBER PERCENT FEMALE PERCENT LUGANDA
EGRA
Learner 4,927 50% 48%
Schools surveyed 264 -- 50%
Teacher 230 67% 52%
Head Teacher 250 30% 48%
12 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
SURVEY NUMBER PERCENT FEMALE PERCENT LUGANDA
SRGBV
Learner 2,904 51% 48%
Primary Caregiver 613 59% 50%
Teacher 151 28% 48%
Head Teacher 80 28% 50%
School Observations 80 -- 50%
Table 8 shows that all 30 planned focus group discussions (FGDs) were completed. The SRGBV FGDs 
were conducted separately by gender, and half of FGDs were conducted with female learners, primary 
caregivers, and senior teachers while the other half with male learners, primary caregivers, and senior 
teachers. 
Table 8. Focus Group Discussions for EGR and SRGBV 
DOMINANT 
LANGUAGE
SRGBV EGR
LEARNERS PRIMARY CAREGIVERS TEACHERS PRIMARY 
CAREGIVERS
Luganda 2 with girls
2 with boys
2 with male caregivers
2 with female caregivers
2 with senior male 
teachers
2 with senior 
female teachers
3
Runyankore/Rukiga 2 with girls
2 with boys
2 with male caregivers
2 with female caregivers
2 with senior male 
teachers
2 with senior male 
teachers
3
TOTAL 8 8 8 6
Finally, we used purposive qualitative range sampling to select 48 respondents for the PE KIIs, including 
informants at the national, district, and school level. At the central level, representatives of 
USAID/Uganda and RTI, as well as MoEs officials in a broad spectrum of directorates, departments, and 
units were interviewed. In the districts of Isingiro, Bukomansimbi, and Masaka, we conducted KIIs with 
education officers, inspectors, tutors, probation and social welfare officers, and child and family 
protection police unit officers. In each language region, 2 T2 schools were visited, and KIIs were 
conducted with the head teacher or deputy head teacher, as well as classroom teachers for P2 and P3 
to discuss EGR, and P4 and P6 teachers to discuss SRGBV. We also interviewed RTI’s subcontractors 
and subgrantees and held meetings with other major players in the EGR and SRGBV spaces. 
3.3 INSTRUMENTS 
A brief summary of instrument development and the final design is presented below. 
EARLY GRADE READING DATA COLLECTION TOOLS 
Early Grade Reading Assessment (EGRA) – The instrument was fielded in the dominant local languages 
(Luganda and Runyankore/Rukiga) as well as English. The local language instruments included subtasks on 
Orientation to Print, Letter Sound Knowledge, Segmenting, Non-word Decoding, Oral Passage Reading, 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 13
Oral Recall and Listening Comprehension. The English EGRA included subtasks on Letter Sound 
Knowledge, Oral Passage Reading, Listening Comprehension and Receptive Vocabulary. Details on each 
subtask can be found in the EGRA instruments in Annex D.
Learner Context Survey – This instrument collects basic information on the learner to complement the 
EGRA. We gather basic demographic information (age, sex, and language), learner’s living arrangements, 
assets in the home, and home literacy environment. The survey also includes questions on school 
attendance for both the learner and his or her teacher. 
Head Teacher Survey – The head teacher survey gathers information from head teachers regarding their 
instructional leadership – including their training and educational background and their support and 
supervision of reading instructional practices in the lower grades. 
Teacher Survey – This survey collects information on the teachers’ demographic characteristics, 
education, experience, and in-service training. The questionnaire also includes questions about support 
supervision received, availability, use, and opinion of teaching materials, and absenteeism. At the end of 
the survey, enumerators counted the number of learners and the availability of reading books in the 
classrooms.
Classroom Observation Tool – To collect both process and input data, NORC designed a tool that 
includes both closed-ended items and open-ended narrative descriptions of the classroom activity. In 
order to gain a ‘thicker’ description in the narrative record, two fieldworkers each produced a 
description of the same lesson. The two descriptions at the point of analysis were then read together. In 
addition, the closed-ended part of the tool was completed after the lesson by both fieldworkers. In this 
way, judgments required in the closed-ended items were subjected to a form of inter-rater reliability at 
the point of data collection. Fieldworkers could also refer back to the written lesson narratives in 
justifying their judgments between each other. The mixed-method approach, in summary, was used to 
obtain a more complete understanding of what was going on in the classrooms.
Caregiver FGD – The protocol focused on a few key questions and sub-questions that are discussed in 
greater detail. The questions prompted caregivers to share who they think is primarily responsible for 
teaching their child to read, and the role of the teacher versus that of the family. It also focuses on the 
caregivers’ confidence and challenges in helping their child learn to read; their child’s access to reading 
materials, and factors affecting their child’s absence from class and their reading skills.
SRGBV DATA COLLECTION TOOLS
Learner Survey – NORC and Panagora developed a survey instrument that adapts the measurements of 
SRGBV experiences depending on gender. The instrument consists of sections on (1) demographic 
information, (2) the learners’ perspectives on school climate, (3) the learners’ general attitudes towards 
gender norms, and (4) personal experiences with SRGBV. Finally, depending on incidents disclosed 
during the interview and some key responses from the interviewer, the learner receives final 
instructions on how to follow-up with a counselor or other support services if they want to speak with 
someone about anything they talked about during the interview. Some reported incidents require an 
immediate referral to services, and these incidences are flagged in the questionnaire.
14 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Teacher Survey – The teacher SRGBV instrument was developed to gather information on teachers’ 
backgrounds, perceptions of their schools’ climate, acceptance of inequitable gender norms, methods to 
teach learners the right behavior or to address a behavior problem, and their opinions on discipline. 
Head Teacher Survey – The head teacher survey covers the background characteristics of the head 
teacher, their perceptions of school climate, attitudes toward disciplinary methods and exposure to any 
SRGBV-related training. 
Primary Caregiver Survey – The primary caregiver survey instrument is designed to collect data on the 
child participating in the LARA program and their home environment (i.e. adverse events during 
childhood, household asset level), as well as establish the attitudes and behavior of the caregiver related 
to disciplinary methods, gender, and violence. 
School Safety Inventory – The school safety inventory is a 10-item observational checklist completed by 
an interviewer at each school. It served as an objective measure of school safety in the context of 
SRGBV as reflected in its relevant infrastructure.
Learner FGD – There were separate FGDs with girl and boy learners. The protocol used participatory 
methods to engage the learners and get their feedback on how they feel on their journey to school, on a
typical day in the classroom, and around the school, play, and latrine areas. Emotion cards and drawings 
by the learners were used to facilitate discussion. Learners were also asked who they can approach if 
they need help while in school and if they had ever had a class discussion about children’s safety in 
school. 
Teacher FGD – This protocol asked teachers about their system of evaluation and reward and 
disciplining difficult learners. It also asked teachers what they know about violence and the mistreatment 
of learners in their school, school safety, if there has been any discussion in school on improving 
learner’s safety and decreasing violence, and teachers’ knowledge of child protection resources.
Caregiver FGD – Caregivers are asked similar questions as teachers concerning the use of discipline in 
school and existing violence and mistreatment of learners. They are asked if they have ever had a 
discussion with teachers or head teachers about safety and violence in school and if they can help their 
learners access child protection resources. The protocol also included a few general questions on the 
role of caregivers in improving school safety and decreasing violence and what other 
interventions/services/programs they think would help prevent violence against children at or around 
the school.
RETENTION AND ATTENDANCE DATA COLLECTION TOOLS
The instrument for this component consists only of the learner’s name (for tracking purposes on follow￾up visits), the learner’s gender, and whether he or she was present in class at the time of the visit. The 
instrument also records if the learner’s teacher is present in-class teaching, present but not teaching, or 
absent from school.
PERFORMANCE EVALUATION DATA COLLECTION TOOLS
National/Kampala-Based KIIs – These KIIs include interview guides for USAID Staff, LARA Staff, MOES,
and MOG officials. These interviews asked USAID and LARA staff questions on their priority of 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 15
information needs and project implementation while asking MoES and MOG officials their perspectives 
on LARA implementation and perceived impacts. 
School-Based KIIs – These KIIs include interview guides for Teachers and Head Teachers who were 
asked about their role in the implementation in LARA and their perceptions on the impacts of both EGR 
and SRGBV components. 
Community-Based KIIs – These KIIs include interview guides for district officials, including education 
officers, inspectors, tutors, probation, and social welfare officers, and child and family protection officers 
as well as LARA subcontractors and sub-grantees. Community members were asked questions about 
their relationship with LARA as well as their perceptions of the program’s impacts and overall 
management. 
 In Annex D: Data Collection Instruments, we include all data collection tools used for the 
P&IE.
3.4 LIMITATIONS
The Evaluation Team encountered some limitations inherent to the design of this evaluation and during 
its fieldwork in Uganda.
 Representativeness of the Sample. The EGR, SRGBV, and R&A samples are
representative of the areas where LARA works within the two language regions in C2. 
Therefore, results are not generalizable at the national level or other geographical areas.
 Response Rate. The sample size of learners is smaller than originally expected because of 
smaller classroom sizes in Luganda-dominant schools. Also, given the sensitive nature of 
questions asked on SRGBV, we experienced some learner’s refusal to respond to the learner 
survey and/or caregivers withdrawing their consent. In addition, some teachers were absent 
on the day of EGR data collection. Although samples are slightly smaller than intended they are
sufficiently large to detect the impacts planned in the EDR.
 Learners Checks. Periodically, during support supervision visits LARA FAs administered 
learners checks - a tool used to assess learners’ reading skills - to a group of randomly 
selected learners. The tool is composed of various EGRA subtasks, letters names, segmenting, 
word recognition, and oral reading fluency, which are administered to learners in different 
grades. The tool is not identical to the EGRA (subtasks are untimed) but it is similar and could 
make learners in treatment schools more used to this type of outside testing, better preparing 
them for EGRAs used in the evaluation. If this is the case, the effects of LARA on EGR 
performance could be overestimated.
The above limitations, however, did not prevent the Evaluation Team from gathering relevant 
information and data needed and do not compromise the integrity of the evaluation in any significant 
way.
16 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
4 FINDINGS
SRGBV FINDINGS
In this section, we present findings using quantitative and qualitative data collected as part of the SRGBV 
evaluation component. We first present the effects of R2 activities on SRGBV intermediate outcomes, 
where we provide an overview of: attitudes towards gender inequality/norms, school safety climate, 
attitudes and practices of violent and non-violent disciplinary methods, learner self-reported past school 
year SRGBV, and child protection resources and availability. We show differences between T1 schools 
that only received EGR activities and T2 schools that received both EGR and SRGBV activities. This is 
followed by the fidelity of implementation, where we discuss if activities are carried out as planned, if 
activities are sufficient and relevant to achieving R2, and factors that accelerated or inhibited 
achievement of LARA results. Throughout the analysis, we use qualitative data to shed light and provide 
context to the quantitative results. All data are primary data collected by NORC. Data from LARA9 is 
included only when we discuss the fidelity of implementation. 
4.1 EFFECTS OF R2 ACTIVITIES ON SRGBV INTERMEDIATE OUTCOMES
GENDER ATTITUDES
The SRGVB learner, primary caregiver, and teacher questionnaires all include a section where the 
interviewers read 14 statements about different aspects of gender inequality and asked the respondents 
if they agreed or disagreed with the statement, or if they were not sure. For 14 statements, agreement 
to the statement indicates a belief in more inequitable gender norms. The 14 statements are used in the 
construction of an inequitable gender attitudes index. Each of the items receives a score on a scale from 
a 0 minimum to 2 maximum value for the response options: “Do not agree,” “Not sure,” “Agree,” 
following the conventions of the Gender Inequitable Men Scale.
10 Higher scores signify greater 
acceptance of inequitable attitudes toward gender norms. The values of all responses are summed, 
allowing for a maximum score of 28. The level of agreement varies substantially across respondent 
types, showing that learners hold more inequitable attitudes toward gender norms than teachers and 
caregivers. Table 9 shows that there is no statistically significant difference in learners’ and caregivers’ 
gender attitudes between T1 and T2 schools. In contrast, teachers in T2 schools show a lower level of 
gender inequitable attitudes than teachers in T1 schools, and the difference is statistically significant.
Table 9. Inequitable Gender Attitude Index1
at Midline
POPULATION T1 AVERAGE T2 AVERAGE DIFFERENCE ADJUSTED DIFFERENCE2
Learners 15.2 15.0 -0.18 -0.21
Teachers 12.0 10.4 -1.56 ** -1.57 **
Caregivers 13.9 13.9 0.00 -0.12
1 Based on 14 statements where each statement received a score on a scale from 0 at a minimum to 2 at a maximum value for the response 
options that include: “Do not agree,” “Not sure,” “Agree.” Following the conventions of the Gender Inequitable Men Scale. A higher score 
signifies greater acceptance of inequitable gender norm attitudes, with a maximum score of 28.2 The adjusted difference takes into account 
region, age, and sex.
2 Notes: *** p<0.01, ** p<0.05, * p<0.1
 
9
From their Activity Monitoring Evaluation and Learning Plan (AMELP).
10 Singh, A., Verma, R., & Barker, G. (2013). Measuring gender Attitude: Using gender-equitable men scale (GEMS) in various socio-cultural 
settings. Making Women Count: An Annual Publication on Gender and valuation by UN Women Multi-Country Office for India, Bhutan, Sri Lanka and 
Maldives, 61. http://promundoglobal.org/resources/measuring-gender-attitude-using-gender-equitable-men-scale-gems-in-various-socio-cultural￾settings/
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 17
In addition, the qualitative data also tends to show similar gender-inequitable attitudes to those found at 
baseline.
CAREGIVERS
In focus groups across both language areas, caregivers expressed frustration with the lack of protection 
that girls have in school. Caregivers perceive that girls are at risk of more violence from both teachers 
and head teachers as well as from individuals they encountered on the journey to and from school. Male 
and female teachers also perceive that girls are more at risk of violence in schools, and expressed 
concerns about girls’ safety. However, teachers and caregivers (male and female) demonstrated attitudes 
that place partial blame on girls for their lack of safety. Notably, such anecdotes were more frequent in 
discussions with female caregivers and female teachers. For example, caregivers with female children 
often reported that raising girls was more work, as girls are more fragile:
I think a boy and a girl cannot face the same violence and mistreatment because the girl is more at-risk. 
They are fragile and get dirty easily and can smell easily if they do not take care of themselves. Yet boys 
require less effort to keep clean so, with poor hygiene, the girls can be mistreated and embarrassed. 
(Female Caregiver, FGD, RR District)
Similarly, male caregivers were very cognizant of the safety challenges girls experience in school, 
including sexual harassment, challenges managing menstrual cycles, and unfair punishment. However, 
male caregivers suggest that girls who get pregnant by older men and drop out do so because they lack 
proper upbringing, or were obsessed with money. Notably, caregivers across both language areas 
recognized that relationships between male teachers and girls were often coercive. However, male 
caregivers placed blame on girls who got pregnant by other men in the community. 
TEACHERS 
Male and female teachers across both language areas also demonstrated similar attitudes. While both 
male and female teachers indicated that girls faced more risks for violence than boys did, there were 
differences in the types of inequitable attitudes expressed. Male teachers felt that girls were given special 
treatment, and prioritized more than boys:
I don’t know why the government is more concerned about girls than others. So, they really try to help 
the girls more than the boys. I don’t know why! (Senior Male Teacher, FGD, RR District)
They [implementer] have done some great work but they should improve on the selection procedure of 
the beneficiaries whereby boys also are given support because sometimes these girls who are supported 
are even better off than most of the boys who are left out (Senior Male Teacher, FGD, L District)
Male teachers believed that such practices were leaving boys behind, and putting girls who already had 
enough help at an advantage.
Female teachers expressed gender inequitable attitudes through victim-blaming, and made moral 
distinctions between girls who were forced into sexual relationships with older men or teachers, and 
those who “invited” advances from older men or teachers. Similar to male caregivers, female teachers 
felt that some girls wanted to buy things, and coerced older men into buying things for them. Similarly, 
female teachers in the Runyankore/Rukiga language area felt that girls who wear makeup were less 
18 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
serious about their schooling. When asked about early warning systems for identifying girls at risk of 
dropping out, one teacher suggests:
Another early warning system is when you find that a primary pupil has begun smearing her face with 
[makeup] and every time she wants to look at her face, then you detect something dangerous in the 
future. (Female Teacher, FGD, RR District)
In interviews, LARA staff also recognized the problem of violence against children in schools is rooted in 
long-standing social norms that uphold harsh physical punishment. Thus, it poses an enormous challenge 
to address this issue within only a couple of years of implementation.
You’re trying to shift norms from time immemorial. [...] It’s a huge rock to push. You can’t just swallow 
a hot potato. You have to cool off and take small bites. It’s a noble cause that we have to push no 
matter how much time it takes. Changes could be small because of time, but they are in the right 
direction. (LARA technical staff member, KII)
SCHOOL CLIMATE
We asked learners about their feelings about safety on the way to and at school. Table 10 shows the 
percentage of learners that agree with each statement. There is no statistically significant difference in 
learner school climate indicators between T1 and T2 schools. The differences between T1 and T2 go in 
the correct direction, with exception of the third statement “Learners feel safe on the way to and from 
school.” However, the differences are very small and not statistically significantly different. 
Table 10. Learner1 School Climate, Percent Agreeing with Statements
STATEMENT
T1 AVERAGE 
PERCENT OF 
AGREEMENT
T2 AVERAGE 
PERCENT OF 
AGREEMENT
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE2 
(PERCENTAGE 
POINTS)
Learners are sometimes afraid to go to 
school for fear of punishment
56% 54% -2 pp -3 pp
Learners feel safe when they are at 
school 
83% 86% 3 pp 3 pp
Learners feel safe on the way to and 
from school 
69% 68% -1 pp -2 pp
Learners fear reporting when someone 
older touches their private parts at 
school 
56% 56% -1 pp -2 pp
Notes:
1
Learners are reporting on perceptions/incidence within their own gender. Thus, girls respond considering the experience of girls, and boys 
considering the experience of boys. 
2 The adjusted difference takes into account region, age, and sex.
Midline qualitative findings reveal several themes related to fear of going to school, feelings of safety to 
and from school, and fear of reporting, outlined below:
Feeling safe while they are at school. Overall, boys and girls feel safe in school, except for specific 
locations within the school compound. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 19
Similar to baseline, both girls and boys feel unsafe around the perimeters of the school compounds, 
where strangers from the community often lurk. In a few cases, boys reported that such people 
confronted them with violence:
In the playground, sometimes the mad person comes around the pitch which makes me afraid while 
outside the classroom because whenever he comes around he throws stones at us when we’re playing in 
school pitch. (Boys, FGD, L District)
Girls had similar fears, expressing fear of being sexually assaulted by strangers roaming around the 
school. Girls across both language areas reported that in many cases, the men roaming near their 
schools were often menacing. In such instances, girls felt that they had to remain on constant alert 
during free time, or they would be taken away.
You may be sleeping there and someone comes and sprays you with chloroform and pulls you into the 
banana plantation. (Girls, FGD, L District)
You may be resting in the field, and someone comes along - especially those old men - and they use 
you…they don’t even have to spray you with chloroform, they can just cover your mouth. No one will 
hear your screams. (Girls, FGD, L District)
Boys and girls recommended that school compounds should be fenced in to promote learner safety.
Latrines. Also similar to baseline, both boys and girls were afraid in the latrines area, albeit for different 
reasons. Girls described latrines as dirty, insecure/without privacy, poorly constructed, and too close to 
boys’ facilities. Boys described latrines in similar ways, focusing on their lack of cleanliness and 
sturdiness. While boys remained afraid of falling in or getting infections, girls remained afraid of peeping 
toms and sexual assault. 
Feelings of safety to and from school. Boys and girls both reported feeling unsafe during the journey 
to and from school. Both boys and girls also reported walking together to school to defend themselves 
in the event of danger. While girls perceive threats from other learners, boda boda men, and rapists, 
boys perceive threats from drunkards and mentally ill people. These findings are similar to baseline, 
where boys and girls expressed fears of violence traveling to and from school.
Fear reporting when someone older touches them. There were limited changes in terms of fear of 
reporting. Boys felt that they could report to head teachers, and the PTA chairman for incidences of 
severe corporal punishment. However, boys did not indicate they felt safer in reporting incidences of 
being inappropriately touched. 
Girls reported that there were trusted female teachers or senior women in whom they could confide, 
but they remained fearful of retaliation from male teachers.
20 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 2. School Climate Indicators by Treatment Arm, Luganda Schools
Figure 3. School Climate Indicators by Treatment Arm, Runyankore/Rukiga 
Schools
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 21
School infrastructure can contribute to creating a safe climate. Figure 2 and Figure 3 show the 
percentage of schools with certain facilities in Luganda and Runyankore/Rukiga language dominant areas
respectively. Three issues are most salient. First, in the majority of the schools, latrines lack locks and 
sometimes they are shared by children and adults. This is an issue that has been mentioned, particularly 
by girls, as problematic in terms of safety and privacy multiple times. Second, water – when available - is 
not close to the school creating opportunities for unsafe situations. Finally, in a good fraction of the 
schools, there is no strict monitoring of the entry of strangers. Learners mentioned feeling threatened 
by people around the schools, and the lack of controls and entry restrictions do not contribute to a 
climate of safety. 
DISCIPLINARY METHODS
We asked caregivers and teachers about violent and non-violent disciplinary methods they use with 
children. Table 11 shows results for caregivers. In general, the use of physical and emotional violence 
remains high. For the most part, there is no statistically significant difference in the approach to 
discipline used between T1 and T2 groups. However, there is a statistically significant reduction of six 
percentage points in the percent reporting taking a child’s privileges away as punishment in T2 vs. T1 
caregivers. This is a reduction in a non-violent discipline approach. 
Table 11. Violent and non-violent disciplinary methods employed by caregivers
DISCIPLINARY METHODS: 
CAREGIVERS
T1 
AVERAGE 
PERCENT 
T2 
AVERAGE 
PERCENT 
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1 
(PERCENTAGE 
POINTS)
Explained to the child why something 
s/he did was wrong 93% 96% 3 pp 3 pp
Gave the child a reward for behaving 
well 58% 60% 2 pp 3 pp
Hit the child with an object such as a 
stick, broom, cane, or belt 69% 69% 0 pp 1 pp
Gave the child something else to do 
to stop or change their behavior 66% 59% -7 pp -6 pp
Shouted, yelled, or screamed at the 
child 66% 64% -2 pp -3 pp
Cursed at the child 17% 17% 0 pp 0 pp
Spanked the child with bare hands 54% 54% 0 pp 0 pp
Took away the child’s privileges 20% 13% -6 pp* -6 pp*
Note: *** p<0.01, ** p<0.05, * p<0.1. 
1The adjusted difference takes into account region, age and sex.
Similarly, there are few statistically significant differences between T1 and T2 teachers. Table 12 shows 
that T2 teachers reduced the use of insults towards learners, not talking to learners, or locking learners 
up as a form of discipline but did not significantly reduce the use of other violent discipline methods or 
increased positive discipline in the classroom.
22 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Table 12. Disciplinary methods employed by teachers
DISCIPLINARY METHODS: 
TEACHERS
T1 
AVERAGE 
PERCENT 
T2 
AVERAGE 
PERCENT 
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1 
(PERCENTAGE 
POINTS)
Explained to a learner why something 
s/he did was wrong 100% 97% -3 pp -3 pp
Gave a learner a reward for behaving 
well 97% 93% -4 pp -4 pp
Told a learner to start or stop doing 
something 97% 96% -1 pp -2 pp
Hit a learner on the buttocks with an 
object such as a stick, broom, cane, or 
belt
63% 60% -4 pp -3 pp
Hit elsewhere not buttocks with an 
object such as a stick broom, cane, or 
belt
29% 27% -2 pp -2 pp
Gave a learner something else to do 
to stop or change behavior 96% 96% 0 pp 0 pp
Threatened to expel a learner 32% 40% 8 pp 9 pp
Shouted, yelled, or screamed at a 
learner 57% 44% -13 pp -13 pp
Threatened to invoke harmful people, 
ghosts, or evil spirits against a learner 18% 15% -3 pp -3 pp
Cursed at a learner 12% 19% 6 pp 7 pp
Spanked a learner with bare hand 31% 29% -2 pp -2 pp
Locked a learner out of school 19% 19% 1 pp 1 pp
Took away a learner’s privileges 13% 17% 4 pp 4 pp
Forbade a learner from going out 
during break time 25% 23% -2 pp -2 pp
Insulted a learner by calling him/her 
dumb, lazy, or other names
27% 14% -13 pp* -13 pp*
Refused to speak to a learner 20% 9% -11 pp* -12 pp*
Blamed a learner for bad things that 
happened in their life 24% 14% -9 pp -9 pp
Locked a learner up or tied him/her to 
restrict movement
8% 1% -6 pp* -7 pp*
Withheld a school meal as a 
punishment 4% 1% -3 pp -3 pp
Used public humiliation to discipline a 
learner 25% 22% -3 pp -3 pp
Note: *** p<0.01, ** p<0.05, * p<0.1
1The adjusted difference takes into account region, age and sex.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 23
Looking specifically at corporal punishment, we find no statistically significant differences between T1 
and T2 schools at baseline or midline when caregivers, teachers, and head teachers were asked whether 
or not they believed corporal punishment was effective as a form of discipline (Figure 4). Few teachers
(around 12 percent at midline) say that corporal punishment is an effective form of discipline, however, 
as seen in Table 12, around 60 percent of them report hitting learners on the buttocks with an object 
such as a stick, broom, cane or belt. To explore this further, at midline a new question was asked 
providing an example of a common type of corporal punishment: hitting a learner with a cane or stick. 
Figure 4. Percent of caregivers, teachers, and head teachers that think corporal 
punishment is an effective form of discipline
Figure 5 shows that the percentage of T1 teachers and head teachers who believe hitting a learner with 
a cane or stick is an effective method of discipline at school is higher than in T2 schools, with the 
difference being statistically significant for teachers. These results, compared to those asking the same 
question using the term “corporal punishment” suggest that there is a disconnect in understanding what 
corporal punishment really is and suggests that teachers do not consider caning as corporal punishment. 
However, in KIIs with teachers in LARA T2 schools the head teachers spoke passionately about ending 
harsh physical punishment of learners, and attributed their newfound skills in alternative discipline 
methods to LARA.
“We learned how to handle children. We used to punish children. I am disgusted with what we used to 
do to punish children. Alternative positive discipline means talking with a child and they apologize. A 
child can always reform. I keep close to the child. We meet and address the best ways to discipline 
children. We have a proper relationship now between pupils, teachers and the community. It was a 
collision before. Now we have a better relationship between teachers and caregivers” (Head teacher, KII, 
T2 school)
Ministry officials in Kampala also have a perception that is different from what the quantitative results
show. They felt that Journeys had contributed to “engaging communications,” and that caregivers are 
“giving up on corporal punishment of children” which is not seen in their responses shown in previous 
tables and figures.
24 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 5. Reports of hitting a learner with a cane or stick being effective at midline
Quantitative findings indicate that physical and emotional violence is still used. Table 12 showed that 
among T2 teachers, 60 percent of the teachers used some object to hit learners, 29 percent spanked 
them with a bare hand, 22 percent use public humiliation and 19 percent cursed learners as a discipline 
method at midline Qualitative data indicate similar findings, despite teachers having received trainings on 
alternative discipline methods.
Teachers across both language areas report having received training on alternative disciplinary methods, 
including counseling and assigning classroom tasks to learners. As a result of training, teachers report 
that they have learned not to reprimand or embarrass learners when they get wrong answers. Notably, 
teachers also acknowledge that the use of alternative discipline methods is better for fostering teacher￾learner relationships, while violent discipline methods often result in contentious relationships, feelings 
of discouragement for learners, or dropping out. However, anecdotes from teachers, caregivers, and 
learners indicate that the use of violent disciplinary methods remains rampant.
At school, we have what we call ‘Conveyor Belt’ whereby about 3 or 4 teachers gather in one class and 
beat up a stubborn learner, or those that have performed badly in a certain subject. (Female Teacher, 
FGD, RR District, T2 school)
Teachers attribute the continued use of physical violence to the behaviors of learners, including 
stubbornness, and a lack of excitement about learning. Male and female teachers express that they don’t 
feel completely confident in their ability to employ alternative discipline methods consistently, and 
should receive refresher trainings to minimize the use of physical, sexual, and emotional violence:
If teachers were sensitized on how to handle such characters maybe some good would come out of it. 
Because a child can behave so badly but if a teacher has some idea on how to handle the child it makes 
a difference…the teacher could angrily beat up the child but if the teacher is aware of some better 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 25
alternative she can avoid torturing the child. So we, the teachers, need refresher courses. (Female 
Teacher, FGD, RR District, T2 school)
Caregivers had similar attitudes about corporal punishment, noting that the use of corporal punishment 
is more effective for “difficult” learners. While male and female caregivers were opposed to extreme 
corporal punishment that could result in significant injuries, they felt that it was an effective method for 
discipline in and outside school. In fact, head teachers in T2 schools highlighted initial resistance from 
caregivers and community members to reducing corporal punishment in schools, stating that they are 
often encouraged to discipline the child harshly. A small set of caregivers in T2 communities revealed a 
change of perspectives and recommended looking more holistically at the child to determine what went
wrong. In these instances, caregivers felt that children could benefit from jointly conceived punishment 
from teachers and caregivers. 
Boys and girls across both language areas reported feeling afraid to go to school because of the 
punishments they would experience. This was most often reported for punishments around tardiness 
and wrong answers.
Across both language areas, boys and girls indicated that when they were tardy, they felt afraid to even 
go to school as they knew punishment was imminent.
You come late and find the teacher teaching. Then he prevents you from entering, and he says if you 
enter he will cane you, then you keep on the door scared up to when his lesson ends, then he comes 
out, and when you enter you can't even keep up with what s/he has taught. (Boys, FGD, RR District)
You come late, they cane you, and even when you sit in class they again cane you. The teacher doesn’t 
ask you what made you late, he just starts caning you that you came late. (Boys, FGD, RR District)
Girls more often reported that the fear of punishment for being late would sometimes make them 
choose not to go to school at all. 
Boys and girls also indicated being most afraid of punishment during class. Boys and girls across both 
language areas revealed that they were afraid of being caned for wrong answers or low marks on exams.
You can do an exam in class and the teacher says that whoever is below 80%, canes! He returns to the 
classroom with the results and canes. That scares us! (Girls, FGD, L District)
When I am in class and the teacher beats me excess canes because of poor handwriting I feel sad. 
(Girls, FGD, L District)
Boys and girls reported similar frequencies of such punishments, with boys reporting harsher 
punishments. While girls often described receiving 5 canes for a given incident, boys described receiving 
up to 20. 
PREVALENCE OF VIOLENCE
One of the main objectives of the SRGBV quantitative data collection effort is to gather reliable first￾person data on the prevalence of gender-based violence in and around schools. Each P4 and P6 learner 
interviewed was asked directly whether specific forms of SRGBV had happened to him or her during the 
26 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
last school year. This section of the questionnaire was divided into four parts: emotional violence, and 
physical violence, sexual violence by a range of perpetrators, and non-verbal disclosure of violence by a 
teacher.
In order to gauge the prevalence of violence, the SRGBV instrument asks a series of questions asking if 
learners have experienced different types of emotional, physical, or sexual violence, for example, “Did 
anyone around school call you rude or hurtful names?” Table 13 shows the percentage of learners that 
report being victims of different types of violence - emotional, physical, and sexual. At midline, the 
prevalence of violence is still very high and there are no statistically significant differences in the 
percentage of learners that were victims of violence between T1 and T2 schools. The vast majority of 
learners report suffering emotional and physical violence at school and around 40 percent report sexual 
violence. 
Table 13. Learners that suffered violence
STATEMENT T1 AVERAGE 
PERCENT 
T2 AVERAGE 
PERCENT 
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1 
(PERCENTAGE 
POINTS)
Percent of learners who 
suffered any incidence of 
emotional violence.
98% 98% 0 pp 0 pp
Percent of learners who 
suffered any incidence of 
physical violence.
91% 94% 3 pp 2 pp
Percent of learners who 
suffered any incidence of 
sexual violence.
40% 38% -2 pp -4 pp
Note: *** p<0.01, ** p<0.05, * p<0.1
1 The adjusted difference takes into account region, age, and sex.
Figure 6 shows the average number of questions that learners answered affirmatively in each violence 
category. Out of 9 possible forms of emotional violence, on average, learners report being victims of 
almost 3 of them. In the case of physical violence, the average learner reports suffering more than 3 
types of violent events out of a list of 10. Learners in T2 schools report suffering slightly fewer types of 
physical violence than their counterparts in T1 schools. Finally, out of 13 types of sexual violence 
situations, the learners report having suffered one of them on average. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 27
Figure 6. Learners’ reported number of types of violence suffered1
1 There are 9 questions about emotional violence, 10 about physical violence and 13 about sexual violence. 
According to the learners, by far the most common form of violence perpetrated against them is getting 
hit with a cane or stick, and this is done mostly by teachers. Although teachers’ reports about using 
canes or sticks to hit learners are not significantly different between T1 and T2 schools, learners’ 
reports show some difference. Comparing T1 and T2 schools, we find that, at midline, learners report a 
prevalence of this type of violence that is 7 percentage points lower in T2 than in T1 schools.
Table 14. Teachers’ use of a cane or stick to hit learners, as reported by learners
T1 AVERAGE 
PERCENT 
T2 AVERAGE 
PERCENT 
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED DIFFERENCE1 
(PERCENTAGE POINTS)
59 52 6 pp ** 7 pp **
Note: *** p<0.01, ** p<0.05, * p<0.1
1 The adjusted difference takes into account region, age and sex.
The frequency in the teachers’ use of a cane or stick to hit learners is the same in T1 and T2 schools.
Twenty-one percent of the learners reported they were hit with a cane or stick once in the past year, 
39 percent indicated that it happened a few times, and the remaining 40 percent indicated that it 
happened many times. A detailed breakdown of each physical, emotional, and sexual violence indicator is 
provided in Annex C: Additional Analysis. 
Qualitative findings also indicate persistent physical and emotional violence for children. Children 
reported experiencing emotional violence in the form of bullying, teachers giving nicknames associated 
with low levels of intelligence, and bullying about their physical features. Girls continue to experience 
bullying from teachers and other learners during their menstrual cycle.
Girls and boys continue to experience physical violence through caning and punishments through 
physically-intensive tasks, including fetching water and carrying other heavy loads.
28 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
As part of good practice interviewing children about past experiences of violence, the learner survey 
included a question that allowed the child to disclose a past incidence of sexual violence by a teacher 
without needing to verbalize the response. The interviewers presented each learner with a sheet of 
paper that showed two illustrations: one of a smiling child, and one of a crying child. The interviewer 
handed the child a pen and asked the child to mark an “X” by the sad face “if a teacher has ever touched 
you on your private body parts, or made you touch their private body parts”, or mark an “X” by the 
happy face if this had not happened to them. The interviewer noted the placement of the “X” as a 
response to the questionnaire. Table 15 below shows a higher percentage of learners reporting teacher 
sexual violence non-verbally compared to those reporting teacher sexual violence verbally. For the 
verbal disclosure, learners responded to the question “if anyone around school touched your private 
parts when you didn’t want them to, what type of person did it?” The result shown is a combined 
response for both male and female teachers being perpetrators. 
Table 15. Non-verbal disclosure of sexual violence perpetrated by teachers
LEARNER DISCLOSURE OF PAST 
SEXUAL VIOLENCE FROM A 
TEACHER
T1 
AVERAGE 
T2 
AVERAGE 
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1 
(PERCENTAGE 
POINTS)
Verbal disclosure of past sexual 
violence by teacher 3.5% 3.0% -.5pp -.5pp
Non-verbal disclosure of past sexual 
violence by teacher 5.1% 4.5% -.6pp -.6pp
Note: *** p<0.01, ** p<0.05, * p<0.1
1 The adjusted difference takes into account region, age and sex.
REPORTING OF VIOLENCE
In both T1 and T2 schools, almost all learners report knowing to whom they can report violence (Table 
16). However, the proportion of learners that report having an adult that they trust to whom to report 
violence is much lower: 49 and 58 percent in T1 and T2 schools respectively. The difference between 
T1 and T2 schools is statistically significant suggesting that LARA activities had an effect; however, the 
proportion of learners that have a trusted adult is still below 60%. 
Table 16. Learner Protection Resources
REPORTING VIOLENCE T1 
AVERAGE
T2 
AVERAGE
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1
(PERCENTAGE 
POINTS)
Learners know who to report to 
when they experience violence at 
school.
98% 97% -1 pp 0 pp
Learners have an adult they trust to 
whom they can report violence.
49% 58% 0.10 pp *** 0.10 pp ***
Notes: *** p<0.01, ** p<0.05, * p<0.1.
1 The adjusted difference takes into account region, age and sex.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 29
Trained as LARA Change Agents, teachers interviewed outlined reporting pathways for child survivor￾disclosed or third-party witnessed SRGBV incidents in schools set up as part of the LARA program. A 
Head Teacher noted that community-based VCCMCs, “tried to put in place control in school discipline, 
and sat with teachers to discuss discipline cases.” In cases of reporting sexual abuse though, a teacher to 
whom a learner disclosed was more likely to report this to the senior woman teacher or senior man 
teacher to “intervene” with the child and bring in the head teacher for further steps. 
“She should report that thing to her caregivers and to the head teacher for further help. I have to report 
to the senior woman teacher or senior man teacher to the head teacher to make a follow-up with the 
senior teacher engaged. The head teacher has to call in the perpetrator for more questions. If proved 
true, we forward the case to the authorities who report to the police.” (Head Teacher, KII)
Some head teachers recognized that cases of severe physical violence with injuries or sexual violence 
against pupils were beyond their capacity or mandate and should be reported to the subcounty 
Community Development Officer (CDO) or the police. Police interviewed said they train DPSWOs, 
CDOs, district education officials, head teachers, as well as held ad hoc community informational 
meetings to provide instructions and answer questions on when a case should be reported to the police. 
Other head teachers sought to investigate and “resolve” these cases on their own. Investigation of and 
decisive ruling over a sexual abuse case is outside of the skills or mandate of a senior teacher or head 
teacher, risking retaliation against and re-victimization of child sexual abuse survivors as well as 
protection of abusers in the school or community. A reporting structure through the head teacher, 
renders head teachers gatekeepers capable of interfering with child violence survivors seeking justice or 
health assistance. 
Key informants widely perceived LARA impacts on SRGBV prevention and response as tied directly to 
head teacher support or lack of support for teachers and pupils taking time and resources to conduct 
Journeys activities in school. Further, follow-up actions after a pupil reported an incident of SRGBV within 
the school often was “left up to the head teacher [regarding] what actions to take” (Male Teacher, 
LARA Change Agent).
Teachers, head teachers, and other school staff lacked skills for effective, non-revictimizing 
communication with a child survivor, particularly survivors of sexual abuse. While many head teachers 
and teachers do their best to care for children, good intentions without technical training, support, and 
mentorship can lead to unintended consequences. Teachers responded simply telling, for example, a 
child sexual violence survivor, “Don’t commit suicide and don’t run away. You’re not the only one this 
has happened to so don’t feel alone. You’ll be fine.” (Senior Woman Teacher, T2 school), which does 
not follow survivor-centered, child-friendly communication practices with child GBV survivors. Teachers 
widely described providing, “guidance and counseling,” to girls and boys who reported SRGBV incidents, 
but when asked what they would say to counsel a child violence survivor, it became clear they had not 
been trained for effective, non-re-victimizing communication in guidance and counseling. “There was a 
lot of victim blaming around sexual violence” (LARA subgrantee). 
Teachers reported that Community Development Officers, with a government child protection 
mandate, “could help solve some disputes, such as if a pupil was defiled, but they have neglected their 
role as CDOs.” When asked about a school register of SRGBV cases, no T2 schools in the performance 
evaluation could produce one. One teacher claimed they had handmade a case-register themselves, but 
could not locate it at the time of the interview.
30 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
EVIDENCE OF FIDELITY OF IMPLEMENTATION 
In this section, we answer two main questions across all SRGBV activities undertaken by LARA. First,
were activities carried out as planned and were they sufficient and relevant to achieving R2? Second, 
what factors accelerated or inhibited achievements of LARA results? Findings here are drawn from both 
the midline impact evaluation and final performance evaluation. 
ACHIEVEMENT OF LARA PERFORMANCE TARGETS, RTI ESTIMATES 
Table 17 below shows LARAs annual progress toward activity targets. Note, however, that LARA 
reports totals for C1 and C2 while the P&IE quantitative information is only for C2 which is the focus of 
this evaluation. Data were not available from LARA for every relevant indicator at the time of the 
external performance evaluation due to limited or unavailable records of teacher and pupil attendance at 
Journeys and UKU activities in schools, for example. According to the P&IE C2 data, 25% of schools had
written feedback on SRGBV prevention and response supportive supervision visits from district 
education officials, LARA’s AMELP shows this (indicator 2102) at 35% in 2019 for C1 and C2 schools. 
While LARA reports having exceeded its targets on the numbers of teachers trained in SRGBV, the 
analysis below spotlights challenges teachers faced in incorporating this training consistently throughout 
school. Insights from LARA staff and other stakeholders in the next sections provide information about 
challenges and opportunities the LARA program faced in implementing Journeys trainings with teachers 
and school staff, with pupils through UKU groups, and with communities.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 31
Table 17: LARA AMELP Indicators
Country 
Development 
Cooperation 
Strategy Links
RESULTS
FY 2016
ANNUAL 
ACTUAL
FY 2017 
ANNUAL 
ACTUAL
FY 2018 
ANNUAL 
ACTUAL
FY 2019 
ANNUAL 
ACTUAL
TARGETS LARA COMMENTS FY 2019 
ANNUAL REPORT
DO3:
IR 3.3:
Sub-IR 3.3.1
Proportion of 
schools 
observed to 
have written 
feedback on 
SRGBV 
response and 
prevention 
from 
monitoring 
and support 
supervision 
actors
-- 17%
(459 /
2,698
T2 schools)
36%
(971/2697
T2 schools)
35%
(895 / 2557
T2 schools)
25% in 2017;
25% in 2018;
35% in 2019
“Monitoring and support 
supervision of school-level 
SRGBV activities by support 
agents (e.g. head teachers, CCTs 
and school inspectors) is still low 
due to limited head teacher 
involvement (e.g., some of them 
see Journeys as extra work, 
excessive head teacher 
absenteeism from school, etc.). 
This is compounded by the 
CCTs’ heavy workload, leaving 
little time to support Journeys 
and limited facilitation for district 
support teams to conduct school 
visits. Our strategy in Year 5 
focuses on strengthening school 
leadership support to Journeys 
activities in schools.” 
DO 2:
IR 2.3:
Sub-IR 2.3.1
Number of 
teachers 
trained in 
SRGBV
3,812 
(1,649 
women, 
2163 men)
2,772 
(1,352 
women, 
1,420 men)
10,369 
(5,734 
women, 4, 
635 men)
12,573
(6,267 
women,
6,306 men)
27,729 
(14,662 
women, 13, 
067 men)
“LARA trained 12,573 teacher 
patrons (7,393 from 1,635 basic 
education schools and 5,180 from 
565 PEPFAR-supported schools. 
The project achieved its target on 
this indicator.”
32 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Country 
Development 
Cooperation 
Strategy Links
RESULTS
FY 2016
ANNUAL 
ACTUAL
FY 2017 
ANNUAL 
ACTUAL
FY 2018 
ANNUAL 
ACTUAL
FY 2019 
ANNUAL 
ACTUAL
TARGETS LARA COMMENTS FY 2019 
ANNUAL REPORT
DO2:
IR 2.3:
Sub-IR 2.3.1
Number of 
schools
implementing 
actions to 
reduce 
SRGBV
-- -- 1,055 / 
2,697 T2 
schools
428 / 2,557
T2 schools
1,027 for 
2017;
1,771 for 
2018;
1,238 for 
2019
“Teachers are still struggling with 
applying the “5 step process of 
inspiring change” which is the 
source of data for this indicator. 
In addition, some schools do not 
document actions they have 
implemented. This makes it 
difficult for the project to report 
without source documents. The 
project intends to address these 
challenges in Yr.5 through 
support supervision visits to 
schools.”
Note: All baseline values were zero; Sources: LARA Annual Reports 2016, 2017, 2018, 2019; and LARA AMELP Year Four 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 33
GENERAL STAKEHOLDER PERCEPTIONS OF LARA
District education officers felt that LARA had contributed skills building on social and emotional learning 
(SEL), self-management, and responsible decision-making in schools. District officials claimed that the 
program has led to changes in their inspection visits, in that now they look for sticks and canes that 
might be used for punishment, in classrooms during inspections. One District Inspector claimed that 
assessment of Journeys and UKU implementation had been added to the District Inspection forms, 
however on review of those forms, the Performance Evaluation team found that the only specific 
mention of any relevant activity was “UKU,” and that this was added to the “comments” section 
grouped together with other co-curricular activities, including EGR, School Management Committees 
(SMCs), and Determined, Resilient, Empowered, AIDS-free, Mentored and Safe (DREAMS) project. In 
no place on the forms were there specific questions or places to put detailed information and 
observations about Journeys and UKU implementation during an inspection visit. In the performance 
evaluation key informant interviews, LARA staff had raised the need for integrating inspection 
monitoring of SRGBV prevention and response activities into district monitoring tools:
“The [LARA] project is doing best with its workforce, but District Officials need to do more, not project 
staff alone. Routine activities should be in District Monitoring Tools, should not be outside of academic 
structural monitoring of activities. The safety of learners comes last. Incorporate SRGBV monitoring tools 
as mandatory to check schools by district officials, DEO, DIs.” (LARA technical staff member, KII)
A LARA technical staff member elaborated further:
“Teachers need to believe alternatives to corporal punishment will work. Teachers need training, but 
once training is over, they won’t do anything about it. The main issues are bullying and corporal 
punishment. Corporal punishment is mainly perpetrated by teachers. Set rules to end bullying in school. 
Strengthen engagement and support to teachers. If they are not supported, they will not see it working 
and they will abandon it. Sustained training is required. When districts are planning supportive 
supervision visits or should do one, they should include EGR activities, but should also include SRGBV, 
Journeys as part of their LoE (Level of Effort) and their performance assessment. [District education 
officials] should set benchmarks and targets on Journeys.” (LARA technical staff member, KII)
District education officers raised the issue that teachers find it difficult to manage learners because of 
the large numbers of pupils in a class, sometimes more than 100, and that this negatively affected 
Journeys and UKU implementation. District officials acknowledged that LARA SRGBV intervention has 
“not been the best implementation because teachers’ workloads are too much” (District Inspector). 
Ministry officials felt that LARA complemented what MoES, their Gender Unit, and the Ministry of 
Gender, Labor and Social Development (MoGLSD) were already trying to do to prevent and respond to 
violence against children. Ministry officials recognized SRGBV as a “very big challenge,” on which LARA 
had “not achieved as much as on EGR.” Uganda already had an enabling, supportive policy environment, 
structures, and strategic plan prior to LARA, according to a MoGLSD official. LARA program 
interventions that aimed to create a friendly school environment fit well with education sector work to 
address challenges of low retention of learners in school, particularly girls, and high violence prevalence. 
Ministry officials reported engaging across a range of key stakeholders to eliminate violence against 
children in all contexts in Uganda, including schools. They highlighted how Journeys has contributed, 
34 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
“engaging communications,” and that caregivers are “giving up on corporal punishment of children.” 
Officials recognized that violence against children in schools leads to dropout and early pregnancy, 
raising issues of sexual exploitation and abuse of girls by teachers and motortaxi (bodaboda) drivers. 
Child labor also was recognized among officials as contributing to violence against children and school 
dropouts, citing reduced school attendance on market days. Ministry officials considered LARA’s lesser 
achievements on safety in schools as “maybe because the job is enormous.”
SBCC CAMPAIGN TO PREVENT CORPOREAL PUNISHMENT, PROMOTE POSITIVE 
ALTERNATIVE DISCIPLINE METHODS IN SCHOOLS AND COMMUNITIES, AND IMPROVE 
TEACHER KNOWLEDGE
Result 2 intervention components included a social behavior change communications (SBCC) campaign 
to prevent corporal punishment in communities and schools. The program approach was to: “Support a 
targeted SBCC initiative that will strengthen the implementation of SRGBV-related policies, build 
widespread awareness about the attitudes and gender power relations that serve to produce and 
maintain SRGBV and to directly equalize gender roles, and to promote actions that serve to prevent 
SRGBV” (LARA Y4 AMELP 2019). LARA partnered with a Ugandan communications organization to 
develop the SBCC campaign11 and with a Ugandan research and evaluation organization to plan and 
conduct an evaluation of a one-month long pilot of the SBCC campaign in two districts. Following the 
pilot, the SBCC campaign was implemented in schools in 15 districts between June and August 2019 and 
was targeted at caregivers and teachers of P1 through P4 learners. 
After about a decade of the corporal punishment ban in schools, the duration of LARA SBCC campaign 
to prevent corporal punishment and promote positive disciplinary methods was simply too short for 
teachers and caregivers to fully grasp and internalize new knowledge and cultivate new social 
expectations in support of building new social norms promoting positive discipline methods. 
“The lifespan of our intervention was limited. Corporal punishment has been done by our caregivers and 
grandparents We cannot address this in one month. The pilot for corporal punishment was 
implemented over three months. It should have been six months, then go back again. Success through 
repetition.” (LARA partner, KII)
The quantitative data confirms the above qualitative findings. Focusing on teacher and head teacher 
knowledge regarding the schools’ codes of conduct and SRGBV being punishable by law, we don’t see 
any statistically significant differences between T1 and T2 teachers and head teachers (
Figure 7 and Figure 8). Both groups of teachers are familiar with teacher codes of conduct. There is a 
difference of seven percentage points between T1 and T2 teachers who know that hitting a learner with 
a cane or stick is punishable by law, however, the difference is not statistically significant. What is 
striking is that even though 90% of T2 teachers know that hitting a learner with a cane is punishable by 
law, 60% of them still do this, suggesting that the law is not enforced (see Table 12).
 
11 LARA brought together caregivers, teachers, head teachers, local government leaders, representatives of community-based theater groups, 
illustrators, and graphic designers to develop the SBCC campaign. The design included drama performances, mobile cinema, interpersonal 
communications, print materials (pocket sized booklets and wall posters), radio advertisements, programs, and jingles. Interpersonal 
communication was led by retired head teachers and teachers trained to use SBCC program materials incorporating counterarguments on the 
benefits of positive discipline, and how they themselves had changed to give up corporal punishment and adopt positive discipline methods
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 35
Figure 7. Teacher knowledge
Figure 8. Head teacher knowledge
TEACHER TRAINING
Table 18 below details results from the Teacher Survey related to types of training teachers have 
received on addressing behavior problems and preventing violence at school. We first asked if teachers 
have ever received training to address behavioral problems. There is a statistically significant difference 
between the percent of T1 and T2 teachers who have received this type of training (18 pp difference). In 
all the categories of this type of training (bullying, physical and sexual violence), we find that more T2 
than T1 teachers received training. Next, we asked about LARA training to prevent violence at school. 
While 82 percent of T2 teachers received training for LARA, only 21 percent of T2 teachers did. This 
suggests there is some level of training leakage to T1 teachers. This may be due to teacher transferring
between schools and/or T1 teachers misreporting on trainings they have received from other sources in 
the past. 
36 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Table 18. Teacher training
STATEMENT T1 
AVERAGE
T2 
AVERAGE
DIFFERENCE 
(PERCENTAGE 
POINTS)
ADJUSTED 
DIFFERENCE1 
(PERCENTAGE 
POINTS)
Have you ever had training for 
learning behavior problems? 73% 91% 18 pp*** 18 pp***
Have you ever had training to prevent 
bullying?
63% 93% 30 pp*** 30 pp***
Have you ever had training to prevent 
physical violence?
75% 92% 18 pp** 18 pp**
Have you ever had training to prevent 
sexual violence?
70% 88% 18 pp** 18 pp**
In the past year, have you participated
in LARA training on preventing 
violence at school?
21% 82% 61 pp*** 61 pp***
Notes: *** p<0.01, ** p<0.05, * p<0.1
1 The adjusted difference takes into account region, age and sex.
Table 19 shows the specific types of training received by the 82 percent of T2 teachers that reported 
participating in LARA training. Most T2 teachers were trained as Teacher Patrons to prevent violence 
for pupils (77 percent); 41 percent were trained as Change Agents and 13 percent as SRGBV lead 
facilitators. A smaller percentage of T2 teachers were trained on SRGBV compliance and referral case 
management (3 percent each) and some were trained as trainers themselves (5 percent). 
Table 19. Type of LARA training received by T2 teachers under R2 activities 
TYPE OF TRAINING
AVERAGE PERCENT OF T2 
TEACHERS THAT RECEIVED 
THE TRAINING
LARA Change Agent training 41%
LARA SRGBV lead facilitator training 13%
LARA SRGBV training of trainers 5%
LARA SRGBV compliance training 3%
LARA SRGBV referral web case management training 3%
LARA UKU Teacher Patron training (violence prevention for pupils) 77%
Note: categories are not mutually exclusive, a teacher could have received more than one type of training.
Indicators on implementation for teachers and head teacher show higher rates of familiarity with 
Journeys materials and experience with Uganda Kids Unite groups (Figure 9). Over 85 percent of 
teachers and head teachers reported having seen Journeys materials in their schools, launching a Uganda 
Kids Unite learner peer support group, and participating in a Uganda Kids Unite meeting or similar 
activity with learners.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 37
Figure 9. LARA SRGBV activities implemented in T2 schools, teachers and head 
teachers
During FGDs, teachers in both language areas indicated some familiarity with Journeys and UKU. In KII, 
however, teachers saw limitations in the capacity of Journeys to contribute to bringing about a positive, 
safe learning environment for learning in that many only received Journeys handbooks in English. This led 
to some inconsistencies in comprehension and implementation fidelity of the content. Teachers 
recommended that LARA provide Journeys in local languages for teachers, pupils, and communities. 
Key informants recognized widely that there was not enough time for teachers to go through training 
covering all Journeys and UKU activities to master them. Too many activities were covered in too short 
of periods. Teachers also did not have sufficient time and space within the expected program duration 
to implement activities, which led to skipping and missing steps. Follow-up training on Journeys was seen 
as too short, only one to three days, with one subcounty workshop per term engaging teachers, and a 
once-per-term training of head teachers by CCTs that also focuses in EGR supportive supervision and 
training. Teachers were concerned that CCTs do not have time, funding, or technical capacity to 
facilitate adequate training. Further, teachers felt they lacked materials, with one school only having 
seven copies of Journeys for pupils in a school of several hundred.
District officials support supervision visits to schools to support SRGBV activities were not happening as 
frequently as expected, according to LARA staff. School staff corroborated this saying that they 
expected to have more support supervision visits and ongoing technical support on SRGBV prevention 
and response through implementing Journeys activities. District officials explained that one District 
Inspector cannot reach all schools in one term, only about half. As they are expected to conduct 
inspections across primary, secondary, and tertiary educational institutions, they must select some at 
each level, with a focus on government and few private institutions.
38 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
JOURNEYS PROGRAM
Teachers reported that the Journeys Program identified change agents within their schools. Following 
this, teachers report having conducted meetings with caregivers to discuss the safety of children. 
Caregivers who are involved in meetings become aware of danger zones their children encounter to and 
from school, as teachers and caregivers agree on the appropriate times for learners to be released from 
school. This was a key concern for caregivers at baseline, who expressed concerns about their children 
leaving school too late in the day.
Teachers and caregivers felt that the Journeys program fostered more communication between teachers 
and caregivers about learner safety. Teachers also reported that the Journeys program had positive 
impacts on child protection reporting procedures. Teachers in both language areas found the inclusion 
of community members and Local Councils (LCs) improved knowledge around reporting procedures, 
and allowed teachers, community members, and government entities to work more cohesively. 
”There are many things in that Journeys books for the teachers. But it is not only for the teachers 
because it even involves the community, the reverend, the chairpersons LC 112, which has helped us to 
know much about the child and a safe school. So in case of anything, now a child can be able to go and 
report even to the LC1 Chairperson in case of any problem.” Senior Female Teachers, FGD, RR District
Female teachers in both language areas reported that Journeys facilitated improved accountability for 
teachers engaging in harsh corporal punishment or sexual harassment of learners. Specifically, senior 
female teachers felt more empowered to hold male colleagues accountable for sexual harassment. In 
these instances, senior female teachers felt that their male colleagues were more open to critical 
feedback and more willing to collaborate on ensuring learner safety.
KIIs with school staff who had received Journeys training confirmed that they acquired new knowledge 
and communication skills for interacting with learners. A head teacher summarized what they learned 
through Journeys as, “When you take away corporal punishment, school becomes friendly to pupils.” 
Head teachers and teachers felt that if support from LARA were continued, then greater impacts could 
have been achieved. LARA staff also highlighted that Journeys provides teachers with Social and 
Emotional Learning (SEL) skills, yet there are challenges in internalizing and practicing these skills within 
a standalone program not integrated into the education system curriculum or teacher continuing 
professional development:
“If we are to prevent VACiS, we are dealing with the human heart. How can we deal with people who 
don’t want to change their heart? Change has to start with me. I [as a teacher] should be able to control 
myself and talk to learners, not raise my hand against them. I want teachers and adults to be stronger 
to use SEL skills to change themselves, so that they model what they want children to do. We’re used to 
ordering children around. How can we incorporate SEL into teacher training and office work?” (LARA 
technical staff member, KII)
“I really want to see things change on SRGBV, but things aren’t changing. I care about this.” (LARA 
technical staff member, KII) 
 
12 Local Council 1 (LC 1) is the lowest level of local elected government within the district, and is responsible for a village, or in the case of 
towns or cities, a neighborhood. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 39
Teachers positively perceived the approach of forming UKU groups of learners to collectively work 
through Journeys materials for pupils with the support of a teacher patron. District officials felt that 
LARA training, using Journeys, enabled both teachers and learners to identify different types of violence, 
which, in turn, allowed teachers and learners to track and report incidences of SRGBV. It did not, 
however, allow for enough time for participants to dialogue through and internalize new knowledge and 
skills for sustainably shifting attitudes, beliefs, practices, and norms concerning SRGBV. 
Ministry officials also felt that the Journeys handbooks were designed to promote reflection among 
teachers and caregivers on actions in need of change, and that a reflective approach, “helps a lot.” 
Officials further added that LARA’s direct engagement with school staff helped a lot with addressing 
school staff perpetration of violence against children. They further highlighted how Journeys: “[...] helped 
to increase children’s awareness of what’s happening to them. Sometimes, they don’t know an act is 
violence. ‘Now they can tell: This is dangerous. This person is not a good friend.’ This, however, 
requires continuity of LARA.”
LARA activities to build community capacity for cultivating a positive and supportive learning 
environment for learners in schools and their communities focused more on the establishment of 
VCCMCs to respond to cases of SRGBV in schools (see discussion above under “Outcomes of 
Interest”), rather than training on and implementation of Journeys activity handbook for communities. A 
LARA technical staff member stated that:
“If had to do it over again, I would not include the community impact of Journeys because it’s 
problematic to implement. Activities need to build one on the other. Community attendance is difficult to 
get consistently. Different people come every time. Community implementation is problematic because it 
is not building on one activity to the next. It is preaching to different cohorts.” (LARA technical staff 
member, KII)
Ministry officials also valued the fact that LARA engaged communities and caregivers as key 
stakeholders, working to help them reflect on their actions. Ministry officials recognized how the 
prevention of violence against children requires teachers, schools, and the community to work together 
to increase child protection and safety. 
“Before everyone had responsibility for child development in a community, now caregivers are alone. 
With Journeys, we all have a role to play. You can disseminate key policy documents, but people don’t 
read. We do though have laws and policies that protect children. We have our own laws here in 
Uganda. It’s not another force pushing.” (Ministry Official, KII)
District officials, however, thought that LARA activities on Journeys were relevant in that they 
contributed to more “gender-friendly environments,” life-skills building, including problem-solving, as 
well as reading, pupil retention, violence tracking in schools, and community mobilization. District 
officials expected schools to keep quarterly documentation on SRGBV reports and prevention and 
response activities in an Inspection File, yet based on performance evaluation findings and the RTI LARA 
AMELP for Year Four, school-level documentation has been lacking. Caregivers also perceived that 
Journeys had positive benefits on school safety and their relationships with teachers. Female caregivers 
felt that learners were being disciplined less harshly, with teachers using “more acceptable methods of 
corporal punishment.”
40 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
UGANADA KIDS UNITE
Participation in Uganda Kids Unite (UKU) or violence prevention activities, and seeing printed materials 
about Journeys are statistically significant and larger in T2 schools than in T1 schools, as expected. 
However, the percentage of T2 learners that participated in these activities is low, and just over half of 
them have seen Journeys materials (Figure 10). 
Figure 10. LARA SRGBV activities implemented in T2 schools, learners
A factor that might explain low levels of UKU undertakings is that Journeys and UKU activities were 
seen by teachers as extracurricular and outside of teachers’ responsibilities to deliver the national 
curriculum. Teachers interviewed highlighted a lack of additional pay to cover teacher change agents’ 
time preparing and carrying out Journeys activities to train other teachers and to serve as teacher 
patrons for UKU Groups. The lack of pay for extra activities and demands on teachers’ time was seen as 
unsustainable. Teachers were busy already preparing lesson plans and trying to ensure that their classes 
produced academic results, which are subject to district-level inspection.
In addition, LARA staff flagged the issue of Journeys activities lacking material resources in schools for 
pupils; UKU sessions needed paper, markers, etc., and these materials are not provided in schools, 
making difficult the implementation of activities. 
Despite evidence of low estimated participation in UKU groups (16% of pupils), and participation in a 
meeting or any other activity focused on violence (25% of pupils), teachers perceived (in 4 of the 6 
schools visited by the PE team) that the UKU group activities using the Journeys Activity Handbook for 
Pupils contributed to a more positive school climate. Teachers observed how participating “children 
turned to the [UKU] group to seek solutions to their problems” (Woman teacher, P4, KII) and 
highlighted how pupils spoke more freely with teachers and trusted them now that teachers knew how 
to speak in a friendlier manner with learners.
Findings from FGDs seem to indicate that UKU has had some benefits for teachers and learners. 
Teachers felt more knowledgeable about resources in the community responsible for children’s welfare. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 41
Learners were also aware of the resources in the community, but remained more trusting of the 
teachers in the school they previously identified. Teachers and learners reported that UKU groups 
allowed learners to speak more freely about issues that were taking place at home. Teachers felt that 
they could more openly talk to learners about their absences, and express concerns about performance. 
Some learners also felt that they could confide in teachers:
I may need shoes but my mother cannot afford to buy me shoes, yet I am contemplating going to dig for 
someone to get money, and yet that would mean that I miss class. So the [UKU] group members can 
contribute to my cause and put money together so that I can at least buy the really cheap shoes in the 
meantime while your parent tries to find the money. (Girl, FGD, L district)
Teachers also reported that having better knowledge of learners’ home environments enabled them to 
better understand the reasons why a learner may be exhibiting behavioral issues, or be absent from 
school. In these cases, teachers reported being less likely to use harsh disciplinary methods for 
infractions related to tardiness.
Each time you sit with your family, they can easily share with you their challenges. For example, that 
maybe we don't eat supper, or I arrive late at school because it is me who does the morning housework 
at home, so you get to understand that this child is coming to school late but not because of their own 
making but because the problem is at home. Instead of sending them to school, they tell them to do 
other things. So you get to know the challenges the child is facing. (Senior Male Teacher, FGD, L district)
4.2 RETENTION AND ATTENDANCE FINDINGS (R&A)
Establishing a positive and supportive school climate and strengthening SRGBV prevention and response
at school and in communities should increase the retention and attendance of pupils.
The following findings summarize data from the panel of 3,502 learners that we followed during eight 
waves of data collection in terms 2 and 3 of 2017 (called ‘Y1T2’ and ‘Y1T3’, respectively), terms 1, 2, 
and 3 of 2018 (called ‘Y2T1’, ‘Y2T2’, and ‘Y2T3’, respectively), and terms 1, 2, and 3 of 2019 (called 
‘Y3T1’, ‘Y3T2’, and ‘Y3T3’, respectively). The panel of learners was created at baseline (term 1 of 2017 
or ‘Y1T1’), when the learners were enrolled in P1 and P4. During the 2018 school year, R&A was being 
primarily carried out in P2 and P5 classrooms where most of the learners in our panel were found. In 
2019, R&A was mainly conducted in P3 and P6 classrooms, but we also visited P2 and P5 classrooms to 
locate learners who were not promoted. In addition, we also monitored the attendance of 142 teachers
assigned to the sampled learners. 
We find statistically significant differences in the learners’ enrollment status by treatment group in Y3T3 
– the last wave – among the Runyankore/Rukiga language area schools, but the data does not suggest 
important differences between treatment groups in the Luganda language area schools. Figure 11 shows 
that learners in T1 and T2 Runyankore/Rukiga area schools tend to stay enrolled in the original school 
at significantly higher rates than those from control schools. By Y3T3, 49 percent of learners in control 
schools remain enrolled in their original school, while the percentage for T1 and T2 is around 14 and 11 
percentage points higher respectively. Moreover, pupils in T1 schools are significantly less likely to 
transfer to other schools, and learners from T2 schools are significantly less likely to drop out, 
compared to those from control schools. One MoES official interviewee mentioned that some dropouts 
are re-enrolling in treatment schools when they hear about how the program is working.
42 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 11. Learners’ enrollment status in Y3T3 by treatment group and region
At baseline, in 2017, all learners were enrolled in P1 or P4 classes. For those that remained enrolled in 
the same schools, we were able to calculate grade retention. If all learners in the sample would have 
been promoted to the next grades, they should be attending P2 and P5 in 2018, and P3 and P6 in 2019; 
this group is defined as ‘enrolled in the corresponding grade.’ However, we observe that some learners 
were not promoted and they are categorized as ‘enrolled in lower grade.’ We also observe a small 
fraction of learners enrolled in a grade higher than the corresponding grade, and we classify them as 
‘enrolled in higher grade’. Figure 12 shows that there are no statistically significant differences in grade 
repetition by treatment group by Y3T3 in either region.
Figure 12. Grade retention in Y3T3 by treatment group and region
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 43
On average, 82 percent of learners still enrolled in the original school13 were present in class during our
unannounced visits. 
Absenteeism remains high among learners and unchanged since baseline. Figure 13 shows an average 
attendance rate of 81 percent in the Luganda speaking region and 83 percent in the Runyankore/Rukiga 
speaking region. We do not find statistically significant differences when comparing attendance rates by 
treatment groups in each language area.
Figure 13. Learners’ average attendance by treatment group and region
In addition, during our learners’ interviews, we asked them if they missed one or more days of schools 
in the week previous to the interview.
Figure 14 shows that, on average, 44 percent of learners answered affirmatively, which was consistent 
with the data from unannounced visits findings; absenteeism was significantly higher in the Luganda (54 
percent) than in the Runyankore/Rukiga (36 percent) region. We also observe that self-reported 
absenteeism is lower in T1 (35 percent) and T2 (31 percent) schools, compared to control schools (40
percent), in the Runyankore/Rukiga region. There are no differences in the Luganda region.
 
13 Attendance rates only consider learners enrolled in P1 or a higher grade, and employ the data collected in the eight school terms covered so 
far.
44 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 14. Learner self-reports being absent at least one school day on the week 
before the survey, by treatment and region 
Around half (50 percent) of those who were absent at least one day on the week before the survey 
stated that they missed school because of an illness (Table 20). Another important reason was staying at 
home doing work, a category mentioned by 22 percent of those who were absent. This is consistent 
with statements from KII respondents, who mentioned that absences tend to be highest during planting 
and harvesting seasons or market days 
Table 20. Learners’ reasons for not attending school1
REASON PERCENT OF LEARNERS
Was sick 50.0
Working at home 21.5
Had to take care of siblings 6.5
Bad weather 3.0
Customs/festivals 2.8
Did not have books or other school materials 2.2
Attended a funeral 2.2
Unpaid school fees 2.1
Other 10.5
1 only for those who self-reported being absent at least one day on the week before the survey
In the FGDs, caregivers widely mentioned that it is important for their children to attend school 
regularly. However, it is understood that there are occasions where a child must not attend school. The 
most commonly cited reasons for a child’s absenteeism include lack of money for school fees (all FGDs), 
although only 2 percent of the children reported this as a reason in our interviews. Caregivers also 
mentioned bad weather (five FGDs), fear of corporal punishment at school (five FGDs), deaths in the 
community, peer pressure to not attend school, dirty uniforms, caregivers needing assistance at home, 
teacher absenteeism, and domestic violence. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 45
In KIIs school staff noted that widespread poverty and needs for household economic strengthening 
affected pupil enrollment, attendance, and retention.
“Boys, in particular, are affected by casual work at home with matoke and cattle. Parents say, “This is 
how I can earn a living.” I cannot stop a parent from earning a living. Parents are very poor, and cannot 
afford notebooks, scholastic books, pens. Children, therefore, are not confident.” (Head Teacher, KII).
School staff further highlighted how girls’ and boys’ school absenteeism increased in the rainy season 
when they were expected to work to sow seeds and pull weeds. Critically, head teachers and teachers 
raised the issue of menstrual hygiene management, and the lack of resources to provide girls with 
sanitary pads to help them remain in school. Girls lacking transportation to and from school, experience 
safety risks on the way to and from school, when motortaxi (bodaboda) drivers try and convince them
“to go to discos and other recreational places.” This contributes to girls’ school dropout through 
absenteeism and early pregnancy.
In KIIs ministry officials felt that LARA contributed to improved enrollment, attendance, and retention 
rates, and reductions in dropout due to the program encouraging learners, “to keep coming.” Some 
highlighted how children who had previously dropped out of school ended up re-enrolling due to the 
materials and activities of the LARA program, often referring to the presence of EGR books as a 
motivating factor. Some of these opinions are corroborated by the evidence in the Runyankore/Rukiga 
area but not in the Luganda language dominant region. 
During the R&A unannounced visits, the field team also monitors the attendance of sampled learners’ 
teachers. Thus, we checked the attendance status of the teachers assigned to P1 and P4 in 2017, P2 and 
P5 in 2018, and P3 and P6 in 2019. 
Figure 15 shows that, on average, slightly over 80 percent of classrooms visited had a teacher present 
during instruction time. In the rest of the cases, the teacher was somewhere else in the school – in 
another classroom or not – or absent. We do not find significant differences between the teachers’ 
attendance status by treatment group in either region.
46 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 15. Teachers’ attendance status by grade
When we asked the pupils, 41 percent of them reported that teachers did not come to school one or 
more days on the week before the survey. Teacher absenteeism, as reported by learners, was 
significantly higher in the Luganda (45 percent) than in the Runyankore/Rukiga (37 percent) region. We 
do not find statistically significant differences between the treatment groups and the control group.
Figure 16. Learners report teacher was absent at least one school day the week 
before the survey
In the KIIs, the most frequent reasons cited by head teachers for teacher absenteeism were the illness of 
the teacher or a family member, or death of a family member. Additionally, MoES officials mentioned 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 47
teachers’ low motivation to attend school every day due to low salaries and/or head teachers’ 
disengagement.
In the FGDs, caregivers had mixed perceptions as to the frequency of teacher absenteeism. While they 
noted that teachers are frequently absent from the school, there are also cases where teacher 
absenteeism has been said to decline. Two group discussions mentioned that teacher absenteeism has 
recently declined after caregivers brought up the issue in school meetings and after the intervention of 
head teachers. According to caregivers, contributing factors to teacher absenteeism include apathy, 
sickness, the loss of a loved one, training in another town, family issues, and prioritization of personal 
businesses during school hours. Caregivers in all FGDs noted that sometimes when teachers are absent, 
substitutes teach their classes and children are not sent home. Often these are other teachers in the 
school or the head teacher. Participants in all FGDs feel that teacher absenteeism negatively affects their 
child’s reading progress. However, while it is a widely recognized problem, caregivers do not always feel 
like it is easy to confront teachers regarding their habits. One respondent expresses a fear that 
remarking on teacher absenteeism would bring retribution to the children:
Some teachers are rude and very tough on children, especially when children report that the teachers 
are missing from school on that particular day. When we follow up as parents, our children are 
victimized. So to keep our children safe, we decide to keep quiet. But teacher absenteeism is high. Some 
teachers can take a period of one week when they are away from school. And when they come to 
school, they do not explain work to the children well and this leaves our children missing (sic) simply 
because they do not understand the work. (Female Caregiver, FGD, RR District)
4.3 EARLY GRADE READING PROFICIENCY FINDINGS
The NORC evaluation team assessed P3 learners’ reading performance in T1, T2, and control schools in 
October 2019. In treatment schools, P3 learners that were promoted to the next grade every year have 
been exposed to LARA interventions since they started P1 in 2017. We examine 10 EGRA subtasks: (1) 
letter-sound identification fluency in local language, (2) letter-sound identification fluency in English, (3) 
segmenting, (4) non-word reading fluency in a local language, (5) oral reading fluency (ORF) in a local 
language, (6) ORF in English, (7) reading comprehension in a local language, (8) reading comprehension 
in English, (9) listening comprehension in local language, and (10) vocabulary in English. We assessed the 
effects of the T1 and T2 interventions separately for the Luganda and Runyankore/Rukiga samples.
Table 21 shows, for each EGRA subtask, the average scores in T1 schools, and the score differences 
between T1 and control schools, and the corresponding effect size14. These differences indicate the 
effect of the T1 intervention (i.e. LARA EGR intervention). We also show the average scores for T2 
schools as well as the scores differences between T2 and T1 schools and the corresponding effect size. 
In this case, the differences show the additional effect that the T2 intervention (i.e. LARA SRGBV 
intervention) had on learners’ reading performance was above and beyond the T1 intervention effect.
We show effects separately by language group. 
We find positive effects of T1 on most subtasks. The exceptions are oral reading fluency for the Luganda 
group, and reading comprehension in English, listening comprehension and vocabulary in English for both 
the Luganda and Runyankore/Rukiga groups Learners from T1 schools perform significantly better than 
 
14 The effect size is calculated as the difference between the T1 and control groups over the pooled standard deviation of the groups at midline.
48 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
their peers from control schools in terms of correct local language letter sounds per minute (effect size
of 0.91 and 0.54 in Luganda and Runyankore/Rukiga, respectively) and correct English letter sounds per 
minute (effect size of 0.81 and 0.62 in Luganda and Runyankore/Rukiga, respectively). We observe the 
same in the case of the percentage of correct phonemes (effect size of 0.26 and 0.18 in Luganda and 
Runyankore/Rukiga, respectively) and correct non-words per minute (effect size of 0.49 and 0.37 in 
Luganda and Runyankore/Rukiga, respectively).
The impacts of the EGR intervention are also significant in oral reading fluency (effect size of 0.54 and 
0.42 in Luganda and Runyankore/Rukiga, respectively) and reading comprehension (effect size of 0.51 
and 0.37 in Luganda and Runyankore/Rukiga, respectively) in local languages, but we find mixed results 
for the same subtasks in English. While there is no impact of T1 on reading comprehension in English, 
there is a positive and significant effect on oral reading fluency in English only for the Runyankore/Rukiga 
group (effect size of 0.2).
We find no significant effects of the T1 intervention on English vocabulary in the Runyankore/Rukiga 
language region, and that learners in control schools outperform those in T1 schools in the Luganda 
language region (effect size of -0.3). Finally, we do not find significant effects of the EGR intervention on 
listening comprehension for either language group. 
Table 21. EGRA Effects on P3 Learners
EGRA SUBTASK
T1 T2
AVERAGE 
T1 
SCHOOLS
INCREASE 
OVER 
CONTROL 
SCHOOLS
EFFECT 
SIZE
AVERAGE 
T2 
SCHOOLS
INCREASE 
OVER T1 
SCHOOLS
EFFECT 
SIZE
Luganda 
Letter sound (clspm) LL 16.6 10.3*** 0.91 13.9 -2.7* -0.23
Letter sound (clspm) EN 18.2 10.8*** 0.81 16.0 -2.2 -0.16
Segmenting (% correct) 72.8 9.9*** 0.26 73.1 0.3 0.01
Non-word reading (cnwpm) LL 12.4 5.0*** 0.49 12.7 0.3 0.03
Oral read. fluency (cwpm) LL 17.2 7.2*** 0.54 17.7 0.6 0.04
Oral read. fluency (cwpm) EN 22.7 2.9 0.16 24.2 1.5 0.08
Reading Comp (% correct) LL 27.1 11.6*** 0.51 26.9 -0.2 -0.01
Reading Comp (% correct) EN 21.9 0.1 0.01 25.5 3.6 0.13
Listening Comp (% correct) LL 91.5 -0.3 -0.02 93.4 1.8** 0.12
Vocabulary (% correct) EN 57.1 -4.2*** -0.30 58.2 1.1 0.08
Runyankore/Rukiga 
Letter sound (clspm) LL 10.5 4.4*** 0.54 12.8 2.4* 0.23
Letter sound (clspm) EN 9.7 5.6*** 0.62 11.3 1.7 0.15
Segmenting (% correct) 64.6 7.2** 0.18 66.9 2.3 0.06
Non-word reading (cnwpm) LL 9.2 3.5*** 0.37 10.7 1.6 0.15
Oral read. fluency (cwpm) LL 15.4 5.6*** 0.42 17.5 2.0 0.14
Oral read. fluency (cwpm) EN 24.5 3.8** 0.20 28.7 4.3** 0.22
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 49
EGRA SUBTASK
T1 T2
AVERAGE
T1
SCHOOLS
INCREASE
OVER
CONTROL 
SCHOOLS
EFFECT 
SIZE
AVERAGE
T2
SCHOOLS
INCREASE
OVER T1
SCHOOLS
EFFECT
SIZE
Reading Comp (% correct) LL 30.4 10.0*** 0.37 34.3 3.9 0.13
Reading Comp (% correct) EN 18.8 2.4 0.10 25.9 7.1*** 0.29
Listening Comp (% correct) LL 93.1 1.1 0.07 94.3 1.2 0.08
Vocabulary (% correct) EN 54.1 -1.3 -0.11 58.2 4.2*** 0.32
Legend: LL: Local language, EN: English, clspm: correct letter sounds per minute, cnwpm: correct non-words per minute, cwpm: correct words 
per minute. 
Notes: *** p<0.01, ** p<0.05, * p<0.1 
Effect size calculated as the difference between T1 and the control group (or T1 and T2) over the 
pooled standard deviation of the groups at midline. 
In general, we find no additional effects on learners’ reading performance due to the T2 intervention. 
Most differences between the T1 and T2 schools are not significant. This means that the SRGBV 
intervention does not provide an additional contribution to better reading outcomes above the EGR 
intervention alone. This is perhaps not too surprising given that, as we have shown, the LARA SRGBV 
activities had null or very small effects. 
Pooling together the T1 and T2 groups, Figure 17 andFigure 18 show that 26 and 30 percent of P3 
learners in the Luganda and Runyankore/Rukiga language regions respectively, cannot read a single 
correct word in the ORF (oral reading fluency) subtask. The percentage of learners from control 
schools with zero oral reading fluency scores is significantly higher: 41 and 45 in the Luganda and 
Runyankore/Rukiga regions, respectively. 
The figures also show that the proportion of learners able to read 20 or more correct words per 
minute is significantly higher in treatment than in schools for both language regions. In the Luganda and 
Runyankore/Rukiga language regions, 49.3 and 43.5 percent of learners in treatment schools read more 
than 20 words at the end of P3, respectively. The proportion of learners reaching this level is still low 
but substantially higher than in control schools, where 20.3 and 22.3 percent of P3 learners in Luganda 
and Runyankore/Rukiga language regions, respectively, do. 
50 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 17. Oral Reading Fluency (cwpm) in local language at midline, Luganda 
region 
Figure 18. Oral Reading Fluency (cwpm) in local language at midline, 
Runyankore/Rukiga region 
EGR EFFECTS FOR GIRLS AND BOYS
In addition, we conducted the above analysis separately for girls and boys. Girls tend to perform slightly 
better than boys in EGRA subtasks. This was true at baseline and it is also the case at midline. The 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 51
effects of the LARA program however are similar across groups. We included a detailed analysis of 
program effects and oral reading fluency performance for girls and boys separately in ANNEX C. 
Findings from FGDs indicate that collaboration between teachers and caregivers on learner reading 
progress has room for improvement. Caregivers had mixed responses as to whether or not they 
worked together with teachers to discuss the learning progress of their children. Some caregivers say 
they only go to schools when invited, and that they are brought in only when their children are not 
doing well or when school fees need to be paid. A number of caregivers also admit that they do not 
always take the initiative to check their child’s performance with teachers. (Five FGDs). 
Teachers do not consistently send reports home either. While most caregivers say they received some 
reports from teachers, and some noted that receiving progress reports was contingent on whether or 
not school fees were paid. Caregivers in two FGDs note that quality of instruction may improve with 
better treatment of teachers by the schools. 
“I can add that on teacher payment; teachers have to be paid in time because teachers cannot 
concentrate on the child when they are not paid well.” (Female Caregiver, FGD L District) 
“As one of my fellow parents spoke, I totally agree with the option and believe that our teachers need to 
be given good care because this translates into the teaching our children well.” (Female Caregiver, FGD L 
District) 
EVIDENCE OF FIDELITY OF IMPLEMENTATION OF EARLY GRADE READING ACTIVITIES 
In this section, we answer two main questions across all EGR activities undertaken by LARA. First, were 
activities carried out as planned and were they sufficient and relevant to achieving R1? Second, what 
factors accelerated or inhibited achievements of LARA results? Findings here are drawn from both the 
midline impact evaluation data and final performance evaluation. 
HEAD TEACHER AND TEACHER TRAINING 
Training head teachers and teachers to improve their reading instruction skills is one of the key activities 
implemented by LARA. The LARA target for “the number of primary school educators who completed 
professional development activities on implementing evidence-based reading instruction” was 16,733. As 
of September 2019, 17,918 teachers had received training in EGR methodology (2019 LARA AMELP, p. 
20). 
LARA’s cascade model started with LARA training the lead facilitators and conducting training-of￾trainers’ sessions for individuals who would provide training to classroom teachers. Trainers were 
identified among CCTs, district education staff, pre-service tutors including deputy principal outreach, 
adding previously trained head teachers and classroom teachers who demonstrated competence in the 
EGR methodology. The trainers then cascaded the training to classroom teachers, initially training P1, 
P2, and head teachers in Cluster 1 schools and over the following three years gradually added P3, P4, 
and deputy head teachers; Cluster 2 teachers starting with P1 and P2 teachers, followed by P3 and P4. 
Similarly, teachers in Cluster 1 control schools were added in Years 3 and 4. Over time, LARA adapted 
the program by refining the content and changing the frequency and duration of training from five 
continuous days to a 3-2-2 program to allow time to absorb the material and time between training 
sessions to practice. LARA made these changes before training P3 teachers. 
52 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Ministry officials in Kampala expressed concern that the quality of the trainers likely had a negative 
impact on teachers’ uptake. They reported that lead facilitators and trainers – selected collaboratively by
the Teacher/Tutor, Instructor Education and Training (TIET), LARA, and the Global Partnership for 
Education (GPE) – were initially selected more based on their position than relevant skills or interest in 
becoming trainers. This means that, for example, CCTs may serve as trainers even if they are trained 
mathematics or science teachers and have no experience teaching reading or with a particular local 
language. LARA is well aware of this and has weeded out weaker trainers over time, using those who 
have demonstrated competence in ERG methodology. LARA has identified ways to improve the training￾of-trainers, including adapting training guides to include more scripting.
In the interviews conducted by NORC, almost 70 percent of head teachers in T1 schools and 82 
percent of head teachers in T2 schools reported their participation in LARA EGR training in 2017 or 
afterwards (Figure 19). The proportion is significantly lower in the case of head teachers in control 
schools: 16 percent. Most likely, these head teachers received the MOES/USAID/RTI training when they 
were head teachers or teachers in a treatment school in previous years.
Figure 19. Head teacher participation in LARA EGR training
In the case of P3 teachers, around 80 and 75 percent of treated schools participated in the LARA 
training and refresher training, respectively (Figure 20). As expected, only a few teachers (3.9 percent) 
from the control schools have received training from LARA, and the difference in the participation rates 
between this and the treatment groups is statistically significant.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 53
Figure 20. P3 teacher participation in EGR training and refresher training
There is some turnover of teachers in the schools. Based on KIIs at the central MoES, districts, and 
schools, the frequency of transfers is a significant issue, as well as a weak handover/takeover practice in 
schools. According to head teachers and teachers interviewed, the capacity of the school is weakened 
when trained teachers are transferred out. We asked P3 teachers if they had taught in the same school 
in 2018, the year when the teachers in T1 and T2 schools were supposed to receive training from 
LARA. Figure 21, shows that while 64.5 percent of the teachers in control schools were in the same 
school in 2018, the figures go up to 85.7 and 76.6 percent in T1 and T2 schools, respectively. 
Figure 21. P3 teachers that taught in same school in 2018
All participants of the LARA EGR training found it useful and learned new things, according to the P3 
teacher survey. Around 93 percent of the respondents who received training from LARA think that the 
LARA EGR teaching approach is better than their usual approaches, and feel better qualified to teach 
***p<0.01, **p<0.05, 
*p<0.1
54 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
EGR after the training. In the KIIs, two MoES officials shared that, based on their belief that the EGR 
program is a success, they are pushing it to be fully integrated into pre-service training offered at 
Primary Teacher Colleges (PTCs). Another MoES official reported that LARA has participated in the 
development of the new EGR curriculum at Kyambogo University, but it has not been fully adopted yet.
Almost 99 percent of P3 teachers who participated in the LARA training say that they are implementing 
the EGR approach in the classroom but the teachers interviewed for the PE were seldom able to 
articulate what they had learned. One exception is that several teachers remember learning to beat the 
syllables in words. When prompted, they came up with other things, such as teaching letter names and 
sounds, repeating a word three times, pointing to pictures, making sounds, and making words.
In general, teachers report that LARA training was too short. Only 15 percent of the P3 teachers that
mentioned participating in the LARA EGR training thought that it was long enough. In the KIIs with 
teachers, they mentioned that the training’s length was not sufficient for them to absorb all the material 
and to practice, let alone master it. An inspector highlighted that training for P4 teachers should have 
been longer to better address the transition issues from the local language in P1-P3 to English as the 
medium of instruction in P4.
TEACHER SUPPORT SUPERVISION
LARA Field Assistants (FAs), who are former teachers, provide supervisory support and mentorship to 
primary teachers trained in EGR and are expected to visit teachers in their classrooms once per school 
term (three times a year in total). However, the LARA team reported that FAs can usually only visit 
two-thirds of all schools in each term as the number of FAs was reduced and those remaining have 
different responsibilities.
According to P3 teachers, support supervision by LARA is not widespread in treated schools. Around 
45 and 64 percent of teachers in T1 and T2 schools respectively were observed at least once teaching a 
P3 reading class by someone from LARA in 2019. In all cases, the LARA FAs provided feedback –
always or sometimes – about the teachers’ instruction practices. The rest of the teachers were not 
observed at any time during 2019 (
Figure 22).
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 55
Figure 22. LARA EGR support supervision for P3 teachers in 2019
More than 90 percent of those who received support and advice on teaching reading believed that it was 
useful or very helpful (Figure 23). On average, 7 percent of the teachers across T1 and T2 schools 
reported that the feedback offered by the LARA coaches needed improvement.
Figure 23. Opinion of P3 teachers about LARA EGR support supervision
In contrast, 73 percent of head teachers in the treated schools reported that somebody from LARA 
came to visit the school in every school term of 2019, and only 6 percent reported that nobody from 
LARA came during the same period. The difference between head teachers and teachers’ reports might 
56 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
be explained by the fact that LARA FAs visited schools during 2019 for other reasons than P3 teachers’
support, such as observing P4 classes. 
In addition to the supervision support provided by LARA directly, the program includes capacity building 
for district support supervision from CCTs. 
Figure 24 shows the frequency in which CCTs observed P3 teachers during 2019. Compared to control 
schools, T1 and T2 P3 teachers were visited by CCTs more frequently during the year, At least, 70 
percent of treatment P3 teachers were observed by CCTs once or more during the year, while this 
figure is around 40 percent in control schools.
Figure 24. Frequency that CCTs observed P3 teachers in 2019
Early in the implementation of LARA, FAs conducted joint monitoring with district education officials to 
build capacity for support visits. These were widely lauded as successful. At the time of our interviews 
with KIs, visits from district officials were less common. One teacher noted that the education official 
occasionally visits, and multiple teachers and head teachers mentioned visits by the district inspector. 
These inspections look at a wide range of things, including a review of lesson plans, but seldom account 
for classroom observation. Inspectors complete a form and write a report, and work with the head 
teacher to develop a school improvement plan. Inspectors also occasionally meet with teachers and 
share their observations. They rarely have time on their short visits to observe classrooms for any 
significant amount of time.
The extent to which instructional support agents within the education system are conducting support 
visits to teachers is related to three main factors, according to KIIs: engagement of head teachers, 
workloads of CCTs and inspectors, and transportation.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 57
Support supervision is a key responsibility of the head teacher, and LARA equipped them to provide 
such supervision. Head teachers participated in the same LARA EGR training as classroom teachers. 
However, based on interviews with head teachers and teachers, we found little evidence that they are 
using that training to support teachers through classroom observation and constructive feedback. The 
lack of engagement is evident in the level of absenteeism among head teachers and the fact that many 
are “caretakers” rather than permanent head teachers.
District officials and the one CCT interviewed reported that CCTs and inspectors must cover a large 
number of schools for support visits. For CCTs, this is in addition to their other responsibilities, such as 
providing in-service training. The CCT noted that he covers 23 government and 20 private schools. He 
typically has 15 days of the month to visit schools, and on a good day he can visit two schools and two 
teachers at each school. A district inspector noted that he and a colleague cover 78 primary, 20 
secondary schools, and two tertiary institutions. Compounding the issue of heavy workloads is transfers 
and attrition, as many CCTs and inspectors are not active or have retired. 
One MoES official noted that there is a renewed focus on improvement, including support supervision, 
emphasis on quality visits, and improved coordination between the MoES and district education offices.
The MoES through the Core PTCs is expected to provide CCTs with motorcycles so they visit schools. 
Unfortunately, not all CCTs are equipped with a motorcycle or given adequate support to fuel, maintain 
and repair their motorcycles. A MoES official acknowledged that schools are far apart and many are hard 
to reach. One CCT escorted the PE team to a school that took more than an hour to reach by car and 
reported that he was responsible for a school that was even further away. When questioned how far 
away the school is, he said he did not know because he has never been there.
TEACHING AND LEARNING MATERIALS
Teacher guides (TG)
LARA aims to distribute EGR TGs to all treated teachers. While LARA has generally been able to keep 
pace with their annual distribution targets, in 2019, they did not meet the target. This was primarily 
because some teachers did not show up for training where the TGs were distributed (2019 LARA 
Annual Report). We found that 10 percent of P3 teachers in T1 and T2 schools reported not having 
them (9 percent had no TG and 1 percent had another type of TG) (Figure 25). Teacher transfers might 
help to explain this issue as, based on KIIs with schools’ staff, some teachers take teaching and learning 
materials with them when they leave a school. In control schools, most teachers (almost 3 out of 4) do 
not have any TG, and no teachers have a LARA EGR TG.
58 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 25. Printed P3 teacher guide for lesson planning
Despite the vast majority of teachers in treatment have a LARA TG, its use in P3 classes is not all too 
common. In our classroom observation study in 2019, only 9 out of 23 teachers used the LARA TG. 
This contrasts with what was observed in P1 classrooms in 2018 where 19 out of 24 teachers used the 
LARA TG during the observed lesson. 
Among P3 teachers that have a LARA TG, almost 85 percent expressed a positive opinion – good or 
very good – of it, while the other 15 percent think the guide is fair or poor. Some teachers mentioned 
in the KIIs that the guide outlines steps, specifies time allocations, identifies questions, and gives 
examples. However, there was a general consensus that the LARA TG is not very user-friendly because 
(1) it requires a lot of cross-referencing within the guide itself, and between the TG and the pupils'
book, (2) the steps are too many to be followed easily, (3) the language in the English version is difficult 
to understand, and (4) it limits the opportunity for teachers’ creativity.
Additional shortcomings were identified in the KIIs with teachers regarding the LARA lesson plans for 
the Literacy Hour that LARA uses (which consists of a half-an-hour of reading and a half-an-hour of 
writing): 30 minutes a day for reading instruction are not sufficient so they often skip lesson steps and 
are not able to respond to the needs of learners. According to some MoES officials interviewed, the TG
complexities make them overwhelming for teachers. Despite the mentioned weaknesses, 90 percent of 
the P3 teachers surveyed that have LARA TG responded that they use them every day; the rest use 
them a few times a week.
The LARA TG training provides scripted lesson plans for each week of the three terms – 11 weeks for 
each term. In considering the macro pacing of the program, our analysis of classroom observation 
focused on whether the week in which the lessons were conducted matched the week specified in the 
LARA program. In 18 P3 treatment classrooms where there was uptake of the training (out of a total of 
23), 12 of the lessons were conducted in the LARA-specified week and 6 lessons were one week behind 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 59
in the program. In general, and despite the LARA TG complexities and shortcomings mentioned by KIs, 
our classroom observations across three years shows that overall adherence to the macro pacing of the 
program - where there is uptake - was not falling behind the plans too much and has improved over 
time, albeit measured at different grade levels. However, 9 of the 18 P3 lessons observed followed half 
or less than half the stipulated lesson steps and only one lesson covered all of the steps. The most 
commonly covered steps (the only ones covered in more than four of the classrooms) were the song 
and the teacher reading of the lesson passage aloud. Thus, although there was good adherence to the 
macro-level pacing of the program, with all lessons within a week of the stipulated week of the program, 
at the micro-level, there was moderate coverage of the lesson plan steps.
PUPIL READING BOOKS 
Pupil reading books were supposed to be distributed to every learner enrolled in treatment schools in 
P1-P4 classes. The end-of-project target for the number of pupil books distributed is 1,976,199 across
both cohorts. This includes books for P1-P4 in English and local languages. As of September 2019, LARA 
had distributed 1,942,497 books, which represents 98 percent of the target. However, we found that for 
a large proportion of P3 classrooms in T1 and T2 schools, not all children had a reading book (42
percent according to the P3 teachers, and 50 percent directly observed by enumerators). According to 
the LARA team, the number of learners in treatment schools has increased due to the success of the 
EGR program, so the classrooms do not have one book per pupil anymore. The same source points out 
the following as additional factors limiting the availability of reading books: locking up books in the 
schools—but not in the classrooms—to preserve them, and teachers allegedly taking books with them 
when they leave the school, and head teachers purportedly taking books (and other teaching and 
learning materials) to the private schools they own. 
According to the teachers, in 35 percent of the P3 classes in control schools, learners have no reading 
books at all (Figure 26). Most teachers (63 percent) indicated that in their classrooms, they have some 
type of reader that learners share. The percentage of teachers in treated schools who mentioned that 
learners have to share or do not have reading books is significantly lower than in control schools but 
still high: 48 and 36 percent in T1 and T2 schools, respectively. Almost all reading books in P3 treatment 
classes (99 and 97 percent of classes that have P3 reading books in the language of instruction in T1 and 
T2 schools, respectively) were brought by LARA.
The MoES’ policy is to put books in the hands of the pupils. Recognizing that teachers and head teachers 
want to prevent damage or loss, LARA trained head teachers and SMCs on book care and storage, and 
that training was cascaded to teachers, pupils, and caregivers. LARA provided each school with a metal 
box and padlock for storage. The books are numbered and inventoried at the end of each school day 
before being returned to the storage box. LARA told head teachers not to penalize caregivers if books 
are damaged or lost. Despite these efforts, there are cases of schools keeping books from pupils to 
preserve them because they are not sure when they will get new ones. There are also reports of 
schools enforcing a policy that if a book is damaged or lost, the caregivers have to pay. Because of this, 
there is a hesitation to take books home even though they might be required for homework.
60 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 26. Percent of classes where P3 learners have reading books in the language 
of instruction, as reported by P3 teacher
NORC enumerators checked the number of reading books in the hands of learners directly by asking
learners in P3 treatment classes to hold up their reading books. In around half the classes, all learners
had their own LARA reading books, and in 12 percent of the classes, no learners were able to show 
LARA reading books. In the rest of the P3 classrooms, only a fraction of the learners had their reading 
books.
Figure 27. Percent of P3 learners observed having LARA reading books
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 61
The fraction of learners that did not have the LARA reading book during class time was 25 and 23 
percent in T1 and T2 schools, respectively (Table 22). It is unlikely that an increase in enrolment alone 
explains this gap. Figure 27 shows that around 50 percent of the children with no reading book at hand 
were in classrooms where no learner had one and another fraction (around 20 percent) were in 
classrooms where more than half of the learners had no reading book. This fact points to a problem 
with the distribution of books to the learners, most likely the result of individual school’s policies, given 
that LARA delivered the materials to the schools. The finding also suggests that, although teachers or 
head teachers that leave the schools may take some materials with them, it is unlikely that they would 
take all the reading books available when leaving.
Table 22. Percentage of learners in class that had their LARA reading book
TREATMENT GROUP P3 LEARNERS IN CLASS
PERCENTAGE OF LEARNERS 
THAT HAD LARA READING 
BOOK
DURING CLASS (N)
T1 schools 3150 75.0% (2364)
T2 schools 3055 77.3% (2361)
The vast majority of teachers in treatment schools like the LARA reading books. Around 39 percent of 
P3 teachers in treatment schools think that the reading books available are very good, while a 
significantly lower share of teachers in control schools (20 percent) believe the same. Similarly, we find 
that around 10 percent of the P3 teachers in treatment classes think that the reading books available are 
fair while almost 29 percent of teachers in control schools think that (Figure 28). According to teachers 
who participated in the KIIs, the LARA reading books are appealing and easy to understand, and they are 
effective because they incorporate EGR principles in the local languages.
Figure 28. P3 Teachers’ opinion about LARA pupil reading book 
CONTINUOUS MONITORING ASSESSMENT (CAM) FORMS
62 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Continuous assessment is meant to inform teacher instruction and guide revision for the learner. The 
CAM forms are intended to support the teachers’ monitoring of learners’ progress in reading. LARA 
teacher training and teacher guides include detailed instructions for CAM forms daily use. However, 
LARA did not distribute the forms, instead, a model of it was included in the teacher guide that teachers 
were supposed to photocopy or use to create their own forms. This may present a difficulty for 
teachers, but more likely the challenge of conducting individualized, continuous assessment is more likely 
to be the issue. KIs’ opinion is that the forms are lengthy, which makes it challenging for teachers to 
use, especially in classrooms with dozens of pupils.
We found that most P3 teachers do not use CAM forms to monitor the progress of their learners. Only 
17 and 20 percent of P3 teachers in T1 and T2 schools respectively had CAM forms available. Among 
this small group of teachers that report having CAM forms, around 40 percent of the T2 teachers use 
them every day, and only 7 percent of T1 teachers do so (
Figure 29).
Figure 29. Frequency of use of CAM forms (for those with CAM forms)
During our P3 classroom observations, we saw little evidence of continuous assessment. Only 2 of the 
treatment classrooms teachers – out of 23 – could produce a CAM form. Both included entries for 
individual learners and also included marks for different components of literacy. Neither of the CAM 
forms was used in the course of the lessons, and only one was dated – for Term 3, 2019. It seems 
unlikely therefore that these are used to record a formative assessment of reading regularly. Although 
teachers were observed listening to individual learners read, the nature of feedback was restricted. One 
of the crucial methods of reading practice and assessment (I do, we do, you do) potentially contributed 
to a repetitive, chorused classroom discourse, empty of evaluative potential. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 63
SUPPLEMENTAL READING MATERIALS
LARA’s supplemental reading materials consist of story cards and reading books in English. The story 
cards are one page, double-sided and laminated for durability. LARA developed eight different story 
cards that are grade-level appropriate: four for P1 and P2, and four for P3 and P4. Field teams of trained 
interpersonal communicators met with caregivers to teach them how to use the story cards, even if the 
caregiver is not literate. LARA also distributed flashcards and caregiver-pupil cards to select schools.
The supplemental reading books in English were intended for pupils to take home to read to/with their 
caregivers. These books were secured through LARA’s cost-share effort and their partnership with 
Cotton on Foundation and Stanbic Bank, as well as through a partnership with Books for Africa. 
Recognizing that there is not a tradition of allowing pupils to take school materials home, LARA 
established a book lending practice and trained teachers on book management and lending practice. The 
one-day training program introduced the lending protocol and tried to reduce teachers’ anxiety about 
books being lost or damaged. However, based on information collected during school visits, it appears 
that there is still work that needs to be done to get teachers and head teacher comfortable lending 
books to pupils to take and use at home.
By the end of Year 4 (September 2019), LARA had distributed 694,158 story cards and 274,094 
supplemental reading books to 791 schools. Only a maximum of eight different story cards was given to 
each participating school – four for P1 and P2, and four for P3 and P4. This is more than they had 
before, but still quite limited exposure to this opportunity to stimulate and challenge young learners. 
Further, the story cards were printed in black-and-white, making them a bit less inviting. Distribution of 
supplemental reading books was even more limited, with only six districts receiving supplemental 
reading books. Again, more than schools had before, but not enough to realize a widespread impact.
During teacher interviews, we asked about the availability of these supplementary reading materials. 
Figure 30 shows the findings. Around half of the P3 classes in treatment schools received LARA 
supplementary reading books in English. We also find that 65 and 81 percent of T1 and T2 schools, 
respectively, received LARA story cards.
64 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Figure 30. Availability of LARA supplementary reading materials in English and 
story cards in P3 classes, as reported by the teacher
Although the story cards were intended to be used in the home rather than the classroom, we asked P3 
teachers about their use during classes, given that the low volume of materials and norms make lending 
difficult and unlikely. About 32 and 42 percent of teachers in P3 classes that have LARA story cards in 
T1 and T2 schools, respectively, use them “every day” or “a few times a week”. However, on average 
around 29 percent of T1+T2 teachers mentioned that they “rarely” or “never” use the LARA story 
cards. 
LANGUAGE USE
Through the provision of resources, the LARA program –in line with the MOES’ policy- promotes local 
language instruction in Literacy 1 in the first four grades. The use of language of instruction during this 
period was thus of interest in the classroom observation studies. Local language was used in all the P3 
lessons that incorporated the lesson plans. In the five lessons that did not, 3 were conducted mostly in 
English (with some use of local language), one was conducted entirely in English, and the other in the 
local language. In the control schools, none of the lessons were conducted in the local language, 4 were 
a mix of English and local language and 2 were in English only. As in the 2017 and 2018 studies, the 
dominance of English in the control lessons suggests the positive impact of the provision of materials and 
lesson plans in the local language as well as its emphasis in the training. There is slightly less use of 
English across the treatment group at the P3 level in 2019, however. 
READING PRACTICES IN THE CLASSROOM
Learners reported more reading activities performed in a regular school day in the treatment schools 
than in the control schools (Figure 31). The rate of P3 learners who silently read to themselves is 
significantly higher in the T1 (77 percent) and T2 (76 percent) groups than in the control group (70 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 65
percent). Similarly, the proportion of learners who read out loud to their classes is significantly higher in 
the T1 (82 percent) and T2 (83 percent) schools than in the control group (77 percent). 
Figure 31. Reading practices of P3 learners in the classroom
There are several factors that KIs believe affect the uptake of the methodology in the classroom. Among 
the most mentioned we found: weak support supervision, the complexity of the teacher’s guide,
teachers’ transfers, teacher attendance, pre-service training, the language of examinations, and quality of 
LARA training. No informant singled out just one factor; most mentioned several possible factors.
Based on the reports obtained during KIIs with teachers, it seems that teachers understand and can 
replicate some of the individual instruction techniques. The most common examples were beating
words and using the “I Do, We Do, You Do” approach, included in the teacher guides. The analysis of 
classroom observations in P1 classes in 2017 and 2018, and P3 classes in 2019 indicates that some of the 
LARA training was incorporated by teachers, with the use of lesson plans and LARA pupil reading books
were common across the majority of classrooms. Most lessons in treatment schools were structured 
by the training and in the majority of treatment classrooms, learners were exposed to text and engaged 
in different types of reading. However, opportunities to read extended text were also constrained, and 
given that Literacy 1 is crucially concerned with learning to read, this should be cause for concern. 
Classroom observations show a limited uptake of the methodologies advanced by the LARA program. In 
the case of each lesson where the training was incorporated, some components were omitted. What 
teachers included from the LARA scripted plans appeared to be those components amenable to 
repetition. Classroom discourse tended to follow patterns of highly repetitive, chorused repetitions of 
sounds, words, and sometimes paragraphs. More expository lessons or parts of lessons entailed a core 
antiphonal structure with ritualized closed questioning and chorused response. What was evident in the 
66 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
instructional content was that lessons entailed a very low level of cognitive demand. This was apparent
in the nature of questioning and classroom discourse, as well as the fact that the more difficult elements 
and steps of the program were left out. Theme discussions, picture discussions, and open-ended 
questions were absent or restricted. Thus, the vagaries (and possibilities) of learner talk were foreclosed 
and opportunities for learners’ oral language development were curtailed. The reading of the text was 
subject to excessive repetition such that it was evident in several settings that learners were learning to 
remember text rather than to read it. 
In addition, very little collaborative learning was evident. There was very little discussion of text or ideas 
and no evidence of scaffolding in the teaching. Continuous assessment was also not evidenced: there was 
a lack of feedback from the teacher on learners’ responses and no use of the LARA Cam form. There 
was no conceptual or linguistic building through the exchange, and no explicit indication when a 
learner’s production was partial or incorrect. Across all the classrooms only one learner was seen to 
ask a question, and this question was deferred by the teacher. 
Despite this, it was striking in the data the contrast between the treatment and control lessons. In the 
treatment classrooms, the LARA program established the potential for systematic teaching and learning 
of reading to occur. Most notable were the predictable pedagogical routines that were focused around 
text, as well as an appropriate progression of content (sounds, words, extended text) written into the 
week-by-week program. The absence of text, structure, and clear purpose in the control classrooms 
was clear, especially in the unpredictable and unfocused whole-class discussions. The lack of specified 
progression for literacy learning was evident in the very low level of content introduced in these 
classrooms. The high level of implementation fidelity in treatment schools potentially offers a good base 
on which to build, to extend teachers beyond excessive repetition, and to encourage more learner talk, 
oral language development, and engagement with the meaning of the text. 
UGANDA LITERACY CAMPAIGN
LARA implemented the Uganda Learning Campaign to promote parental involvement in their children’s 
reading practice. The campaign, which targeted caregivers, ran for three months from February to April 
2018. LARA trained field teams on interpersonal communication to engage with caregivers on home 
visits on four thematic areas of supporting their children’s reading: the value of education, the role of 
the caregiver, creating time for reading, and the caregiver-child relationship. LARA provided sub-grants 
to implement campaigns in 29 districts which included at least 20 drama shows in each C2 district, as 
well as home visits with caregivers. At the national level, the LARA team implemented two television 
commercials running daily for two weeks, one radio program weekly for one month, and four radio 
advertisements daily for one month.
Despite this, we found that almost 60 percent of head teachers across the treatment groups were not 
familiar with the ULC, and only around one third know its main message (it encourages caregivers and 
others to read with children at home). As expected, the proportion of head teachers in the control 
group who are familiar with the ULC and know its main message is significantly lower (8 percent)
(Figure 32). Similarly, most P3 teachers in treated schools are not familiar with the ULC: only 17 and 43 
percent of teachers in T1 and T2 schools, respectively, responded that they are familiar with ULC. On 
average, almost all – 93 percent – of those familiar with the ULC know its main message, although, in 
the case of the teachers who participated in the KIIs, none knew about the ULC. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 67
Figure 32. Head teachers’ awareness of the ULC
Even though the ULC targeted caregivers and it included home visits, caregivers participating in FGDs 
were not widely informed about LARA. Most had a faint idea of LARA activities, but there was one 
group where all but one person had no familiarity with LARA. Five focus group discussions agree that 
LARA carried a good message, and at least one member from all focus group discussions note that they 
had enacted some change in their behavior due to LARA. These changes included improved Luganda 
reading skills for learners, increased availability of reading material at home, and encouraging older 
learners in the household to support younger ones. Other benefits include caregiver motivation to 
undertake financial planning for their children. One caregiver from Luganda notes that LARA “…helped 
me to plan early from the meager resources at my disposal and pay school fees for the children in time…” 
(Male Caregiver, FGD, L District)
According to teachers interviewed, caregivers believe it is the teachers’ responsibility to teach children 
to read and the caregivers’ responsibility to pay fees for teachers to do their job. Also, according to a 
sub-grantee interviewed, caregivers are too busy in the evening hours preparing food and with other 
domestic chores to actively engage in their children’s learning. Therefore, any effort to engage caregivers 
must start with helping them understand the benefits of reading to/with their children so the children 
can practice what they are learning in school.
AT-HOME READING
We do not find the LARA intervention has an effect on the availability of books at home or at-home 
reading practices. On average, across all schools, 3 in 4 learners have books that they can read at home,
apart from schoolbooks. Also, around 94 percent of all learners interviewed reported reading at home, 
either every day or sometimes. Moreover, on average, 58 percent of learners have someone at home 
reading to or with them. This finding is consistent with the figures reported by the LARA team. 
68 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
According to KIIs with teachers, learners are not always allowed to take the story cards home. Even 
though LARA trained teachers on a lending protocol, head teachers and teachers fear that the story 
cards would get “spoiled” or lost so they do not feel comfortable lending them. Also, the lack of 
functional libraries in most schools makes it difficult to use any existing lending practices. 
Caregivers indicated that some children bring home textbooks and/or English books from the school to 
read. One caregiver noted that the textbooks provided by the school to take home are helpful when 
learners have assignments, or read by themselves. One focus group discussion respondent mentioned
buying school books at a market, but they were subsequently stolen. 
Two focus group discussions explicitly note that schools do not lend out books for learners to take 
home, saying that lending was restricted to older learners. Caregivers from both of these FGDs suggest 
that this was due to poor conditions at homes that would prevent the books from remaining intact. One 
FGD respondent went in-depth about these conditions:
“Some of the homes have leaking roofs so when it rains, books from school may be spoilt, others have 
rats, others have small babies who will play with the book and tear it apart while others just don’t know 
how to keep them … So some teachers are cautious” (Female Caregiver, FGD, RR District)
The aforementioned lending policy is particularly problematic for families that live far from school, and 
aren’t able to reference materials discussed in class for their homework. One caregiver notes:
“My children want to read; they have an interest at reading. However, they can only access the 
textbooks at school and when they get home, there is nothing for them yet they have an interest.”
(Female Caregiver, FGD, RR District)
At home, caregivers did not seem to possess a large variety of engaging and age-appropriate reading 
materials for their children. Bibles, hymn books, and old newspapers were frequently mentioned as 
being available to readers, with two groups saying they only possessed these materials and nothing 
supplemental. Other groups spoke of having books bought in-town or reading charts available to the 
learners in their households.
One of the most frequently repeated factors enabling reading (explicitly or otherwise) appeared to be 
the presence and involvement of caregivers who are invested in their child’s learning progress. 
Characteristics and behaviors of involved caregivers include checking learner homework/books at home, 
providing reading material and proper meals at home, having good relationships and communication with 
children, and engaging in income-generating activities or extra work to raise money for their children's
education. Even caregivers who admitted to lacking in education themselves noted that they still 
attempted to encourage their children to learn to read.
Challenging and prohibitive factors related to reading widely centered on household poverty and the 
frustrations that it brings. Lack of ability to pay school fees and general financial constraints were 
repeatedly mentioned, as well as the lack of food available for learners. Poor lighting in the evenings was 
also brought up frequently, as alternatives to candles such as paraffin, electricity, and solar power are 
preferred but are often too expensive to procure. Domestic responsibilities assigned to learners often 
get in the way of reading and homework time as well. Other issues that get in the way of reading 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 69
progress include fear of corporal punishment from teachers at school, an unstable home environment, 
and a lack of age-appropriate reading materials at home. 
“It is a greater booster for them to learn that we are walking the road with them and that we take their 
studying very seriously. Even if some of us may not read, the mere fact that you can ask a child to read 
for you or to look at his or her book in itself sends a powerful statement that you value and take what 
the child is doing at school with a lot of seriousness.” (Female Caregiver, FGD, RR District) 
There also seems to be relatively low participation in reading events by learners. Only one focus group 
discussion respondent mentions their child’s involvement in such events. However, it does appear that 
children have other opportunities to do public speaking in their communities in both Luganda and 
Runyankore/Rukiga. Members of four FGDs note that their children read aloud at church services, and 
the three Runyankore/Rukiga FGDs note that their children participate in debates at school. 
It should be considered, however, that caregivers who took the initiative to attend focus group 
discussions could be more involved and interested in their children’s academic progress and overall 
development than the typical caregiver. Caregivers in all focus group discussions explicitly state that 
promoting reading at home helps them in their efforts build a positive relationship with their child. 
70 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
5 CONCLUSIONS 
5.1 SRGBV 
Overall we see a very modest effect of R2 activities implemented by LARA. There is no statistically 
significant difference in learners’ and caregivers’ gender attitudes between T1 and T2 schools. In 
contrast, teachers in T2 schools show more gender-equitable attitudes than teachers in T1 schools, and 
the difference is statistically significant. The qualitative data also tends to show similar gender-inequitable 
attitudes to those found at baseline. 
We asked learners whether they felt safe on the way to and at school, asking if they agree with specific 
statements about their school climate. There is no statistically significant difference in school climate 
indicators between T1 and T2 schools. The differences between T1 and T2 go in the correct direction, 
that is, they are favorable for T2 schools. The exception is the statement “Learners feel safe on the way 
to and from school” where a slightly higher percent of learners say they feel safe in T1 schools 
compared to T2 schools. However, the differences are very small and not statistically significantly 
different. The qualitative data indicate that both girls and boys feel unsafe around the perimeters of 
school compounds that are unfenced, and afraid in the latrines area, albeit for different reasons. While 
boys are afraid of falling in or getting infections, girls remained afraid of peeping toms and sexual assault. 
Both genders reported feeling unsafe during the journey to and from school. 
We asked caregivers and teachers about violent and non-violent disciplinary methods they use with 
children. There is no statistically significant difference in the approach to discipline used by caregivers 
between T1 and T2 groups. Among teachers in T2 schools, there is a statistically significant decrease in 
the use of insults towards learners, refusing to speak to learners, or locking learners up as a form of 
discipline. The percentage of T1 teachers and head teachers who believe that hitting a learner with a 
cane or stick as an effective method of discipline at school is higher than in T2 schools, with the 
difference being statistically significant for teachers. These results, compared to those asking the same 
question using the term “corporal punishment” suggest that there is a disconnect in understanding what 
corporal punishment really is and suggests that teachers do not consider caning as corporal punishment. 
However, in KIIs with teachers in LARA T2 schools the head teachers spoke passionately about ending 
the harsh physical punishment of learners, and attributed their newfound skills in alternative discipline 
methods to LARA. 
We see no statistically significant difference in the total prevalence of violence between T2 and T1 
schools. When looking at the prevalence of different types of violence, we do see some statistically 
significant differences between T1 and T2 learners, though when analyzing so many outcomes (types of 
violence reported) we expect to have a certain percentage of false effects. For example, 59 percent of 
learners in T1 schools report being hit in school with objects such as a cane, stick, belt, or book by a 
teacher. In T2 schools, it is 52 percent and the difference is statistically significant. The frequency of this 
type of behavior is identical in T1 and T2 schools. Qualitative findings also indicate persisting physical 
and emotional violence for children. Children reported experiencing emotional violence in the form of 
bullying, nicknames associated with low levels of intelligence from teachers, and bullying about their 
physical features. Girls continue to experience bullying from teachers and other learners during their 
menstrual cycle. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 71
In both T1 and T2 schools, almost all learners report knowing to whom they can report violence. 
However, the proportion of learners that report having an adult that they trust to whom to report 
violence is lower: 49 and 58 percent in T1 and T2 schools respectively. The difference between T1 and 
T2 schools is statistically significant suggesting that LARA activities had an effect; however, the 
proportion of learners that the activity has affected positively is still below 60 percent. 
Evidence of implementation of Journeys activities in T2 schools, however, is very far from universal. For 
example, only 25 percent of learners participated in any activity related to violence prevention and only 
52 percent had seen any materials related to Journeys. Indicators on implementation for teachers and 
head teacher show higher rates of familiarity with Journeys materials and experience with Uganda Kids 
Unite groups. 
5.2 EGR 
At the end of 2019, LARA shows positive effects on the reading performance of cohort 2 P3 learners in 
Luganda and Runyankore/Rukiga dominant language regions. The effects of the program are similar in 
each region, giving us great confidence in the findings and, as expected, LARA had stronger effects in 
local languages than in English. We found no difference in reading performance of P3 learners between 
T1 and T2 schools which indicates that SRGBV-related program components brought no additional 
benefit to learners’ reading ability. 
LARA has benefited learners with both low and high reading skills. An improvement in reading 
performance was found across groups of learners with different reading abilities. EGRP reduced the 
number of zero scores among P3 learners and also increased the percentage of learners that reach at 
least 20 cwpm in the oral reading fluency subtask. This is an important achievement; however, reading 
performance remains, on average, quite low for the treated learners at the end of P3. Learners decoding 
skills are low; on average, learners in T1 and T2 schools, can identify less than 14 correct letter sounds 
in a minute. Reading ability is low as well. Over a quarter of P3 learners benefitted by LARA are non￾readers – they cannot read a single word from a short grade 2 level paragraph – and on average oral 
reading fluency is around 17 words per minute, which is far from the levels needed to be able to 
comprehend the text read. 
Nevertheless, the improvement due to LARA EGR activities is significant. The progress achieved by the 
program is a good base on which to build. With that aim, we explored in detail the different 
components of LARA ERG activities to identify what is working well and what needs to be improved. 
The strengths of the LARA program are evident when comparing instructional reading practices 
between treatment and control schools. Many, although not all, treatment classrooms implemented the 
program on the day of classroom observation. Having LARA reading books in the lessons supported a 
greater engagement by learners with text and more opportunities for reading extended text. Learners in 
control classrooms and treatment classrooms not implementing the program read no extended text, but 
rather engaged in choral recitation of single words or short text written on the blackboard. LARA also 
offered opportunities to engage with literacy-specific skills and content around the phonetic, semantic, 
and syntactic aspects of language. Lessons tend to focus on the letter, syllable, word, sentence, and 
extended text levels, although around half completed 50 percent or less of the lesson plan for the day. 
The more challenging aspects of the program were left out. P3 teachers are more likely to follow 
aspects of the program that resembled closely the structure of P1 and P2 lessons than those that 
72 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
require additional and more challenging work on vocabulary and language structure. Lessons lack 
opportunities for learners to read an extended text, particularly to read silently and independently. 
The fact that these aspects were often not exploited does not diminish the importance of explicit 
instruction of learners in how to read in the program. In control and treatment classrooms not 
implementing the program, instruction focused on content topics (health, culture, gender) where the 
informational content relevant to these topics was emphasized, rather than literacy-specific skills and 
content. 
Neither CAM forms nor alternative methods are used by P3 teachers to conduct a continuous 
assessment of learners. Sometimes, teachers listen to individual learners read, but their feedback is quite 
restricted. One of the crucial methods of reading practice and assessment (I do, we do, you do) 
potentially contributed to a very repetitive, chorused discourse, empty of evaluative potential. 
The above observations suggest that further work is needed during teaching training and support 
supervision visits to instill rich, engaging, and motivating language in the classroom as the fundamental 
basis for becoming literate. Support supervision visits are less frequent than they should be. In particular, 
support supervision by CCTs and district education officers is very low. This seems to be a consequence 
of lack of time, lack of means to reach the schools, and competing responsibilities. Head teachers are 
not engaged in support supervision either. This creates an important challenge that was already present 
in USAID/SHRP – the activity that preceded LARA. If support supervision is not properly and credibly 
embed in the education system, it will not be sustainable and most likely will disappear once LARA ends. 
LARA distributed plenty of reading books among treatment schools that were very well-received. 
However, in many classes, reading books are not in the hand of learners as they should. This seems to 
be the result of school policies that keep all or some of the learner’s books in storage perhaps to avoid 
them being damaged, undermining the full potential of the program. The lack of opportunities that 
learners have to handle books and experience individual reading is reinforced by the lack of reading 
materials at home. Supplementary reading materials and reading cards are scarce and the lending system 
for the few materials available does not work well. It is hard to overestimate the importance of having 
access to quality, print text in the development of children’s literacy and learning, and in the creation of 
positive behaviors and habits that support reading development. 
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 73
6 RECOMMENDATIONS 
A number of recommendations stem from our findings: 
SCHOOL-RELATED GENDER-BASED VIOLENCE
Improve the school climate. Make the school climate safer for learners is within the domain of head 
teachers and teachers’ responsibilities. More can be done here in two key areas. 
 School infrastructure: Many learners continue to feel unsafe in school due to a lack of boundary wall 
and latrines that lack locks on the doors. Future programs on SRGBV that include funding for a 
boundary wall may result in quick payoffs. At a minimum, it is easy to insist that school latrines have 
functioning locks to ensure program support. 
 Change head teacher and teacher behavior: Train all teachers in the school on positive, non-physical 
disciplinary methods, SRGBV prevention and response, not victim-blaming survivors of violence, and 
effective, non-re-victimizing communication through guidance and counseling. Going beyond an 
SBCC approach, start with teacher discussion groups to reflect on gender norms, as well as school 
and community expectations for teachers’ behaviors that underpin SRGBV. Groups can work 
toward developing a shared understanding of the negative effects of SRGBV on child development 
and academic achievement and work toward new, positive norms that define a safe, supportive, 
nurturing, and reliable teacher. Embedded within a norm change approach could include a nested 
SBCC campaign that sends a clear message that it is acceptable to talk with children respectfully, use 
non-violent, positive disciplinary methods, and not use harsh physical or psychological punishment. 
On-going and in-service training for teachers is required beyond the limited implementation of one 
program; the cascading model of training does not seem to be working.
Include SRGBV prevention activities as part of the school hours, rather than having SRGBV as an 
extracurricular activity. Undertake and test a pilot that changes the approach towards SRGBV, for 
example all grades P1-P7 could have one hour of SRGBV prevention per week.
SRGBV encompasses a difficult and large range of issues to address as one of multiple program 
components. Future programs that take on this topic should have this as an explicit focus, with activities 
that address the problem from several angles directly, including social and gender norm change 
strategies at a peer reference group level, and school-wide and community levels specifically addressing 
norms that underpin and perpetuate SRGBV in order to evaluate their contributions toward change in 
the incidence of pupil SRGBV exposure. Journeys focuses largely on socio and emotional learning with 
SRGBV interspersed throughout; a more targeted and streamlined message with more time to get 
through the material is critical. 
Revise the program that relies on district support supervision. District officials do not have the time and 
resources to visit the schools for support supervision for SRGBV activities. Given that the current 
system is not working and there are challenges to changing this system, work with the head teachers to 
empower them and have them provide support supervision within their schools. 
Work at the national level with MoES: Ministry officials understand the importance and seem open to 
the idea of including SRGBV prevention and response in their primary teacher college curriculum that 
focuses on teacher development and management. This is important and will have long term impact on 
reducing SRGBV. Additionally, explore working with MoES to change the current Reporting, Tracking, 
74 | UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY - USAID.GOV
Referral and Response process whereby only head teachers report cases of SRGBV cases to the LC or 
district. If all school teachers are allowed to report an incidence it can prevent head teachers from 
becoming gatekeepers and interfering with child violence survivors seeking justice. 
EARLY-GRADE READING
Teacher training and support: The high level of implementation fidelity in treatment schools potentially 
offers a good base on which to build, to extend teachers beyond excessive repetition and to encourage 
more learner talk, oral language development and engagement with meaning of text. We recommend to 
include more discussion and more demonstration of these activities during training and during support 
visits. In particular, conducting demonstrations in the actual classroom could be very valuable for 
teachers. Some MoES officials, based on their belief that the EGR program is a success, are pushing it to 
be fully integrated into pre-service training offered at Primary Teacher Colleges (PTCs) and Universities. 
This type of action is promising and should be carefully considered in future programs.
Improve teacher support supervision: A larger fraction of teachers in treatment schools received more 
frequent support supervision than those in control schools; however, there are still many teachers that 
do not receive supervision at all or do not receive it frequently enough to make it useful. In particular, 
support supervision by head teachers, CCTs and district education officers is very low. Evidence 
suggests that including follow-up classroom visits and teachers support increases learning gains (see for 
example, 2018 World Development Report). We recommend exploring this challenge and focus on 
how to effectively scale support supervision within the education system to ensure sustainability of the 
program, given that the current approach is not working.
Continuous Assessment. LARA or future programs need to revise its approach to train teachers in 
conducting continuous assessment of learners. CAM forms are not being used, teachers are not creating 
alternatives and even oral feedback to learners seems insufficient.
Putting reading books in learners’ hands. Additional work and training needs to be done with head 
teachers and teachers to persuade them that reading books and supplemental materials are only useful if 
they are in the hands of the children. This should be reinforced during support supervision visits. 
Reading materials to take home are insufficient and schools are reluctant to lend them to the children. In 
the future, it would be worth considering alternatives to create inexpensive products, for example 
newsprint materials, which even if not designed to last through the years, can be given to children to 
read at home or in school.
USAID.GOV - UGANDA PERFORMANCE AND IMPACT EVALUATION FOR LITERACY ACHIEVEMENT AND RETENTION ACTIVITY | 75
7 POLICY BRIEFS 
We present in this section four policy briefs that summarize the P&IE’s findings and recommendations 
regarding the following topics:
 The High Absenteeism of Teachers and Learners in Uganda 
 Why Lule still can’t read at the end of P3? 
 Teachers’ implementation of LARA in classrooms
 Ending Teacher-to-Learner Violence as an Entry-point to Addressing School-related Gender￾based Violence 
Acoli 
Ateso 
Leblango 
Lhukonzo 
Luganda 
Lugbarati 
Lugwere 
Lumasaaba 
Lusoga 
Ngakarimojong 
Runyankore/Rukiga 
Runyoro/Rutooro 
50% 
54% 
64% 
59% 
49% 
57% 
62% 
45% 
55% 
49% 
49% 
34% 
43% 
53% 
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
SHRP Schools 2016 LARA Schools 2019 
UGANDA
 
The High Absenteeism of 
Teachers and Learners In Uganda 
Learners’ and teachers’ absenteeism are negatively associated with school performance. Previous 
work in Uganda and other countries have established this link. Learner absenteeism in Uganda 
is very high. Figure 1 shows the proportion of learners that reported having been absent at least 
one school day in the week previous to the interview by language dominant area among sample 
schools. The percentages of absenteeism reported are very high in all the language subgroups. 
Figure 1. Percentage of learners that report being absent at least one day in the 
previous week, by language dominant area 
SHRP and LARA schools do not overlap. Sources: Menendez (2017) and Menendez et al. (2020)1 
More than half the learners in the LARA sample that self-reported being 
absent in the previous week indicated that they were absent two or 
more days (57 and 59 percent in Luganda and Runyankore/Rukiga respectively). 
1
The School Health and Reading Program (SHRP: 2012 to 2017) and the Literacy Achievement and Retention Activity (LARA: 2016 – 2021) 
were early grade reading programs in Uganda funded by USAID. NORC at the University of Chicago (NORC) conducted the Performance 
and Impact Evaluations for both these programs as an external third party evaluator funded by USAID.
1 | the high absenteeism of teachers and learners in uganda 
 
19% 16%
Learners absent in Learners absent in Source: 
Luganda areas Runyankore/Rukiga areas Rojas et al. (2020) 
In addition to asking the learners about their school attendance, NORC conduced unannounced visits to 
schools of the LARA sample every term (there are 3 in a year) during the period 2017-2020 in the Luganda and 
Runyankore/Rukiga language dominant areas. Figure 2 shows the percentage of learners that were absent on 
the visit day by language dominant area. We found that in any given day, almost 18 percent of learners were 
absent. Learner absenteeism was slightly worse in the Luganda language dominant area but both regions are 
similar. These findings from the unannounced visits are consistent with the learners’ self-reports. 
Figure 2. Daily learner absenteeism 
Daily absenteeism is similar for boys (18 percent) and girls (17 percent). 
It is also quite similar across all primary grades. 
In addition to high absenteeism, it is known that many learners do not start school on the first day 
of classes -particularly in the first term of the year- and that in the last weeks of the year classes are 
interrupted by days devoted to exams and also extra-curricular activities and therefore learner’s 
absenteeism increases. This is so common that data collection in schools is never done during the first 
or last weeks of the school year to avoid biased samples. This also indicates that there are behaviors 
regarding school attendance that are not justified by work, illness, weather, etc. 
Cause of absenteeism 
Learners mostly report missing school because of illness, followed by needing to do work at home 
and market-day activities. Caregivers, in contrast, mention reasons such as children stopping on 
the way to school, flooded roads, lack of scholastic materials, lack of money for school fees, 
school climate, and teachers’ absence. Some caregivers also admit to keeping their children at 
home so they can help during planting and harvesting periods. Teachers also mention farm 
work as reason to miss school days. However, with exception of the beginning and end of the school 
year, we did not find differences in attendance rates through the year, suggesting that sowing and 
harvesting work are not the main reasons behind the absence of almost 20 percent of the students 
every day. If working at the farm were the driver, absenteeism would be more evident at certain 
times of the year and among older learners. 
Surely, work at the farm and home, market-days, illnesses and bad weather are all factors that 
preclude children from attending school every day; however, there is lot of room to improve 
attendance if all stakeholders are persuaded of the importance of every child going to school 
every day, from the first to the last day of classes. 
2 | the high absenteeism of teachers and learners in uganda 
 
Acoli 
Ateso 
Leblango 
Lhukonzo 
Luganda 
Lugbarati 
Lugwere 
Lumasaaba 
Lusoga 
Ngakarimojong 
Runyankore/Rukiga 
Runyoro/Rutooro 
35% 
43% 
28% 
34% 
30% 
41% 
30% 
32% 
39% 
22% 
41% 
46% 
56% 
35% 
57% 
63% 
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
Percentage reported by: Learners, SHRP 2016 Learners, LARA 2019 Teachers, LARA 2019 
Teacher absenteeism is also very high. Learners in our studies not only reported on their own 
absenteeism but also their teachers’. Specifically, they were asked whether there was any day in the last 
week when their teacher was not present. If the majority of learners reported that the teacher was absent 
in the previous week, we coded him/her as absent and otherwise as present. We also asked teachers 
to report if they had missed at least one day of class in the previous week. In Figure 3 we show teacher 
absenteeism as reported by learners and teachers, by language. Figures indicate that absenteeism 
among teachers is high and consistent with previous studies. 
Figure 3. Percentage of teachers absent at least one day in the previous week 
Teacher is considered absent if more than 50% of the learners in the classroom report so. SHRP and LARA schools do not overlap. 
Sources: Menendez (2017) and Menendez et al (2020) 
Around one in three teachers reported being absent at least one day in the 
week previous to the interview (Luganda and Runyankore/Rukiga dominant 
language areas, 2019). 
3 | the high absenteeism of teachers and learners in uganda
Teacher in Assigned teacher Assigned teacher in Assigned teacher 
classroom teaching another class school, not teaching absent 
83% 2% 7% 8%
 
World Bank research based on unannounced visits to Ugandan public schools found that less than 40 
percent of the teachers were actually in the classroom teaching learners, and 24 percent of the teachers 
were not present in the school. During the unannounced visits that NORC conducted for the LARA 
evaluation, we also gathered information about teachers’ presence in schools (Figure 4). In our school 
sample –which is not nationally representative- the average daily teacher absenteeism was 8 percent, 
which is high although below the national average in the World Bank report. 
Figure 4. Teachers’ attendance and presence in the classroom on the day of 
NORC school visit 
It is a big deal! 
High absenteeism implies that the probability of having the teacher and learners 
simultaneously in the classroom is very low. An optimistic calculation using the information 
from the NORC’s unannounced visits in Luganda and Runyankore/Rukiga dominant language 
areas suggests a probability of 70 percent. 
Absenteeism First and last 
weeks of the 
school year 
180 
School days 
126 
School days 
116 
School days 
This means that if the school year has approximately 180 school days, absenteeism 
reduces it to only 126 days. If the first and last week of the year are also missed, then 
the school year is further reduced to only 116 days. In some other areas of the country, 
absenteeism rates are higher and therefore there are even fewer effective school days. 
Should we be surprised when the learners do not perform at the expected grade level? 
4 | the high absenteeism of teachers and learners in uganda 
Non-monetary incentives could be used to recognize learners and their families that 
start the school year on time and show high attendance. 
The teachers noted that they are not recognized for their contribution to learning 
which has led to low teacher morale. Teachers and head teachers could be motivated 
through non-monetary incentives like provision of certificates of attendance and 
recognition on open days. In addition, on time salary payment should be the norm. 
Recommendations 
A lot of progress has been made in increasing school enrollment. Now, while we try to improve 
education quality, is the time to work hard to reduce absenteeism. 
References 
Najjumba, Habyarimana, and Bunjo (2013) Improving Learning in Uganda, Volume 3: School-Based 
Management: Policy and Functionality. Washington DC. World Bank. 
Menendez, Alicia (2017), USAID/Uganda School Health and Reading Program (SHRP): Result 1 School Level 
Interventions. Impact Evaluation-Final Report, NORC at the University of Chicago 
Menendez, A., R. Nayyar-Stone, I. Rojas, C. Fierros, L. Onyango, and S. Downey (2020) “Uganda Performance 
and Impact Evaluation for Literacy Achievement and Retention Activity (LARA) Midterm impact and final 
performance evaluation report” NORC at The University of Chicago. USAID. 
Rojas, I., A. Menendez and R. Nayyar-Stone (2020), Retention and Attendance Data Report. NORC at the 
University of Chicago 
Wane, Waly and Gayle H. Martin (2013) “Education and Health Services in UGANDA Data for Results and 
Accountability Service Delivery Indicators, Education and Health” Washington DC. World Bank 
For more information, contact NORC at the University of Chicago: 
Ritu Nayyar-Stone, nayyarstone-ritu@norc.org | Alicia Menendez, menendez@uchicago.edu 
5 | the high absenteeism of teachers and learners in uganda 
Efforts should be directed to raise awareness about the importance of attending 
school every day since the very early grades. Production and broadcasting of 
messages to parents and communities supporting school attendance are strongly 
recommended. This should be an intensive and long campaign. 
UGANDA
79% 
87% 
75% 
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
2017 (P1) 2018 (P1) 2019 (P3) 
87% 
87% 
79%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
of primary school through the implementation of an early grade reading program. Predictable 
instructional routines focused on graded texts and an appropriate progression of content 
(sounds, words, extended text) written into the week-by-week program. NORC at the 
University of Chicago conducted a third-party performance and impact evaluation of 
aimed to improve the mother tongue and English reading abilities of children in the first four years 
The USAID/Uganda Literacy Achievement Retention Activity (LARA: April 2016 - April 2021) program 
Why Lule still can’t read at 
the end of P3? 
? 
LARA and found that after three years of exposure to the program, there was a positive 
effect on the reading performance of P3 learners in the treatment schools. However, 
the level of reading was found to be still very low. Over a quarter of LARA P3 learners 
could not read a single word from a short P2 level paragraph. At the end of P3, 
learners’ decoding skills were still very low and the average oral reading fluency (ORF) 
among P3 treated learners was 17 correct words per minute (cwpm) in their local language, 
far from the levels needed to be able to comprehend the text read (Menendez et al., 2020). 
Why is it that Lule still can’t read? 
NORC’s evaluation included a classroom-based research study (in term 3 of 2017, 2018 and 2019) 
to assess the implementation of LARA in schools and understand how it shapes teaching practices 
in P1 and P3 classrooms. The results of this study shed some light on why, despite relatively high 
implementation fidelity, the program has yet to show results in getting learners to read. 
High fidelity 
Percentage of treatment lessons implementing the LARA program 
Pecentage of treatment lessons conducted within one week of LARA’s proposed timeline 
(NORC 2018, Hoadley 2019 and Hoadley 2020) 
1 | why lule still can’t read at the end of p3? 
5 | the high absenteeism of teachers and learners in uganda
 
 
We identify five classroom-based issues that potentially undermine reading acquisition. 
1 | Lule doesn’t have the opportunity to read on his own 
When it occurred, reading in all classes was constituted as a communal, oral performance with the 
teacher emphasizing pronunciation. It was highly repetitive leading to memorizing of the text. This is 
distinctly different from what the ultimate goal of reading is – to individually and silently retrieve meaning 
from text. In P3, individual ‘whisper’ reading is one of the steps of the program, but was seen in only two 
classrooms. Linked to this communal reading, the blackboard was the dominant text learners read from 
in 60 percent of the treatment classrooms. Further, opportunities to read extended text (longer than a 
sentence) was limited. In nine of the 23 P3 treatment lessons, there was no reading of extended text by 
learners in the course of the lessons and they were not read to by the teacher. Opportunities to read 
extended text were far fewer at the P3 level in both treatment and control school lessons than they were at the 
P1 level in 2018 (Hoadley, 2020). Only through opportunities to read on his own will Lule learn to read. 
There is now widespread agreement that 
learning vocabulary, developing oral language 
skills, and acquiring background knowledge are 
as important to learning to read as the tasks of 
learning letters, sounds, decoding and fluency. 
These aspects were undeveloped or absent in 
the observed LARA lessons. 
2 | Lule’s lessons focus on constrained skills 
It is relatively straightforward and rewarding to teach young children phonological awareness and 
phonics, as these are constrained skill domains and children can generally master them. Skill domains 
like vocabulary and comprehension are much larger, less well-defined, and thus less likely to be taught 
in many instances. They are, however, crucial determinants of long-term literacy success. The more 
difficult and open-ended elements and steps of the program were left out. In P1, the dominant focus 
was on phonics and the extensive repetition of sounds (NORC 2018 and Hoadley 2019). These lower 
order decoding skills took the place of teaching vocabulary, engaging in understanding the meaning 
of text and discussing themes and background knowledge. At the P3 level, vocabulary and language 
structures were taught superficially. Rather than understanding the meaning of word segments 
(morphological awareness), the focus was on identifying and counting syllables. Comprehension 
questions were infrequent, closed and low-level. Emphasis on constrained skills for Lule came at the 
cost of developing fluency and comprehension. Extensive practice in phonemic awareness and phonics 
also didn’t translate into high levels of decoding skills in assessments at the end of P3 (Hoadley 2020). 
2 | why lule still can’t read at the end of p3? 
3 | Classroom talk is restricted and repetitive 
Classroom discourse generally consisted of simple chained sequences of teacher initiation and learner 
chorused response. In P1, teacher initiation entailed a closed question or close phrase requiring a 
single word response from learners. This was a highly ritualized form of instruction, with the two￾move exchange structure precluding feedback on learner responses (Hoadley 2019 and NORC 2018). 
Likewise, most of the talk at P3 level consisted of repetitive chants of sounds, words and on occasion a 
paragraph. There was very little discussion of text or ideas. Across all the classrooms only one learner 
was seen to ask a question, and this question was deferred by the teacher (Hoadley 2020). Low levels 
of oral language ability and poor overall language skills have consistently been linked to reading 
comprehension difficulties (Snow, 2017). Lule needs to engage in rich classroom talk that provides 
opportunities for meaning-making, extends understanding and models rich vocabulary. 
Across observations of 
93 lessons, only once did a 
learner ask a question! 
Lara Pupil book. Term 1, Week 10. 
4 | Lule is not motivated to read 
The reading of passages in the classrooms was very repetitive, with little 
time expended on getting learners to engage in the contextual meaning 
of the text and its relation to learners’ prior and new knowledge. In one classroom the class read a 
single passage 23 times. By the end, most had memorized it. In another lesson, the teacher began 
the lesson by asking pupils to close their eyes and recite the passage from the previous day. Reading 
was often construed in this way – as remembering text rather than retrieving meaning from it, thus 
undermining Lule’s motivation to read. Further, the themes and topics across lessons were very local, 
reflecting Lule’s everyday world. Relying so heavily on the community’s knowledge base has limited 
interest and motivation for Lule (Menon et al, 2019). Community-based knowledge should be invoked 
in order to move Lule beyond his everyday understandings of the world to give him the background 
knowledge and vocabulary to read widely and think beyond his existing experience. 
3 | why lule still can’t read at the end of p3? 
 
A novel research methodology 
When measuring classroom practices at scale, it is common to use closed-ended instruments 
requiring relatively low inference judgments on a range of features of instruction. The NORC 
classroom observation studies from 2017 to 2019 used a novel methodology that collected 
both process and input data by designing a tool that included both closed-ended items and 
open-ended narrative descriptions of classroom activity. To gain a ‘thicker’ description in the 
narrative record, two trained observers each produced a description of the same lesson. The two 
descriptions at the point of analysis were then read together. In addition, the closed-ended part 
of the tool was completed after the lesson by both fieldworkers. The mixed method approach was 
used to obtain a more complete understanding of what was going on in the classrooms. At the 
point of analysis, it was used to confirm the quantitative measures with qualitative accounts and 
to explain some of the quantitative results. 
5 | Lule’s teacher doesn’t know that he can’t read 
No informal or formal continuous assessment monitoring (CAM) of reading was seen in any of the 
classrooms. Across the years many teachers could not produce a CAM form for their learners. While 14 
teachers produced a CAM in 2017 and 15 in 2018, only 2 teachers produced a CAM in 2019. In addition, 
given that most reading and oral activity occurs through whole class chorusing, it would be difficult for 
teachers to identify individual learners’ reading levels. Although the teachers were observed to circulate 
amongst learners at times while they read, the purpose was invigilatory rather than instructional. Lule’s 
teacher did not know that he could not read. 
References 
NORC (2018) “UGANDA Performance and Impact Evaluation for Literacy Achievement and Retention Activity 
(LARA) Classroom Observation Report 1” NORC at The University of Chicago. USAID 
Hoadley (2019) “UGANDA Performance and Impact Evaluation for Literacy Achievement and Retention Activity 
(LARA) Classroom Observation Report 2” NORC at The University of Chicago. USAID 
Hoadley, U. (2020) “UGANDA Performance and Impact Evaluation for Literacy Achievement and Retention 
Activity (LARA) Classroom Observation Report 3” NORC at The University of Chicago. USAID 
Menendez, A., R. Nayyar-Stone, I. Rojas, C. Fierros, MC. Schulte, L. Onyango, and S. Downey (2020) “Uganda 
Performance and Impact Evaluation for Literacy Achievement and Retention Activity (LARA) Midterm impact 
and final performance evaluation report.” NORC at The University of Chicago. USAID. 
Menon, S., Kutty, S. S., Apte, N., Basargekar, A., & Krishnamurthy, R. (2019). Powerful reforms in early language and 
literacy instruction in India. In Improving Early Literacy Outcomes (pp. 191-206). Brill Sense. 
Snow, C. (2017). Early literacy development and instruction: An overview. In The Routledge international 
handbook of early literacy education : A contemporary guide to literacy teaching and interventions in a 
global context, eds. Natalia Kucirkova, Catherine E. Snow, Vibeke Grøver, and Catherine McBride-Chang, 5-13. 
Abingdon, Oxon; New York, NY: Routledge. 
For more information, contact NORC at the University of Chicago: 
Ursula Hoadley, uk.hoadley@uct.ac.za | Alicia Menendez, menendez@uchicago.edu 
4 | why lule still can’t read at the end of p3? 
UGANDA
 
 
Treatment Control
Language Local Use of local language was 
dominant EN Use of English was 
dominant 
Exposure to text The vast majority of No Learers had any 
learners had their own reading books 
reading book 
Learning reading 
opportunities
Choral reading of text by 
learners (declines in P3) 
No learners read any 
extended text 
implementation of an early grade reading program. To this end, the program includes the provision 
of technical assistance to teachers and structured materials for use in classrooms – including a Pupil 
Book and a Teacher Guide. In the randomized control trial conducted by NORC at the University of 
tongue and English reading abilities of children in the first four years of primary school through the 
The USAID/Uganda Literacy Attainment and Retention Activity (LARA) aims to improve the mother 
Teachers’ implementation of 
LARA in classrooms 
Chicago, significant differences were found between treatment and control groups in relation to 
learner outcomes (Menendez et al., 2020). A classroom observation study (n=31) conducted across 
three years also found large differences in teachers’ instructional practices across the groups and 
found relatively high implementation fidelity. And yet, reading levels remained very low amongst the 
treatment group, with over a quarter of P3 learners not able to read a single word from a short P2 level 
paragraph after three years of exposure to the program. 
How did LARA classrooms differ from controls? 
Differences between treatment and control classrooms, 2017, 2018, and 2019 
1 | TEACHERS’ IMPLEMENTATION OF LARA IN CLASSROOMS 
5 | the high absenteeism of teachers and learners in uganda
P1 Classrooms P3 Classrooms 
• Phonics and phonemic awareness •	 Focus on syllabic awareness, but 
formed the focus of most lessons. pertaining to sounds rather than the 
•	 The dominant practice was rote meaning of word segments. 
recitation of syllables and sounds •	 Identification of single letters was also 
decontextualized from words or evident at this level as well as double 
sentences. consonant blends, even though these are 
not part of the lesson plans at this level. 
 
In the LARA classrooms, most notable were the predictable pedagogical routines that were focused 
around text, as well as an appropriate progression of content (sounds, words, extended text) written 
into the week-by-week program. The absence of text, structure and clear purpose in the control 
classrooms was clear, especially in the unpredictable and unfocused whole class discussions. The lack 
of specified progression for literacy learning was evident in the very low level of content introduced in 
these classrooms. In short, lessons in LARA classrooms were more structured, coherent and contained 
appropriate grade-level content. 
A novel classroom observation methodology 
When measuring classroom practices at scale, it is common to use closed-ended instruments 
requiring relatively low inference judgments on a range of features of instruction. The study 
designed a tool that included both closed-ended items and open-ended narrative descriptions 
of classroom activity. In order to gain a deeper description in the narrative record, two 
fieldworkers each produced a detailed description of the same lesson. The two descriptions 
at the point of analysis were then read together. In addition, the closed-ended part of the 
tool was completed after the lesson by both fieldworkers. In this way judgments required in 
the closed-ended items were subjected to a form of inter-rater reliability at the point of data 
collection. The result? A more complete understanding of what was going on in the lessons 
and more reliable judgements on particular dimensions of instructional practices. 
What did LARA teachers focus on? 
What teachers included from the LARA scripted plans were constrained skills and those 
components amenable to repetition. Different steps of the program were delivered 
in a mechanistic, procedural way through very repetitive chorused routines around 
sounds, syllables, words and sentences. 
2 | TEACHERS’ IMPLEMENTATION OF LARA IN CLASSROOMS 
? 
What did LARA teachers leave out of the program? 
Very repetitive readings of text dominated the classrooms, with little time for oral language 
development. The focus in the texts on local topics and learners’ everyday worlds meant that 
the development of new background knowledge and vocabulary was limited, and learners’ 
understanding of the world was circumscribed to their local environment. Theme discussions, picture 
discussions and open-ended questions were absent or restricted. Thus, the vagaries (and possibilities) 
of student talk were foreclosed and opportunities for students’ oral language development was 
curtailed. Excessively repetitive reading of text took up the majority of time in most LARA classrooms. 
There was a significant decline in the opportunities for learners to read, for learners to be read to and 
for the teacher to hear learners read across the three years. The form of reading also did not change 
as learners progressed – for example, from the teacher reading aloud to learners to learners reading 
individually silently. All forms of reading declined from 2017 to 2019. ‘Whisper reading’, a key element 
at the P3 level focused on individual silent reading, was practiced in only a few classrooms. 
Teachers avoided the more challenging and complex steps in the lesson plans. Where these were 
attempted, the teacher’s own knowledge (in say, retrieving the central idea from a text or identifying 
a grammatical feature) appeared lacking. Although the teachers appeared familiar with most of 
the instructional routines, they did not appear confident in the instructional content. The very 
procedural and often mechanical way in which the program is delivered suggests that teachers need 
support in understanding the different program components and their underpinning principles. 
Teachers leave out Summary of teacher uptake 
the more complex, •	 No differentiation of pupils • Very little discussion of text challenging and •	 Very little collaborative • Limited development of open-ended parts of learning	 print awareness 
the lesson plans. •	 Restricted assessment • Highly repetitive chorused 
of the reading of text by readings of short text 
teacher 
How did the teachers understand the core LARA methodology? 
The LARA program adopts a particular methodology called ‘I do, we do, you do’. The idea is that 
new content is modelled for the learners by the teacher ‘I do’. Then learners and teachers practice 
the content as a class ‘we do’. Finally, individual learners practice the content on their own ‘you do’. 
The focus of ‘we do’ and ‘you do’ is to assess whether learners are able to produce the given content 
and identify where support is required. What occurs in practice is a very repetitive practicing (of 
phonics, words, a passage) by teachers and learners. The ‘gradual release’ aspect of ‘I do, we do, you 
do’, methodology becomes repetitive rather than progressive in the learning process, and is 
stripped of its evaluative component. It is likely that this method is grafted onto existing classroom 
discourse, which is repetitive and involves extensive oral chorusing. Thus, what was observed in many 
of the classrooms was repetitive reading of text in a call-and-response pattern reminiscent of the 
ritualized chorus exchange found across many similar classroom contexts (Hoadley 2018; Kewaza and 
Welch 2013; and Ssentanda 2014). Observations of learners producing or practicing lesson content on 
their own or with peers was very rare. 
3 | TEACHERS’ IMPLEMENTATION OF LARA IN CLASSROOMS 
 
Reading or remembering? 
At the beginning of a P1 lesson in 2017 most of the learners are unable to 
read the short passage introduced from the Pupil book. The teacher reads 
the passage with the class in chorus 7 times. She corrects words that are 
mispronounced. She then asks groups of learners to read, and the passage 
vs. is read another 4 times. Finally, she selects individuals and then rows to read 
11 times. Finally, the passage is read successfully by the whole class with the 
teacher. The passage is read a total of 23 times, so that by the end the whole 
class is able to recite the passage. It is likely that those who were unable to 
read the passage at the start have now memorized the words rather than 
developed the ability to read the passage. 
How can instructional practice be improved? 
Program principles and methodology Teachers need further development in 
two key areas. First, understanding the principles underpinning the program – and 
the purpose behind the different steps. Second, understanding the gradual release 
methodology and its difference from mere repetition. 
Individual reading Pupils need more opportunities to read on their own, 
especially at P3 level. The teacher also needs to hear and assess pupils read on their 
own and a simple mechanism to record reading levels needs to be made available to 
the teacher. 
Oral language development Teachers need support in conducting theme 
discussions and asking questions. Developing their own background knowledge 
could foster greater confidence in allowing more open discussion in the classroom. 
References 
Hoadley, U. (2018). Pedagogy in Poverty: Lessons from 20 years of curriculum reform in South Africa. London: 
Routledge. 
Kewaza, S. & Welch, M. (2013). Big class size challenges: Teaching reading in primary classes in Kampala, 
Uganda’s central municipality. US-China Education Review, 3, 15, 283-296. 
Menendez, A., R. Nayyar-Stone, I. Rojas, C. Fierros, MC. Schulte, L. Onyango, and S. Downey (2020) “Uganda 
Performance and Impact Evaluation for Literacy Achievement and Retention Activity (LARA) Midterm impact 
and final performance evaluation report.” NORC at The University of Chicago. USAID. 
Ssentanda, M. (2014). The Challenges of Teaching Reading in Uganda: Curriculum guidelines and language 
policy viewed from the classroom. Journal of Applied Language Studies, 8, 2, 1– 22. 
For more information, contact NORC at the University of Chicago: 
Ursula Hoadley, uk.hoadley@uct.ac.za | Alicia Menendez, menendez@uchicago.edu 
4 | TEACHERS’ IMPLEMENTATION OF LARA IN CLASSROOMS 
Physical Violence 81% 
88% 
Emotional Violence 75% 
81% 
40% 
39% 
Sexual Violence
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
 Girl learners (n=1,485) Boy learners (n=1,419) 
UGANDA
 
Ending Teacher-to-Learner Violence 
as an Entry-point to Addressing 
School-related Gender-based Violence 
There is high prevalence of school related gender based violence (SRGBV) in Ugandan primary 
schools, including emotional, physical, and sexual violence, mostly from teachers and peers 
(Menendez et al., 2020). The Ugandan Penal Code (Amendment) Act of 2007 abolished corporal 
punishment and made sexual violence and defilement of persons under 18 years of age a criminal 
offence subject to life imprisonment on conviction. All forms of violence are detrimental to learners 
and need addressing in Ugandan schools. This brief makes the case that that ending teacher-to￾learner violence, which correlates strongly with learner exposure to all forms of SRGBV by peers, can 
become a catalytic entry point for SRGBV prevention and accountability, within a holistic framework 
of strategies at policy, district, school, and community levels. 
SRGBV is highly prevalent in Uganda’s primary schools 
As much as 81 percent of girls and 88 percent of boys reported experiencing physical violence at 
school by any type of perpetrator in the previous year. Proportions of learners reporting experiencing 
school-related emotional violence were only slightly lower, at 75 percent for girls and 81 percent for 
boys. A similar percentage of boys and girls reported sexual violence exposure, with both teachers 
and peers reported as perpetrators. Girls reporting of sexual violence exposure increased with age, 
and the share of 11 to 12-year-old girls reporting sexual violence exposure (38 percent) is 4 percentage 
points lower than 13 to 16 years old girls (42 percent). 
Figure 1. Percentages of learners who experienced SRGBV in the past year 
1 | ENDING TEACHER-TO-LEARNER VIOLENCE AS AN ENTRY-POINT TO ADDRESSING SCHOOL-RELATED GENDER-BASED VIOLENCE 
57% 
67% 
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
Girls Boys (Menendez et al., 2020) 
58% 
72%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
Women Men
Figure 3. Percentage of girls and boys surveyed corroborated high levels of physical 
violence by a teacher 
5 | the high absenteeism of teachers and learners in uganda
Corporal Punishment is a form of physical violence 
More than 75 percent of nine to 16-year olds reported past-year physical violence from a teacher 
in Uganda’s National Violence Against Children Survey (Ministry of Gender, Labour, and Social 
Development, 2015). Yet, as noted above, legislation abolished corporal punishment in Uganda in 
2007. In 2015, the Ministry of Education and Sports issued a circular stating violence against children 
in schools, institutes, and colleges had been outlawed in the Constitution, the Children Act, the Penal 
Code, the Domestic Violence Act and the Employment Act. The circular set a nation-wide education 
sector requirement that schools replace corporal punishment with positive, non-violent discipline 
and learning actions. Regulations have been under development since 2016 to implement laws 
against corporal punishment in schools. 
An external evaluation by NORC at the University of Chicago of the Literacy Achievement and 
Retention Activity (LARA: April 2016 – April 2021) in Uganda, found that 88 percent of teachers reported 
believing that corporal punishment is not an effective form of discipline (Nayyar-Stone et al., 2016; 
Menendez et al., 2020). However, many schools continue to tolerate and normalize teacher-to- learner 
physical violence. 
Figure 2. Percentage of women and men teachers admitting to hitting learners 
with an object such as a stick, cane, or belt in the previous year 
Violence against children in schools is gendered, including corporal punishment (Schulte, 2020). 
In LARA schools, girls and boys expressed fear of being caned for giving wrong answers, having 
low grades, or arriving late (Menendez et al., 2020). Girls reported lower frequencies of physical 
punishment from a teacher and often described that teachers caned them with five strokes for a
	
given incident, versus boys who reported up to 20. Still, girls reported more often than boys that they
	
avoid attending school at times due to fear of harsh physical punishment. Higher proportions of girls 
than boys reported experiencing sexual violence by a teacher, while boys may be yet more likely to 
stay silent about sexual violence. 
2 | ENDING TEACHER-TO-LEARNER VIOLENCE AS AN ENTRY-POINT TO ADDRESSING SCHOOL-RELATED GENDER-BASED VIOLENCE 
-
Physical Violence 62% 
70% 
72% 
78% 
Emotional Violence 
Sexual Violence 35% 
36%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 
 Girl learners (n=1,485) Boy learners (n=1,419) 
Teacher-to-learner physical violence correlates strongly with learner 
exposure to all forms of SRGBV by peers 
Teachers and school staff are the most commonly reported perpetrators using physical violence 
against learners. Where teacher-to-learner physical violence is highly prevalent, learners report high 
levels of exposure to all forms of SRGBV (emotional, physical, sexual) by peers. 
Figure 4. Percentages of learner exposure to peer perpetrated violence in schools 
LARA evaluation results indicate that high levels of teacher-to-learner physical violence correlate 
strongly with high levels of peer-perpetrated violence in Ugandan primary schools. Similar studies 
have found associations between attitudes supportive of violence against children from school staff 
and learner exposure to physical and emotional violence by peers (Wandera et al., 2017). 
Harsh physical punishment in schools normalizes cycles of power abuse and violence by authority 
figures as societally condoned ways to resolve conflict. Teacher-to-learner violence creates a school 
climate that tolerates and perpetuates SRGBV by other teachers and learners’ peers. Community 
norms that frame caning or beating children as necessary for child development, need to be 
replaced with new positive norms valuing non-violent disciplinary methods, positive teacher-learner 
communication, emotionally supportive classroom management techniques, and physical and 
mental health in child development. Given that a high proportion of teachers surveyed reported 
seeing corporal punishment as an ineffective form of discipline, many teachers may be receptive to 
and uphold positive changes in classroom practices for learner safety. 
What can make schools safer for learners? 
Previous evaluations in Uganda and Jamaica provide strong evidence that teacher-to-learner 
physical violence can be prevented within a holistic, multi-component program. In Uganda, 
an evaluation of the Good School Toolkit, a complex behavioral intervention designed by 
Raising Voices, reported a 42 percent reduction in the likelihood of primary school learners 
suffering physical violence from school staff (Devries et al., 2015). In Jamaican pre-schools, an 
evaluation of the Irie Classroom Toolbox reported, “fewer counts of violence against children 
by teachers in the intervention schools compared with control schools at post-intervention” 
3 | ENDING TEACHER TO-LEARNER VIOLENCE AS AN ENTRY-POINT TO ADDRESSING SCHOOL-RELATED GENDER-BASED VIOLENCE 
(Baker-Henningham et al., 2019). Devries and Naker (2021), called for building a programs -
based, policy-informing research agenda to reduce teacher-to-learner physical violence. 
Prioritizing education sector planning, policy and program evaluation to focus on reducing 
teacher-to-learner violence is an evidence-based first step in reducing SRGBV. 
Recommendations 
Recommendations will require close cooperation between the Ministry of Education and Sports, and 
the Ministry of Gender, Labour, and Social Development, for coordinated investment and action within 
and beyond the education sector at national, district, school, and community levels. 
1. Create a nation-wide advocacy campaign promoting new norms supporting 
positive, non-violent teacher communication and disciplinary methods. 
Raise awareness of the positive benefits for healthy child development and 
achievement. Clearly communicate that teachers caning and beating children in 
schools is illegal, socially unacceptable, and harmful to child development. Co-design 
the campaign with diverse range of communities in multiple languages, and use 
print, broadcast and online media for community-level campaign reach. 
2. Plan for in-service teacher trainings on positive, non-violent behavioral 
interventions and learner support, including several full-day workshops spread 
over one school year for all teachers and head teachers, not just senior teachers. 
Combine trainings with SRGBV prevention and accountability-dedicated in-class 
technical support and mentorship to each teacher at least once per month to 
promote safe classrooms and academic success. 
3. Build accountability pathways for teacher-to-learner violence: 
•	 Create targeted 2-3 day compulsory trainings for teachers who use violence 
against learners. Make the training interactive, using role-play, and continuous 
standardized assessment. Facilitate sessions covering laws and regulations 
abolishing physical and sexual violence against children; positive, non-violent 
disciplinary methods; and emotionally supportive classroom management 
techniques. Make successful training completion a requirement to retain the
 
license to teach in schools, colleges, and institutions.
 
•	 Implement and enforce laws and regulations against violence against children in 
schools. Pursue strategic litigation and judicial sentencing for emblematic cases of 
severe teacher physical and sexual violence against learners. Ensure child violence 
survivors’ confidentiality, recovery support services (e.g., psychosocial counseling, 
medical treatment, legal aid, safe house), and protective accompaniment 
throughout the investigation and court proceedings. Stop transferring teachers 
who use severe forms of violence against learners to another school, as it only 
shifts the violence and its harmful effects on children elsewhere. 
4 | ENDING TEACHER-TO-LEARNER VIOLENCE AS AN ENTRY-POINT TO ADDRESSING SCHOOL-RELATED GENDER-BASED VIOLENCE 
 
References 
Baker-Henningham, H., Bowers, M., Francis, T., Vera-Hernández, M., Walker, S.P. 2019. Evaluation of a Violence￾Prevention Programme with Jamaican Primary School Teachers: A Cluster Randomised Trial. Int. J. Environ. 
Res. Public Health 16, no. 15: 2797. https://doi.org/10.3390/ijerph16152797 
Devries, K.M., Knight, L., Child, J.C., Mirembe, A., Nakuti, J., Jones, R., Sturgess, J., Allen, E., Kyegombe, N., 
Parkes, J., Walakira, E., Elbourne, D., Watts, C., Naker, D. (2015). The Good School Toolkit for Reducing 
Physical Violence from School Staff to Primary School Students: A Cluster Randomised Controlled Trial in 
Uganda. Lancet Global Health. 3:7; e378-e386. https://doi.org/10.1016/S2214-109X(15)00060-1 
Devries, K.M., and Naker, D. 2021. Preventing Teacher Violence Against Children: The need for a research 
agenda. The Lancet. Vol. 9. April 2021. https://www.thelancet.com/pdfs/journals/langlo/PIIS2214-
109X(21)00093-0.pdf 
Menendez, A., Nayyar-Stone, R., Rojas, I., Fierros, C., Schulte, M.C., Onyango, L., Downey, S. (2020) Uganda 
Performance and Impact Evaluation for Literacy Achievement and Retention Activity (LARA) Midterm impact 
and final performance evaluation report. NORC at The University of Chicago. USAID. 
Ministry of Gender, Labour and Social Development. (2015) Violence against Children in Uganda: Findings 
from a National Survey. The Republic of Uganda. UNICEF. https://www.unicef.org/uganda/media/2156/file/ 
Violence%20Against%20Children%20Survey%202018.pdf 
Nayyar-Stone, R., Menendez, A., Schulte, M.C., Pancratz, S., Owen, R., Onyango, L. (2016). Baseline Results from 
the Evaluation of USAID Uganda Literacy Achievement and Retention Activity (LARA). NORC at the University 
of Chicago. USAID. 
Schulte, M.C. (2020). Corporal Punishment in Schools in Low-income Countries: An Evidence Review. Select 
Gender-based Violence Literature Reviews. NORC at the University of Chicago. USAID. https://pdf.usaid.gov/ 
pdf_docs/PA00X1XZ.pdf 
USAID. 2018. Education Policy. https://www.usaid.gov/sites/default/files/documents/1865/2018_ 
Education_Policy_FINAL_WEB.pdf 
Wandera, S.O., Clarke, K., Knight, L., Allen, E., Walakira, E., Namy, S., Naker, D., Devries, K. (2017). Violence 
Against Children Perpetrated by Peers: A Cross-sectional school-based survey in Uganda. Child Abuse & 
Neglect. 68 (2017) 65-73. https://doi.org/10.1016/j.chiabu.2017.04.006 
For more information, contact NORC at the University of Chicago: 
Ritu Nayyar-Stone, nayyarstone-ritu@norc.org | Marie-Celine Schulte, marie-celine.schulte@charite.de 
5 | ENDING TEACHER-TO-LEARNER VIOLENCE AS AN ENTRY-POINT TO ADDRESSING SCHOOL-RELATED GENDER-BASED VIOLENCE 
U.S. AGENCY FOR INTERNATIONAL DEVELOPMENT
1300 Pennsylvania Avenue, NW
Washington, DC 20523