July 2017
This publication was produced at the request of the United States Agency for International Development. It was 
prepared independently by NORC at the University of Chicago. 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.
EVALUATION
BASELINE REPORT: USAID/LIBERIA READ 
LIBERIA IMPACT EVALUATION
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | i
USAID/LIBERIA READ 
LIBERIA IMPACT 
EVALUATION 
BASELINE REPORT 
November 2017 
PN 7617.016.01 
DISCLAIMER 
The author’s views expressed in this publication do not necessarily reflect the views of the United States 
Agency for International Development or the United States Government. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | ii
TABLE OF CONTENTS 
Acronyms .......................................................................................................................................vi
Executive Summary ...................................................................................................................... 1
Evaluation Purpose and Evaluation Questions ....................................................................................................... 1
Project Background ...................................................................................................................................................... 1
Evaluation Design, Methods and Limitations .......................................................................................................... 1
Findings and Conclusions ............................................................................................................................................ 2
Evaluation Purpose ....................................................................................................................... 4
Evaluation Purpose, Audience and Intended Use .................................................................................................. 4
Evaluation Indicators .................................................................................................................................................... 4
Early Grade Reading Assessment ......................................................................................................................................... 4
Additional Instruments ............................................................................................................................................................. 5
Project Background ...................................................................................................................... 5
Evaluation Methods & Limitations .............................................................................................. 6
Evaluation Questions .................................................................................................................................................... 6
Methodology Overview ............................................................................................................................................... 6
Experimental Design ................................................................................................................................................................ 7
Limitations ................................................................................................................................................................................... 7
School and Student Sample ......................................................................................................................................... 8
School Selection ......................................................................................................................................................................... 8
Student Selection ....................................................................................................................................................................... 8
Sample Size Calculation .......................................................................................................................................................... 8
Baseline Data Collection ...........................................................................................................................................10
BASELINE RESULTS ................................................................................................................. 11
Baseline Descriptive Results from the Baseline Survey .....................................................................................11
Student Sample Description .................................................................................................................................................11
EGRA Scores .............................................................................................................................................................................12
Student Reading Activities at Home and School .............................................................................................................17
Students Absenteeism and General School Performance .............................................................................................21
Parent Attitudes and Household Context .........................................................................................................................22
Teacher Characteristics and In-Class Activities................................................................................................................25
Principal and School Questionnaire ....................................................................................................................................35
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | iii
The Determinants of EGRA Scores ....................................................................................................................................39
Conclusions & Recommendations ............................................................................................. 41
References .................................................................................................................................... 42
Annexes ........................................................................................................................................ 43
Annex I: Evaluation Statement of Work ................................................................................................................44
Annex II: Evaluation Methods and Limitations .....................................................................................................55
Annex III: Data Collection Instruments .................................................................................................................57
Annex V: Disclosure of any Conflicts of Interest ............................................................................................. 137
Annex VI: Full Regression Results ........................................................................................................................ 141
Annex ViI: egra subtasks distrbution graphs ...................................................................................................... 144
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | iv
LIST OF FIGURES
Figure 1: Distribution of student age by boys and girls. .........................................................................................11
Figure 2: Correct Familiar Words per Minute, Oral Reading Fluency, and Reading Comprehension 
score (primary axis) and percentage of zero scores (secondary axis). ................................13
Figure 3: Distribution of Oral Reading Fluency, correct words per minute. ....................................................14
Figure 4: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), and 
Reading Comprehension score (secondary axis) by boys and girls. ......................................14
Figure 5: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), and 
Reading Comprehension score (secondary axis) by urban and rural areas. .......................15
Figure 6: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), and 
Reading Comprehension score (secondary axis) by age group. .............................................16
Figure 7: Student is read aloud and reads to others at home, by area, gender and age group. ...................17
Figure 8: Student is read to aloud and reads to others at home in English, by area, gender and age 
group. ....................................................................................................................................................18
Figure 9: Teacher reads aloud to student and practice letter sounds with student, by area, gender 
and age group. ....................................................................................................................................19
Figure 10: Teacher makes student read silently and aloud, and assigns reading homework to 
students by area, gender and age group. ......................................................................................20
Figure 11: School has a library and student has books at school to take home to read by area, 
gender and age group. ......................................................................................................................21
Figure 12: Books at home and books in English at home by area, gender and age group. ............................21
Figure 13: Student absenteeism in the previous week by area, gender and age group. ................................. 22
Figure 14: Grade repetition by area, gender and age group. ................................................................................22
Figure 15: Parents knowledge of student performance, according to student report by area, gender 
and age group. ....................................................................................................................................23
Figure 16: English as main language at home by area, gender and age group. ...................................................23
Figure 17: Parent’s literacy and parent uses English as main language for reading and writing by area, 
gender and age group. ......................................................................................................................24
Figure 18: Household factors related to students’ socio-economic status. ......................................................24
Figure 19: Distribution of teaching certificates by gender, area and type. .........................................................26
Figure 20: Training and professional development activities for all and volunteer teachers. ........................26
Figure 21: Distribution of number of grades taught by gender, area, and teacher type.................................28
Figure 22: Distribution of frequency of developing a lesson plan, by gender, area, and teacher type. .......28
Figure 23: Teacher had a lesson plan for the day of the interview, and teacher has a schedule time 
to prepare lesson plan during school day, by gender, area and teacher type. ....................29
Figure 24: Distribution of frequencies for the different in-class activities. ........................................................30
Figure 25: Frequency of use of official reading curriculum in classroom by gender, area and teacher 
type........................................................................................................................................................30
Figure 26: Teacher has teacher guide and how useful teachers find them by gender, area and 
teacher type. .......................................................................................................................................31
Figure 27: Teachers activities to measure students’ progress and expected reading skills. ..........................32
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | v
Figure 28: Amount of extra effort made compared to a normal year, by gender, area and teacher 
type........................................................................................................................................................33
Figure 29: Reasons why teacher exerted special effort to improve reading. ....................................................34
Figure 30: Percentage of teachers receiving reading training and thinking it might be important 
because children were tested, by gender, area and teacher type. .........................................34
Figure 31: Type of respondent to the principal questionnaire and distribution by gender and area. .........35
Figure 32: Respondents to the principal survey education degree, by gender and area. ...............................36
Figure 33: Percentage of respondents to the principal survey that received training to teach reading, 
by gender and area. ...........................................................................................................................36
Figure 34: Review of teachers’ lessons plans, by gender and area.......................................................................37
Figure 35: Frequency of observation of teachers’ lessons in grade 1 and 2, by gender and area. ...............37
Figure 36: Principal made a special effort to improve reading at their school by gender and area. ............38
Figure 37: Reasons why respondents to principal survey exerted special effort to improve reading 
at their school this past year. ..........................................................................................................39
LIST OF TABLES 
Table 1: Skills and corresponding sub-tasks included in baseline EGRA tool. ...................................... 5
Table 2: Sample size by county and area. ....................................................................................................... 9
Table 3: Data Collection Activities for the RL: Baseline and Endline Dates. ......................................10
Table 4: Students average age and location, by gender. ...........................................................................11
Table 5: Proportion of observations by County and selected variables. ..............................................12
Table 6: Summary of EGRA mean scores. ...................................................................................................13
Table 7: Remaining EGRA subtasks, by gender. .........................................................................................15
Table 8: Remaining EGRA subtasks by urban and rural area. ................................................................. 16
Table 9: Remaining EGRA subtasks by age group. .....................................................................................17
Table 10: Household factors related to students socio-economic status, by area, gender and 
age group. ............................................................................................................................................25
Table 11: Teacher gender, volunteer status and average years of teaching experience.....................25
Table 12: Training and professional development, by area and gender. .................................................27
Table 13: Average years in principal (vice-principal) position. ..................................................................35
Table 14: Explanatory variables for EGRA scores. ......................................................................................40
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | vi
ACRONYMS 
EGR Early Grade Reading
EGRA Early Grade Reading Assessment
EMIS Education Management Information System
IE Impact Evaluation
IRB Institutional Review Board 
KII Key Informant Interview
LTTP Liberia Teacher Training Program
MDES Minimum Detectable Effect Size 
M&E Monitoring and Evaluation
RCT Randomized Control Trial
RL Read Liberia
SOW Scope of Work
USAID U.S. Agency for International Development
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 1
EXECUTIVE SUMMARY 
EVALUATION PURPOSE AND EVALUATION QUESTIONS 
The main purpose of the impact evaluation will be to assess the causal impact of the Read Liberia (RL) 
activity on reading performance of second grade students. 
The main goal of the IE will be to answer the following question: 
What is the impact of the Read Liberia activity on the reading fluency and comprehension of 
second grade students? 
PROJECT BACKGROUND 
Liberian students’ reading performance is below the desired levels. A recent program, the Liberia 
Teacher Training Program (LTTP), measured oral reading fluency in treated schools at only 14 words 
per minute. This was accompanied by poor performance in other literacy abilities, such as reading 
comprehension. To address this issue, USAID is partnering with the Ministry of Education to implement 
Read Liberia (RL), a 5-year activity designed to improve early grade reading skills of Liberian students in 
grades one and two and pilot activities designed to develop emergent literacy skills for Liberian students 
in public kindergarten schools in six targeted counties (Lofa, Bong, Grand Bassa, Nimba, Margibi, and 
Montserrado) out of 15 counties. 
NORC at the University of Chicago, through the USAID Reading and Access Evaluation Contract, has 
been charged with conducting the external impact evaluation (IE) of the Read Liberia (RL) program. The 
primary interest is to measure the degree to which the RL activity will increase the proportion of 
students who, by the end of two grades of primary schooling, demonstrate that they can read and 
understand the meaning of grade level text. 
EVALUATION DESIGN, METHODS AND LIMITATIONS
The NORC team decided that an impact evaluation is feasible and necessary to properly answer the 
agreed upon evaluation question. An impact evaluation serves to assess the causal effect of a specific 
intervention on a set of outcomes. 
The evaluation team is using an experimental or randomized control trial (RCT) approach to answer the 
evaluation question, with schools randomly assigned to treatment and control groups. The fact that the 
activity has not started yet and that the total number of schools is greater than the number of schools 
to be treated in the six counties, allows for an RCT. 
The evaluation measures reading outcomes using subtasks of the Early Grade Reading Assessment 
(EGRA), a widely used tool to measure various aspects of reading proficiency. The EGRA includes 
several subtasks that measure higher-order skills (e.g., fluency and comprehension) and lower-order 
skills (e.g., phonemic awareness, letter name knowledge, and decoding), which were all conducted in 
English, given that English is the official language of instruction in Liberia. In addition to the EGRA tool, 
the baseline data collection included: (1) a short student background questionnaire which was 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 2
administered immediately following the EGRA to every student sampled, (2) a teacher questionnaire 
which was administered to the Grade 2 teacher of each sampled class, and (3) a questionnaire 
administered to the school’s principal.
NORC will estimate the causal impact of the program on Early Grade Reading Assessment (EGRA) 
scores using regression models. 
The evaluation faces several risks: 
1. The design of the evaluation preceded the award of the Read Liberia contract, which potentially 
poses a threat to the evaluation in the event that the eventual implementation of the activity 
diverges from the assumptions on which the design is based. 
2. Other risks are common to most evaluations and include the possibility of contamination across 
treatment and control groups; the evaluation team will attempt to detect and account for such 
contaminations prior to and during endline data collection. 
3. A final risk that USAID’s Goal 1 team has raised are concerns about test security related to the 
EGRA instrument being used for this evaluation, since it is the same assessment tools that was 
used in 2011 and 2015. However, there is no evidence of test leakage and, given time and 
budget constraints, NORC and USAID decided the creation and testing of new tools was not 
warranted. 
FINDINGS AND CONCLUSIONS 
In this report we present findings from the data collection activity that took place in May 2017 in order 
to establish the baseline reading level of Grade 2 students prior to the beginning of the 5 year RL 
activity. 
 The baseline data collection was successful and went as planned. Tests conducted during 
fieldwork show very high reliability and consistency in the assessment work. 
 Students are substantially older than what is expected for second graders. The official primary 
school age is 6-11, but starting school at six is rare and only 5.5% of six-year-old children are 
enrolled in school. Attention should be paid to this phenomenon. 
 Student absenteeism and grade repetition, as reported by the learners, are high. Addressing 
absenteeism is very important as most activities planned for Read Liberia will take place in 
school and absenteeism could reduce its potential impact. 
 Average scores in EGRA subtasks are in general, low for second graders. 
EGRA subtasks
Average 
Score
Standard 
Error % of Zeros
Correct Words per Minute (CWPM)* 10.2 0.6 17%
Oral Reading Fluency (ORF)* 14.6 0.9 35%
Reading Comprehension [max=5]* 0.8 0.1 58%
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 3
EGRA subtasks
Average 
Score
Standard 
Error % of Zeros
Correct Letter Names per Minute (CLPM) 67.5 1.3 1%
Correct Non-words Per Minute (CNONWPM) 1.4 0.1 81%
Orientation To Print [max=3] 2.2 0.1 10%
Phonemic Awareness [max=10] 3.8 0.1 10%
Listening Comprehension [max=3] 1.6 0.1 17%
*main impact outcomes 
On average, students in the sample can correctly identify 10.2 familiar words per minute and can 
correctly read 14.6 words per minute from a connected paragraph (Oral Reading Fluency). More than a 
third of the students could not identify a single word in the paragraph. On average, learners answered
correctly less than one comprehension question based on the passage read aloud and more than half 
(58%) were not able to answer any of the 5 comprehension questions correctly. Students in the sample 
could identify 67.5 correct letter names per minute with only 1% unable to identify any letter names. On 
the other hand, decoding skills are low. Students could read, on average, 1.4 invented words correctly 
and, on average 3.8 out of 10 phonemes correctly. Most of the students (81%) were not able to read 
any invented words at all. Students were asked to identify where a sentence started, in what direction 
they would read, and where the next line starts as part of the Orientation to Print subtask. Students 
were able to perform 2.2 of these activities successfully with 10% of them unable to perform any of the 
activities. Finally, after listening to a short story students were asked 3 comprehension questions and, on 
average, they were able to respond1.6 questions correctly, with 17% of them not able to understand 
what was being read to them at all. 
 On average, girls tend to score lower than boys in all EGRA subtasks and rural students’ 
performance is substantially lower than that of urban learners. 
 A third of the Grade 2 teachers interviewed are volunteer teachers. Their reported teaching 
practices and behaviors are similar to those of paid teachers. Having a volunteer teacher rather 
than a paid teacher does not seem associated with differential student performance. 
 Teacher education levels and their reported classroom practices do not correlate with 
assessment scores. 
 Most teachers report following the official curriculum, having teacher guides and preparing 
lessons plans. Principals report reviewing teachers’ lesson plans and a large majority claims to do 
so every week or more often. However, less than half the teachers were able to show the 
interviewers the written lesson plan for the day of the visit. In the future, it would be useful to 
conduct classroom observations designed to capture elements specific to the Read Liberia 
pedagogy and include both low inference and higher inference items. Additional monitoring and 
other incentives could be included in the RL activity to improve actual behavior if needed. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 4
EVALUATION PURPOSE 
EVALUATION PURPOSE, AUDIENCE AND INTENDED USE 
The main purpose of the impact evaluation will be to assess the causal impact of the RL activity on 
reading performance of second grade students. 
Despite the different sub-activities that this intervention will include, the focus of this evaluation is to 
measure the effect of Read Liberia as a whole. NORC has been instructed by USAID to plan for an 
impact evaluation focusing only on Grade 2 students. To this end, NORC assessed Grade 2 students 
before the end of the 2016-17 academic year (May 2017) and interviewed their teachers and principals. 
These data serves as the baseline for the evaluation. 
This evaluation will contribute to the increasing body of evidence on Early Grade Reading Activities in 
developing countries. The results of the baseline and the impact evaluation will inform USAID, 
USAID/Liberia Mission, the Government of Liberia –in particular the Education Sector-, the activity 
implementer, the research and practice community, and the donor community. 
EVALUATION INDICATORS 
Early Grade Reading Assessment 
The evaluation measures reading outcomes using subtasks of the Early Grade Reading Assessment 
(EGRA), a widely used tool to measure various aspects of reading proficiency. The EGRA subtasks 
included in the assessment are described in Table 1 below and all were conducted in English. 
We used an existing version of EGRA, created for Liberian students in 2011, to collect baseline data. 
The assessment was not modified, given the short time available before data collection. The skills tested 
with the EGRA tool and their corresponding sub-task are indicated in Table 1: Skills and corresponding 
sub-tasks included in baseline EGRA tool. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 5
Table 1: Skills and corresponding sub-tasks included in baseline EGRA tool. 
Skill Sub-task & Measurement
Orientation to print Awareness of text direction, where to start reading, how to read down a page. 
Total of 3 questions.
Alphabet knowledge Number of letters names, out of 100, correctly identified in 60 seconds 
Phonological awareness Identify words that start with a different sound. Number of correctly identified 
sounds, out of 10 
Decoding Number of nonsense words, out of 50, correctly decoded in 60 seconds
Word Recognition Familiar word reading, number of correct words, out of 50, read in 60 seconds
Oral Reading Fluency Oral passage reading, number of words fluently read (with accuracy), out of 60,
from a reading passage in 60 seconds
Reading Comprehension Number of questions answered correctly, out of 5, about a passage read aloud by
the student
Listening Comprehension; 
oral language
Number of questions answered correctly, out of 3, about a passage read aloud to 
the student by the enumerator
Additional Instruments 
In addition to the EGRA tool, the baseline data collection included: (1) a short student background 
questionnaire which was administered immediately following the EGRA to every student sampled, (2) a 
teacher questionnaire which was administered to the Grade 2 teacher of each sampled class, and (3) a 
questionnaire administered to the school’s principal. 
NORC used validated questionnaires that had been previously used in Liberia along with the 2011 
EGRA tool. While the EGRA assessment tool was not altered, we slightly modified questions in the 
Student Background, Teacher, and Principal questionnaires where necessary to ensure relevancy to the 
objectives of the activity as it was defined in March 2017. 
The principal survey included questions about their education and experience, coaching and monitoring 
of teachers, interaction with education officers, parents, school resources, etc. The teacher 
questionnaire contained questions about the teachers’ education and experience, interactions with the 
principal or others regarding coaching and monitoring, instructional practices related to reading, etc. 
Finally, the student context interview focused on the child’s home language, reading practices at home, 
some instructional practices in the classroom, and household possessions. 
PROJECT BACKGROUND 
Liberian students’ reading performance is below the desired levels. A recent program, the Liberia 
Teacher Training Program (LTTP), measured oral reading fluency in Grade 2 treated schools at only 14 
words per minute while the recommended benchmark for Liberia is 35 words per minute1 and the 
international standard for English is 45-60 words per minute. This was accompanied with poor 
performance in other literacy abilities, such as reading comprehension. To address this issue, USAID is 
 
1 LTTP II: Endline Assessment of the Impact of Early Grade Reading and Mathematics, Annex G, Interventions 
http://www.earlygradereadingbarometer.org/files/EGRA%20in%20Liberia.pdf 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 6
implementing READ Liberia (RL), a 5-year activity designed to improve early grade reading skills of 
Liberian students in grades one and two and pilot improvements in developing emergent literacy skills 
for Liberian students in public kindergarten schools in six targeted counties (Lofa, Bong, Grand Bassa, 
Nimba, Margibi, and Montserrado). 
NORC at the University of Chicago, through the USAID Reading and Access Evaluation Contract, has 
been charged with conducting the external impact evaluation (IE) of the READ Liberia (RL) program. 
The main purpose of the IE will be to assess: What is the impact of the RL activity as a whole on the reading 
fluency and comprehension of second grade students2? The evaluation will measure reading outcomes using 
the Early Grade Reading Assessment (EGRA), a widely used tool to measure various aspects of reading 
proficiency. 
EVALUATION METHODS & 
LIMITATIONS
EVALUATION QUESTIONS 
The main goal of the IE will be to answer the following question: 
What is the impact of the Read Liberia activity on the reading fluency and comprehension of 
second grade students? 
As explained above, the measurement of reading fluency and comprehension will be done using the 
EGRA in English, specifically designed for Liberia. 
METHODOLOGY OVERVIEW 
The NORC team decided that an impact evaluation is feasible and necessary to properly answer the 
agreed upon evaluation question. An impact evaluation serves to assess the causal effect of a specific 
intervention on a set of outcomes. It allows us to attribute changes in an outcome to a specific 
intervention or set of interventions by answering the counterfactual question “What would have 
happened to activity participants in the absence of the intervention?” Ideally, this is done by observing 
the same program participants both with and without the intervention at the same point in time. Of 
course, this is not possible; at any given time, a participant either receives the intervention or does not. 
Therefore, we can never directly observe the counterfactual and instead need to create a comparison 
group to serve as the counterfactual proxy. Identifying a credible comparison group is a critical aspect of 
an impact evaluation and there are several approaches to doing so. 
The evaluation team used an experimental or randomized control trial (RCT) approach to answer the 
evaluation question, with schools randomly assigned to treatment and control groups. The fact that the 
 
2 At the time of writing this report, the Read Liberia activity is still in the procurement phase and an implementing firm has not 
been selected. Therefore, the details of the actual activity and its components are unknown. As such, the design of the IE is 
based on the information available in the Read Liberia RFP and agreements reached during conversations with USAID/Liberia 
and USAID/E3/ED. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 7
activity has not started yet and that the total number of schools is greater than the number of schools 
to be treated in the six counties, allows for an RCT design for this evaluation. As mentioned above, we 
want treatment and control groups to be as similar as possible; the primary advantage of a randomized 
design is that it minimizes the potential for selection bias- that is, the possibility that observed 
differences in outcomes between the treatment and control groups are due not to the impact of the 
project, but to other external and possibly systematic differences between groups that the evaluation is 
not able to account for. Another advantage is that the randomizing approach tends to be fairly simple to 
explain to stakeholders and policy makers, and it is a fair allocation when not all schools can receive 
treatment at once. 
Randomization into treatment and control groups will occur at the school level. Randomization at the 
school level is possible given that all sub-activities included in IR2, IR3 and IR4 will be implemented at the 
school level, while the IR1 sub-activities are common to all schools in the country. 
Experimental Design 
We use an experimental approach to create treatment and control groups of schools. We will estimate 
the causal impact of the program on EGRA scores using regression models explained in Annex VI. 
Although the evaluation question focuses on reading fluency and comprehension outcomes, we plan to 
estimate the impact of RL on all sub-tasks included in the EGRA tool. 
In addition to estimating the overall impact of Read Liberia on reading outcomes, we will also conduct a 
sub-group analysis to examine whether the impacts were different between boys and girls. Depending 
on the findings derived from baseline data, we can potentially explore some interactions between school 
characteristics and the RL activity. For example, if we find that student characteristics and performance 
are different between rural and urban schools, we could study whether the effects of the activity are 
different accordingly to the school location. 
Limitations 
The evaluation faces several risks. Primary among them, is the fact that the design of the evaluation 
completely preceded the award of the Read Liberia contract, which potentially poses a threat to the 
evaluation in the event that the actual implementation of the activity diverges from the assumptions on 
which the design is based. 
Another risk, which is common to most evaluations, is the possibility of contamination across treatment 
and control groups; evaluation plans were designed in order to minimize these risks and the evaluation 
team will attempt to detect and account for such contaminations prior to and during endline data 
collection. 
A final risk -that USAID’s Goal 1 team has raised- relates to test security of the EGRA instrument being 
used for this evaluation, since it is the same assessment tools used in 2011 and 2015. However, there is 
no evidence of test leakage and, given time and budget constraints, NORC and USAID decided the 
creation and testing of a new tool was not warranted. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 8
SCHOOL AND STUDENT SAMPLE 
Our methodological approach to answering the evaluation question requires collecting data before the 
start of the activity’s implementation in order to ensure treatment and control schools are, on average, 
statistically identical. Data collection will take place again after two to three years of RL implementation. 
We opted to use a repeated cross section sample, collecting data from a random sample of Grade 2 
students in May 2017, as requested by USAID/Liberia, and again from a random sample of Grade 2 
students in the same schools in two to three years3. 
School Selection 
USAID/Liberia, in collaboration with the Ministry of Education, has defined a set of criteria that was used 
to identify schools that are eligible to receive the Read Liberia assistance. These criteria include all 
schools that: 1) are public; 2) are located in the six selected counties (Montserrado, Margibi, Bong, 
Grand Bassa, Nimba and Lofa); and 3) have both grade one and grade two classes. The evaluation sample 
will be drawn from schools who meet these same criteria. 
For the evaluation, NORC requested an additional criterion, that the schools have at least 20 learners in 
Grade 2 (according to the EMIS) to assure finding enough learners for the assessment. The EMIS list of 
schools provided by the MOE includes designations for urban and rural schools. Using these data, the 
sample is assigned proportionally to the number of urban and rural schools within each county. As a 
result, the sample is representative of the population of public schools which have grade 1 and grade 2 
classes for each of the 6 counties in the sample. 
Student Selection 
Sixteen second grade students - 8 girls and 8 boys - were selected randomly in each school. All learners 
were eligible for selection unless they refused to participate or severe disabilities precluded them from 
participating.4
 
Sample Size Calculation 
The sample size calculation was made based on the following assumptions. 
 α is the significance level of the test, or probability of Type I error. We use the standard value 
of 0.95. 
 β is the power of test, where (1-β) is the probability of Type II error. We use the standard 
value of 0.8. 
 
3 The exact length of the evaluation and the year in which endline data will be collected will be decided at a later date, and 
taking into account the activity details. There are trade-offs to consider: going sooner minimizes the risk of teachers switching 
between treatment and control schools while postponing the endline gives the activity more time to be fully implemented and 
adjusted. 
4 The request for 16 students addresses the need to maximize the probability of collecting data from a minimum of 16 students 
(eight girls and eight boys) per school. This provides enough observations to give the evaluation sufficient power to detect 
impact, given the number of schools in the sample. The NORC evaluation team specified 20 enrolled students, rather than 16,
to account for absenteeism which is quite common.
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 9
 ρ is the intracluster correlation coefficient, or ICC. The ICC in the present case is a measure of 
how much variability lies between schools and how much lies within schools. Based on King et 
al. (2015) which describes Liberian EGRA data, we assume an ICC of 0.3.5
 r
2 is the proportion of the variation in the outcome due to the covariates anticipated in the 
regression analysis. In our case, these covariates will include a range of household and individual 
characteristics, as well as school fixed effects. For this parameter, we assume an approximate 
value of 0.2. 
 MDES is the minimum detectable effect size. The MDES is the smallest impact of the activity on 
the outcome variable that the evaluation will be able to detect. The selected MDES is 0.34 of a 
standard deviation, a medium size effect. This corresponds to the standardized effect for Oral 
Reading Fluency found in the midline evaluation of LTTP (The NORC team estimated this 
standardized effect using our own calculations and data from the Midline LTTP report.6,7 ) 
During a phone call (February 2017) USAID/Washington and USAID/Liberia considered it a 
reasonable effect to be expected given previous experiences and the intensity of the 
intervention. 
Based on these parameters, the required sample size is 45 schools in each study group (treatment and 
control) with 16 students in each school, for a total sample of 1,440 students. The teacher and head 
teacher sample consist of one Grade 2 teacher and one head teacher per school and has the objective 
of completing the picture of the learners’ education context and helping understand the mechanisms 
behind the impact effects. The sample size by county is presented in 
Table 2. 
The number of schools in the sample is exactly as planned. The number of learners per school is 
sometimes lower than planned due to lower enrolment than anticipated or due to absenteeism. In other 
cases, it is higher given that it was decided to assess all learners in schools that had 20 learners or less. 
This decision was taken for two reasons, first to compensate for cases where the number of learner was 
insufficient and second to avoid leaving out of the exercise a very small number of learners which could 
have felt rejected by such an action. Overall, we have a sufficient number of schools and learners in our 
sample to conduct the IE as planned. 
Table 2: Sample size by county and area. 
County Area Number of Schools Number of students
Planned Observed Planned Observed
Bong Urban 2 2 32 35
Rural 15 15 240 262
 
5 This figure corresponds to the one in Annex H for grade 2 in the Endline report of LTTP (King, Simon, Medina Korda, Lee 
Nordstrum and Susan Edwards (2015). Liberia teacher training program: Endline assessment of the impact of early grade 
reading and mathematics interventions) 
6 DeStefano, Joseph, Timothy Slade, and Medina Korda (2013). Liberia Teacher Training Program (LTTP): Midterm Assessment 
of the Impact of Early Grade Reading and Math Interventions. 
7 We divide the DID for ORF and grade 2 in table 8 of the midline report, by the standard deviation that we calculated using 
the data in table 2 of the midline report. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 10
County Area Number of Schools Number of students
Planned Observed Planned Observed
Grand Bassa Urban 2 2 32 35
Rural 6 6 96 85
Lofa Urban 2 2 32 36
Rural 17 17 272 292
Margibi Urban 1 1 16 17
Rural 6 6 96 90
Montserrado Urban 7 7 112 119
Rural 6 6 96 100
Nimba Urban 5 5 80 88
Rural 21 21 336 363
Total Urban 19 19 304 330
Total Rural 71 71 1136 1192
TOTAL SAMPLE SIZE 90 90 1440 1522
BASELINE DATA COLLECTION 
NORC worked closely with subcontractors School-to-School International and the Khana Group (TKG) 
to conduct all data collection activities. The data collection schedule is as follows: 
Table 3: Data Collection Activities for the RL: Baseline and Endline Dates. 
Survey Type Baseline Date Endline Date
Early Grade Reading Assessment May/June, 2017 May/June 2019 or 2020
Student Survey May/June, 2017 May/June 2019 or 2020
Teacher Survey May/June, 2017 May/June 2019 or 2020
Head Teacher Survey May/June, 2017 May/June 2019 or 2020
There were no significant challenges identified during baseline data collection. In addition to the 
extensive training and practice assessors received prior to starting fieldwork, they continued to be 
tested for reliability and consistency during data collection. NORC, organized a protocol for monitoring 
inter-rater reliability during baseline data collection. The percentage of agreement between assessors 
for different EGRA subtasks was very high, between 98 and 99%. The Kappa statistic ranged from 0.83 
to 0.98. The Kappa strength of agreement indicates the extent to which two different ratings of the 
same subject could have happened by chance. Kappa values range from -1.0 to 1.0 where higher values 
indicate lower probability of agreement by chance. According to Fleiss (1981)8 Kappa measures greater 
than 0.75 are excellent, and according to Landis and Koch (1977) over 0.81 is almost perfect. Our 
indicators consistently show excellence in the quality of data collection. 
 
8 Fleiss, J. L. (1981). Statistical methods for rates and proportions (2nd ed.) New York: John Wiley. 
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 11
BASELINE RESULTS 
BASELINE DESCRIPTIVE RESULTS FROM THE BASELINE SURVEY 
Student Sample Description 
The student sample is composed of 52% boys and 48% girls however this difference is not statistically 
significant. We also see that more boys attending school in rural areas (81%) than girls (76%), and that
difference is statistically significant. The distributions are presented in Table 4. 
Table 4: Students average age and location, by gender. 
Boys Girls Difference
All 52% 48% 3%
Rural (proportion) 81% 76% -5%
Age (average) 12.6 12.4 -0.3
Observations 786 736
The average age of boys is 12.6 years old and for girls is 12.4 years old, and the difference in age is also 
statistically significant. Figure 1 shows the distribution of student age by gender. We note that there are 
observations of second grade students between 15 and 20 years old among both genders. 
Figure 1: Distribution of student age by boys and girls. 
A set of characteristics of the sample by County are presented in Table 5. The sample is predominantly 
rural (78% overall) and representative of the target population. Most counties contain more than 70% 
rural students with the exception of Montserrado (46%). The distribution of genders remains similar 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 12
across counties with a consistent skew toward boys, except in Montserrado where girls comprise 53% 
and boys represent 47% of the sample. The majority of the students are between 11 and 15 years old 
(75%). Around 10% of the Grade 2 students are 16 years or older in Bong (13%), Grand Bassa (11%), 
Margibi (9%), and Nimba (9%). 
Table 5: Proportion of observations by County and selected variables. 
Variable Bong
Grand 
Bassa Lofa Margibi
Mont￾serrado Nimba Total
Urban 12% 29% 11% 16% 54% 20% 22%
Rural 88% 71% 89% 84% 46% 80% 78%
Boy 54% 52% 52% 53% 47% 52% 52%
Girl 46% 48% 48% 47% 53% 48% 48%
6-10 years 12% 8% 29% 10% 27% 13% 17%
11-15 years 75% 81% 68% 82% 72% 78% 75%
16 and more years 13% 11% 3% 9% 1% 9% 7%
Total Observations 297 120 328 107 219 451 1522
EGRA Scores 
The different subtasks included in the assessment are described in Table 1. Our focus is on the main 
impact indicators (correct familiar words per minute (CWPM), oral reading fluency (ORF), and reading 
comprehension (RC)), but we also analyze the rest of the subtasks assessed (Correct Letter Names per 
Minute (CLPM) and Correct Non-Words per Minute (CNONWPM), Orientation to Print (OTP), 
Phonemic Awareness (PA), and Listening Comprehension (LC)). The aggregated results of the different 
EGRA subtasks are presented in Table 6. 
On average, Liberian students in these 6 counties can correctly identify 10.2 familiar words per minute 
and can correctly read around 14.6 words per minute from a paragraph (Oral Reading Fluency9). 
Depending on how much of the paragraph the student is able to read, a certain number of 
comprehension questions about it are asked. It is important to bear in mind that there are 5 
comprehension questions based on the reading passage but students are only asked the questions 
corresponding to the part of the passage that they were able to read. Less than half of the students 
were asked at least one comprehension question, while the rest did not read enough to particiapte. On 
average the learners that participated in the reading comprehension task answered less than one 
question correctly and more than half (58%) were not able to answer any of the comprehension 
questions correctly. These results are presented graphically in Figure 2. 
 
9 Oral Reading Fluency The oral reading fluency subtask is timed and measures speed and accuracy in terms of the number of 
correct words read per minute. The number of words per minute is calculated as= (Total words correct – Total incorrect) / 
[(60 – Time remaining on device) / 60] 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 13
Figure 2: Correct Familiar Words per Minute, Oral Reading Fluency, and Reading 
Comprehension score (primary axis) and percentage of zero scores (secondary axis). 
Liberian students in the sample could identify 67.5 correct letter names per minute (CLPM) with only 
1% unable to identify any letter names. They could identify, on average, 1.4 invented words correctly 
(81% were not able to read any invented words at all). The assessment asked students to identify where 
a sentence started, in what direction they would read, and where the next line starts as part of the 
Orientation to Print subtask. Students were able to perform 2.2 of these activities correctly with 10% of 
them unable to perform any of the activities. Students could identify on average 3.8 out of 10 phonemes 
correctly. Finally, related to listening comprehension, students were able to answer, on average, 1.6 out 
of 3 questions correctly, with 17% of them not able to understand what was being read to them at all. 
Table 6: Summary of EGRA mean scores. 
EGRA subtasks Average
Standard 
Error
Percentage 
of Zeros
Correct Words per Minute (CWPM)* 10.2 0.6 17%
Oral Reading Fluency (ORF)* CWPM from connected paragraph 14.6 0.9 35%
Reading Comprehension [max=5]* 0.8 0.1 58%
Correct Letter Names per Minute (CLPM) 67.5 1.3 1%
Correct Non-words Per Minute (CNONWPM) 1.4 0.1 81%
Orientation To Print [max=3] 2.2 0.1 10%
Phonemic Awareness [max=10] 3.8 0.1 10%
Listening Comprehension [max=3] 1.6 0.1 17%
* Main Impact Evaluation indicators. 
Figure 3 shows the complete distribution of ORF scores. While 22% of the students read at the 
proposed benchmark for second grade of 35 words per minute or even faster, the rest do not reach the 
level. As mentioned before about 35% of students are non-readers; this means that they were not able 
to read a single word from the reading paragraph. 
10.2
14.6
0.8
17%
35%
58%
0%
10%
20%
30%
40%
50%
60%
70%
0
2
4
6
8
10
12
14
16
Correct Words per Minute Oral Reading Fluency Reading Comprehension
Number of correct answers
Percentage of Zero scores
Score % of Zeros
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 14
Figure 3: Distribution of Oral Reading Fluency, correct words per minute. 
There are statistically significant differences between boys and girls in the number of correct words per 
minute and oral reading fluency, but not for reading comprehension. In the first indicator, boys are able 
to read 11.3 words correctly per minute, while girls only 9 per minute. For oral reading boys read, on 
average, 15.8 words correctly per minute while girls correctly read only 13.2. In terms of reading 
comprehension, boys and girls perform similarly (0.8 and 0.7 correct answers respectively). These 
results are presented in Figure 4. 
Figure 4: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), 
and Reading Comprehension score (secondary axis) by boys and girls. 
35%
20%
18%
9% 8%
4% 3%
1% 1% 0% 0% 0% 0% 0% 0% 0%
0%
5%
10%
15%
20%
25%
30%
35%
40%
% of students
correct words per minute
11.3
9.0
15.8
13.2
0.8
0.7
0.0
0.2
0.4
0.6
0.8
1.0
0.0
3.0
6.0
9.0
12.0
15.0
18.0
Boys Girls
Number of correct answers
Number of correct words read
Correct Words per Minute Oral Reading Fluency Reading Comprehension
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 15
Boys tend to perform better than girls across the remaining EGRA subtasks, with the exception of 
phonemic awareness and listening comprehension where no statistically significant difference was found. 
For example, boys can identify 69.3 letters correctly per minute while girls can identify 65.5 letters 
correctly per minute. The differences are even greater when it comes to identifying invented words; 
boys can identify 2 non-words correctly per minute while girls could only identify 0.8 non-words 
correctly per minute on average. Finally, boys can perform the tasks under orientation to print (2.2 
tasks out of 3) slightly better than girls (2.11 out of 3). These results are presented in table 7. 
Table 7: Remaining EGRA subtasks, by gender. 
EGRA Subtask Boys Girls Difference
Correct Letter Names per Minute (CLPM) 69.3 65.5 -3.8*
Correct Non-words Per Minute (CNONWPM) 2.0 0.8 -1.2*
Orientation To Print [max=3] 2.2 2.1 -0.1*
Phonemic Awareness [max=10] 3.9 3.7 -0.2
Listening Comprehension [max=3] 1.6 1.6 0.0
* Statistically significant at 5%. 
Students in urban areas performed better than those in rural areas (i.e. all differences are statistically 
significant). For instance, students in urban areas could read 13.4 familiar words correctly per minute 
while those in rural areas could read 9.3. Also, urban students are able to read 19.5 words correctly 
from a reading paragraph in a minute, while rural students read 13.2. Finally, urban students can, on 
average, answer 1.0 comprehension questions correctly while rural students can answer 0.7 questions 
correctly. 
Figure 5: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), 
and Reading Comprehension score (secondary axis) by urban and rural areas. 
There are also statistically significant differences by urbanity in scores from the correct letter names per 
minute and listening comprehension subtasks. Students in urban areas can identify 74.1 letter names 
correctly per minute while students in rural areas identify 65.6 letter names correctly. In the case of 
13.4
9.3
19.5
13.2
1.0
0.7
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.0
3.0
6.0
9.0
12.0
15.0
18.0
21.0
Urban Rural
Number of correct answers
Number of correct words read
Correct Words per Minute Oral Reading Fluency Reading Comprehension
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 16
listening comprehension, students in urban areas can answer 1.8 questions correctly while students in 
rural areas answered 1.5 questions correctly on average. There are no statistically significant differences 
between urban and rural areas in the remaining EGRA subtasks. The results are presented in Table 8. 
Table 8: Remaining EGRA subtasks by urban and rural area. 
EGRA Subtask Urban Rural Difference
Correct Letter Names per Minute (CLPM) 74.1 65.6 -8.5*
Correct Non-words Per Minute (CNONWPM) 1.7 1.3 -0.4
Orientation To Print [max=3] 2.2 2.2 0.0
Phonemic Awareness [max=10] 4.0 3.8 -0.2
Listening Comprehension [max=3] 1.8 1.5 -0.3*
* Statistically significant at 5%. 
There were no statistically significant differences in the main impact indicators between the youngest 
students (6-10 years old) and oldest students (16 years old or more) when compared to students 
between 11 and 15 years old. Students between 6 and 10 years old correctly read 10 familiar words per 
minute on average, while those between 11 and 15 read 10.5 familiar words per minute and those 16 
and older read 9.8. Oral reading fluency among students between 6 and 10 years old is 15.2 words on 
average, while those between 11 and 15 read 14.9 words and those 16 and older read 13.7 words. 
Finally, students between 6 and 10 years old correctly answered one question on average, while those 
between 11 and 15 correctly answered 0.7 questions and those 16 and older answered correctly 0.9 
questions. The results are presented in Figure 66. 
Figure 6: Correct Words per Minute (primary axis), Oral Reading Fluency (primary axis), 
and Reading Comprehension score (secondary axis) by age group. 
There were no statistically significant differences in the remaining EGRA subtasks when comparing the 
younger (6-10 years old) and the older students (16 years old or more) with those 11 to 15 years old, 
with two exceptions. First, students between 6 and 10 years old correctly identified 61.6 letters per 
minute, while those between 11 and 15 identified 69.6 letters per minute. Second, students between 11 
10.0 10.5 9.8
15.2 14.9 13.7
1.0
0.7
0.9
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.0
3.0
6.0
9.0
12.0
15.0
18.0
6-10 years old 11-15 years old 16 years old and more
Number of correct answers
Number of correct words read
Correct Words per Minute Oral Reading Fluency Reading Comprehension
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 17
and 15 correctly answered 1.6 questions in listening comprehension, while those 16 and older correctly 
answered an average of 1.4 questions. These differences are statistically significant at conventional levels. 
All the results are presented in Table 9. 
Table 9: Remaining EGRA subtasks by age group. 
EGRA Subtask
6-10
(A)
11-15
(B)
16 +
(C)
Difference
(A)-(B)
Difference
(C)-(B)
Correct Letter Names per Minute (CLPM) 61.6 69.6 69.0 -8.1* -0.6
Correct Non-words Per Minute (CNONWPM) 1.6 1.4 1.5 0.2 0.1
Orientation To Print [max=3] 2.0 2.2 2.2 -0.2 0.0
Phonemic Awareness [max=10] 3.6 3.9 3.9 -0.3 0.0
Listening Comprehension [max=3] 1.6 1.6 1.4 0.0 -0.2*
* Statistically significant at 5%. 
Student Reading Activities at Home and School 
Reading Practices at Home: The majority of students reported that someone at home reads aloud to 
them (61.9%) and reported that they read to others at home (62%). The results are presented in Figure 
77. There are small differences between urban and rural students and between boys and girls. Younger 
students report that someone reads to them aloud in larger proportion than older students. 
Figure 7: Student is read aloud and reads to others at home, by area, gender and age 
group. 
Almost all the students that reported being read to and read to someone at home, report that the 
language used is English. The results are presented in Figure 88. 
61.9% 62.6% 61.7% 63.4% 60.3% 66.1% 61.8% 55.0%
62.0% 62.1% 61.9% 63.4% 60.5% 63.2% 62.6% 59.5%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
Does anyone read aloud to you at home? Do you practice reading aloud to someone at home?
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 18
Figure 8: Student is read to aloud and reads to others at home in English, by area, gender 
and age group. 
Reading Practices in Class: Almost all students (90.9%) reported that their teacher reads aloud to them. 
There are no large differences in the reports between urban (92.4%) and rural (90.5%) students or 
between boys (90.2%) and girls (91.6%). Younger students (6-10 years old) and older students (16 years 
old or older) reported that the teacher reads aloud to them slightly less than those between 11-15 
years old (88.4% and 88.2%, vs. 91.9%). The results are presented by the bars in Figure 99. 
95.4% 97.6% 94.8% 95.1% 95.7% 94.6% 95.6% 100.0%
96.0%
99.5%
95.1% 95.9% 96.2% 96.2% 95.8%
100.0%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
Read aloud: English Read to others: English
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 19
Figure 9: Teacher reads aloud to student and practice letter sounds with student, by area, 
gender and age group. 
Around 3 out of 4 students reported that teacher practiced letter sounds with them. There are 
differences between urban (70.8%) and rural (77.2%) areas, and between the younger students (6-10 
years old – 78.5%), and the older students (11-15 years old – 75% -, and 16 years or older – 72.7%). 
Those results are represented by the circles in Figure 9 above. 
The majority of students reported that their teacher makes them read both silently (88.5%) and aloud in 
class (92.3%). The results are presented in Figure 10. Report of reading aloud are more common among 
urban students than among their rural counterparts. There are no important differences in the reports 
by age or gender. 
90.9% 92.4% 90.5% 90.2% 91.6% 88.4% 91.9% 88.2%
75.8%
70.8%
77.2% 76.4% 75.1%
78.5%
75.0%
72.7%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
Teacher reads aloud to students Teacher practice letter sounds with student
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 20
Figure 10: Teacher makes student read silently and aloud, and assigns reading homework 
to students by area, gender and age group. 
Similarly, the majority of students (88.8%) reported that teacher assigns reading homework to them. 
There are differences between urban (85.7%) and rural (89.6%) areas, between boys (87.6) and girls 
(90%), and between the younger students (6-10 years old – 92.4%), and the older students (11-15 years 
old – 87.6% -, and 16 years or older – 90.1%). The results are presented by the circles in The majority 
of students reported that their teacher makes them read both silently (88.5%) and aloud in class (92.3%). 
The results are presented in Figure 10. Report of reading aloud are more common among urban 
students than among their rural counterparts. There are no important differences in the reports by age 
or gender. 
Figure 10School Library and Books: Around a third of the students (37.3%) reported that their school 
has library. This proportion is higher for urban students (43.6%) than for rural students (35.6%), and for 
boys (38.1%) than girls (36.5%). There are no significant differences by age. The results are represented 
by bars in Figure 1111. 
Around 2 out of 3 students reported having books at school that they could take home to read. There 
are large differences between urban (56.8%) and rural areas (69%). More boys (68.6%) report having 
books to take home than girls (63.9%) do however, there are no differences between age groups. 88.5% 83.9% 89.8% 89.1% 87.9% 85.6% 88.9% 89.1% 92.3% 93.9% 91.8% 91.2% 93.4% 90.0% 92.5% 93.6% 88.8% 85.7% 89.6% 87.6% 90.0% 92.4% 87.6%
90.1%
0%
20%
40%
60%
80%
100%
All Urban Rural Boys Girls 6-10 11-15 16+
Teacher makes you practice reading silently Teacher makes you practice reading aloud
Teacher assigns reading homework
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 21
Figure 11: School has a library and student has books at school to take home to read by 
area, gender and age group. 
Books at Home: Two thirds of students reported having reading books at home. There are differences 
between urban (59.7%) and rural (64.6%) areas and between boys (66.9%) and girls (59.9%), but not by 
age. Almost all students (95%) have reading books in English. The results are presented by the bars in 
Figure 1212. 
Figure 12: Books at home and books in English at home by area, gender and age group. 
Students Absenteeism and General School Performance 
Around 1 out of 3 students missed at least one day of school in week prior to being interviewed. There 
is more absenteeism among urban students (36.8%) than students in rural (33.9%) areas and absenteeism 
seems to increase with age. However, there was no difference by gender. The results are presented in 
Figure 1313. 
37.3% 43.6% 35.6% 38.1% 36.5% 38.8% 36.9% 37.0%
66.4%
56.8%
69.0% 68.6%
63.9% 66.8% 66.4% 67.6%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
School has library Student has book at school to take home
63.5% 59.7% 64.6% 66.9% 59.9% 64.0% 63.4% 64.5%
94.7% 97.5%
94.0% 95.2% 94.1% 94.4% 95.0%
100.0%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
Student has reading books at home Reading books in English
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 22
Figure 13: Student absenteeism in the previous week by area, gender and age group. 
Around 4 in 10 students report having repeated a grade and of those, 2% of them have repeated 2 or 
more grades. These results are presented in Figure 1414. The proportion of students who have 
repeated is lower in urban areas (37.2%) compared with rural areas (43.9%) but is more common to 
repeat more than one grade in rural areas than in urban areas (2.4% vs. 0.8%). Boys and girls have the 
same proportion of repeating grades (42.4% and 42.5% respectively). As expected, younger students 
have lower rate of repeating grades than older students. 
Figure 14: Grade repetition by area, gender and age group. 
Parent Attitudes and Household Context 
Parents Knowledge of Student Performance: Around 57.2% of the students reported that their parents 
knew when they did not do well on a test or assignment, while 80.6% students reported that their 
parents knew when they did well. The results are reported in Figure 1515. Urban area students are 
36.8%
33.9%
34.5%
34.6%
31.6%
34.3%
37.3%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
Urban
Rural
boy
girl
6-10
11-15
16+
42.4% 37.2%
43.9% 42.4% 42.5% 37.1% 42.5% 48.1%
2.1%
0.8%
2.4%
2.8%
1.3%
1.1%
2.0%
3.9%
0%
1%
1%
2%
2%
3%
3%
4%
4%
5%
0%
10%
20%
30%
40%
50%
60%
All Urban Rural Boys Girls 6-10 11-15 16+
Student repeated grade Repeated 2 grades or more
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 23
more likely to say their parents knew when they did not do well than rural area students, while the 
opposite holds for parents knowing when they did well 
Figure 15: Parents knowledge of student performance, according to student report by 
area, gender and age group. 
Boys report that their parents know their performance in higher proportion than girls (not doing well 
60.9% vs. 53.2%; doing well 83.5% vs. 77.5%). Also, younger students are more likely to report that their 
parents knew when they did not do well and when they did not than older. 
Figure 16: English as main language at home by area, gender and age group. 
Language Spoken at Home: The survey included questions about the student’s context at home. The 
most common language students speak at home is Kpelle (25.1%), followed by English (20.6%). Around 
27.5% of urban households primarily speak English, while only 18.6% of rural households primarily speak 
English. There are some differences by student’s age, with older students living in home where English is 
less commonly spoken. 
57.2%
58.2%
56.9%
60.9%
53.2%
60.5%
56.9%
52.3%
80.6%
78.1%
81.3%
83.5%
77.5%
80.9%
81.1%
77.3%
100% 80% 60% 40% 20% 0% 20% 40% 60% 80% 100%
All
Urban
Rural
Boys
Girls
6-10
11-15
16+
Parents knew about NOT doing well on test or assignment
Parents knew about doing WELL on a test or assignment
27.5%
18.6%
21.0%
20.1%
19.6%
21.3%
15.6% All, 20.6%
0% 5% 10% 15% 20% 25% 30% 35%
Urban
Rural
boy
girl
6-10
11-15
16+
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 24
Figure 17: Parent’s literacy and parent uses English as main language for reading and 
writing by area, gender and age group. 
Parents’ Literacy: The majority of parents can read and write and they are able to do so in English. The 
results by area, gender and age group are shown in Figure 1717. Literacy is slightly higher among 
between urban (88.8%) than rural (82.0%) parents. No differences are notable between gender and age 
groups. 
Socio-economic Status: Approximately 8 out of 10 students listen to radio at home but only 3 out of 10 
students watch TV or have electricity at home. Around 60% of students eat before coming to school 
and 63.2% ate at school the day before. The results are presented in Figure 1818. 
Figure 18: Household factors related to students’ socio-economic status. 
83.4% 88.8% 82.0% 83.7% 83.2% 86.7% 83.6% 83.8%
83% 87%
82% 84% 82% 79%
85% 83%
0%
20%
40%
60%
80%
100%
All Urban Rural boy girl 6-10 11-15 16+
Literate Parents Parent reads/writes in English
79.9%
28.7% 30.7%
60.0% 63.2%
0%
20%
40%
60%
80%
100%
Listen to radio at
home
Watch TV at home Electricity at home Eat before coming
to school
Eat at school
yesterday
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 25
As we show in Table 10, there are differences between urban and rural household. Urban households 
are more likely to have electricity and TV sets. Urban students also report eating before going to school 
in a higher proportion than rural students. Girls seem to listen to radio and watch TV at home in slightly 
larger proportions than boys. The percentage of younger students reporting watching TV and eating 
before school is higher than older students. 
Table 10: Household factors related to students socio-economic status, by area, gender 
and age group. 
Category
Listen to 
radio at 
home
Watch TV 
at home
Electricity 
at home
Eat before 
coming to 
school
Ate at 
school 
yesterday
Urban 80.8% 45.6% 47.9% 65.3% 64.8%
Rural 79.6% 24.0% 25.9% 58.5% 62.8%
Boys 78.6% 27.8% 29.4% 60.3% 62.7%
Girls 81.2% 29.6% 32.1% 59.7% 63.8%
6-10 years old 79.8% 33.3% 33.6% 66.8% 66.5%
11-15 years old 80.3% 28.2% 30.6% 58.6% 62.6%
16 years old or more 81.1% 21.6% 27.0% 54.1% 60.4%
Total Observations 1,512 1,517 1,515 1,514 1,517
NOTE: Total sample size of students is 1522. Response totals to individual questions vary due to item non-response. 
Teacher Characteristics and In-Class Activities 
The teachers of the selected grade 2 classes answered the teacher questionnaire. Select teacher 
characteristics are presented in Table 11. Twenty-two percent of teachers are female and 34% are 
volunteer teachers. The average number of years of experience is 11.
Table 11: Teacher gender, volunteer status and average years of teaching experience. 
Characteristic Observations
All Teachers
Female Teacher (%) 22% 89
Volunteer Teacher (%) 34% 89
Urban (%) 21% 89
Number of Years of Experience Teaching (Average) 11.1 89
Volunteer teacher only
Female Teacher (%) 7% 30
Urban (%) 7% 30
Number of Years of Experience Teaching (Average) 4.9 30
We have a sample of only 30 volunteer teachers; among them 7% are women and 7% live in urban areas. 
Their average experience is 4.9 years. These results are presented in the bottom portion of Table 11. 
No further disaggregation is possible, given the small number of observations. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 26
Teachers Certificate: The majority of teachers (84%) have some form of certificate (C certificate 68%, B 
certificate 6%, AA certificate 2%, and other 8%) while only 16% say they have no formal teaching 
certification. The results are presented in figure 19. All paid teachers had some sort of certificate 
however this is true for just over half of the volunteer teachers. Urban teachers (89%) are slightly more 
likely to report having a certificate of some sort than rural teachers (83%). 
Figure 19: Distribution of teaching certificates by gender, area and type. 
Teacher Training: 7 out of 10 teachers attended at least some in-service training or professional 
development sessions last year. Sixty-four percent of teachers in grade 2 have received training on how 
to teach reading at some point and 19% have received support visits at school within the past year on 
how to teach reading. Nine percent of teachers received training this school year, with an average 
duration of 15.4 hours. The results are presented in Figure 2020. 
Figure 20: Training and professional development activities for all and volunteer teachers. 
0% 20% 40% 60% 80% 100%
Overall
Male
Female
Rural
Urban
Paid
Volunteer
C Certificate B Certificate AA Certificate Other None
70%
64%
9%
19%
63%
50%
13%
17%
0% 10% 20% 30% 40% 50% 60% 70% 80%
Attended any in-service training or professional
development sessions last year.
Received training on how to teach reading.
Receive training for this school year.
Received support visits at school this past year on
how to teach reading.
All teachers Volunteer teachers
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 27
Among volunteer teachers, the percentage that received training is lower. In part, this could be because 
volunteer teachers have on average fewer years of experience. Sixty-three percent attended at least one 
in-service training or professional development session last year. Half of volunteer teachers in grade 2 
have received training on how to teach reading and 17% have received support visits at school in the 
past year on how to teach reading. 13% of teachers received training this school year, with an average 
duration of 5.5 hours. 
Table 12: Training and professional development, by area and gender. 
Training and professional development activity Male Female Rural Urban N
Attended any in-service training or professional 
development sessions last year.
70% 70% 70% 68% 89
Received training on how to teach reading. 59% 80% 66% 58% 89
Receive training for this school year. 12% 0% 11% 0% 57
Received support visits at school this past year on how 
to teach reading.
22% 10% 21% 11% 89
Around 70% of men, women, rural and urban teachers attended an in-service training or professional 
sessions last year. Around 59% of men received training on how to teach reading, while 80% of women 
received this type of training. Twenty-two percent of male teachers, compared to 10% of female 
teachers, received support visits at school on how to teach reading this past year. Sixty-six percent of 
teachers in rural areas received training on how to teach reading, while 58% of teachers in urban areas 
received this training. Twenty-one percent of rural, compared to 11% of urban teachers, received 
support visits at school this past year on how to teach reading. 
Multi-grade teaching: More than half (57%) of the teachers in the sample are teaching 2 or more grades 
this school year. Sixty-one percent of male teachers teach 2 or more grades, while 45% of female 
teachers teach 2 or more grades. Teaching 2 or more grades is more common in rural areas (67%) than 
in urban areas (21%) and slightly higher among paid teacher than volunteer teachers (46% vs 37%). The 
results are presented in Figure 2121. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 28
Figure 21: Distribution of number of grades taught by gender, area, and teacher type. 
Lesson Plans: Sixty-three percent of teachers develop a lesson plan daily, 29% develop one weekly, 2% 
develop one biweekly, and 4% do so monthly. Male teachers tend to develop lesson plans daily more 
often than women, and teachers in rural areas also report preparing lesson plans every day in higher 
proportion than their urban colleagues. Figure 22 shows details. 
Figure 22: Distribution of frequency of developing a lesson plan, by gender, area, and 
teacher type. 
Despite the reports, only 44% of teachers had a lesson plan that they could show to the enumerator on 
the day of the interview. There were some differences between male (46%) and female (37%) teachers, 
and between rural (41%) and urban (58%) teachers. There was a very small difference between paid 
(45%) and volunteer (43%) teachers. The results are presented in Figure 2323. 
39%
55%
79%
33%
46% 37%
61%
45%
21%
67%
54% 63%
0%
20%
40%
60%
80%
100%
Male Female Urban Rural Paid Volunteer
1 grade 2 or more grades
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Overall
Male
Female
Rural
Urban
Paid
Volunteer
Daily Weekly Bi-weekly Monthly Never
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 29
Figure 23: Teacher had a lesson plan for the day of the interview, and teacher has a 
schedule time to prepare lesson plan during school day, by gender, area and teacher type. 
Additionally, 48% of teachers say they have time during school hours to prepare a lesson plan. There are 
differences between male (54%) and female (26%) teachers, between rural (54%) and urban (26%) 
teachers, and between paid (36%) and volunteer (70%) teachers. The results are presented in Figure 
2323. 
Teaching Practices: Teachers were asked to report the frequency in which they perform seven in-class 
activities over the last 5 school days: i) Class repeats letters or words; ii) Students sounded unfamiliar 
words; iii) Students read aloud; iv) Students learn meanings of words; v) Students retell stories; vi) 
Students read during school time; and vii) Students have reading homework. 
Between 67% (students sounded unfamiliar words) and 81% (students have reading homework) of 
teachers report doing all activities at least sometime over the past 5 days. The least frequent activity 
reported by teachers (22% of them reported never doing it) is students sounding out unfamiliar words, 
while the most frequent is students learning the meanings of words (20% reported doing this every day). 
The results are presented in Figure 2424. 
46% 37% 41%
58%
45% 43%
54%
26%
54%
26%
36%
70%
0%
20%
40%
60%
80%
100%
Male Female Rural Urban Paid Volunteer
Teacher has a lesson plan for today Time during school time scheduled to prepare lesson plan
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 30
Figure 24: Distribution of frequencies for the different in-class activities. 
Reading Curriculum Use: Seventy-six percent of teachers reported using the official reading curriculum 
at least sometimes. There are differences between male (78%) and female (70%) teachers, between rural 
(80%) and urban (63%) teachers, and between paid (78%) and volunteer (73%) teachers. The results are 
presented in Figure 2525.
Figure 25: Frequency of use of official reading curriculum in classroom by gender, area and 
teacher type. 
Teacher Guides: 8 out 10 teachers have teacher guides (Figure 26, Panel A) and all of them find them 
very or moderately useful (Figure 26, Panel B). There are some differences in having teacher guides 
between male (84%) and female teachers (75%), between rural (86%) and urban (68%) teachers, and a 
small difference between paid (83%) and volunteer (80%) teachers. 
0% 20% 40% 60% 80% 100%
Class repeats letters or words
Students sounded unfamiliar words
Students read aloud
Students learn meanings of words
Students retell stories
Students read during school time
Students have reading homework
Never Sometimes Frequently Every day
78% 70%
80%
63%
78% 73%
22% 30%
20%
37%
22% 27%
0%
20%
40%
60%
80%
100%
Male Female Rural Urban Paid Volunteer
Sometimes or more often Never
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 31
Figure 26: Teacher has teacher guide and how useful teachers find them by gender, area 
and teacher type. 
Panel A 
Panel B 
Most teachers find the guides useful, although they are less popular among urban teachers. 
Student Assessments Used by Teachers: almost all teachers (97%) use written tests to measure 
students’ progress, followed by oral evaluations (70%) and homework (49%). These results are 
presented by the gray bars in Figure 2727. 
82% 84%
75%
86%
68%
83% 80%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Overall Male Female Rural Urban Paid Volunteer
79% 75% 81%
68%
81% 73%
4% 0%
4%
0%
2%
7%
0%
20%
40%
60%
80%
100%
Male Female Rural Urban Paid Volunteer
Very Useful Moderately Useful
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 32
Expected Reading Skills: teachers were asked about the reading skills that they expect their students to 
have. Most of them (80%) mentioned to read grade level stories, but only 44% of them mentioned that 
they expect their students to understand the stories they read. Not many (20%) teachers mentioned 
sounding out unknown words. Even fewer (10%) mentioned that they expect their students to know 
letter names. We speculate that this pre-literacy skill did not come to many teachers’ minds because it is 
a basic skill which most grade 2 students have. In fact, the average number of correct letter names that 
students identified was 65 per minute. These results are presented by the blue bars in Figure 2727. 
Figure 27: Teachers activities to measure students’ progress and expected reading skills. 
Additional Efforts to Improve Reading: around 73% of teachers say they have made special efforts to 
improve reading skills. There are differences in whether the teachers made a special effort to improve 
reading between male (77%) and female teachers (60%), between rural (79%) and urban teachers (53%), 
and between paid (69%) and volunteer teachers (80%). The results are presented in Figure 2828. 
10%
19%
21%
44%
80%
2%
12%
26%
49%
70%
97%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
 Know letter names
 Other
 Sound out words they don't know
 Understand stories that they read
 Read grade level stories
Their portfolios and other projects
Other
End of term evaluations
Their homework
Oral evaluations
Written tests
Expected reading skills Measuring Students' Progress
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 33
Figure 28: Amount of extra effort made compared to a normal year, by gender, area and 
teacher type. 
Among those teachers who made special efforts to improve reading, more than half report exerting a 
good bit or a lot more effort than they had in past years. Slightly more female teachers exerted a good 
bit or a lot of effort (58%), while half of male teachers did the same. The difference is greater by 
urbanity with 56% of rural teachers saying they exerted a good bit or a lot effort while only 30% of 
urban teachers did the same. There are no large differences between paid teachers (51%) and volunteer 
teachers (52%). The results are presented in Figure 2828. 
Additionally, around 29% of teachers suggested that they exerted special effort because training showed 
them better ways to teach reading and 17% suggested that it was important because the children were 
being tested. Only around 5% of teachers suggested that they saw other schools doing it.10 The results 
are presented in Figure 2929. 
 
10 63% of teachers suggested other reasons for exerting the special effort. This might be an important issue to address in 
questionnaires moving forward. 
77%
60%
79%
53%
69%
80%
0%
20%
40%
60%
80%
100%
Male Female Rural Urban Paid Volunteer
A little Some A good bit A lot
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 34
Figure 29: Reasons why teacher exerted special effort to improve reading. 
Some teachers were less likely to say they received training on how to better teach reading as 26% of 
male teachers compared to 42% of female teachers offered this explanation, urban teacher (10%) are 
less likely than rural teachers (33%), and volunteer (21%) are less likely than paid teachers (34%) to say 
they received this training. The results are presented in Figure 3030. 
Figure 30: Percentage of teachers receiving reading training and thinking it might be 
important because children were tested, by gender, area and teacher type. 
Additionally, there were some differences among teachers who suggested they thought it might be 
important because the children were being tested between rural (20%) and urban (0%) teachers, and 
between paid (20%) and volunteer (13%) teachers. There was no difference between male and female 
teachers. The results are presented in Figure 3030. 
5%
9%
9%
17%
29%
0% 20% 40% 60% 80% 100%
We saw other schools doing it
The Ministry told us to do it
An NGO told us to do it
We thought it might be important because the
children were tested
Teachers got teachers training that showed them
how to teach reading better
26%
42%
33%
10%
34%
21% 17% 17% 20%
0%
20%
13%
0%
10%
20%
30%
40%
50%
Male Female Rural Urban Paid Volunteer
Teachers got teachers training that showed them how to teach reading better
We thought it might be important because the children were tested
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 35
Principal and School Questionnaire 
The survey asked the principal or highest level administrator available (if the principal was not available 
at the time of the visit) about his/her characteristics and activities to improve reading activities in their
school. The principal was interviewed 78% of the time, 16% of respondents were vice principals, and 7% 
of these interviews were conducted with someone from another administrative position. In the case of 
the principals, 94% of them were men and among vice principals, 71% were men. These results can be 
found in Figure 3131. 
Figure 31: Type of respondent to the principal questionnaire and distribution by gender 
and area. 
The average number of years in the position of the principal or vice principal is 6.6 years. The tenure is a 
longer for men (6.9 years) than for women (4.3 years), and for rural areas (7 years) than for urban areas 
(5.2 years). The results can be found in Table 13. 
Table 13: Average years in principal (vice-principal) position. 
Average Observations
Male 6.9 81
Female 4.3 9
Urban 5.2 19
Rural 7 71
Total 6.6 90
The majority of respondents to the principal survey have C, B, or AA Certificate (75%). Only 16% have 
a bachelor’s, master’s, or other type of college degree. About 4% of male principals a have senior high 
school degree only, as well as 4% of rural principals. The results are presented in Figure 3232. 
94%
71% 83%
6%
29% 17%
83%
64% 67%
17%
36% 33%
0%
20%
40%
60%
80%
100%
Gender Area Gender Area Gender Area
Principal (77.8%) Vice Principal (15.6%) Other (6.7%)
Male Female Rural Urban
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 36
Figure 32: Respondents to the principal survey education degree, by gender and area. 
About 68% of respondents to the principal survey have received or taken courses to teach or support 
the teaching of reading. There are some differences between male principals (69%) and female principals 
(56%), and between urban (42%) and rural (75%) principals. The results are presented in Figure 3333. 
Figure 33: Percentage of respondents to the principal survey that received training to 
teach reading, by gender and area. 
Nearly all (97.8%) respondents report reviewing the lesson plans of first and second grade teachers and 
86% of respondents say that the lessons plans are reviewed every week or more often. There are no 
substantial differences between genders nor between areas. 
0% 20% 40% 60% 80% 100%
Total
Male
Female
Urban
Rural
Senior High School
Associate
Certificate C
Certificate B
Certificate AA
Bachelors Degree
Master or other
68% 69%
56%
42%
75%
0%
20%
40%
60%
80%
100%
Total Male Female Urban Rural
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 37
Figure 34: Review of teachers’ lessons plans, by gender and area. 
Furthermore, 93% of respondents say that teachers’ lessons in grade 1 and 2 are observed in their 
classrooms. 
When asked about the frequency in which lessons are observed, there are differences by the gender of 
the respondent; 87.5% of female respondents suggested they observe lessons at least once a week or 
more often, while 94.6% of male respondents suggested the same. All urban respondents say they 
observe lessons at least once a week or more often, while 92.2% of rural respondents say they do so 
this frequently. The results are presented in Figure 3535. 
Figure 35: Frequency of observation of teachers’ lessons in grade 1 and 2, by gender and 
area. 
98% 98%
100% 100%
97%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Total Male Female Urban Rural
62.70%
61.30%
75.00%
57.90%
64.10%
31.30%
33.30%
12.50%
42.10%
28.10%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Total
Male
Female
Urban
Rural
Daily Once every week Once every two weeks Once every month Once per year
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 38
Seventy percent of the respondents say they have made special efforts to improve reading at their 
schools the past year. There are differences by respondent gender where 74% of male respondents say 
they made a special effort to improve reading, while only 56% female respondents say the same. Also, 
there are differences by area where 68% of urban respondents say they are making a special effort to 
improve reading, while 73% of rural respondents say the same. The results are presented in Figure 
36:36. 
Figure 36: Principal made a special effort to improve reading at their school by gender and 
area. 
Around 19% of respondents claim to have exerted special effort because they thought it might be 
important due to testing, while 15% say that the ministry told them to do it, and 12% say that teachers 
got teacher training which showed them better methods for teaching reading. Only around 6% of 
teachers say they increased their own effort because they saw other schools doing it. Forty-six percent 
of the respondents suggested other reasons for exerting the special effort.11 The results are presented in 
Error! Reference source not found.37. 
 
11 This might be an important issue to address in questionnaires moving forward and understand motivation. 
72% 74%
56%
68% 73%
0%
20%
40%
60%
80%
100%
Total Male Female Urban Rural
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 39
Figure 37: Reasons why respondents to principal survey exerted special effort to improve 
reading at their school this past year. 
The Determinants of EGRA Scores 
In addition to the descriptive analysis presented above, it is possible to use the baseline data to look into 
how student characteristics, household context, school and teacher characteristics correlate with 
observed differences across EGRA scores. In this section, we present results for a series of statistical 
processes for estimating the relationships (regressions) among available variables and EGRA subtask 
scores. 
A large number of variables could be related to EGRA scores. We experimented with a range of 
specifications that include student characteristics and behavior, household characteristics, proxies for 
socio-economic status of the household, teacher characteristics and classroom practices, and school 
characteristics. In particular, we have included 
1. Student’ characteristics: age, sex, rural or urban location 
2. Home environment: home language, parents’ literacy status, household owns a working TV, 
radio, has electricity, student ate before going to school on the day of the assessment. 
3. Home reading practices: student reads aloud to someone at home, someone reads to the 
learner at home 
4. Teacher characteristics: sex, years of experience, paid or volunteer status, education degree, 
teacher received training to teach reading, teacher has lesson plan available 
5. School characteristics: sufficient materials/textbooks for Grade 2, has library or reading room. 
Each of these variables details are presented in Table 14 below. 
2%
6%
12%
15%
19%
0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20%
An NGO told us to do it
We saw other schools doing it
Teachers got teacher training that showed them
how to teach reading better
The ministry told us to do it
We thought it might be important because children
were tested
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 40
Table 14: Explanatory variables for EGRA scores. 
Variable Mean SE Obs. Min Max
Girl 48% 1% 1522 0 1
Rural 78% 4% 1522 0 1
Student age 12.5 0.1 1481 6 21
Student reads aloud to someone at home 62% 2% 1509 0 1
Someone reads to the student at home 62% 2% 1506 0 1
Language Home: English 20% 2% 1522 0 1
Parents cannot read/write 17% 2% 1522 0 1
Do you watch television at home? 29% 2% 1517 0 1
Do you listen to radio at home? 80% 2% 1512 0 1
Do you have electricity at home? 31% 2% 1515 0 1
Did you eat before coming to school today? 60% 2% 1514 0 1
Female Teacher 23% 5% 1509 0 1
Volunteer Teacher 34% 5% 1522 0 1
Teacher's years of experience 11.2 0.9 1509 0 37
Teacher receive training to teach reading 64% 5% 1509 0 1
Teacher had lesson plan available 44% 5% 1491 0 1
Teacher has Junior/Senior High School 37% 5% 1509 0 1
Teacher has Certificate degree 57% 5% 1509 0 1
Teacher has BA degree 2% 2% 1509 0 1
Do you have sufficient resource materials/
textbooks for Grade 2?
8% 3% 1522 0 1
Do you have a library or reading room? 23% 4% 1522 0 1
While the full specifications of the regression analyses can be found in Annex VI: Full Regression Results, 
here we summarize the results. 
Being a girl has a negative correlation across scores when compared to boys. For instance, when all else 
is equal, girls correctly identify 4.5 fewer letter names, 2.6 fewer familiar words, and read 2.8 fewer 
words from a paragraph than boys. Similarly, students in rural areas show a large disadvantage when 
compared to their urban counterparts. Urban learners identify 10 fewer letter names, 4 fewer familiar 
words than urban students, 6.6 fewer words from a connected paragraph than urban students. These 
results could be reflecting different aspects associated with rurality, such as lower socio-economic 
circumstances, but could also signal less exposure to print, or lower school quality, etc. We include in 
our analysis other variables such as parental literacy, having TV, radio and electricity at home, or eating 
before going to school to take into account the learner socio-economic background but rural status 
In general teacher characteristics other than experience are not significantly associated with students’ 
performance. Teacher sex, volunteer or paid status, training to teach reading, having written lessons 
plans or education degree have no significant correlation with EGRA scores. Only years of experience is 
associated with better scores but only in two EGRA subtasks, familiar word reading and oral reading 
fluency. Similar results have been found in other studies (for example, Luschei & Carnoy (2010), Rockoff 
(2004), Clotfelter et al. (2007)) 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 41
Reading aloud at home is correlated with higher reading scores. The association does not prove 
causality from reading at home to better performance as reverse causality could be present. This is, it 
could be the case that only those learners that are able to read fairly well, read aloud at home. 
In general, having materials, a library or a reading room in the school does not appear significantly 
associated to student performance. 
CONCLUSIONS & 
RECOMMENDATIONS
In this report we present the findings from the Early Grade Reading Assessment of Grade 2 students 
and complementary teachers and principal surveys, which constitute the baseline for the Impact 
Evaluation of the Read Liberia Activity. The baseline data collection did not face significant challenges and 
the tests show very high reliability and consistency in the assessment work. 
Appropriate ages for grade 2 are 7–8 years. However, very few students in Liberia start primary school 
on time. Students in our sample are substantially older (12.5 years on average) than what is expected for 
second graders reflecting this phenomenon. More research should be conducted to find how to address 
the problem. Issues related to school entry age, attendance, and repetition should receive attention. 
Student absenteeism and grade repetition, as reported by the learners, are high. Addressing absenteeism 
is very important as most activities planned for Read Liberia will take place in school and absenteeism 
could reduce the activity potential impact. 
Average scores in EGRA subtasks are in general low for second graders. Students know letter names 
but their decoding skills are low. This is evident in the high percent (81%) of learners that are not able 
to sound a single non-word. Reading ability is low as well. Thirty-five percent of the students are non￾readers –they cannot read a single word from a short grade level paragraph- and on average oral reading 
fluency is 14.6 words per minute which is far from the levels needed to be able to comprehend the text 
read. Excluding non-readers, the average oral reading fluency of the rest of the students is 22.4 words 
per minute, and only 22% of them read 35 or more words per minute. On average, girls tend to score 
lower than boys in all EGRA subtasks and rural students’ performance is substantially lower than that of 
urban learners. 
A third of the teachers are volunteer teachers. Their reported teaching practices and behaviors are 
similar to those of paid teachers. Having a volunteer teacher rather than a paid teacher does not seem 
associated with differential student performance. Teacher education and reported practices do not 
correlate with assessment scores either. Most teachers reported following the official curriculum, 
having teacher guides and preparing lessons plans. Despite that almost all teachers reported preparing 
daily lesson plans every day or once a week, less than half the teachers were able to show the 
interviewers the written lesson plans for the day of the visit. It is interesting to also note that almost all 
respondents to the principal questionnaire reported reviewing teachers’ lesson plans and a large 
majority claimed to do so every week or more often. The contrast between reports and lesson plans 
available suggests desirability bias in the answers of teachers and principals regarding their classroom 
practices. In the future, it would be useful to collect more information through classroom observations. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
BASELINE REPORT | 42
The classroom observations can be designed to capture elements specific to the Read Liberia pedagogy 
and contain both low inference and higher inference items. Low inference items can include 
management tools, lesson plans, presence and quantity of instructional materials, use of them; 
examination of timetables to assess time allocated to reading specifically; examination of learners’ 
workbooks and exercise books to assess curriculum coverage; etc. High inference measures should 
include pedagogical process variables. Based on the results, additional monitoring and other incentives 
could be included in the activity to improve actual behavior if needed. 
REFERENCES 
Clotfelter, C.T., Ladd, H.F. & Vigdor, J.L. 2007. Teacher credential and student achievement: Longitudinal 
analysis with student fixed effects. Education of Education Review. 26:673-682. 
DeStefano, Joseph, Timothy Slade, and Medina Korda (2013). Liberia Teacher Training Program (LTTP): 
Midterm Assessment of the Impact of Early Grade Reading and Math Interventions 
Fleiss, J. L. (1981). Statistical methods for rates and proportions (2nd ed.) New York: John Wiley. 
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. 
Biometrics, 33, 159–174. 
Luschei, T.F., & Carnoy, M. 2010. Educational production and the distribution of teachers in Uruguay. 
International Journal of Educational Development. 30(2):169-181
Rockoff, J.E. 2004. The impact of individual teachers on student performance: Evidence using panel data. 
American Economic Review. 94(2):247-252. 
ANNEXES | 43
ANNEXES 
ANNEX 1: EVALUATION STATEMENT OF WORK | 44
ANNEX I: EVALUATION STATEMENT OF WORK
Contract No. AID-OAA-M-13-00010 
Activity: Read Liberia Baseline and Impact Evaluation 
October 3, 2017 
Statement of Work 
Budget Estimate: $905,000 
Period of Performance: January 1, 2017-September 2020 
INTRODUCTION: 
This Statement of Work (SOW) outlines the details of a baseline data collection and impact evaluation 
of the upcoming Read Liberia activity, to be carried out by NORC at the University of Chicago, the 
prime contractor of the Reading and Access contract. The Reading and Access Contract This SOW 
covers contextual information as well as sub-activities already carried out by NORC and those activities 
to be handled in the remaining time on this evaluation activity. This abbreviated SOW takes the place of 
a formal Statement of Work, which was not developed prior to the inception of this activity. The 
information in this SOW draws from conversations held among USAID/Liberia, USAID/Washington 
(E3/ED), and NORC around the design of the evaluation and decision memos created prior to the 
development of an Evaluation Design Report and Data Collection. The Evaluation Design Report also 
provides additional details to the approach being taken and informed the development of this SOW. 
In the remaining sections of this SOW, we provide the following information: 
● Background Information on Read Liberia Activity 
● Evaluation Objectives 
● Evaluation Questions and Methodology 
● Deliverables 
● Period of Performance and Budget Estimate 
Background Information: 
USAID/Liberia is in the process of procuring a new basic education activity, designed to provide 
technical assistance to improve early grade reading skills for Liberian grade one and two students in in 
six targeted counties. The activity will also pilot test emergent literacy skills for Liberian students in 
selected public kindergarten schools. USAID/Liberia has an expected award date in September 2017. 
The new activity will be informed by lessons learned and experience from previous reading initiatives in 
Liberia, including follow-on to the Liberia Teacher Training Program, (LTTP II). Details of the approach 
to be taken will be finalized with the implementing partner post-award. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 45
Project Context: 
Since 2005, a number of education reforms aimed at restoring the educational infrastructure and 
reconstructing schools to accommodate the increasing population of school-age children have been 
championed by the Government of Liberia. The Government of Liberia’s overarching goal of the 
education emergency response in 2005 was to get as many children, as quickly as possible, into school 
and deal with the quality of education issues in the future. As a result, the number of students enrolled 
in schools has increased significantly. Per the 2014 Census, the Ministry of Education now estimates a 
gross enrollment rate of 57.7% in grades one to six, this is a large improvement over the 31.8% gross 
enrollment in 2012, and prior years. 
While enrollment has increased the issues of educational quality and efficiency at all levels of the system 
are a major challenge. Dropout is high, repetition is frequent, and student outcomes on system exams 
remain far from satisfactory. English is the official language of instruction in Liberia public schools, aimed 
at fostering a sense of national unity. 
The Government of Liberia has implemented a robust policy framework to address these persistent 
hindrances to quality. The Education Sector Plan (ESP) (2010-2020); the Education Reform Law of 2011; 
the Roadmap for System Transformation; the National Agenda for Transformation (NAfT); and the 
MOE three-year Operational Plan (OP, 2014-2017) all highlight the importance of ensuring that Liberian 
children receive the “minimum stipulated quality (ESP).” Specific, reading-related initiatives incorporated 
in these plans include: 
• The establishment of Liberia as a middle income country by 2030 (Education Reform 
Law and Poverty Reduction Strategy) 
• A plan to decentralize the education system and create a management structure that is 
more locally focused (Education Reform Act of 2011- See Section J, J.4) 
• The introduction of annual learning achievement tests in early grade reading (ESP) 
• Analysis of test scores and their use to inform policy decisions as well as action (ESP) 
• The prioritization of human development to address capacity and inclusiveness, and 
the improvement of the quality of basic education through soliciting community 
oversight (NAfT) 
• A plan to Eradicate illiteracy to reduce the risk of conflict (NAfT) 
• Results-based planning (MOE three-year plan -“Quality education for all-redirecting 
our future”) 
• The monitoring of all training activities with emphasis on mainstreaming early grades 
reading and math (OP, Objective 1, Output 2, Activity 2.2) 14 
• Supplementary readers provided to public and community schools grades 1-4 (OP, 
Objective 1.4, Output 2) 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 46
Even with these ambitious initiatives, student results remain very low. As the USAID/Liberia Teacher 
Training Program II (LTTP II) results indicate, students make more progress when they receive regular, 
well-designed reading lessons. Therefore, the new activity will be structured to build on and supplement 
these initiatives. 
Read Liberia: 
While the Read Liberia Activity has not yet been awarded, the request for proposal (RFP), articulated a 
clear problem statement and hypothesis. The problem statement is as follows:12
The USAID/Liberia Development Objective hypothesis for Read Liberia states that: 
IF: 
1. Kindergarten students are taught the oral vocabulary lexicon needed for emergent literacy in 
English; 
2. Official time allocations for reading instruction in the early grades are increased and enforced; 
 
12 The problem statement and hypothesis are stated on pages 11 and 12 of Request for Proposal (RFP) # SOL-669-
17-000004, Read Liberia! 
PROBLEM STATEMENT: EARLY GRADE READING
1. Liberian students in grade one and two are not learning to read with fluency 
and comprehension because:
 Children in kindergarten are not being exposed early to oral vocabulary 
needed for emergent literacy.
 Time allocations for reading instruction in the early grades during the school 
day are insufficient
 Data to drive system-wide decision-making about early grade reading is not 
regularly collected
2. An evidence-based reading curriculum, with accompanying materials 
appropriate for reading instruction in the early grades, exists, but has not been 
incorporated into the primary grade comprehensive curriculum
3. Teachers’ levels of literacy are very low. More time and resources are needed 
to coach teachers to become better readers.
4. Teachers receive insufficient training in reading instruction.
5. Monitoring and coaching systems for early grade reading instruction are 
under-developed.
6. Students’ reading skills are not routinely assessed in the early grades.
7. Some children have limited access to print in their home environments.
8. Some parents and other family members lack the skills and time to support 
children’s acquisition of reading skill.
9. Investment by the Government of Liberia and partners in early grade reading 
remains low.
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 47
3. Data about early grade reading is collected and used to drive system-wide decision-making; 
4. Reading intervention is refined and simplified with the key components needed for children to 
learn to read words and to understand what they read are preserved and improved if necessary; 
5. Texts appropriate for early grade reading instruction are improved and additional leveled texts 
created; 
6. Teachers receive intensive training and systematic coaching in effectively teaching the early grade 
reading curriculum; 
7. Trained teachers receive evidence based, scripted reading lessons and materials to support 
reading improvement in the early grades; 
8. Teachers are monitored and supported in their classrooms; 
9. Children’s reading skills in the early grades are routinely assessed and children provided 
opportunities to practice their reading skills at home; 
10. More parents and other family members learn how to support their children in learning to read; 
and Private and public sources of funding to support early grade reading progress are identified; 
THEN: Students will be able to read with fluency and comprehension at the end of second grade. 
In addition to the aforementioned problem statement and development objective, the RFP for the Read 
Liberia activity outlines the intended results and provides illustrative examples of the sub-activities the 
implementing contractor may implement to reach the intended outcomes.13 The intermediate results for 
Read Liberia targets include the following: 
● Increased government commitment to and support of evidence-based reading instruction 
● Improved early grade reading classroom instruction 
● Improved service delivery systems 
● Increased parent, community and private support for Early Grade Reading. 
The RFP also includes a table of technical deliverables that outlines specific activities such as coaching, 
the development of new materials, and an MOE approved teacher's manual. 
It is important to emphasize that the manner in which the Mission structured the Read Liberia RFP 
allows for the development of sub-activities that will help the Mission reach its intended outcomes. This 
is important because it could potentially place some limitations on the design of the evaluation described 
herein. 
 
13 See Appendix A for SOW (Section C) of Read Liberia RFP. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 48
School Selection: 
USAID/Liberia, in collaboration with the Ministry of Education, has defined a set of criteria that will be 
used to identify schools that are eligible to receive the READ Liberia activities. These criteria include 
schools that: 1) are public; 2) are located in the six selected counties (Montserrado, Margibi and Bong, 
Grand Bassa, Nimba and Lofa); and 3) have grade one and grade two classes. To strengthen the design 
of the Impact Evaluation, NORC has asked that the schools have at least 20 students in grade 2. 
During the data collection process, the evaluation team learned that there was an agreement with the 
Liberian government that there be no overlap between the sample of schools selected for Read Liberia 
and for the Read Liberia Evaluation and the Partnership Schools for Liberia (PSL) treatment and control 
schools. 
It will be essential for USAID/Liberia and its implementing contractor to follow these criteria as they 
serve as the basis for the sampling frame and the sample size calculations. If modifications are made, it 
could jeopardize the study. 
EVALUATION OF READ LIBERIA 
Evaluation Objectives, Intended Audience and Questions:
USAID/Liberia had two primary objectives when it reached out to USAID/Washington and NORC. 
 OBJECTIVE ONE: undertake a baseline study that will inform USAID/Liberia and their 
contractor about the population with which they will be working-including schools, teachers and 
students. 
 OBJECTIVE TWO: Conduct an impact evaluation to measure the impact of the Read Liberia 
activity on reading outcomes. 
Objective One: Informing Activity Design 
The baseline will be used to provide the Mission and Implementing Partner descriptions of the schools, 
students, and teachers included in the population of schools they will target to help them finalize the 
design of the sub-activities under Read Liberia. While both the information gathered about students’ 
performance on literacy tasks and all the information gathered through the supplemental data collection 
(teacher survey, principal survey, and brief student survey) can be a helpful resource for understanding 
the population that will be served, USAID/Liberia are particularly interested in a several topics. First, 
they want to understand, from the students’ perspectives, if teachers are working with students on 
letter sounds and if students are taking books home to read. Second, they want information about the 
qualifications of the teachers and whether or not the current teachers are government paid teachers or 
volunteers. 
Objective Two: Examining the Results/Impact of Read Liberia 
In addition to understanding the background of the activity participants and the context in which the 
activity will occur, USAID/Liberia also wants to assess the results of the Read Liberia activity. More 
specifically, the Mission would like to examine the impact of the activity on grade 2 students’ ability to 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 49
read and understand grade level text in English after two years of schooling. Students’ literacy gains will 
be assessed using the Early Grade Reading Assessment (EGRA). 
During the design phase, USAID and the evaluation team identified three potential options for the 
impact evaluation. 
1. An impact evaluation that examines the teacher to coach ratio needed for the activity to be 
effective and the effectiveness of teachers with different levels of coaching on reading fluency 
and comprehension of second grade students. 
2. An impact evaluation with two treatment arms. One arm would target schools with a new 
kindergarten sub-activity. The second arm would target schools that received the remaining 
treatment without the kindergarten sub-activity. This design would allow us to assess the degree 
to which the activity as a whole had an impact on reading fluency and comprehension of second 
grade students, and the difference that adding a kindergarten component would have on impact, 
if any. 
3. An impact evaluation with a treatment and control that examines the impact of all Read Liberia 
interventions, as a package, on reading fluency and comprehension of second grade students. 
USAID/Liberia, USAID E3/ED, and the NORC team determined that the third was the best option due 
to a number of factors. First, USAID/Liberia concluded that the first option was not likely the best use 
of resources because they wanted to make sure that there weren’t schools that were deprived of 
coaches or a potentially effective coach:teacher ratio. The second option was of great interest, especially 
given the political situation in Liberia and the role KG has played there and the lack of evidence we have 
on the effectiveness of literacy activities in kindergarten. Despite this, with a total budget of $900,000, it 
was not feasible to undertake an impact evaluation with three distinct arms. However, if additional funds 
became available, the team may consider adding an arm and undertake baseline data collection for this 
component in either the fall or spring. USAID will give NORC at least five months notice if they decide 
to add an arm for kindergarten so that NORC has sufficient time to prepare a new tool to assess 
literacy skills. 
Taking the third approach also allows some flexibility in the design of the sub-activities, which is 
especially important because the Read Liberia contract has not yet been awarded. 
To meet this objective, NORC will answer the following evaluation questions: 
What are the features of the population that will be served through Read Liberia? 
What is the impact of the Read Liberia activity on the reading fluency and comprehension of 
second grade students? 
Evaluation Design and Methods: 
Based on the objectives and evaluation questions identified above, NORC implemented a baseline data 
collection during the 2016-2017 school year. The sampling approach taken for the baseline data 
collection will allow for an impact evaluation of the Read Liberia activity as well as a description of the 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 50
population served. It is crucial, however, that the Mission and Implementing partner adhere to the same 
selection criteria used for the baseline in order to select schools. If differing criteria are used, this could 
jeopardize the design, as the sample may no longer represent the activity as a whole. The sections below 
provide details related to the expectations for the design, data collection, and analysis. 
Design: 
NORC was responsible for proposing an evaluation design. The design and analysis plan proposed 
consists of the following features: 
● The primary outcome of interest for the IE is oral reading fluency and reading comprehension 
● Randomized control trial of Read Liberia activity, with a single treatment arm 
● Two rounds of data collection - at the baseline and at the endline 
● Point of randomization at the school level 
● Endline to occur either two or three years after implementation has commenced 
● Data will be collected using four tools: 
○ EGRA 
○ Student Survey 
○ Teacher Survey 
○ Principal Survey 
● The EGRA instrument previously used for the Liberia Teacher Training Program (LTTP) and 
follow-on, LTTP II will be used with minimum modification 
● Power calculations to determine sample size rely on the following assumption: 
○ α of .95 
○ β of .80 
○ ρ (or ICC) of .3 
○ r² of .20 
○ MDES of .34 
● NORC in collaboration with STS and a local subcontractor will undertake enumerator training 
and data quality control. 
● At a minimum, NORC’s analysis plan should include an estimate of the impact of Read Liberia 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 51
on reading fluency and reading comprehension, as well as sub-group analyses of the impact of 
these indicators on boys and girls. Other sub-group analysis and analysis of other key 
dimensions of the EGRA may be proposed by NORC and approved by USAID/Liberia. 
While the NORC evaluation team will be responsible for carrying out the baseline and impact 
evaluation, it will be up to USAID/Liberia and their contractor to carry out effective activity monitoring 
and activities such as data collection on the fidelity of implementation, classroom observations, and the 
collection of other performance data, such as annual learning assessments. 
This is important because the impact evaluation will be supplemented with descriptive data gathered 
throughout the duration of Read Liberia implementation. These data will be gathered by the contractor 
that will be implementing the activity and will not only supplement the impact evaluation, but will also 
provide information that can help guide programming as the activity progresses. Data that will be 
systematically collected include, but are not limited to the following: 
○ LQAS 
○ Classroom observation 
○ Other relevant implementation data (including teacher retention and turnover) 
If implementing contractor does not gather this information, it will not be included in the evaluation. 
Additionally, USAID/Washington and NORC expect that USAID/Liberia and its contractor to keep 
NORC aware of any potential contamination issues such as employment of coaches at control schools. 
Instrument and Data Collection: 
While some raised concerns regarding test security of the EGRA version that used in 2011, the team 
elected to move forward using the same version of the test due to limitations in the budget and timeline. 
The evaluation team collected the data electronically. NORC and their partners were responsible for 
training local enumerators, soft piloting of the instrument and undertaking data quality checks. 
NORC undertook data collection for the baseline in June 2017. The evaluation team will collect endline 
data either in April/May, 2019 or April/May, 2020. NORC will use the same version of the instruments 
used at baseline. 
NORC gained approval for the research from both the NORC IRB at the University of Chicago and the 
University of Liberia, and has worked in close collaboration with the Ministry of Education. The NORC 
evaluation team trained all team members from the local data collection firm in ethics and all team 
members committed to comply with child protection policies. 
Deliverables and Audience: 
Table 1.0 below reflects the deliverables as required Section F of the Read and Access Contract, 
Contract Number AID-OAA-M-13-00010. Descriptions of the standards for each deliverable are 
located under the results in Section C.4 of the contract. Table 1.0 also includes notes regarding the 
status of the deliverable. 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 52
USAID/Washington provided a report template to NORC for the Baseline report. USAID expects all 
reports to follow guidance in the Reading and Access contract. The evaluation team should develop 
clearly written reports in plain English with a non-technical audience in mind. USAID will identify specific 
targeted audiences prior to the creation of the deliverable so that the evaluation team may target that 
audience. 
Deliverabl
e Number
Type of Written 
Deliverable/Repor
t 
Due Date Audience Format Notes
1.1 Consultation Plan Within 30 
calendar days 
after written 
notification of an 
evaluation task
N/A Email An outlined plan 
was not 
communicated. The 
process unfolded 
more organically as
a formal SOW for 
the task was not 
issued
1.2 Draft Evaluation 
Plan
Within 60 
calendar days of 
receiving written 
notification of the 
evaluation task or 
as agreed upon 
by CO
USAID/Washing
ton, 
USAID/Liberia
Email Received and 
accepted
1.3 Final Evaluation
Plan
Within 90 
calendar days of 
receiving written 
notification of the 
evaluation task or 
as agreed upon 
by the contractor 
and COR
USAID/W, 
USAID/Liberia
Email Received and 
accepted
1.4 Data Collection 
Instruments for 
Data collection 
and Analysis for 
Impact Evaluation
30 days before 
pilot testing 
begins
USAID/W, 
USAID/Liberia
Email Received and 
accepted
1.6 Evaluation Baseline 
Report and Data 
Files
Within 60 days of 
completion of 
baseline data 
collection or as 
USAID/W, 
USAID/Liberia, 
MOE, 
Implementing 
Email Data collection 
completed prior to 
June 1. Deadline for 
the baseline report 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 53
agreed upon by 
the contractor 
and COR
contractor, 
Development 
Partners
extended to allow 
for USAID to 
provide additional 
guidance on what 
was to be included 
in the report. 
Report Draft 
submitted August 
25, 2017
1.8 Final Evaluation 
Report and/or 
Summative 
Evaluation Report 
and Data Files
Within 60 
calendar days of 
end-line data 
collection or as 
agreed upon by 
the contractor 
and COR
Email End line currently 
slated to take place 
in 2019 or 2020. 
2.1 Evaluation Results: 
Briefing 
documents, 
summaries and 
analysis
Meeting with IP 
following award
Two that are 
suggested, one is a 
meeting with IP 
once award made 
and second one 
after endline to 
debrief Mission on 
findings 
Evaluation Budget and Period of Performance: 
The estimated budget for the evaluation is $905,000. The current Reading and Access contract was 
modified and is now active through September 29, 2021. The period of performance for this specific 
task, Impact Evaluation of Read Liberia runs from January 1, 2017 through September 2020. 
Under this task order and at the request of USAID/Liberia, the baseline data collection occurred prior 
to the end of the 2016-2017 school year and prior to awarding the Read Liberia contract. This is 
important to note as it may create challenges down for the impact evaluation. However, USAID/Liberia 
chose to take this approach so that the information gathered during the baseline could be used to make 
programmatic decisions. 
While the period of performance runs through 2020, the evaluation team may complete the work in 
2019. This is because the evaluation team will collect endline data either two years after baseline, in 
2019, or three years after the baseline data in 2020. During the planning phase, USAID/Liberia, E3/ED 
and NORC weighed the pros and cons of waiting three years before collecting endline data. The 
advantage of waiting until 2020 is that it maximizes the time for the intervention to have affect. This 
could increase the impact that the intervention has on outcomes and also allows necessary flexibility 
should the Read Liberia activity start too late in the year. However, there are conditions that may cause 
BASELINE REPORT: USAID/LIBERIA READ LIBERIA IMPACT EVALUATION 
ANNEX 1: EVALUATION STATEMENT OF WORK | 54
a decline in outcomes after two years. If we see a substantial change in staff, with the number of 
experienced staff departing and new staff coming on board, the length of exposure to the treatment 
could decrease which could diminish the size of the impact on learning outcomes. 
Building in midline data collection could help mitigate this risk. However, due to budgetary constraints, 
this is not an option for this evaluation. To reduce this risk, USAID/Liberia and their contractor will 
monitor the turnover of teaching staff and alert NORC if they are losing a substantial number of staff. If 
this occurs, the evaluation team may undertake the endline a year earlier, in 2019. 
ANNEX II: EVALUATION METHODS AND LIMITATIONS | 55
ANNEX II: EVALUATION METHODS AND LIMITATIONS 
Experimental Design 
The evaluation team will use an experimental or randomized control trial (RCT) approach to answer the 
evaluation question, with schools randomly assigned to treatment and control groups. The fact that the activity 
has not started yet and that the total number of schools is greater than the number of schools to be treated in 
the six counties, allows for an RCT. We want treatment and control groups to be as similar as possible; the 
primary advantage of a randomized design is that it minimizes the potential for selection bias- that is, the 
possibility that observed differences in outcomes between the treatment and control groups are due not to the 
impact of the project, but to other systematic differences between the two groups that the evaluation is not able 
to account for. Another advantage is that the randomizing approach also tends to be fairly simple to convey to 
stakeholders and policy makers and fair when not all schools can receive treatment at once. 
We will estimate the causal impact of the program on EGRA scores using regression models of the following 
form: 
Yijt = β0 + β1δtime + β2𝐗ijt + γ(δtime × δtreat) + φj + εijt
where: 
Yijt is the EGRA sub-task score for student i in school j at time t, 
δtime is a dummy variable equal to one at endline, 
δtreat is a dummy variable equal to one for observations in the treatment group, 
Xijt is a vector of control variables, 
φj
 is a school-level fixed or random effect, 
εijt is a random error term, and 
β0, β1 , the vector β2 and γ are parameters to be estimated. 
The parameter of interest is γ, the coefficient on the interaction term between the time and treatment 
dummies. The value of γ will provide an estimate of the impact of Read Liberia relative to the control group on 
EGRA scores. 
The parameter of interest is γ, the coefficient on the interaction term between the time and treatment 
dummies. The value of γ will provide an estimate of the impact of RL relative to the control group on EGRA 
scores. 
Although the evaluation question focuses on the reading fluency and reading comprehension scores, we plan to 
estimate the impact of RL on all the sub-tasks include in the EGRA tool. 
Note also that the models will include a set of control variables Xijt reflecting student characteristics that will be 
obtained from short surveys of students in the course of administering EGRA. Potential control variables include 
age, home language if different from English, whether child reads at home regularly, and some basic questions 
about household assets to estimate wealth status, among others. Incorporating these control variables into the 
analysis will improve the precision of the impact estimates and help controlling for any potential change in 
characteristics of learners between baseline and endline as well. 
ANNEX II: EVALUATION METHODS AND LIMITATIONS | 56
In addition to estimating the overall impact of Read Liberia on reading outcomes, we will also carry out a sub￾group analysis to examine whether the impacts were different for boys and girls. Depending on the findings 
derived from baseline data, we can explore some interactions between school characteristics and the RL activity. 
For example, if we find that student characteristics and performance are different between rural and urban 
schools, we could study whether the effects of the activity are different accordingly to the school location. 
Limitations 
The evaluation faces several risks. Primary among them, is the fact that the design of the evaluation preceded 
the award of the Read Liberia contract, which potentially poses a threat to the evaluation in the event that the 
eventual implementation of the activity diverges from the assumptions on which the design is based. 
Other risks are common to most evaluations and include the possibility of contamination across treatment and 
control groups; the evaluation team will attempt to detect and account for such contaminations during endline 
data collection. 
A final risk that USAID’s Goal 1 team has raised concerns about test security related to the EGRA instrument 
being used for this evaluation, since it is the same assessment tools used in 2011 and 2015. However, there is no 
evidence of test leakage and, given time and budget constraints, NORC and USAID decided the creation and 
testing of a new tool was not warranted. 
ANNEX III: DATA COLLECTION INSTRUMENTS | 57
ANNEX III: DATA COLLECTION INSTRUMENTS 
Liberia pre-visit questionnaire 
1 School Code
2 County
3 District
4 Address (Village or Town?)
5 GPS Coordinates Lat: Long:
6 School Name
7 Is the school sometimes referred 
to locally by a different name than 
the one listed in the official 
registry? 
Yes: ____ No: _____
8 [If yes]: What is this name?
9 Interview Date (dd/mm/yyy) Day: ___ Month: _____ Year:_______
10 Principal Name First Name: ___________ Last Name: _______________
11 Principal Telephone
12 Principal’s email address 
13 Alternate Contact First Name: ___________ Last Name: _______________
14 Alternate Contact’s Title
15 Alternate Telephone
16 Alternate email address
17 Are Grade 2 students in a 
multigrade classroom?
Yes ______ No _____
18 Number of Classrooms for Grade 
2 students
Drop down with options for [1-10]
19 Classroom Size Grade 2 ______ students 
20 Number of Grade 2 Students Boys: ________ Girls: __________ TOTAL: ___________
21 Time school day starts Grade 2 ____ hour ____ minute AM/PM
22 Time school day ends Grade 2 ____ hour ____ minute AM/PM
23 Please indicate any dates this 
school will be closed due to 
holidays, teacher training, or other 
events in the next month. 
(dd/mm/yyyy) 
If more than 1 date, add field for each additional date the school 
will be closed. 
24 At the end of the 
questionnaire ask the 
principal to sign the consent 
form. [only for the 
enumerator]
Did the principal sign the consent?
Yes _____ No _____
ANNEX III: DATA COLLECTION INSTRUMENTS | 58
Student Participation – Consent Statement from Principal 
Written Consent 
Hello, my name is _________. I work with The Khana Group in Liberia.
 We are trying to understand how children learn to read and, in collaboration with 
the Ministry of Education, we are conducting the Early Grade Reading Assessment 
(EGRA) during May 2017 in 90 schools.
 Your school was selected randomly.
 We will randomly select 16 students in Grade 2 to assess their reading skills using a 
brief reading assessment. We will also ask these students a few questions about 
some of their normal school activities, reading practices at home, as well as home 
asset ownership. 
 Each interview should take around 40 minutes and the entire interview process for 
all 16 students will last around 4 hours. The interviews will take place during the 
school day.
 Selected students need only participate if they wish. 
 No names will not be mentioned anywhere in reports based on this exercise. 
 The assessments and responses will NOT affect the school, you, the teachers or 
students in any way. 
Do you provide consent for the Grade 2 students to participate in this assessment and 
survey?
 Yes
 No
Signature: ______________________________________
Print Name: _____________________________________
Date:___________________________________________
ANNEX III: DATA COLLECTION INSTRUMENTS | 59
Student Instrument 
General Instructions
It is important to establish a playful and relaxed rapport with the children 
to be assessed, via some simple initial conversation among topics of 
interest to the child. The child should perceive the following assessment as a game to be 
enjoyed rather than a severe situation. After you have finished, thank the child for his/her time 
and effort. 
Verbal Consent 
Read the text in the box clearly to the child: 
My name is _________. I work with the Ministry of Education in Liberia. 
 We are trying to understand how children learn to read. You were picked by chance, like in a 
raffle or lottery.
 We would like your help in this. But you do not have to take part if you do not want to.
 We are going to play a reading game. I am going to ask you to read letters, words and a short 
story out loud. 
 Using this stopwatch, I will see how long it takes you to read. 
 This is NOT a test and it will not affect your grade at school. 
 I will also ask you questions about your family, like what language your family uses at home 
and some of the things your family has. 
 I will NOT write down your name so no one will know these are your answers. 
 Once again, you do not have to participate if you do not wish to. Once we begin, if you would 
rather not answer a question, that’s all right. 
 Can we get started?
Check box if verbal consent is obtained: YES
(If verbal consent is not obtained, thank the child and move on to the next child, using this same form) 
○
ANNEX III: DATA COLLECTION INSTRUMENTS | 60
○
D. Unique School 
code :
J. Class section:
E. School shift : ○ 1 = Full day
○ 2 = Morning 
○ 3 = Afternoon
K. Student’s month and year of 
birth : 
Month : ______
Year : _________
Age: ___________
G. Teacher name 
(important!)
M. Student’s gender ○ 1 = girl 0 = boy
Task 1. Orientation to Print 
Show the child the paragraph segment on the last page of the student assessment (Section 6). 
Read the instructions in the gray boxes below, recording the child’s response before moving to the next 
instruction. 
I don’t want you to read this now. On this page, where would you begin to read? Show me 
with your finger.
[Child puts finger on the top row, left-most word] ○ Correct ○ Incorrect
○ No 
Response
Now show me in which direction you would read next.
[Child moves finger from left to right] ○ Correct ○ Incorrect
○ No 
Response
When you get to the end of the line, where would you read next? 
[Child moves finger to left-most word of second line] ○ Correct ○ Incorrect
○ No 
Response
ANNEX III: DATA COLLECTION INSTRUMENTS | 61
Section 2. Letter Name Knowledge 
Show the child the sheet of letters on the first page of the student assessment. Say, 
Here is a page full of letters of the alphabet. Please tell me the NAMES of as many letters as 
you can--not the SOUNDS of the letters, but the names. 
1. For example, the name of this letter [point to O] is “OH”. 
Now you try: tell me the name of this letter [point to V]: 
 [If correct:] Good, the name of this letter is “VEE.”
 [ If incorrect:] The name of this letter is “VEE.” 
2. Now try another one: tell me the name of this letter [point to L]: 
 [If correct:] Good, the name of this letter is “ELL.”
 [If incorrect:] The name of this letter is “ELL.” 
Do you understand what you are supposed to do? When I say “begin,” name the letters as 
best as you can. I will keep quiet and listen to you, unless you need help. Ready? Begin.
Set the timer on 1 minute. Start the timer when the child reads the first letter. Follow along 
with your pen and clearly mark any incorrect letters with a slash ( ). Count self-corrections as correct. Stay 
quiet, except when providing answers as follows: if the child hesitates for 3 seconds, provide the name of the 
letter, point to the next letter and say “Please go on.” Mark the letter you provide to the child as incorrect. 
WHEN THE TIMER REACHES 0, SAY, “stop.” Mark the final letter read with a bracket ( ). 
Early stop rule: If the child does not give a single correct response on the first line, say “Thank you!”, draw a line 
through the letters in the first row, discontinue this exercise, check the box at the bottom, and go on to the next exercise.
L i h R S y E O n T 10
i e T D A t a d e W 20
h O e m U r L G R U 30
g R B E i f m t s R 40
S T C N p A F c a E 50
ANNEX III: DATA COLLECTION INSTRUMENTS | 62
y s Q A M C O t n P 60
e A e s O F h u A T 70
R G H b S i g m i L 80
L i N O e o E r p X 90
N A c D d I O j e N 100
Time left on stopwatch if student completes in LESS than 60 seconds: __________ 
Exercise was discontinued as child had no correct answers in the first line. 
ANNEX III: DATA COLLECTION INSTRUMENTS | 63
Task 3. Phonemic Awareness 
This is NOT a timed exercise and THERE IS NO STUDENT SHEET. Read aloud each set of words once 
and have the student say which word begins with a different sound. Read these instructions to the child: 
This is listening exercise. I’m going to say THREE words. ONE of them begins with a different 
sound, and you tell me which word BEGINS WITH A DIFFERENT SOUND
1. For example: 
 “lost”, “map”, “like”. Which word begins with a different sound?
 [If correct:] Very good, “map” begins with a different sound. 
 [If incorrect:] “lost”, “map”, “like”. “map” begins with a different sound than 
“lost” and “like.” 
2. Now try another one: “train”, “trip”, “stop”. Which word begins with a different sound?
 [If correct:] Very good, “stop” begins with a different sound. 
 [If incorrect:] “train”, “trip”, “stop”. “stop” begins with a different sound than 
“train” and “trip.” 
Do you understand what you are supposed to do?
Pronounce each set of words once slowly (about 1 word per second). If the child does not respond after 3 
seconds mark it no response and move on. 
Early stop rule: If the child gets the first 5 sets of answers incorrect or no response, draw the line through 
each of the 5 first rows, discontinue this exercise, check the box at the bottom of this page and go on to the 
next exercise. 
Which word begins with a different sound? [repeat each set ONCE]
1 boy ball cat [cat] Correct Incorrect No Response
2 man can mad [can] Correct Incorrect No Response
3 pan late pin [late] Correct Incorrect No Response
4 back ten tin [back] Correct Incorrect No Response
5 fish fat cat [cat] Correct Incorrect No Response
ANNEX III: DATA COLLECTION INSTRUMENTS | 64
6 boat bit coat [coat] Correct Incorrect No Response
7 day bag dot [bag] Correct Incorrect No Response
8 can girl cold [girl] Correct Incorrect No Response
9 run race sand [sand] Correct Incorrect No Response
10 leg make lay [make] Correct Incorrect No Response
Exercise was discontinued as child had no correct answers in the first five sets of words.
ANNEX III: DATA COLLECTION INSTRUMENTS | 65
Task 4. Familiar Word Identification
Show the child the sheet of words on the second page of the student assessment. Say, 
Here are some words. I would like you to read me as many words as you can (do not spell the words, 
but read them). 
For example, this word is: “CAT”.
1. Now you try: [point to the word “mat” and say ]please read this word: 
 [If correct]: Good, this word is “mat.”
 [If incorrect]:This word is “mat.” 
2. Now try another one: [point to the word “top”] please read this word :
 [If correct]: Good, this word is “top.”
 [If incorrect]: This word is “top.” 
Do you understand what are you supposed to do? When I say “begin,” read the words as best as you 
can. I will keep quiet and listen to you, unless you need help. Ready? Begin.
Start the timer when the child reads the first word. Follow along with your pencil and clearly mark 
any incorrect words with a slash ( ). Count self-corrections as correct. Stay quiet, except when providing 
answers as follows: if the child hesitates for 3 seconds, read the word, point to the next word and say “Please 
go on.” Mark the word you read to the child as incorrect. 
WHEN THE TIMER REACHES 0, SAY, “stop.” Mark the final word read with a bracket ( ). 
Early stop rule: If the child gives no correct answers on the first line, say, “Thank you!”, discontinue this exercise, 
draw the line through the words in the first row, check the box at the bottom of the page, and go on to the next exercise. 
but time in the also 5
make no its said were 10
came very do after long 15
water as all for even 20
her was three been more 25
ANNEX III: DATA COLLECTION INSTRUMENTS | 66
that must can around it 30
another words back called work 35
could an him on see 40
than get not where what 45
you if their through when 50
Time on stopwatch if student completes in LESS than 60 seconds: __________ 
Exercise was discontinued as child had no correct answers in the first line.
ANNEX III: DATA COLLECTION INSTRUMENTS | 67
Task 5. Simple unfamiliar nonword decoding
Show the child the sheet of nonwords on the third page on the student form. Say, 
Here are some made-up words. I would like you to read me as many made-up words as you can (do 
not spell the words, but read them). 
For example, this made-up word is: “ut”.
1. Now you try: [point to the next word: “dif’ and say] please read this word
 [If correct]: “Very good: dif”
 [If incorrect]:This made-up word is “dif.”
2. Now try another one: [point to the next word: mab and say] please read this word. 
 [If correct]: “Very good: mab”
 [If incorrect]:This made-up word is “mab.”
Do you understand what you are supposed to do? When I say “begin,” read the words as best as you 
can. I will keep quiet and listen to you, unless you need help. Ready? Begin.
Start the timer when the child reads the first word. Follow along with your pencil and clearly mark 
any incorrect words with a slash ( ). Count self-corrections as correct. Stay quiet, except when providing 
answers as follows: if the child hesitates for 3 seconds, provide the word, point to the next word and say 
“Please go on.” Mark the word you provide to the child as incorrect. 
WHEN THE TIMER REACHES 0, SAY, “Stop.” Mark the final word read with a bracket ( ). 
Early stop rule: If the child gives no correct answers on the first line, say “Thank you!”, discontinue this exercise, draw 
the line through the words in the first row, check the box at the bottom of the page, and go on to the next exercise.
loz ep yat zam tob 5
zom ras mon jaf duz 10
tam af ked ig el 15
tig pek dop zac ik 20
uf ral ep bab vif 25
ANNEX III: DATA COLLECTION INSTRUMENTS | 68
lut sig zop zar jaf 30
ruz huf wab ak jep 35
wub dod ik vus nux 40
pek zel bef wab hiz 45
wof ib dek zek vok 50
Time left on stopwatch if student completes in LESS than 60 seconds: __________ 
Exercise was discontinued as child had no correct answers in the first line.
ANNEX III: DATA COLLECTION INSTRUMENTS | 69
Task 6. Passage reading and Comprehension. 
Show the child the story on the last page of the student form. Say, 
Here is a short story. I want you to read this aloud. When you finish, I will ask you some questions about what you have read.
Do you understand what are you supposed to do? When I say “begin,” read the story as best as you can. I will keep quiet and listen to you, 
unless you need help. Ready? Begin.
Set the timer on 1 minute. Start the timer when the child reads the first word. Follow along with your pencil and clearly mark any incorrect words with a slash ( ). Count 
self-corrections as correct. Stay quiet, except when providing answers as follows: if the child hesitates for 3 seconds, provide the word, point to the next word and say “Please go on.” 
Mark the word you provide to the child as incorrect. WHEN THE TIMER REACHES 0, SAY,, “stop.” Mark the final word read with a bracket ( ). If the child gets the entire first 
line incorrect, discontinue this exercise – both reading and comprehension questions -, check the box below and go on to the next exercise. 
STOP THE CHILD AT 0 SECONDS AND MARK WITH A BRACKET ( ).
Take the text away from the child after they read it. Read instructions to the child. Then read each question slowly and clearly. After you read each question, give the child at most 15 
seconds to answer each question. Mark the answers to the questions as correct or 
incorrect. 
Now I am going to ask you a few questions about the story you just read. 
Try to answer the questions as best as you can.
ANNEX III: DATA COLLECTION INSTRUMENTS | 70
Kemah lives near the big river. 6 Where does Kemah live? 
[near the big river] 
 Correct Incorrect No Response
There is a big tree by the river where Kemah lives. Kemah likes 
to sit in the tree.
24 Where does Kemah like to sit when she goes to the river? 
[in the tree, in the big tree near the river] 
 Correct Incorrect No Response
Every day after school, she stops by the tree and looks for a 
place to sit. She climbs the tree and sits on a branch.
49 What does Kemah do after she climbs the tree? 
[she sits on a branch, she finds a place to sit in the tree] 
 
 Correct Incorrect No Response
She looks at the fish in the river. 57 What does Kemah do when she sits in the tree?
[she looks at the fish in the river, looks at fish] 
 Correct Incorrect No Response
Kemah is happy. 60 Why is Kemah happy when she sits in the tree? 
[ she likes to look at fish in the river, she likes the tree] 
 Correct Incorrect No Response
 Time left on stopwatch: _______________________ Test Discontinued because child read NO words on the first line: 
ANNEX III: DATA COLLECTION INSTRUMENTS | 71
Task 7. Listening Comprehension 
This is NOT a timed exercise and THERE IS NO STUDENT SHEET. The administrator reads aloud the following 
passage ONLY ONE TIME, slowly (about 1 word per second). Say, 
I am going to read you a short story aloud ONCE and then ask you some questions. 
Please listen carefully and answer the questions as best as you can. 
Do you understand what are you supposed to do? 
Musu goes to the Bong Town School every day on a motor bike. One 
day, Musu could not get a motor bike to take her to school because it 
was raining and they were all busy. Musu did not want to get wet. Then, 
one old man said, “You can have my son’s raincoat.” Musu was happy. 
She did not have to be wet at school. 
1. How does Musu usually get to school? 
 [On a motorbike] Correct Incorrect No Response 
2. Why did the old man give Musu a coat? 
[Because it was raining] Correct Incorrect No response 
3. Why was Musu happy at school that day? 
[Because she did not have to be wet at school] Correct Incorrect No Response 
ANNEX III: DATA COLLECTION INSTRUMENTS | 72
Student Context Interview 
Say to the child: Thank you very much. Now, I am going to ask you some questions about your family and about reading. 
S1 What language/dialect does your family speak most 
often at home?
English = 1
Others = 2 [Specify main one]__________
Don’t know = 88
Refuse/No Answer= 99
S2 What language/dialect do your parents read or write 
in?
Cannot read and write = 0
English = 1
Others =2 [Specify main one]__________
Don’t know = 88
Refuse/No Answer = 99
S3 Do you have any reading books at home?
(If no, skip to S5.)
No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S4 If answer to S3 is Yes, in what language/dialects? English = 1
Other = 2 [Specify main one]________
S5 Does anyone read aloud to you at home?
(If No, skip to S7.)
No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S6 If answer to S5 is Yes, in what language/dialects do 
they read to you?
English = 1
Other = 2 [Specify main one]________
S7 Do you practice reading aloud to someone at home? 
(If No, skip to S9.)
No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S8 If answer to S7 is Yes, in what language(s) do you
read?
English = 1
Other = 2 [Specify main one]________
Don’t know = 88 Refuse/No answer = 99
S9 Have you ever repeated a grade? 
If yes, which ones? (CIRCLE the grades repeated.)
No = 0
Yes, Grade 1 = 1
Yes, Grade 2 = 2
Yes, Grade 3 = 3
Don’t know = 88
Refuse/No answer = 99
S10 Does your current teacher ever practice letter 
sounds with you? 
[Give student example of /k/ and /m/].
Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
S11 Does your teacher ever read aloud to you? Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
S12 Did you eat lunch at break time at school yesterday 
[or last school day]?
No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S13 Did you miss any school days last week? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
ANNEX III: DATA COLLECTION INSTRUMENTS | 73
S15 The last time you did not do well on a test or 
assignment in school, did your parent(s) find out? 
If yes, what did s/he do? (CIRCLE all that apply. Do 
not read the choices to the child.)
No/Never learned about it = 0
Learned but did nothing = 1
Helped/encouraged me to do better = 2
Punished me physically = 3
Criticized me verbally = 4
Discussed with the teacher = 5
Argued with the teacher = 6
Other (Don’t specify) = 7 
Don’t know = 88
Refuse/No answer = 99
S16 The last time you did well on a test or assignment in 
school, did your parent(s) find out?
(If No, skip to S18.)
No/Never learned about it = 0
Yes = 1
Don’t know = 88 
Refuse/No answer = 99
S17 If the answer to S16 is Yes, what did your parent(s) 
do?
Learned but did nothing = 0
Congratulated or encouraged me = 1
Other = 3__________________________
S18 Do you have a library at your school? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S19 Do you watch television at home? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S20 Do you listen to radio at home? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S21 Do you have electricity/current at home? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S22 Did you eat before coming to school today? No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S23 Do you have books at school that you can take home 
to read? 
No = 0 Yes = 1
Don’t know = 88 Refuse/No answer = 99
S24 Does your teacher make you practice silent reading 
in class?
Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
S25 Does your teacher make you practice reading out 
loud in class?
Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
S26 Does your teacher assign reading for you to do at 
home?
Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
S27 Does your teacher ever make you re-tell a story 
during class?
Never = 0
Often = 1
Always = 2
Refuse/No answer = 99
Thank the student by shaking his/her hand! 
ANNEX III: DATA COLLECTION INSTRUMENTS | 74
THE EGRA ASSESSOR CHEAT SHEET 
SUBTASK 1 – LETTER NAME 
KNOWLEDGE 
Remain quiet, except if the child hesitates for 3
seconds. Then point to the next letter and say, 
“Please go on.” Mark the skipped letter as 
incorrect. 
Stimuli REQUIRED - TIMED 
 Auto-stop = No correct responses in the first 
line (10 letters)
 Skipped Letter – is marked incorrect. If the 
child self-corrects, mark as correct
 Skipped Row – If the child skips an entire row, 
the entire row is marked incorrect
SUBTASK 2 – PHONEMIC AWARENESS
 Read the prompt and then pronounce the 
words, but do not overemphasize the 
beginning sound of each word
 Pronounce each word only once
 Remain quiet, except if the child hesitates for 
3 seconds. Then, mark the item as no 
response
NO Stimuli - NOT Timed
Auto-stop = No correct responses in the first 5 
items
SUBTASK 3 – FAMILIAR WORD 
IDENTIFICATION
Remain quiet, except if the child hesitates for 3
seconds. Then point to the next word and say, 
“Please go on.” Mark the skipped word as 
incorrect. 
Stimuli REQUIRED - TIMED 
 Auto-stop = No correct responses in the first 
line (5 words)
 Skipped Word – is marked incorrect. If the 
child self-corrects, mark as correct 
 Skipped Row – If the child skips an entire 
row, the entire row is marked incorrect 
SUBTASK 4 – SIMPLE UNFAMILIAR 
NONWORD DECODING
Remain quiet, except if the child hesitates for 3
seconds. Then point to the next word and say, 
“Please go on.” Mark the skipped word as 
incorrect. 
Stimuli REQUIRED - TIMED 
 Auto-stop = No correct responses in the first 
line (5 words)
 Skipped Word – is marked incorrect. If the 
child self-corrects, mark as correct 
 Skipped Row – If the child skips an entire 
row, the entire row is marked incorrect 
SUBTASK 5A - ORAL READING 
FLUENCY
 Remain quiet, except if the child hesitates for 
3 seconds. Then point to the next letter and 
say, “Please go on.” Mark the skipped 
word as incorrect 
 After the child has finished, REMOVE the 
passage from in front of the child 
Stimuli REQUIRED - TIMED 
 Auto-stop = No correct responses in the first 
6 items
 Skipped Word – is marked incorrect. If the 
child self-corrects, mark as correct 
 Skipped Row – If the child skips an entire row, 
the entire row is marked incorrect
ANNEX III: DATA COLLECTION INSTRUMENTS | 75
SUBTASK 5B – ORAL READING NO Stimuli - NOT timed
COMPREHENSION Answers – In comprehension questions potential 
 Mark the child’s response and continue to answers are found in parenthesis below each 
the next question question. Be careful not to read these answers to 
 Give the child 15 seconds to answer the the pupil. If the pupil responds in a local dialect 
question. Mark the child’s response and or language, ask the pupil one time if they can 
move on answer in English
 Mark as correct only responses that are
same or similar in meaning to the answers 
provided 
SUBTASK 6 - LISTENING NO Stimuli - NOT timed
COMPREHENSION Answers – In comprehension questions potential 
 Read the entire passage out loud to the child answers are found in parenthesis below each 
ONE TIME ONLY. Read slowly (about 1 question. Be careful not to read these answers to 
word per second) with expression. the pupil. If the pupil responds in a local dialect 
 Ask all of the questions. Do not allow the or language, ask the pupil one time if they can 
child to look at the passage or the questions answer in English
 If the child hesitates for 15 seconds, mark as 
incorrect and move on to the next question 
Administration Procedure Reminders
 Examples – Always read instructions and conduct examples before starting 
 Timed tasks – 1) start the timer when learner begins to read the first letter or word, 2) When 
timer ends, mark the final letter or word with ]
 Pronunciation – Dialect. EGRA is a reading test, not a pronunciation test. If children in a particular 
area pronounce a word in their regular speech differently than everywhere, they should not be 
marked incorrect. This is not a reading error 
 Testing Materials – Hold the tablet out of view of the child. The pupil stimuli is blank side up 
except when the child is reading it. Have materials easily accessible from the start 
 Pupil interview – It should be administered verbally. Do NOT read aloud any responses unless 
protocol prompts to do so when text is in ALL CAPS
 Vertical Reading – If the child begins to read vertically, immediately stop the child from reading 
and point with your finger to show the child that they should read from right to left. Restart the 
exercise.
ANNEX III: DATA COLLECTION INSTRUMENTS | 76
What is Tangerine? 
Tangerine is an Android-based app designed to collect data electronically for a 
variety of types of surveys. It was designed specifically to conduct the Early Grade 
Reading Assessment. These training materials have been adapted from the 
Tangerine users’ guide developed by RTI as well as the Galaxy Tab 4 user manual. 
Rules of Tablet Use 
Each enumerator has received a tablet after signing a tablet user agreement. Here are the rules outlined 
in the signed agreement: 
 Use materials and equipment correctly and only as directed. In particular, this includes use of a 
tablet and all accessories (stylus, charger and cord). 
 Sign the tablet “out” and “in” each day.
 Keep the device in a safe place at all times. The tablet is the sole property of NORC and 
should not be removed from the training venue.
 Maintain the tablet and accessories in good working condition and keep them in a safe place at 
all times. 
 Take care not to drop the device. 
 Save and upload data as directed. 
Each enumerator takes responsibility for any damages to, or loss of, the tablet during the time that the 
enumerator uses the tablet. 
Tablet Basics 
Here are basic instructions about how to operate the Samsung Galaxy Tab 4. 
Turning your device on: Press and hold the Power/Lock Key. 
Turning your device off: 
1. Press and lock the power/lock key. 
2. On the Device options menu, touch Power Off  OK.
ANNEX III: DATA COLLECTION INSTRUMENTS | 77
Securing Your Device 
By default, the device locks automatically when the screen times out or you can lock it manually. Unlock 
the tablet by swiping your finger across the screen. 
Unlocking the Device 
Press the Power/Lock Key. The Lock screen displays. Swipe across the screen as 
shown. 
Battery 
The Battery icon in the Status Bar shows the battery power level. When battery 
power is 15% or less, your device prompts you to charge the battery. If you 
continue to operate the device without charging, the device powers off. 
Battery full Battery charging Battery low 
Opening Tangerine 
To open the Tangerine application, find the Tangerine icon on the tablet screen and 
touch it. 
Logging into Tangerine on tablets 
When Tangerine opens, the login screen appears. Notice there are two tabs in the login screen: Login 
and Sign Up. Login is for users who have already registered a user account on the tablet on the device 
and on the most up-to-date version of Tangerine that is currently installed. 
How to Sign Up: New users should tap the ‘Sign Up’ tab to establish an account on the tablet. It is
best to create user names and passwords that are simple to remember. Keep a record of your username 
ANNEX III: DATA COLLECTION INSTRUMENTS | 78
and password in a safe place. It is best to choose a username that can be associated with an individual. 
Typically, it will require enumerators’ usernames to consist of the first letter of their first name, followed 
by their surname. Therefore, user John Smith’s username would be jsmith and his password would be
something simple and easy for him to recall. 
Login: Users who have already signed up on the tablet (and on the version of the app that is installed) 
can enter their credentials to login. If a user has lost or forgotten the username or password, they can 
simply create a new login with the ‘Sign Up’ tab. All data from all users on each tablet is uploaded – there 
is no data lost when a user needs to create a new account. 
Managing Instruments on Tablets 
After you have logged on to Tangerine, you will see all of your active assessments from your online 
Wizard group listed on the home screen. The Tangerine app home screen may look like this: 
Under “Assessments”, the enumerator is able to see the different assessments that can be run. Clicking 
the “Run” button will start the assessment. The “Results” button will bring the enumerator to a 
screen that shows the data collected thus far, and allows the option for interrupted assessments or surveys 
to be resumed. 
Entering data on tablets 
Launch your assessment. Before entering data, you must open the assessment that you wish to
conduct. 
Click the run icon to begin collecting data on any given instrument. 
Types of input. Input in Tangerine will either be done by tapping response items or entering text. Once 
you have mastered the types of input per each item in your instrument, each subtest should be simple to
administer. Below are descriptions of the different types of input. 
1. Buttons: Most input is done through buttons. To use a button, simply touch the button in the
center and release with a short tap or a firm, but short, press. Tangerine buttons are used to: 
ANNEX III: DATA COLLECTION INSTRUMENTS | 79
 Advance from screen to screen with the “Next” button 
 Assign a student ID 
 Start or stop the timer 
 Indicate incorrect items in an assessment grid 
2. List selection. The school name field is an example of an
automated list selection input type. Enumerators must select 
items from drop-down menus. 
3. Automatically generated fields. Certain field values are 
automatically filled in. These are the date and time fields, which 
are generated according to the date and time that the device is
set to when the assessment begins; and the unique student ID
field, which is filled in when the enumerator presses the 
“Generate” button. 
Timed subtests. On a timed grid page, you will not see the grid items 
until you press “Start.” When you press “Start,” the timer will start 
counting down, and you will see the items appear in the grid (see 
screenshot shown at right). Once you have started you should not stop 
and restart except for exceptional circumstances. If a restart is needed, 
there is a button at the bottom of the page for that purpose. 
Auto-correction: By default, all grid items are “correct.” Mark items “incorrect” by tapping on that 
item, causing a line to be struck through that grid item. If a student first answers incorrectly, but then 
corrects himself or herself, mark the answer as correct. You can re￾mark an incorrect answer as correct by tapping the button a second 
time. 
Marking an entire line as incorrect: Use the asterisk button on
the right-hand side of the grid to mark a whole row incorrect. Re￾tap the line wrong icon to correct the line all at once. 
Ending a subtest: There are three ways to end a timed subtest:
1. Autostop: The subtest may end automatically if a student 
answers a pre-determined number of questions at the 
beginning of the test incorrectly. If this occurs, the screen 
will flash red, and a message will appear that says 
“Autostop activated. Discontinue test.” The next step for 
the enumerator is to press “Next” to move to the next 
exercise. 
2. The timer runs out: If the timer runs out while the child 
ANNEX III: DATA COLLECTION INSTRUMENTS | 80
is still reading, the screen will flash red, and a message to “select the last item read” will 
appear (see below). The enumerator should ask the child to stop reading, then touch the 
last item attempted by the child when the timer went off. If it is necessary to mark the last 
item read as incorrect, use the “mode” feature (see below). Otherwise, the next step for 
the assessor is to press the “next” button. 
3. The enumerator clicks the stop button: If a child 
completes all of the grid items before the allotted time 
runs out, the assessor will stop the timer by using the 
“stop” button. This will automatically place the ‘Last
Item Read’ bracket around the final grid item, as the 
timer should only be stopped when the child has 
attempted the final grid item. If need be, the ‘Last Item 
Read’ bracket could be moved away from the final grid 
item. 
Input mode: The “Input mode” feature (above, at right) exists 
because buttons behave differently depending on what “mode” the 
application is in. When the timer has started, the buttons are in
the “Mark” mode that expects a click to mean “mark this item 
incorrect.” After the timer is stopped, the mode automatically 
changes to “Last attempted” and the application expects that the 
next click will mean “mark this as the last item read” (above). In
some cases, the last item read is both the last item AND an incorrect item. Therefore, the enumerator 
must toggle between modes in order to mark the button accordingly. 
Completing an assessment. Upon reaching the end of the assessment 
(after the last subtest or interview question), you will see a confirmation 
page that the test has been completed, and an overview of results for each 
subtest (select the details button on “Subtests completed”). This is for the 
enumerator‘s benefit as a way to verify that all of the subtests were 
completed. 
Enumerators should press ‘Save Result’ to finalize their entry. This button 
will then become “Perform another assessment.” You can either press this 
button to begin assessing another child on the same instrument, or you can 
use the Tangerine icon button in the upper-left hand corner to get back to
the list of available tests and start a new assessment. 
To check the number of assessments collected, return to the list of
assessments. Click on the data/results icon located to the right of any 
assessment and you will see a list of assessments conducted by that assessor 
for that test. Click on each assessment to see details. 
ANNEX III: DATA COLLECTION INSTRUMENTS | 81
Key Reminders 
 At the end of each test, make sure to click “Save Results” on the Results screen. 
 To stay in the Tangerine application, NEVER click on any of the buttons on the tablet, including the 
Home button, back key, and recent key on the bottom, because doing so brings the user to the 
home screen of the tablet and aborts ongoing processes in Tangerine. 
 While conducting an assessment, YOU CANNOT navigate pages using the Back button at the 
bottom of the tablet.
Troubleshooting 
Question: What should I do if a subtest doesn’t respond when I’m in the middle of conducting an 
assessment or survey?
Answer: It is possible to resume the assessment. To do so, follow these steps: 
1. Touch the Tangerine icon in the upper left hand corner of the tablet. You will return to the 
assessments page. 
2. Press the data/results icon located to the right of the assessment you were conducting. 
3. Find the assessment you were conducting. Press “details” and then “resume.” You will then go 
to the subtest at which you left off. 
Question: What happens if the steps above do not work because the Tangerine application is frozen? 
Answer: Exit the application by touching the home key at the bottom of the tablet. Go to Settings > 
Apps, then swipe right-to-left to get to a listing of Apps under “All.” Select Tangerine and then 
“Force Stop.”
Question: What should I do if I accidentally stop a timed subtest? 
Answer: If the child is still answering questions when the timer is stopped, tell the student to stop 
and to resume the test from the beginning when you say to start. Also, tell the student that he or she 
has to resume the test because of a mistake you made. 
Question: What if I mark an answer incorrectly, and the time runs out before I can correct my 
mistake? 
Answer: It is possible to correct a mistake even if time has run out. After the timer stops and you 
mark the last letter attempted, switch input mode at the bottom of the grid from “last attempted” to 
“mark,” and then correct your mistake. 
ANNEX III: DATA COLLECTION INSTRUMENTS | 82
TEACHER Instrument
Verbal Consent 
My name is _________. I work with the Ministry of Education in Liberia. 
 We are trying to understand how children learn to read. Your school was selected
through the process of statistical sampling. We would like your help in this. But you
do not have to take part if you do not want to.
 Your name will not be mentioned anywhere in reports based on this survey. Results
of this survey will be presented in an aggregated format so that your specific
responses will not be identifiable The information acquired through this instrument
will be shared with the Ministry of Education in the hopes of identifying areas where
additional support may be needed.
 Your responses will NOT affect you in any way. They will not have any impact on
your employment or your pay.
 If you agree, I would ask you some questions regarding your normal activities at
school, including your interactions with school staff, Ministry office staff, students,
and parents.
 Then, with your assistance, I would randomly select 16 students in Grade 2 to assess
their reading. I would also ask these students about some of their normal school
activities, school assets, language use, and reading practices at home, as well as home
asset ownership. Selected students need only participate if they wish. I will spend
about 40 minutes interviewing each child. My interview with you will take 15 to 20
minutes.
 Are you willing to participate? Once again, you do not have to participate if you do
not wish to. Once we begin, if you would rather not answer a question, that’s all
right.
Can we get started?
Check box if verbal consent is obtained : YES NO
(If verbal consent is not obtained, thank the teacher and select the next one)
ANNEX III: DATA COLLECTION INSTRUMENTS | 83
Section 1. Teacher Interview 
SC1 School Name
SC2 School Code
T1 Interviewer Name
T2 Interviewer Code
T3 Starting Time of Interview ____:____AM
T4 Ending Time of Interview ____:____PM
T5 Interview Date (dd/mm/yyyy) Day:_____Month:______Year:______
T6 Interview Status Refused =1
Partially Completed = 2
Complete = 3
T7 [Is the teacher male or female?] Male=0, Female=1 
T8 Teacher name (make sure the same name is used in the questionnaires 
of the children who are taught by this teacher, so the questionnaires 
can be linked)
T9 What type of teaching certificate do you 
have? 
C Certificate = 0
B Certificate = 1
AA Certificate = 2
Other =3
Don’t know = 88
Refuse/No Answer = 99
T10 What is your highest level of education? Elementary=0
Junior High School =1
Senior High School =2
Associate = 3
Certificate C = 4
Certificate B = 5
Certificate AA = 6
Bachelors Degree = 7
Masters degree or other = 8
Other (Do not specify) = 9
Don't Know = 88
Refuse/No Answer = 99
T11 How many years of teaching experience do you have (Enter number)
T12 Have you attended any in-service training 
or professional development sessions 
such as workshops in the last year?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T13 Have you ever received training on how 
to teach reading?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T14 If the answer to T13 is Yes, did you 
receive training for this school year?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
ANNEX III: DATA COLLECTION INSTRUMENTS | 84
T15 If the answer to T14 is Yes, about how many hours? (Enter the hours)
T19 Have you received support visits at 
school this past year on how to teach 
reading?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T20 What grade or grades do you teach in 
this school year? [CIRCLE all that apply]
 Grade 1
 Grade 2
 Grade 3
 Grade 4
 Grade 5
 Grade 6
T21 Have you been teaching the same class 
since the beginning of the school year?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T22 Do you keep an attendance record of 
students?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T23 How often do you develop lesson plans? Daily = 1
Weekly = 2
Bi-weekly = 3
Monthly = 4
Don't Know = 88
Refuse/No Answer = 99
T24 Where do you develop lesson plans? Home = 0
School = 1
Don't Know = 88
Refuse/No Answer = 99
T25 Do you have a written lesson plan for 
today?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T26 Could I please see it? Was not able or willing to show it = 0
Could show it but it did not look very good = 1
Was able to show it and it looked pretty good = 3
T27 Do you have any scheduled time during 
the school day for lesson planning? 
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
I'm going to ask you about different activities you might do with your students. Think about the last 5 school 
days and tell me how frequently the following activities took place
T28.1 The whole class repeated letters or 
words that you said first, when teaching 
reading skills
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T28.2 Students sounded unfamiliar words they 
are learning
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
ANNEX III: DATA COLLECTION INSTRUMENTS | 85
T28.3 Students read aloud to teacher or 
another student
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T28.4 Students learned the meaning of new 
words
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T28.5 Students had to retell a story that they 
read during the week
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T28.6 Students were assigned reading to do on 
their own in school time
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T28.7 Students were assigned reading to do at 
home
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
Now we are going back to some questions about your overall work.
T29 Do you use the official reading 
curriculum in your classroom lessons? 
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
T30 Do you have teacher guides? No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T31 If answer to T30 is Yes, “How useful do 
you find them?”
Not very useful = 0
Moderately useful = 1
Very Useful = 2
Don't Know = 88
Refuse/No Answer = 99
T32 How frequently does your principal 
observe (your) classes when they were in 
session? 
Never = 0
Once a year = 1
Once every 2-3 months = 2
Once every month = 3
Once every two weeks = 4
Once every week = 5
Daily = 6
Don’t Know = 88
Refuse/No Answer = 99
T33 Within the last year, did you receive an 
external inspection or support visit from 
Education Officers? 
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
ANNEX III: DATA COLLECTION INSTRUMENTS | 86
T34 How do you measure your students' 
progress? [Do not read options and 
CIRCLE those mentioned.]
Written tests = 1
Oral evaluations = 2
Their portfolios and other projects = 3
Their homework = 4
End of term evaluations = 5
Other (Do not specify) = 6
T35 What reading skills should your children 
have at the end of school year? 
[Do not read options. Just CIRCLE the 
number for those mentioned.]
Read grade level stories = 1
Sound out words they don't know = 2
Understand stories that they read = 3
Know letter names = 4
Other (Do not specify) = 5
T41 Do students take books from school to 
read at home?
Never = 0
Almost never = 1
Sometimes = 2
Almost always = 3
Always = 5
Refuse/No Answer = 99
T42 Do the teachers at this school work 
together as teams to solve problems 
related to teaching?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T43 Have you made any special efforts to 
improve reading at your school this past 
year?
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T44 If the answer to T43 is Yes, ask “Why did 
you do that?” Do not read the answers, 
just CIRCLE the ones that apply.
We saw other schools doing it = 1
We thought it might be important because the children 
were tested = 2
The Ministry told us to do it = 3
An NGO told us to do it = 4
Teachers got teacher training that showed them how 
to teach reading better = 5
Other = 6 (Do not specify)
T45 If the answer to T43 is Yes, ask “How 
much more effort did you make that in a 
normal year?”
A little = 1
Some = 2
A good bit = 3
A lot = 4
Refuse / No Answer = 99
T48 Are you a volunteer teacher? No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
T49 If Yes to T48, ask: ‘Will you continue 
teaching in September 2011? 
No = 0 Yes = 1
Don't Know = 88 Refuse/No Answer = 99
THANK YOU VERY MUCH
General Impression _________________________________________________________________
ANNEX III: DATA COLLECTION INSTRUMENTS | 87
Principal Instrument 
Verbal Consent 
My name is _________. I work with the Ministry of Education in Liberia. 
 We are trying to understand how children learn to read. Your school was selected
through the process of statistical sampling. We would like your help in this. But you
do not have to take part if you do not want to.
 Your name will not be mentioned anywhere in reports based on this survey. Results
of this survey will be presented in an aggregated format so that your specific
responses will not be identifiable The information acquired through this instrument
will be shared with the Ministry of Education in the hopes of identifying areas where
additional support may be needed.
 Your responses will NOT affect you in any way. They will not have any impact on
your employment or your pay.
 If you agree, I would ask you some questions regarding your normal activities at
school, including your interactions with your staff, Ministry office staff, students, and
parents.
 Then, I would randomly select 16 students in Grade 2 to assess their reading skills. I
would also ask these students about some of their normal school activities, school
assets, language use, and reading practices at home, as well as home asset ownership.
Selected students need only participate if they wish. The entire interview process for
all 16 students will last around 4 hours (40 minutes per each child). My interview with
you will take 15-20 minutes. Finally, I would spend about 15-20 minutes interviewing
the Grade 2 teacher.
 Are you willing to participate? Once again, you do not have to participate if you do
not wish to. Once we begin, if you would rather not answer a question, that’s all
right.
 Can we get started?
Check box if verbal consent is obtained: YES NO 
(If verbal consent is not obtained, thank the principal and terminate the exercise in this school) 
ANNEX III: DATA COLLECTION INSTRUMENTS | 88
Section 1. Principal Interview
School Name
School Code
Principal Name
Interviewer Name
Starting Time of Interview ____:____AM/PM
Ending Time of Interview ____:____PM/PM
Interview Date (dd/mm/yyyy) Day:___Month:___Year:____
Interview Status Refused = 1
Partially Completed = 2
Complete = 3
What is your position at this school? 
(circle all that apply)
Principal = 1
Vice Principal = 2
Teacher = 3
[Is the principal male or female?] Male = 0 Female=1
How many years have you been a 
principal? Enter years:
What is your highest level of education? Elementary=0
Junior High School =1
Senior High School =2
Associate = 3
Certificate C = 4
Certificate B = 5
Certificate AA = 6
Bachelors Degree = 7
Masters degree or other = 8
Other (Do not specify) = 9
Don't Know = 88
Refuse/No Answer = 99
What grades are taught at this school this 
year? (Enter the first grade taught and 
the last grade taught.)
First grade taught
Last grade taught
Have you received special training or taken 
courses that prepared you to teach reading or 
support teachers to teach reading ?
Yes 
No
Don’t know
Refuse/No Answer
At what grade do you expect all of your 
students to read fluently? (Enter the 
grade:)
Think about first and second grade teachers, 
does anyone review those teachers’ lessons 
plans? 
Yes 
No skip to question
Don’t know  skip to question 
Refuse/No Answer skip to question
ANNEX III: DATA COLLECTION INSTRUMENTS | 89
Who is the main person that reviews the 
teachers' lesson plans? 
Principal = 1 
Vice Principal = 2
Other
Don’t know (DK)
Refuse/No Answer
If other person, specify _______________________________
How often are these plans reviewed? Once per year 
Once every term 
Once every month 
Every week or more often 
Don't Know (DK)
Refuse/No Answer
In your school, does anyone observe Grade 1 
and Grade 2 teachers’ lessons in their 
classrooms? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Who mostly observes those teachers’ practices 
in the classrooms?
Principal = 1 
Vice Principal = 2
Other
Don’t know
Refuse/No Answer
If other person, specify ____________________________
If yes, ask how often: Once a year = 1
Once every 2-3 months = 2
Once every month = 3
Once every two weeks = 4
Once every week = 5
Daily = 6
Don’t Know (DK) = 88
Refuse/No Answer = 99
Within the last year, did you receive an 
external inspection or support visit from 
Education Officers? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Do you have a feeding program at 
school? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Does your school observe student dress 
code? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Do you have sufficient resource 
materials/ textbooks for Grade 1? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
ANNEX III: DATA COLLECTION INSTRUMENTS | 90
Do you have sufficient resource 
materials/ textbooks for Grade 2?
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Do you have a library or reading room? No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Do you hold regular PTA meetings? No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
If answer to HT20 was Yes, how many 
parents come to the PTA meetings
Few = 1
Some = 2
Most = 3
All = 4
In one year, how often do you send 
teachers for in-service teacher training, 
on average? 
Less than 5 = 1
5 to 10 times = 2
More than 10 times = 3
Do you have lockable book storage at 
your school?
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
Do you keep records of teachers’ 
attendance?
No = 0 Yes = 1
DK= 88 Refuse/No Answer = 99
If answer to HT24 is Yes, say “Could I 
see the attendance records please?”
Was not able or willing to show = 1
Was able to show but badly kept (out of 
date, incomplete) = 2
Was able to show and in good shape = 3
Have you made any special efforts to 
improve reading at your school this past 
year?
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
If the answer to HT31 is Yes, ask, “Why 
did you do that?” Do not read the 
answers, just CIRCLE the ones that 
apply.
We saw other schools doing it = 1
We thought it might be important 
because the children were tested = 2
The Ministry told us to do it = 3
An NGO told us to do it = 4
Teachers got teacher training that 
showed them how to teach reading 
better = 5
Other = 6 (Do not specify)
Does your school implement the 
Accelerated Learning Program (ALP) or 
Non Formal Education Program (NFE)
No = 0 Yes = 1
DK= 88 Refuse/No Answer = 99
Are you currently teaching in this school? No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
If answer to HT35 is Yes, “Ask: ‘What 
grades are you teaching? [CIRCLE all that 
apply]
 Grade 1
 Grade 2
 Grade 3
 Grade 4
 Grade 5
ANNEX III: DATA COLLECTION INSTRUMENTS | 91
 Grade 6
Are any of your teachers (other than 
you) also principals or vice-principals? 
No = 0 Yes = 1
DK = 88 Refuse/No Answer = 99
How many Grade 1 teachers does your 
school have? 
[write the number]
How many Grade 1teachers are currently 
volunteer teachers? 
[write the number]
How many Grade 2 teachers does your 
school have? 
[write the number]
How many Grade 2 teachers are 
currently volunteer teachers? 
[write the number]
 
THANK YOU VERY MUCH 
General Impression _________________________________________________________________
 
 
 ANNEX III: DATA COLLECTION INSTRUMENTS | 92
APRIL 2017
Baseline Enumerator Training Manual
USAID/Liberia Impact Evaluation for the 
Read Liberia Activity
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 93
TABLE OF CONTENTS 
1. Training Program .................................................................................................................... 97
2. Background and Objectives.................................................................................................. 101
2.1 Who We Are ..................................................................................................................................................... 101
2.2 Our Client ........................................................................................................................................................... 102
2.3 How Do We All Work Together? ............................................................................................................... 102
2.4 The Intervention Under Study ....................................................................................................................... 102
2.5 Our Survey.......................................................................................................................................................... 102
2.6 Baseline Data Collection ................................................................................................................................. 103
3. Methodology .......................................................................................................................... 103
3.1 Where Will Our Survey Be Conducted? .................................................................................................... 103
3.2 Sample Size.......................................................................................................................................................... 103
3.3 How Will Our Questionnaires Be Structured? ......................................................................................... 104
4. General Overview ................................................................................................................. 105
4.1 Team Structure .................................................................................................................................................. 105
4.2 Roles and Responsibilities ............................................................................................................................... 105
4.2.1 Enumerator – 3 per team ......................................................................................................................................105
4.2.2 Supervisors/Field SUPERVISOR (FS) – 1 per team ..........................................................................................107
4.3 Quality Control ................................................................................................................................................. 108
5. Basic Interviewing Techniques............................................................................................. 110
5.1 Gaining Cooperation ........................................................................................................................................ 110
5.1.1 Gaining cooperation techniques and Frequently Asked Questions ...............................................................111
5.1.2 FAQ ...............................................................................................................................................................................113
5.2 Conducting the Interview................................................................................................................................ 115
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 94
5.2.1 Informed Consent .....................................................................................................................................................115
5.2.2 Interviewing techniques ...........................................................................................................................................118
5.2.3 Bias ...............................................................................................................................................................................120
5.2.4 Probing .........................................................................................................................................................................121
5.2.5 Asking Questions .......................................................................................................................................................123
5.2.6 Recording Responses on Tablets ...........................................................................................................................125
5.2.7 Checking for Internal Consistency ........................................................................................................................126
6. Field Procedures.................................................................................................................... 128
6.1 Preparation before Going to the Field ......................................................................................................... 128
6.2 Whom to Interview and Contacting Strategy ............................................................................................ 128
6.2.1 Contacting Strategy ..................................................................................................................................................129
6.3 Tablets .................................................................................................................................................................. 130
6.3.1 Rules of Tablet Use ..................................................................................................................................................130
6.3.2 Tablet settings............................................................................................................................................................131
6.4 Data Security ...................................................................................................................................................... 132
6.4.1 Confidentiality ............................................................................................................................................................132
6.4.2 Tablet Safety ..............................................................................................................................................................132
6.5 Validation (Direct Obseravtions and Spot-Checks).................................................................................. 132
7. Our Questionnaires .............................................................................................................. 133
7.1 Language .............................................................................................................................................................. 133
7.2 Concepts and Definitions ................................................................................................................................ 133
7.3 Questionnaire Overview ................................................................................................................................. 133
7.3.1 Questionnaire Number Codes ...............................................................................................................................134
7.3.2 Look out for these common mistakes .................................................................................................................134
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 95
FOREWORD 
On behalf of the United States Agency for International Development (USAID), NORC is serving as an 
independent evaluator for the Read Liberia! project administered in the counties of Lofa, Bong, Grand 
Bassa, Nimba, Margibi, and Montserrado. Read Liberia! is a 5-year intervention aimed at improving the 
reading skills of students in grades one and two through: (1) increased government commitment to, and 
support of, evidence-based reading instruction (Result 1); (2) improved early grade reading classroom 
instruction (Result 2); (3) improved service delivery systems in early grade reading (Result 3); and (4) 
increased parent, community and private support for early grade reading. 
NORC is serving as the independent evaluator of this program, and will conduct the impact evaluation of 
the Read Liberia! activity. 
To evaluate the impact of Read Liberia! we are conducting a randomized controlled trial, where schools 
are assigned randomly to treatment and control groups. The details of the program intervention are still 
being developed, but the general research questions of interest are: 
 What is the impact of early grade reading (EGR) activities on reading performance and retention 
rates? 
 What is the total impact that Read Liberia! activities have on reading performance and retention 
rates? 
The May 2017 data collection will support the evaluation of these questions and will establish baseline 
data on early grade reading skills of students. This data collection will also collect background information 
on the students, teachers, and principals in order to develop a more complete picture of the schools 
sampled at baseline, before any intervention activities occur. 
This document has been prepared to serve as a guide to interviewers and supervisors, who will help to 
interview teachers, principals and students for evaluation purposes. 
NOTE ON CONFIDENTIALITY 
We wish at the beginning of this manual to emphasize the confidential nature of your task as an 
Enumerator. The success of the survey depends on the willingness and cooperation of the respondents. 
This in turn depends in large part on whether the respondents believe that the information they give will 
be kept confidential. 
You must know the provisions made by NORC, STS and TKG to keep information confidential. In this 
manual, the subject of confidentiality will be raised several times. Study these passages carefully and resolve 
from the outset that you will in no way breach the confidentiality provisions. 
Pledge of Confidentiality 
Every person involved in data collection, data entry and data analysis is required by NORC to sign a pledge 
of confidentiality. As an enumerator, supervisor, or quality control officer, you must therefore sign this 
pledge as well. During training, we will read this document together and you will be asked to sign it. If you 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 96
decide not to sign the pledge of confidentiality, you will be unable to participate in data collection activities 
for this project. 
As an enumerator, you are TKG’s representative in the field. You are required to handle all documents 
carefully, and conduct your work in a manner that will ensure confidentiality. Remember you have taken 
an oath of security. Legal action could be taken against you if you release or disclose any information on 
the survey. 
a. Handling of Electronic Data during Interview: 
You must follow all safety procedures related to using the tablet outlined during this training. Do not 
share passwords with anyone, or send data to another party. 
b. Reference to Other Respondents: 
Do not refer to names and answers obtained from respondents while interviewing another respondent. 
Do not let a respondent see a completed questionnaire of another respondent. 
c. Keeping of Survey Materials at Home: 
Keep your tablet, along with field control sheets, in safe and secure places. Preferably, keep them in boxes 
or cup-boards with locks when not in use. 
d. Keeping Survey Data from Members of Family and Friends 
Do not discuss survey data with members of family nor friends. You must not permit any unauthorized 
person, not even a member of your family, to see a completed questionnaire. You must not solicit for, or 
accept assistance from non-survey personnel in your enumeration duties. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 97
1. Training Program 
Welcome to the Read Liberia! IE Baseline Survey training! This manual outlines the information needed 
by Enumerators, Supervisors, and Quality Control Officers for this project. It is to be used as a training 
guide as well as a reference manual for you to review your duties and to enhance efficiency and effective 
data collection. 
As a field staff member of the Read Liberia! IE Baseline Survey, you will be involved in a data collection 
activity to support the development of a detailed impact evaluation. By conducting this survey completely 
and with integrity, you will play a pivotal role in this program and, ultimately, in decisions that are made 
regarding the future of the program. 
Training Program
Training comprises both classroom and field practice. You should study the training program and take 
note of the topics covered each day. Before each training session, study the manual and write down any 
questions you may have. Ask the questions during the training sessions and make sure you understand all 
terms used and the procedures to be followed. 
High Standards Expected of You 
When you are carrying out the enumeration, you will deal with all kinds of people who will expect you to 
be professional and, in particular, be knowledgeable about the purpose of the survey and about the 
questionnaire. This manual and training course will adequately prepare you for that. It is therefore 
important that you pay attention during training in order to learn the survey concepts and procedures. 
Training Continues in the Field 
Formal training provides you with basic knowledge and information regarding the survey, the 
questionnaires, and field procedures. Continued observation and supervision during the field work builds 
on this. As you encounter situations which you did not cover in training, discuss them with your 
supervisor. Your experience may alert us (NORC and TKG) to a new aspect of the survey and may assist 
with the interpretation of data. 
Your training as an enumerator does not end when the formal training period is completed. Each time 
your supervisor meets with you to discuss work in the field, your training is being continued. While in the 
field, take every opportunity you can to improve your skills as an enumerator to ensure we collect the 
highest quality data. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 98
Full Training Agenda 
Day 1 – Tuesday, April 24, 2017
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30 Introduction of participants and objectives of the training
9:30 - 10:30 Client introduction and Overview of Project
10:30 - 11:00 Tea Break
11:00 - 11:30 Overview (continued)
11:30 - 13:00 Overview of the enumerators’ manual, principles of data collection
13:00 - 14:00 Lunch
14:00 - 14:45 Ethics Overview – Cofidentiality, Anonymity, and Field Staff Responsibilities
14:45 - 15:30 Interviewing Children: Ethics and Special Consisderations
15:30 - 15:45 Tea Break
15:45 - 16:45 Interviewing Children: Theory and Practical Techniques
16:45 - 17:00 Closing
Day 2 – Tuesday, April 25, 2017
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30 What is EGRA? Importance of quality education, why early grade, why reading
9:30 - 10:30 Overview of the enumerators’ manual, principles of data collection
10:30 - 11:00 Tea Break
11:00 - 11:30 Overview (continued)
11:30 - 13:00 Overview and Intent of Teacher, Principal, and Student Background Surveys
13:00 - 14:00 Lunch
14:00 - 15:00 EGRA Administration Principles
15:00 - 15:45 Distribution of Tablets and EGRA Verbal Assent and Biographical Info
15:45 - 16:00 Tea Break
16:00 - 16:45 Initial practice with consent and tablets 
16:45 - 17:00 Closing
Day 3 – Wednesday, April 26, 2017
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30
Recap of day 1 
Day 2 Agenda
9:30 - 10:30 Task 2 Phonemic Awareness 
10:30 - 10:45 Tea Break
10:45 - 12:00 Task 1 Letter Name
12:00 - 13:00 Task 3 Familiar Word
13:00 - 14:00 Lunch
14:00 - 15:00 EGRA practice in pairs
15:00 - 15:45 Task 4 Non-Word Decoding
15:45 - 16:00 Tea Break
16:00 - 16:45 Task 5 Oral Reading Fluency/Reading Comprehension
16:45 - 17:00 Closing
Day 4 – Thursday, April 27, 2017
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 99
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30
Recap of day 2 
Day 3 Agenda
9:30 - 10:15 EGRA Task 6 Listening Comprehension
10:15 - 10:30 Tea Break
10:30 - 12:00 EGRA practice in pairs, small groups and modeling in plenary
12:00 - 13:00 EGRA review in group
13:00 - 14:00 Lunch
14:00 - 15:00 Assessor Accuracy Test #1 
15:00 - 15:45 Student Questionnaire
15:45 - 16:00 Tea Break
16:00 - 16:45 Student Questionnaire practice
16:45 – 17:00 Closing
Day 5 - Friday, April 28, 2017
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30
Recap of Day 3
Day 4 agenda
9:30 - 10:15 Assessor Accuracy feedback
10:15 - 10:30 Tea Break
10:30 – 12:00 EGRA practice in pairs, small groups and modeling in plenary
12:00 - 13:00 Assessor Accuracy Test #2
13:00 - 14:00 Lunch
14:00 - 15:00 IRR Protocol in the Field
15:00 - 15:45 EGRA practice in pairs, small groups and modeling in plenary
15:45 - 16:00 Tea Break
16:00 - 16:45 EGRA Jeopardy! Game
16:45 - 17:00 Closing
Day 6- Saturday, April 29, 2017
8:30 - 9:00 Arrival and registration of participants
9:00 - 9:30
Recap of Day 4
Day 5 agenda
9:30 - 10:30 Assessor accuracy feedback
10:30 - 10:45 Tea Break
10:45 - 11:45 Teacher Questionnaire
11:45 - 13:00 Principal Questionnaire
13:00 - 14:00 Lunch
14:00 - 15:00 Tangerine Troubleshooting
15:00 - 15:45 EGRA practice in pairs, small groups and modeling in plenary
15:45 - 16:00 Tea Break
16:00 - 16:45 Assessor Accuracy Test #3
16:45 - 17:00 Closing
Day 7 – Monday, May 1, 2017 
7:00 - 15:00 Full Pilot Day – School Visit 
Day 8 – Tuesday, May 2, 2017
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 100 
8:30 - 9:00 Arrival and registration of participants
9:00 – 10:30 Recap of Pilot Day
10:30 – 10:45 Tea Break
10:45 – 13:00 Selection of enumerators and supervisors; break into teams and discuss logistics
13:00 – 14:00 Lunch
14:00 – 15:45 Paper EGRA protocol for select enumerators; roles of supervisors during training
15:45 – 16:00 Tea Break
16:00 – 16:45 Paper EGRA practice / roles of supervisors (cont.)
16:45 – 17:00 Closing
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 101 
2. Background and Objectives 
2.1 WHO WE ARE 
NORC at the University of Chicago is an independent 
research organization that collaborates with government 
agencies, foundations, educational institutions, nonprofit 
organizations, and businesses to develop enduring 
knowledge that supports evidence-based decisions. NORC applies academic excellence, innovative and 
rigorous methods, and a collegial philosophy to the most pressing and complex global issues in fields such 
as health, education, economics, crime, justice, energy, security, and the environment. NORC’s 70 years 
of leadership and experience in data collection, analysis, and dissemination—coupled with deep subject 
matter expertise—provides the foundation for effective solutions. For more information, you can visit: 
www.norc.org. 
NORC is running the impact evaluation of the Read Liberia! program. NORC designed the evaluation, and 
is working with STS and TKG to collect data from respondents in Liberia. Once data is collected, NORC 
will analyze the data and report on the impact of the Read Liberia! program. 
School-to-School International (STS) is a non￾profit organization based in the San Francisco Bay 
Area. STS specializes in providing technical services 
in education in three main areas: (1) research and evaluation, (2) curriculum and training, and (3) 
policy and planning. For over a decade, they have partnered with individuals, government ministries, 
non-governmental organizations, private foundations, and research institutions to improve 
educational outcomes. Currently, STS is operating in 19 countries across the globe. 
STS will serve as a technical specialist on EGRA assessments and EGRA-related data collection for 
the baseline survey. STS will provide EGRA training to data collection enumerators, and carry out 
data collection oversight of EGRA. 
The Khana Group (TKG) is a research organization that was 
founded in 2008 and conducts over 20 research projects annually, 
including about 4 large nationally representative surveys (1000+). TKG 
recruits and trains over 300 data collectors/researchers annually, 
manages multiple development projects/programs and provide support to various projects/programs. 
TKG is in charge of collecting the data needed for the impact evaluation, including arranging logistics for 
data collection, helping NORC make sure the instruments are appropriate for the Liberian context, 
overseeing the quality of data collection, and hiring staff like you to collect the data! You work for TKG, 
but are part of the entire Read Liberia! Evaluation team. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 102 
2.2 OUR CLIENT 
The United States Agency for International Development is 
the lead U.S. Government agency that works to end extreme global 
poverty and enable resilient, democratic societies to realize their 
potential. Working with the Liberian Government, cooperating 
partners, and the people of Liberia, USAID’s work in Liberia focuses on creating more effective, 
accountable, and inclusive governance; providing sustained, market-drive economic growth to reduce 
poverty; improving the health status of Liberians; and ensuring Liberians are better education. 
USAID is funding both the intervention itself, and the evaluation study of the Read Liberia!
2.3 HOW DO WE ALL WORK TOGETHER? 
It is important that you understand how all these organizations and stakeholders work together so you 
can best represent TKG/NORC while in the field. These are the things you should remember: 
1. USAID engaged NORC to do the data collection for this survey and analyze the data. NORC in turn 
engaged TKG to implement the data collection in the field. 
2. While TKG and NORC work together to implement data collection, you are an employee of TKG. 
NORC, STS and TKG are conducting this training together but when you are in the field, you will be 
wearing TKG name tags. In order to avoid confusing respondents, you will tell them that you are from 
TKG. 
3. Our client is USAID. In the end, we will deliver the dataset and analysis to USAID. 
2.4 THE INTERVENTION UNDER STUDY 
In order for you to feel confident talking to the respondent and conducting the interview, it is very 
important that you are as familiar as possible with not only the questionnaire but also the education 
intervention under study. 
Read Liberia! is a 5-year USAID-funded initiative to improve reading skills for public primary students in 
first and second grade as emergent literacy skills for kindergarten students in the six targeted counties in 
Liberia. Read Liberia! will work through the Government of Liberia and in collaboration with other 
partners, and will concentrate on four results: (a) increased government commitment to, and support of, 
evidence-based reading instruction; (b) improved early grade reading classroom instruction; (c) improved 
service delivery systems in early grade reading; and (d) increased parent, community and private support 
for early grade reading. 
2.5 OUR SURVEY
Our objective is to collect baseline data (prior to the implementation of the Read Liberia! intervention) 
for the evaluation of the Read Liberia! project. For this baseline, we will be collecting 3 types of data: 
1. Early Grade Reading Assessment: Collects data on the reading levels of young learners. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 103 
2. Teacher Survey: Collects information on the learner’s teacher, such as level of education and 
professional qualifications. 
3. Principal Survey: Collects information on the school, such as the size of the school and the 
quality of the resources. 
2.6 BASELINE DATA COLLECTION
The purpose of collecting data from learners, teachers, and principals is to understand the reading levels 
of students prior to the beginning of the program. Your team in May 2017 will be speaking to learners, 
teachers, and principals to understand how learners in Liberia are taught to read, the home environments 
of the learners, and the school characteristics. The data you collect will help us answer the research 
questions in the evaluation. 
We will visit again the same schools at endline to interview students, teachers and principals. Comparing 
baseline and endline data provides us with a “before” and “after” snapshot of respondents, and allows us 
to better understand how reading scores and practices have changed over time, if at all. 
3. Methodology 
3.1 WHERE WILL OUR SURVEY BE CONDUCTED? 
Our survey will be conducted in villages that are located in the counties of Monteserrado, Margibi, Bong, 
Grand Bassa, Nimba, and Lofa. We will be visiting 90 schools. We are interviewing 16 Grade 2 students, 
one teacher, and one principal in each school—for a total of 1,440 students, 90 teachers and 90 principals. 
We randomly chose the 90 schools in order to have a representative sample. 
The study we are conducting is a randomized controlled trial—this means half of the schools will receive 
the Read Liberia! intervention (called the treatment group), and the other half will not receive this 
particular intervention at this time (called the control group).Each school has an equal chance of being 
assigned to a treatment or control group. The assignment into treatment or control groups will not 
happen until after the baseline assessment is complete, which means nobody knows which schools will be 
assigned to these two groups right now. For this reason, it is extremely important that you do not say 
anything to teachers, principals, or even students about what will happen with the program in their school. 
If you are asked by the schools staff about the program intervention in general or about specific activities 
of the program, it is important to let them know that at this point your team is only conducting the 
baseline research and that you do not have answers to those questions at this time. The answers to those 
specific questions will be developed by USAID and the Ministry of Education over the coming months as 
they develop additional details for this program intervention. 
3.2 SAMPLE SIZE 
The term “sample size” refers to the number of respondents that we will survey. Our target sample size 
is approximately 1,440 students—16 for each school, in addition to the 90 teachers and 90 principals—
one for each school. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 104 
The sample has been selected so that it best represents the population under study. Each learner or 
each teacher represents 100's and sometimes 1000's of other learners or teachers in Liberia who would 
answer in the same way. It is therefore very important that we reach that target number and that we 
do not substitute the selected respondent for someone else since it might misrepresent the 
population and bias the results. Later in the training, we will talk about the importance of gaining 
cooperation and avoiding refusals. 
3.3 HOW WILL OUR QUESTIONNAIRES BE STRUCTURED? 
The Early Grade Reading Assessment is an assessment administered to learners to measure their reading 
levels in English. The EGRA is broken down into several subtasks, presented below: 
Subtask Description
1 ORIENTATION TO PRINT
2 LETTER NAME KNOWLEDGE
3 PHONEMIC AWARENESS
4 FAMILIAR WORD IDENTIFICATION
5 NONWORD DECODING
6 ORAL PASSAGE READING
7 LISTENING COMPREHENSION
Our student questionnaire is short (10 minutes) but collects detailed information about learners’ reading 
practices at home, availability of print materials, socioeconomic status, and other topics. Our teacher 
questionnaire is also short (20 minutes), and collects detailed information about teachers’ backgrounds, 
education levels, teaching credentials, and other information. Our principal questionnaire (also 20 
minutes) collects information on principals’ background, interaction with teachers and the school 
characteristics. 
Lastly, per the guidance from the University of Liberia ethics committee we must collect written consent 
from the school principal to for the school to participate in the study. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 105 
4. General Overview 
4.1 TEAM STRUCTURE 
There will be a total of 6 Supervisor-led teams, with 1 supervisor and 3 enumerators on each team. The 
TKG structure for this project is outlined below with each of the 6 teams being led by a supervisor. 
4.2 ROLES AND RESPONSIBILITIES 
4.2.1 ENUMERATOR – 3 PER TEAM 
The most important role of the enumerator is to collect data so the researchers can analyze it and 
produce the statistics used by scientists, policy makers, academics, corporate executives, government 
officials and others. Enumerators must precisely follow the guidelines presented in this manual so 
that we can maintain a high level of accuracy. Sampling procedures must be followed exactly in order 
for the information to accurately represent the designated population. Finally, it is essential for the 
accuracy of data that the enumerator does not introduce any bias into the survey (techniques to 
remain neutral and avoid bias will be covered later in the training). 
As an enumerator, it is your responsibility to: 
 Follow Instructions: It is your responsibility to clearly understand the manual and follow 
instructions on the questionnaires and the manual. It is necessary that each enumerator ask the 
survey questions the same way, and use identical definitions and procedures. Instructions must be 
thoroughly learnt and strictly followed. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 106 
 Maintain Confidentiality: It is your responsibility to keep the information you collect 
confidential. Your tablets at all times be kept in a secure place, used discretely, locked, and 
password protected. Once you have checked all questionnaires, you should let your team’s 
Supervisor know so that he or she can perform validation checks to ensure data is accurate. 
 Complete Your Assignment: You are expected to work hard and complete the area assigned 
to you in good time. In case you become ill or incapacitated while working on the survey you 
should report immediately to your supervisor so that he/she can arrange a replacement.
The enumerator is responsible for carrying out the following tasks: 
 Home study prior to and during training, which may require reading and completion of exercises 
 Administering/fielding the Read Liberia! Baseline Survey and managing the case load so that all 
appropriate schools, students, and teachers are surveyed in the allotted time 
 Travelling to respondent's schools, during the school day
 Gaining the cooperation of the respondent and reassuring those who are skeptical or 
reluctant so they feel comfortable and agree to participate. 
 Documenting the status of particular cases and assigning disposition codes. 
 Routine communication, as specified by the project, with the supervisor to report progress 
on the assignment, set goals, and discuss strategies in approaching difficult cases. 
 Accommodating respondents as well as changes in project guidelines. 
 Document any refusals in the tracking sheet 
 Support team supervisors to make first contact with the school 
 Be available for daily start up and end-of day meetings with the rest of the team 
 Correctly filling out the tracking sheet and survey tool 
The process of becoming a good enumerator requires both discipline and commitment. We count on 
you for collecting the highest quality data possible! 
What you can NOT do 
The continuation of your appointment is conditional on your satisfactorily carrying out the duties of an 
enumerator and maintaining a high standard of professionalism and work ethic. In this respect, keep in 
mind that you are representing NORC and TKG and your behavior could also have an impact on our 
client, USAID. Your conduct must therefore be professional at all times. 
Note that NORC/TKG relies on the goodwill and cooperation of respondents. Any enumerator who is 
aggressive or disrespectful will be dismissed from the survey. Other grounds for dismissal include: 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 107 
 Performing Other Activities During Enumeration: You are prohibited from engaging in 
activities such as petty trading, political or religious propaganda during enumeration. People who 
are in employment elsewhere cannot take part in the fielding of the survey. 
 Allowing Unauthorized Persons to Help You in Your Work: You are not permitted to 
delegate your work to anyone. No one can help you except your supervisor or another 
enumerator as directed by your supervisor. 
 Discussion of Survey Work with Unauthorized Persons: You are not to discuss your work 
with anyone except your supervisor or an authorized Survey staff. It is fine to discuss the general 
purpose of the survey and the nature of your job with your family members or friends but you 
must not discuss the respondent’s data that you will be collecting.
 Absenteeism: Any person who is absent from the training or enumeration without prior 
approval from the survey project manager or his/her supervisor will be dismissed from the survey. 
 Disruptive Influence: Any enumerator who creates a disruptive influence will be dismissed from 
the survey. 
 Breach of Confidentiality: Disciplinary proceedings will be taken should you breach the 
confidentiality requirements of the survey. In addition, you may face legal prosecution. 
If you break any of the above rules disciplinary action such as withholding of allowance, or summary 
dismissal will be taken against you. 
4.2.2 SUPERVISORS/FIELD SUPERVISOR (FS) – 1 PER TEAM 
Each team will have 1 supervisor. The supervisor plays an important role in the continued training of 
enumerators and in ensuring the quality of the survey. The duties of the supervisor include: 
 Locating the schools in the sample (with the help of GPS coordinates when available) 
 Introducing/explaining the purpose of the data collection to a responsible member of the 
community before work commences—in this case, likely the principal and a village head.
 After introduction lead the process of setting up appoint ents with schools 
 Giving out daily assignments to enumerators and ensuring that daily data collection activities run 
smoothly 
 Coordinating and integrating the teams driver into the process of data collection and inform driver 
of pick-up and drop off time and location. 
 Tracking and reporting on field issues to TKG’s research coordinator, and following their 
instructions as you solve any problems related to finding the listed villages, understanding the 
concepts in the questionnaire, or with difficult respondents 
DO NOT MAKE EMPTY PROMISES TO RESPONDENTS
During data collection, you will visit many schools. It is of the utmost importance 
that you introduce the study in the same manner to all schools and that you do not promise 
to any respondent that any specific activity will take place. As a research firm, we must 
make sure that we do not set expectations in order to avoid potential conflicts and 
resistance of respondents to participate in future studies. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 108 
 Administering one teacher and one principal survey at each school 
 Performing 10% direct observations of the team’s interviews to ensure that enumerators are 
conducting interviews in the right manner, properly recording results, and interpreting concepts 
correctly. 
 Conducting 15% random spot-check—where you randomly pop in and out while interviews are 
being conducted to make sure they are being conducted correctly 
 Coordinating with Quality Control Officers to ensure they are able to locate you and your team 
at your school each day. 
 Cooperate and assist Quality Control Officers so they are able to perform direct observations of 
at least 10% of your team’s interviews and 15% spot check (these are in addition to the direct 
observations conducted by yourself of your team, as described above) 
 Assigning disposition codes to each case 
 Completing the central office receipt control sheets 
 Meeting with the enumerators on a daily basis to discuss performance and challenges 
 Reporting to assigned research coordinator daily 
 Ensure enumerators have functional, fully-charged tablets and other survey materials 
 Communicate with TKG if there become a need to replace a school 
 At the end of field work, develop a report of field activity, documenting challenges, success and 
lessons learned. 
 Review all completed questionnaires for completeness 
4.3 QUALITY CONTROL 
NORC and TKG consider rigorous quality control crucial to conducting superior survey research. Over 
the years, our field supervisors, enumerators and Central Office staff have shared efforts to produce data 
of consistently high quality and we are proud of this cooperation. High quality begins with training and 
continues throughout all our work. Listed below are some of the protocols we have implemented to 
ensure high quality standards. 
Quality control begins with training: All supervisors, quality control officers, data managers and 
enumerators will participate in survey training to familiarize themselves with the survey and all of its 
procedures. 
 Questionnaire mock interviews and assessments: Each enumerator will be required to complete 
several practice interviews and assessments during training before going out into the field. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 109 
 In-training pilot test: The full teams will conduct one day of pilot testing of the EGRA in the field 
during training. This will be followed by a day of team debriefing, and subsequent modifications of 
instruments or protocols (if required) before the full data collection process begins. 
 Data review: All data that is transmitted from you will be reviewed by your Field Supervisor and 
Central Office staff for completeness and accuracy. Overall, the data review process will entail 
employing software for monitoring interviews and data, daily observations of enumerators by the 
team leaders, and review of raw data for consistency as it arrives.
 Quality control is a two-way street. If you find errors, have difficulties with the survey, or are 
unclear about procedures at any time during the field period, please bring them to your 
supervisor’s attention right away. Enumerators have sharp eyes for problems we did not 
anticipate, and we value your comments. If you are encountering a problem, chances are others 
are as well and the sooner your supervisor is notified, the faster the problem can be addressed 
for everyone on the field team. 
More detailed data quality control procedures will be described later in this manual. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 110 
5. Basic Interviewing Techniques 
5.1 GAINING COOPERATION 
Successful interviewing is an art and should not be treated as a mechanical process. Each interview is a 
new source of information; so we have to make it interesting and pleasant. The art of interviewing develops 
with practice but there are certain basic principles which are followed by every successful enumerator. In 
this section we will review a number of general guidelines on how to build rapport with the respondent 
and conduct a successful interview. 
Your role as an enumerator is vital to the success of this study. This section focuses on gaining respondent 
cooperation and then discusses techniques to use once an interview has started. 
How Gaining Cooperation Is Accomplished 
Gaining cooperation is a combination of many factors of behavior and sensitivity toward the individual or 
individuals with whom enumerators make contact. The successful enumerator is organized, prepared 
logistically and mentally, and LISTENS in order to know exactly what to say based on the circumstance of 
the contact. 
Here are some rules to follow that will help you gain cooperation: 
 Be Organized 
 Be Prepared Logistically: Have all materials at hand. 
 Be Prepared Mentally: Before making any respondent contact, review your materials, 
especially data relating to the study. This would include this Training Manual, advance letters, the 
questionnaire and QxQs (Question by Question Specifications). Be prepared to answer any 
questions you may be asked from the respondent or questions you may need to ask the 
respondent. 
 Know The Facts: About NORC, STS, TKG, USAID, and social science research. 
 Know The Study 
o Who is sponsoring the research 
o The reason it’s being conducted 
o The uses of the data 
o Why we need the respondent’s cooperation 
 Be Sold On The Study: You must believe in the survey's value yourself; otherwise, you'll be 
unable to convey the value of the study to the respondent 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 111 
 Assume a good attitude: Your attitude directly impacts the respondent. 
 Be Confident: Sound interesting and be interested. 
 Be Professional, But Balance That With Some Friendliness 
 Be Assertive, But Not Pushy 
 Let Human Qualities Show Through 
 Be Flexible 
 Smile! 
5.1.1 GAINING COOPERATION TECHNIQUES AND FREQUENTLY ASKED 
QUESTIONS 
In order to gain cooperation, it is important to set the right tone from the first contact with a potential 
respondent. Your supervisor will also be meeting with the principal when you arrive in a new school in 
order to facilitate entry and gain cooperation from target schools. 
Your efficiency is important. We want to disrupt the school day as little as possible! 
a. Introducing yourself and building rapport
You will be working with pupils and with teachers. They do not know you so it is important to give a good 
impression and establish good rapport from the get-go by being courteous and professional. 
1) When introducing yourself to the teacher you can say: 
“My name is …. I am with TKG, a research organization from Monrovia. We are conducting a 
survey to better understand how children in Liberia learn to read and their reading practices 
and we need YOUR help.”
Never adopt an apologetic manner, and try to avoid sentences like “could you spare a few minutes?” or 
“would you mind answering some questions?” Ours is an important national assignment, there is no need 
to be apologetic. On the other hand, don’t try to be over bearing. For example, say “I would like to ask 
you some questions” or “I would like to talk to you for a moment.” rather than “You must answer the 
following questions.”
2) When working with a child it is very important that you build a positive rapport with the pupil you are 
assessing. When a pupil feels at ease with you, the pupil will be more motivated to continue the assessment 
and more likely to demonstrate his or her best efforts. The following list offers useful guidelines about the 
rapport-building period: 
Spend approximately 1–2 minutes developing rapport. Less could be too short to make the child 
comfortable with you and longer will delay the assessment too much. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 112 
Be self-assured and positive about the experience. Do not apologize for taking the pupil to do an 
assessment. Many pupils find the assessment like a game and have fun doing it. Present it this way in a 
smiling and relaxed manner. Pupils will take their cue from you regarding what the experience will be like. 
Be positive and warm all through the test regardless of the learner performance. Avoid facial expressions, 
impatience, or comments about the responses. You want to encourage the child to keep on completing 
the assessment. Compliment the effort and not the correct answers. 
b. Listen attentively to what the respondent is saying
If either the teacher or the learner have questions, please listen attentively and answer to their 
questions. Always remember to thank the respondent for his or her time and to make the appropriate 
notes on the questionnaire. 
c. If the respondent does not seem to want to participate 
Stress the Importance of the Respondent and the Survey 
Stress that the respondent is providing his/her information so that others might learn from him/her. An 
effective psychological tool in gaining cooperation is to give the respondent a vested interest in this study 
by impressing on the respondent that he/she cannot be replaced and that he/she represents many others 
like himself/herself. He/she represents many other people in Liberia. Make the respondent feel a sense of 
importance because he/she has been selected to participate and emphasize the importance of his/her 
opinions. This will require that you are familiar with the survey (see FAQ below). For example you could 
say: 
Interacting with children Dos and Don’ts
DO:
 Set a friendly tone 
 When introducing yourself tell the truth while talking about things which are relevant 
for a young child
 Listen to the child carefully, look at him/her when they talk
 Answer the child’s questions in a nice but concise way
 SMILE!
DON’T:
 Do not rush through questions if the child has not understood what you are asking 
them to do
 DO not talk about issues that are difficult to understand by a young child (“I have a 
diploma in public accountancy” might be a difficult concept; you might want to say “I 
finished university/college”)
 Do not talk with the child without looking at him/her
 DO not use harsh words! 
 Do not make comments on the child’s answers (both in EGRA and when 
administering the questionnaire) – remain neutral in your responses to the child
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 113 
“Your answers are very important in order to give policy makers the information they 
need to make informed decisions about the future of children in Liberia. We are asking 
you to participate in the survey so that your voice can be heard and so that we may 
learn more about individuals similar to you.” 
Empathize 
For example, if the teacher or principal is concerned about the amount of time required for the interview, 
explain that you understand how difficult it is to manage all of the time demands he/she is under. Ask if 
there is a better time for you to come back. 
Don’t Argue! 
Even if the respondent does not want to participate, maintain a pleasant, friendly attitude and stress the 
positive. Stress how important the survey is and how important it is for each respondent to participate. 
Make the respondent feel very special and important. A useful technique is getting the respondent to say 
“yes” to something, like “Do you remember why you were asked to join in this survey?” or “Do you 
remember what the study is about?” Once a respondent says “yes” to a question, he or she is much more 
inclined to continue cooperating. 
d. Know the study 
Practice, Practice, Practice 
 Become very familiar with the study, the questionnaire, and the uses of the data. 
 Anticipate respondent’s questions and/or concerns.
 Practice your answers out loud in front of a mirror. It really works. 
 Ask a family member to listen to your explanation of the survey (but do not discuss the 
respondent’s data)
 Review your points with your Field Supervisor. 
 Learn all you can about confidentiality and the handling of respondent data. 
5.1.2 FAQ 
These are questions that the respondent might ask you. You should be prepared to answer them at any 
time. 
Q: What is the study about? 
A: The study is collecting baseline data on children’s reading knowledge and practices to understand how 
to improve children’s reading performance in Liberia.
Q: Who is financially supporting the study? 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 114 
A: The study is financially supported by the USAID which has recruited TKG to conduct the interviews. 
Q: What will you do with the data? Will my data be kept confidential? 
A: We will compile all of the data from all the assessments and interviews we are conducting and then 
analyze it. The data is therefore in aggregated form. This means that we will NOT identify you in any of 
our reports and your data will always be kept confidential. If you feel hesitant about responding to any 
questions, you may choose to skip the question. 
Notes: In order to reinforce what you are saying about confidentiality, it is important that you keep your 
tablet secure. It is also important that you do not mention the names of other respondents that you’ve 
already interviewed. Your careful handling of questionnaires will do much to build up the respondent’s 
confidence. 
Q: What is the survey about? How long is the survey? 
A: We will ask you questions about your teaching experience, about teaching early grade reading, things 
teachers do during class, school characteristics, and how often the pupil reads in class. The survey should 
take about 20 minutes to complete. 
Notes: If the respondent is not comfortable with the length of the interview, try to stress the importance 
of the information we are collecting. 
Q: Will I get anything if I participate? 
A: The answers that you provide are very important since they will help us understand the well-being of 
different types of individuals. These answers will inform future decisions made by policy makers regarding 
programs that might help your community and the rest of Liberia. 
Q: Who can I contact if I have any questions about the survey? 
A: You may contact Charles Nnabugwu at TKG. His phone number is +231-770-431-133.
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 115 
5.2 CONDUCTING THE INTERVIEW 
In this section, we will review some tips and information to keep in mind once you have gained respondent 
cooperation and have begun an interview. 
5.2.1 INFORMED CONSENT
In order to protect the rights of respondents it is essential to first allow them to consent to the interview. 
The following text must be used when introducing the survey as it has been approved by NORC’s 
Institutional Review Board (our board of ethics) and the University of Liberia’s Institutional Review Board.
We want to know if: 
- The participant is willing to participate in the survey
Survey Participation 
First, we ask the respondent if he or she wants to participate in the survey. In the case of learners, you 
will read out the following text: 
My name is _________. I work with The Khana Group. 
 We are trying to understand how children learn to read. You were picked by chance, like 
in a raffle or lottery. 
 We would like your help in this. But you do not have to take part if you do not want to. 
 We are going to play a reading game. I am going to ask you to read letters, words and a 
short story out loud. 
 Using this stopwatch, I will see how long it takes you to read. 
 This is NOT a test and it will not affect your grade at school. 
 I will also ask you questions about your family, like what language your family uses at home 
and some of the things your family has. 
 I will NOT write down your name so no one will know these are your answers. 
 Once again, you do not have to participate if you do not wish to. Once we begin, if you 
would rather not answer a question, that’s all right. 
You will then ask the pupil the following: 
Do you have any questions? Can we get started? 
In the case of teachers, you will read out the following text: 
My name is _________. I work with The Khana Group in Liberia. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 116 
 We are trying to understand how children learn to read. Your school was selected 
through the process of statistical sampling. We would like your help in this. But you do 
not have to take part if you do not want to. 
 Your name will not be mentioned anywhere in reports based on this survey. Results of 
this survey will be presented in an aggregated format so that your specific responses 
will not be identifiable The information acquired through this instrument will be shared 
with the Ministry of Education in the hopes of identifying areas where additional 
support may be needed. 
 Your responses will NOT affect you in any way. They will not have any impact on your 
employment or your pay. 
 If you agree, I would ask you some questions regarding your normal activities at school, 
including your interactions with school staff, Ministry office staff, students, and parents. 
 Then, with your assistance, I would randomly select 16 students in Grade 2 to assess 
their reading. I would also ask these students about some of their normal school 
activities, school assets, language use, and reading practices at home, as well as home 
asset ownership. Selected students need only participate if they wish. I will spend about 
40 minutes interviewing each child. My interview with you will take 15 to 20 minutes. 
 Are you willing to participate? Once again, you do not have to participate if you do not 
wish to. Once we begin, if you would rather not answer a question, that’s all right. 
Can we get started? 
And for principals, you will read the following text: 
My name is _________. I work with The Khana Group in Liberia. 
 We are trying to understand how children learn to read. Your school was selected 
through the process of statistical sampling. We would like your help in this. But you do 
not have to take part if you do not want to. 
 Your name will not be mentioned anywhere in reports based on this survey. Results of 
this survey will be presented in an aggregated format so that your specific responses 
will not be identifiable The information acquired through this instrument will be shared 
with the Ministry of Education in the hopes of identifying areas where additional 
support may be needed. 
 Your responses will NOT affect you in any way. They will not have any impact on your 
employment or your pay. 
 If you agree, I would ask you some questions regarding your normal activities at school, 
including your interactions with your staff, Ministry office staff, students, and parents. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 117 
 Then, I would randomly select 16 students in Grade 2 to assess their reading skills. I 
would also ask these students about some of their normal school activities, school 
assets, language use, and reading practices at home, as well as home asset ownership. 
Selected students need only participate if they wish. The entire interview process for all 
16 students will last around 4 hours (40 minutes per each child). My interview with you 
will take 15-20 minutes. Finally, I would spend about 15-20 minutes interviewing the 
Grade 2 teacher.. 
 Are you willing to participate? Once again, you do not have to participate if you do not 
wish to. Once we begin, if you would rather not answer a question, that’s all right. 
Can we get started?
DO NOT PROCEED UNTIL THE RESPONDENT SAYS “YES”
The respondent must explicitly answer “yes” to this last question before you are 
allowed to proceed with the interview. You will find this same text at the beginning of our 
questionnaire with space to mark the respondent’s consent. You must mark down that the 
respondent answered “yes” before conducting the interview. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 118 
5.2.2 INTERVIEWING TECHNIQUES 
a. Basic Rules 
Remain Neutral 
When interviewing the respondent, you must be disciplined in your reaction to the information you are 
given. Do not by your expression or your reply register disbelief in the data he has given you. Be calm and 
neutral whenever you need to indicate that the data is inconsistent with other data previously given. Later, 
we will talk about the different types of bias that should be avoided and techniques for remaining neutral. 
Don’t Estimate for the Respondent
In many cases we will be asking respondents to make estimations. For example, we will ask how many 
times the principal observes the Grade 2 teacher lessons. In such cases you can tell the respondent ‘Please 
give me your best estimate’ or ‘I understand you might not know 
the exact number, we are still interested in your best estimate 
though.’ Remember it is the respondent who is to make the 
estimate, not you. Earlier respondents may have given you very 
different answers; we are interested in what each respondent 
thinks, regardless of the correctness or incorrectness of the 
answer. 
 If a respondent cannot answer a certain question, for 
example how many times a week they spend reading 
with the students, ask them to provide an estimate or 
guess 
 Unless we say that multiple answers are accepted, there 
is just one single answer per person 
It is also possible that the respondent thinks that more than one 
answer is appropriate to a single question, encourage the 
respondent to choose the one response that is most appropriate 
or the one response that is more important. 
If the respondent refuses to answer a question, reassure the 
respondent that his/her answers are confidential and emphasize 
how important it is that s/he tries to answer all of the questions. Do not try to determine an answer 
based on a statement that a respondent made on a previous response. There is a fine line between 
encouraging the respondent to answer and being too pushy. As an enumerator, you have to provide the 
respondent with the opportunity to answer, but you also should not force them to answer if you truly 
feel that they do not want to. If the respondent truly does not know the answer to a question or does 
not want to answer, skip this item and proceed to the next one. 
Three key traits of a good enumerator:
1. Being Professional: Enumerators 
perform a delicate balancing act between 
being congenial with a respondent and 
maintaining a professional distance. 
Maintaining a distance is essential to 
avoid biasing the interview. 
2. Being Non-Judgmental: Enumerators 
must be non-judgmental, show no 
emotion or reaction to a respondent's 
answer and not lead the respondent 
toward a possible response. It is never 
appropriate to comment on an answer to 
a question.
3. Keeping Control: Occasionally a 
respondent will enjoy giving his/her 
opinions, which may ultimately lead to 
taking control of the interview to tell 
stories. If this should occur, the 
enumerator must tactfully, in an 
assertive manner, regain control. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 119 
Do Not Change the Sequence of the Questions 
Questions should be asked in the order they appear on the tablet-based survey application. Ensure that 
the respondent has answered a question sufficiently before trying to move to the next. Moreover, all 
answers should be noted on the survey itself at the time of the interview. Do not take notes on a pad and 
transcribe them to the survey later. 
Don’t Allow the Respondent to Hurry the Interview 
Respondents sometimes try to hurry the interview. Politely say that you have to get answers to each 
question, and calmly proceed with the interview. If you get lost or stumped, take your time to read any 
instructions or to find your place. Don’t hurry. Let the respondent know what you are doing. Once you 
have completed the interview, it will be difficult to go back and correct errors or edit. 
In Summary: 
 Always be polite and patient 
 Never show disbelief with the respondent’s answers
 Never argue. Be professionally persistent. 
b. Additional Techniques and Concerns 
To supplement the best practices for interviewing, here are tips for good in-person interviewing: 
 Appropriate eye contact during the interview is very important to establish comfort and trust 
with the respondent. Use your best judgment as eye contact with elders and gesturing with your 
left hand may also be considered rude in Liberia. 
 There are some questions that a respondent may hesitate to answer if other people are present. 
Use your common sense, tact, and ingenuity to avoid/elude nosy neighbors, fussy children, curious 
friends or spouses, and other such annoyances. For example you could say to the third party: “I 
am interviewing [respondent name] for an important survey conducted by TKG. It is important 
that I conduct this interview in private. Would you mind coming back at a later time when we are 
done with the interview?” If all else fails and you cannot rid yourself of a third party, try again to 
request privacy when you come to a sensitive question. With luck, the third party will become 
bored and will leave on his or her own. 
 In some cases the teacher or principal may request help from another person in answering a 
question. This is okay in instances where the question is factually based and someone else in the 
school may have more accurate information. 
o For example, if you ask the principal “In one year, how often do you send teachers for in￾service teacher training, on average?” it is possible there is an administrative assistant in 
the school who has more knowledge (perhaps they keep a record) of the number of days 
teachers take part in these trainings. It is fine for the principal to consult with the 
administrative assistant in order to get a more accurate response to a question like this. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 120 
 However, if you ask an opinion question or a question that relates to a respondent’s behaviors or 
activities, you should encourage the respondent to share THEIR opinion on the question. 
o For example, if you ask a teacher “What reading skills should your children have at the 
end of the school year?” we are interested only in THEIR opinion on the topic. If they 
consult with someone else on a question like this, you should politely remind them “for 
this, we are just interested in your opinion. There are no right or wrong answers, so 
please just tell me how you personally feel about this.”
 For students, especially when they are being administered the EGRA, it is NOT alright for them 
to get assistance because we are specifically testing the student’s abilities and we do not want 
anyone to provide any assistance to ensure we are getting an accurate measure of that student’s 
skills. And for the student context questions, it is unlikely that anyone else can better answer 
questions about the students’ homes than the students themselves, so you should again encourage 
the student to give THEIR best answer. 
5.2.3 BIAS 
While you are conducting an interview, it is very important that you are aware of how you approach the 
interviewee. Enumerators collect data reflecting the attitudes, opinions, and behaviours of our 
respondents. Because of this, you should be aware of the ways in which you might influence respondents’ 
answers. There are three main types of bias: 
 style 
 nonverbal 
 verbal 
a. Style bias 
Style bias is a bias by first impression. Tone of voice, facial expression, appearance, 
actions, and attitudes will influence the respondent's welcome. 
It is important to present a pleasant but neutral appearance. You do not want to scare a 
potential respondent by presenting yourself in a way that offends or possibly belittles the 
respondent. You should always be aware of the image conveyed by your appearance and 
attitude and how respondents of different social and economic backgrounds might react to this. 
b. Nonverbal bias 
Nonverbal bias is revealed in body language and facial expressions. 
Nonverbal bias might influence a respondent’s answers through body language or facial 
expressions. It could be just a raised eye brow or physically pulling back when the 
respondent answers or asks a question, such as showing surprise or shock. Even a knowing smile can 
influence the way the respondent will react to further questions. As an enumerator, you must maintain 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 121 
self-awareness, keep a neutral face, and avoid reacting to respondent’s answers throughout the 
interviewing process to avoid nonverbal bias.
c. Verbal bias 
Verbal bias is influencing the respondent through your words. 
Verbal bias can happen during the first few seconds of your 
introduction, when the respondent is trying to decide if he/she wants to participate in the study. 
In this case your words may influence the respondent’s willingness to participate in the study or 
not. Verbal bias can also happen during the actual interview as you influence a respondent’s 
responses. During the interview, you should read questions as written, stressing only indicated words. 
As such, you should verbally practice introducing the study to become familiar with correct pronunciation, 
the flow of the words and the intent of the study. And you should never make any side comments about 
the respondent or the respondents’ answers.
5.2.4 PROBING 
What is a probe? 
A probe is a question or comment by the enumerator which encourages a clearer or more complete 
response without biasing the answer. A good probe is always controlled and non-directive or neutral and 
helps the respondent focus on the question. 
When is it necessary to probe? 
 For clarity 
 For completeness 
 For relevance - to return to the point of the question 
 To elicit more information from a "Don't know" response 
Six basic types of probes 
1. The silent probe: pausing 
A silent probe is often effective. Simply stop writing and look expectantly at the respondent. 
2. Re-reading the question 
Focus on the specific needs of the question. Stress key words. 
Example: When asking a pre-coded question the responses are included in the question and the 
enumerator is to code one of the predetermined options. A respondent may answer a pre-coded question 
with a term that is not one of the given options included in the question. The enumerator then must train 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 122 
the respondent to select one of the given options. The best probe is to re-read the entire question, 
including the answer options. 
3. Asking for more information 
 “Please tell me more.”
 “What else?”
 “What else do you mean that?” 
4. Stressing generality 
“In general, would you say...” -- then repeat the question. 
5. Stressing subjectivity 
 “In your opinion, [...]”
 “What is your best estimate?” 
6. Zeroing in 
 “What was the season?” 
 When the response is a range of numbers, the enumerator needs to zero in for a precise response. 
 When the response is in units other that what the question calls for, the enumerator needs to 
zero in for the correct units. For example: The question asks for the amount of time spent reading 
in minutes and the response is hours. 
Probing Don't Know Responses 
The NORC policy is to probe all “don't know” at least once. There are 3 basic reasons why people use 
“don't know” as a response: 
1. The respondent is thinking over his/her reply and says, "I don't know" to fill the silence. Silent 
probe: The best probe is to pause and wait expectantly while the respondent comes up with a 
codeable answer. 
2. The respondent is unsure and reluctant to answer for fear of giving the wrong response. Stressing 
subjectivity: The best probe is to assure the respondent that there is no right or wrong answers 
and we would like his opinion or best estimate. The respondent is very likely to come up with a 
codeable response. 
3. The respondent truly does not know how to respond and cannot give an answer under any 
circumstances. If an enumerator determines this to be the situation after probing once, it is 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 123 
acceptable to record the response as “don’t know” (codes for “refused” and “don’t know” will 
be discussed later). 
DO NOT PROBE DURING EGRA SUBTASKS! 
5.2.5 ASKING QUESTIONS
Survey questions and introductory or transitional statements are to be read exactly the way they are 
written with no changes in wording. Each question is carefully composed and tested so that most people 
will understand the meaning. Even small changes in the way questions are worded may have significant 
effects on the way questions are answered. Questions are written so as not to suggest answers and lead 
to biased results. 
It is important that you practice reading aloud the questions many times prior to the first interview. During 
training, the opportunity is presented to practice through several mock interviews to learn the 
questionnaire and rehearse reading the questions. Questions should be read slowly and in a conversational 
tone emphasizing the appropriate words. Remember, this is the respondent's first exposure to the 
questions and it is important that he/she hears it correctly the first time. 
Read all questions the way they are written 
Every question must be read to the respondent even though it may have already been answered in the 
context of a previous question or in general conversation. It is appropriate to acknowledge the previous 
information with a comment such as: “You may have answered this question earlier but I need to ask it 
at this time. I want to make sure I have it right." Another question may ask about something that seems 
obvious to you by observation, such as asking if the respondent has a bicycle and you can see one. DO 
NOT ASSUME. The bicycle may have been borrowed. 
Occasionally a respondent may ask you to define a word. When this happens, don't try to explain the 
meaning in your own words: simply reread the question. There is danger in trying to interpret and 
summarize meanings. If hundreds of interviewers each give their own version, the results could be similar 
to those in the old-fashioned game of “telephone” or "gossip”, where the last person repeating the original 
sentence gives a hilarious statement which bears little resemblance to the first. 
Four Types of Questions 
For this survey, you will encounter 4 main types of questions: 
 1. Numeric 
Some questions in our survey ask for a numeric value. You will be given a field such as: 
in which the number will be entered on your tablet.
How many years have you been a principal?
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 124 
2. Pre-coded (read out response options) 
The choice of responses is included in the text of the question and you have to read the response options 
to the respondent. Researchers have defined the range of possible responses. No verbatim is required. 
The appropriate answer is simply coded by the enumerator. 
3. Field-Coded (do NOT read out response options) 
The choice of responses is not part of the question but there are code categories to select. Researchers 
have some idea of the range of possible answers, and list them in the questionnaire. The enumerator then 
enters the code that corresponds the closest to what the respondent answered. 
The enumerator may need to probe to zero in on the right category. 
Think about the last 5 school days and tell me how frequently your students learn the 
meaning of new words
Never = 0
Sometimes = 1
Frequently = 2
Everyday = 3
Don’t know = 88 
Refuse/No Answer == 99
Why have you made special efforts to improve reading at your school this past year? Do 
not provide response options. Tick all that apply.
1. We saw other schools doing it
2. We thought it might be important because the children were tested
3. The Ministry told us to do it
4. An NGO told us to do it
5. Teachers got teacher training that showed them how to teach reading better
6. Other_________________________
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 125 
3. Verbatim (Open-ended) 
When the respondent gives an answer that is not one of the pre-coded answers, you will have to enter 
verbatim (word for word) what the respondent says. For all of these “other” options, you must specify 
exactly what the respondent says. 
For example: 
5.2.6 RECORDING RESPONSES ON TABLETS 
Different questions ask for various kinds of information and several recording techniques are used. 
Because we are using tablets to collect data, we have put in controls so that you do not accidentally type 
in letters when you were supposed to type in numbers, or do not select more than one response. 
PAY ATTENTION to what you click on! 
1. Open-ended responses (Other: Specify) 
If the question calls for a response to be recorded verbatim (i.e., exactly as said), type exactly what the 
respondent says word for word. The question option will not be available for you on the tablet, and 
instead you will need to type the answer down. You should not paraphrase or summarize the respondent's 
answer. 
Because most people speak faster than most enumerators can type, you need to develop comfortable 
techniques to handle verbatim situations. Here are some acceptable techniques: 
 Begin to record as soon as the respondent begins to answer. 
 Tell the respondent what you are doing. 
 Repeat the answer as you record it. 
 Slow down the respondent if you get lost. Read aloud where you left off and let him/her take it 
from there. 
 Use standardized abbreviations.
What reading skills should your children have at the end of the school year? 
01 Read out grade level stories
02 Sound out words they don’t know
03 Understand stories that they read
04 Know letter names
05 Other_________________________
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 126 
2. Reserve Codes 
Reserve codes are included with every applicable question. These codes will be answer choices at the 
bottom of every page of your questionnaire for 
- “Don’t know” – the respondent does not know the answer to the question, after you have tried 
to probe 
- “Refused/No Answer”—the respondent does not want to answer the question, after you have 
tried to probe 
- These response types are not to be offered to the respondent, but only used if respondent gives 
them in return to a question! 
5.2.7 CHECKING FOR INTERNAL CONSISTENCY 
As an enumerator you must also use your common sense and periodically conduct consistency checks. It 
is your job to record the respondent’s answer exactly as he/she communicates them but not necessarily 
to blindly do so. While we have stressed the importance of remaining neutral, not estimating for the 
respondent and not arguing with the respondent, you must also remain alert in order to catch blatant 
errors or inconsistencies in the respondent’s answers.
Here is an example of an inconsistency: 
10) How many years have you been teaching overall? 
| 5 | YEARS INTERVIEWER: Enter a number from 0 to 99.
11) How many years have you been teaching as a trained teacher? 
| 7 | YEARS INTERVIEWER: Enter a number from 0 to 99.
It looks like the teacher has been teaching for 5 years overall, but then spent more time teaching as a 
trained teacher. This does not make sense, since you would expect the answer to #11 to be less than or 
equal to the answer to #10. 
The respondent’s answers raises questions in terms of logic and consistency. In these cases, what should 
you do as an enumerator? 
What to do in case of inconsistencies/illogical answers? 
Repeat the question 
The first thing you want to do is repeat the question to make sure that the respondent heard and 
understood the question correctly (this is probe type #2). 
Re-reading the answers to the respondent 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 127 
If repeating the question does not help, you may want to re-read the answers to the respondent to make 
sure that you understood the respondent’s answers correctly (only if the question allows the enumerator 
to read out answers). 
The conversation may go like this: 
- ENUMERATOR: “Let me make sure that I understood your answers correctly – you said that 
you have been teaching for 5 years? 
- RESPONDENT: “Yes, that’s correct”
- ENUMERATOR: “And you’ve been teaching as a trained teacher for 7 years?”
- RESPONDENT: at this point, the respondent may be re-thinking his answers and realizing that 
the answers do not make sense. If the respondent just says “yes it’s correct”, then record the 
answers as stated and make a note that the answers are inconsistent but that you did probe.
Your Supervisor will also review your questionnaire for internal consistency and completeness. He or she 
may come to you with questions if some answers appear illogical. 
RECAP: INTERVIEWER TECHNIQUES AND QUESTIONNAIRE EDITING
 Remain neutral, do NOT estimate for the respondent, and do NOT argue
 Avoid bias (style, verbal and nonverbal)
 Use probes when needed (there are 6 basic types of probes)
 Ask questions the way they are written
 Respect skip patterns
 Check for internal consistency
 Use reserve codes
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 128 
6. Field Procedures 
6.1 PREPARATION BEFORE GOING TO THE FIELD 
Before leaving to conduct interviews, make sure you have the following: 
 Your TKG identity card 
 Principal consent forms (enough for 1 per school plus extras as backup) 
 Letter of introduction from MOE and ethical approval from UL-PIRE 
 Your tablet—FULLY CHARGED! 
 A charger for the tablet 
 Sample list 
 One extra tablet per team 
 A backup copy of the questionnaires (1 P, 1 T, and 16 EGRAs) to be used in case technology fails 
 2 pens 
 1 pencil, 1 eraser, 1 sharpener, 2 clipboards 
 Tracking sheets (One tracking sheet per school) 
 Your instruction manual and EGRA reference documents 
 Cell phone, fully charged and switched to silence during interviews, with supervisor and field 
manager numbers programmed 
 Sufficient phone credit for the day 
 Bag to carry all materials 
Your Appearance 
Everybody on the survey team should be neatly dressed. You should not wear fancy clothing. You should 
be cheerful, polite and confident. Remember that your appearance can influence the success of an 
interview (style bias). 
6.2 WHOM TO INTERVIEW AND CONTACTING STRATEGY 
A total of 1,440 pupils and 90 teachers and 90 principals will be asked to interview for this project in 90 
schools. Interviews will be conducted on school grounds. In each school, there are 16 pupils, 1 Second 
Grade teacher, and 1 principal that we need to interview. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 129 
6.2.1 CONTACTING STRATEGY 
We have the names of school, GPS coordinates, and contact information for the principals in advance, 
which we gather during baseline preparations. You will be provided with this information prior to going 
to the field. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 130 
Step 1 
Call the school at least 2 days in advance to schedule a time to meet with them and to obtain consent 
for student survey and assessment. We will go through a detailed script later in the training for conducting 
this call. 
If no one answer the phone, you can leave the following message in the local language: 
Hi, my name is [YOUR NAME] from TKG. Your school has been selected to participate in a study aimed 
at understanding children’s reading performance and practices in Liberia. You may remember us contacting 
you earlier. I will be in your school on ___DATE___. Can I interview you then? What time? Can we meet 
at the school? 
When you have a response, record down the date and time. 
6.3 TABLETS 
Aside from information you will record on your tracking sheet, responses to questions will be recorded 
on tablets. Tablets are a great way to collect data—they help you do your job quickly and more accurately, 
instead of having to record everything on paper and pencil. However, tablets are expensive—TKG and 
NORC are trusting you to take good care of the tablets while you are out in the field, and to not abuse 
your access to these tablets. 
6.3.1 RULES OF TABLET USE 
Be very careful with the tablets 
 Protect tablets from dirt, sun, water, and rain 
o If it begins to rain during your interview, move under a roof immediately so the tablet 
does not get wet 
o This will also protect the tablet from the sun 
 Keep the tablet in a bag and out of sight to protect it from theft
 Never remove your tablet from its case 
 Dropping your tablet will damage it 
 If you notice any anomalies, report them to your supervisor immediately 
Supervisors are in charge 
 Supervisors are responsible for distributing and collecting tablets every day. 
 In the morning, the supervisor will give your tablet to you. 
 At the end of the day, return your tablet to your supervisor. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 131 
 Supervisors will transmit all data to central office and charge all tablets for the next day. 
Prohibited Uses 
 You are prohibited from using the tablets for anything other than this project. 
 For example, you cannot use your tablets to: 
 Text 
 Make personal phone calls 
 Use social media (Facebook, Twitter) 
 Check email 
 Surf the web 
 Play video games 
 Install other applications than the one already provided in the tablet 
 Take pictures with the tablet’s camera
6.3.2 TABLET SETTINGS 
Tablet settings are important for many reasons. Most importantly, proper settings help conserve battery 
power. Tablet settings should be monitored constantly by both enumerators and supervisors to ensure 
that the proper settings are always applied. Using improper settings will shorten the battery life of the 
tablet and impair your ability to complete your assignments. 
Display 
The display consumes the most power of any component: 
 Turn OFF automatic brightness
 Keep brightness as low as possible (brighter screens consume more power) 
 Conduct interviews in the shade so that you can set the screen to a lower brightness. 
 A roof can also provide shade, which will also protect the tablet from rain! 
GPS 
 Like the display, the GPS consumes a lot of power. To conserve battery power, always keep 
the GPS turned OFF whenever not in use.
 You may turn the GPS on/off by swiping down on the status bar. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 132 
 The GPS is only needed on the Supervisor’s tablet. It will be recorded on the Teacher and 
Principal interviews and those recordings of coordinates will provide us with the location of the 
school, which will be used to verify the team was in the correct school that day. 
WiFi 
 You do not need WiFi while in the field. 
 Proper WiFi setting – WiFi should remain OFF at all times. 
 Only supervisors are allowed to turn on WiFi when necessary to upload data (i.e. when there is 
no cell coverage) 
6.4 DATA SECURITY 
6.4.1 CONFIDENTIALITY 
Throughout the whole data collection period, you must do your best to protect the respondent’s 
confidentiality. This means that you cannot discuss respondent’s answers with anyone outside the team, 
and disclose the name of respondents whom you’ve interviewed to other respondents. This also means 
that you must handle the paper questionnaires very carefully. 
Do not discuss survey data with members of family nor friends. You must not permit any unauthorized 
person, not even a member of your family, to see a completed questionnaire. You must not solicit for, or 
accept assistance from non-survey personnel in your enumeration duties. 
6.4.2 TABLET SAFETY 
Keep all survey materials, especially documents, the tablet in safe and secure places when traveling. 
Preferably, keep them in a closed satchel when not in use to avoid unwanted attention. 
6.5 VALIDATION (DIRECT OBSERAVTIONS AND SPOT-CHECKS) 
In addition to the data quality control procedures which have already been described (consistency checks, 
reserve codes, disposition codes), the data review process will entail: 
 10% direct observation of interviews from start to finish by supervisors; 
 15% spot-checks of partial interviews by supervisors;
 10% direct observation of full interviews by Quality Control Officers (for EGRA/student, principal 
and teacher interviews);
 15% spot-checks of partial interviews by Quality Control Officers.
These validation checks serve 2 purposes: (1) ensure that the interview did take place, (2) verify the quality 
of the collected data. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 133 
Every day, the supervisor will randomly select at least 10% of all EGRA student assessments to observe 
from start to finish (ideally, each supervisor should directly observe two full interviews each day). The 
supervisor will follow along throughout the interview to make sure the enumerator is asking every 
question, following protocols around timing of tasks, coding correctly, and following all other rules and 
guidelines related to proper administration of the interviews. Each QC officer will also select two different 
interviews to observe within the school each day. These will follow the same protocols and the 
combination of Supervisor and QC Officer checks will result in at least 20% of all interviews being directly 
observed and evaluated throughout the sample. 
In addition to the direct observation of full interviews, both Supervisors and QC officers will also spot 
check at least 15% of interviews each day. For these, the Supervisor or QC officers can observe a few 
minutes of an interview to make sure protocols are being properly followed. These will ensure that every 
enumerator is observed, either through a full observation of their interview or for a portion of an 
interview during a spot check, at least once every day by their supervisor and/or their QC Officer. 
Each Quality Control Officer will be assigned two teams to monitor throughout the field period. 
Depending on the locations of their teams, we understand that in some cases the QC Officer may only 
be able to visit one of their teams in a given day. However, whenever it is logistically possible, the QC 
Officer should attempt to visit both of their teams each day. 
7. Our Questionnaires 
7.1 LANGUAGE 
All of the surveys will be conducted in English. 
7.2 CONCEPTS AND DEFINITIONS 
Acronyms 
For your reference, below is a list of acronyms that are commonly used throughout the questionnaire and 
this training manual: 
Acronym Definition
TKG The Khana Group
USAID United States Agency for International Development
7.3 QUESTIONNAIRE OVERVIEW 
The key topics of interest for the current survey are: child reading skills, the teacher’s attitudes and 
practices about teaching the child to read.
The current survey will provide data to aid in answering the question of whether the Read Liberia! 
Intervention leads to increases in reading outcomes among project beneficiaries. 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 134 
7.3.1 QUESTIONNAIRE NUMBER CODES 
On your sample sheet, NORC/TKG will provide sampled schools with ID numbers already assigned for 
all schools, you should use this code exactly as it is. Make sure to choose the proper school and code 
from the drop-down menu in the tablet and double-check that you have entered it in correctly prior to 
proceeding with the questionnaire. 
We will also provide a student ID, a teacher ID, and a principal ID: these will be auto-generated by the 
tablet. 
Here is a description of the different kinds of identifying information and corresponding codes that we 
will be using in our questionnaire: 
School ID: Each learner is associated with a school. Copy the school’s name and unique ID onto the 
Field Control Sheet exactly as it is. 
Student ID: Each learner has an ID generated in Tangerine. Copy this information down onto the Field 
Control Sheet exactly as it is displayed on the tablet. 
Teacher ID: Each learner is also associated with a teacher, who will also have an ID generated in 
Tangerine. Copy this information down onto the Field Control Sheet exactly as it is. 
Principal ID: Each learner is also associated with a principal, who l also have an ID generated in 
Tangerine. Copy this information down onto the Field Control Sheet exactly as it is. 
7.3.2 LOOK OUT FOR THESE COMMON MISTAKES 
Review these tips to help you avoid common mistakes and more accurately record responses. Make sure 
to double-check these points as you fill out the survey: 
 Pay attention to categorizing responses. Oftentimes, the codes you enter are field-coded, 
meaning you have to listen for the respondent’s answer, then refer to the questionnaire to select 
the appropriate response. When you do this, be careful to record the correct response. For 
Yes/No answers, this may be easier than more complex answers where we ask you to listen to 
the respondent and select the appropriate response. 
 Pay attention to units of measurement: If the respondent provides the answer “2” to the 
question “How long have you been a teacher?”, you should ask what unit they are referring to. 
Months? Years? PAY ATTENTION to what we are asking for—which, in this case, are the number 
of years, not months.
o In this case, confirm with the respondent: “Ok, so 6 hours is half a year” 
 Pay attention to timeframes. Often, we ask respondents “how often” they read, or their child 
reads. The options we give are: 
o FOUR DAYS A WEEK OR MORE 
NORC USAID/Liberia Impact Evaluation for the Read Liberia Activity
 ANNEX III: DATA COLLECTION INSTRUMENTS | 135 
o TWO OR THREE DAYS A WEEK 
o ONE DAY A WEEK 
o ONCE OR TWICE A MONTH
o LESS THAN ONCE A MONTH
o NEVER 
If a teacher says “every Thursday”, then that means they read one day a week. Make sure you 
understand whether or not the teacher is referring to days, weeks, or months. Repeat back the 
response to the respondent as you categorize the responses.
 If it does not seem right, probe! Trust your instinct—if a response does not make sense follow￾up with respondents. 
ANNEX IV: SOURCES OF INFORMATION | 136
ANNEX IV: SOURCES OF INFORMATION 
FIELD IMPLEMENTATION SCHEDULE 
County # schools Position Name Movement Pattern (Districts)
Montserrado
Margibi
Grand Bassa 28 QCO Decontee Peters
Bong
Lofa 36 QCO Anita S.K. Marshall
Nimba 26 QCO J.Momolu Paasewe
Supervisor King Dennis Wolokoluwuo
Assessor A Wilhemina Y. Harris
Assessor B Tiangay J. Williams
Assessor C Samuel Paye
Supervisor Allison M. Paye
Assessor A Edwin W. Deshield
Assessor B Georgina C. Davis
Assessor C Henry W. Scott
Supervisor William N. Kpoto
Assessor A Paul Sahr Johnson
Assessor B Mac Kornorboi
Assessor C Romeal P. Burton
Supervisor Arthurline S. Joloe
Assessor A Momsieh.B. Mallet
Assessor B Dweh N.Nyenmoh 
Assessor C Edwin M Bonar
Supervisor Jokua Gbassie
Assessor A Joseph N. Davis
Assessor B Emmanuel J.King
Assessor C Priscillia D. Chae
Supervisor Benjamin Y. Everett
Assessor A Samuel O . Moore
Assessor B Clarence F. Wordsworth
Assessor C Prince O. Gboneh
Read! Liberia EGRA Baseline Team Schedule
Suacoco, Kokoyah, Panta Kpai, Zota, Sanoyea, Salala, Fuamah
District 1, District 2, District 3, District 4, District 5, 
kakata, Gibi, Marshal
Salaye, Zorzor, Kolahun, Vahun, Foya, 
Greater Monrovia 1, Greater Monrovia 2, MCSS, 
Todee, Caresburg, Right Bank, Left bank
Lofa 19
Montserrado 13
Bong 17
Grand Bassa & 
Margibi
15 
(8 Grand 
Bassa & 7 
Margibi)
Saclepae 1, Saclepae 2, Yarwin-Mensonah, Buu-Yao, Yarpea-Mah
Bain Garr, Sanniquelle-Mah, Gbehle-geh, Tuah river, Zoe-Geh
Nimba 2 13
Nimba 1 13
ANNEX V: DISCLOSURE OF ANY CONFLICT OF INTEREST | 137
ANNEX V: DISCLOSURE OF ANY CONFLICTS OF INTEREST 
Name Alicia Menendez
Title VP International Projects
Organization NORC at the University of Chicago
Evaluation Position? X Team Leader Team member
Evaluation Award Number (contract or other 
instrument)
AID-OAA-M-13-00010 – Impact Evaluation of
READ Liberia Project
USAID Project(s) Evaluated (Include project 
name(s), implementer name(s) and award number(s), if 
applicable)
USAID/Liberia Read Liberia Impact Evaluation
I have real or potential conflicts of interest 
to disclose.
 Yes X No
If yes answered above, I disclose the 
following facts:
Real or potential conflicts of interest may include, but are not limited 
to:
1. Close family member who is an employee of the USAID operating 
unit managing the project(s) being evaluated or the implementing 
organization(s) whose project(s) are being evaluated.
2. Financial interest that is direct, or is significant though indirect, in 
the implementing organization(s) whose projects are being 
evaluated or in the outcome of the evaluation.
3. Current or previous direct or significant though indirect experience 
with the project(s) being evaluated, including involvement in the 
project design or previous iterations of the project.
4. Current or previous work experience or seeking employment with 
the USAID operating unit managing the evaluation or the 
implementing organization(s) whose project(s) are being 
evaluated.
5. Current or previous work experience with an organization that 
may be seen as an industry competitor with the implementing 
organization(s) whose project(s) are being evaluated.
6. Preconceived ideas toward individuals, groups, organizations, or 
objectives of the particular projects and organizations being 
evaluated that could bias the evaluation. 
I certify (1) that I have completed this disclosure form fully and to the best of my ability and (2) that I will update 
this disclosure form promptly if relevant circumstances change. If I gain access to proprietary information of other 
companies, then I agree to protect their information from unauthorized use or disclosure for as long as it remains 
proprietary and refrain from using the information for any purpose other than that for which it was furnished. 
Signature
Date 8/16/2017
X
X
ANNEX V: DISCLOSURE OF ANY CONFLICT OF INTEREST | 138
Name Varuni Dayaratna 
Title VP International Projects 
Organization NORC at the University of Chicago 
Evaluation Position? Team Leader X Team member 
Evaluation Award Number AID-OAA-M-13-00010 - Impact Evaluation 
(contract or other instrument) of READ Liberia Project 
USAID Project(s) Evaluated USAID/Liberia Read Liberia Impact 
(Include project name(s), implementer name(s) and award Evaluation 
number(s), if applicable)
I have real or potential conflicts of interest to Yes X No
disclose. 
If yes answered above, I disclose the following 
facts: 
Real or potential conflicts of interest may include, but are not limited to: 
1. Close family member who is an employee of the USAID operating unit 
managing the project(s) being evaluated or the implementing organization(s) 
whose project(s) are being evaluated. 
2. Financial interest that is direct, or is significant though indirect, in the 
implementing organization(s) whose projects are being evaluated or in the 
outcome of the evaluation. 
3. Current or previous direct or significant though indirect experience with the 
project(s) being evaluated, including involvement in the project design or 
previous iterations of the project. 
4. Current or previous work experience or seeking employment with the USAID 
operating unit managing the evaluation or the implementing organization(s) 
whose project(s) are being evaluated. 
5. Current or previous work experience with an organization that may be seen 
as an industry competitor with the implementing organization(s) whose 
project(s) are being evaluated. 
6. Preconceived ideas toward individuals, groups, organizations, or objectives of 
the particular projects and organizations being evaluated that could bias the 
evaluation. 
I certify (1) that I have completed this disclosure form fully and to the best of my ability and (2) that I 
will update this disclosure form promptly if relevant circumstances change. If I gain access to proprietary 
information of other companies, then I agree to protect their information from unauthorized use or 
disclosure for as long as it remains proprietary and refrain from using the information for any purpose 
other than that for which it was furnished. 
Signature 
Date 8/16/2017 
ANNEX V: DISCLOSURE OF ANY CONFLICT OF INTEREST | 139
Name Sebastian Monroy-Taborda 
Title Senior Research Analyst 
Organization NORC at the University of Chicago 
Evaluation Position? Team Leader X Team member
Evaluation Award Number AID-OAA-M-13-00010 - Impact Evaluation of 
(contract or other instrument) READ Liberia Project 
USAID Project(s) Evaluated USAID/Liberia Read Liberia Impact Evaluation 
(Include project name(s), implementer name(s) and 
award number(s), if applicable)
I have real or potential conflicts of interest to Yes X No 
disclose. 
If yes answered above, I disclose the following 
facts: 
Real or potential conflicts of interest may include, but are not limited to: 
1. Close family member who is an employee of the USAID operating unit 
managing the project(s) being evaluated or the implementing 
organization(s) whose project(s) are being evaluated. 
2. Financial interest that is direct, or is significant though indirect, in the 
implementing organization(s) whose projects are being evaluated or in 
the outcome of the evaluation. 
3. Current or previous direct or significant though indirect experience with 
the project(s) being evaluated, including involvement in the project design 
or previous iterations of the project. 
4. Current or previous work experience or seeking employment with the 
USAID operating unit managing the evaluation or the implementing 
organization(s) whose project(s) are being evaluated. 
5. Current or previous work experience with an organization that may be 
seen as an industry competitor with the implementing organization(s) 
whose project(s) are being evaluated. 
6. Preconceived ideas toward individuals, groups, organizations, or objectives 
of the particular projects and organizations being evaluated that could 
bias the evaluation. 
I certify (1) that I have completed this disclosure form fully and to the best of my ability and (2) that I 
will update this disclosure form promptly if relevant circumstances change. If I gain access to proprietary 
information of other companies, then I agree to protect their information from unauthorized use or 
disclosure for as long as it remains proprietary and refrain from using the information for any purpose 
other than that for which it was furnished. 
Signature 
Date 8/16/2017 
ANNEX V: DISCLOSURE OF ANY CONFLICT OF INTEREST | 140
Name Brian Kirchhoff
Title Research Scientist
Organization NORC at the University of Chicago
Evaluation Position? Team Leader X Team member
Evaluation Award Number (contract or other 
instrument)
AID-OAA-M-13-00010 – Impact Evaluation of 
READ Liberia Project
USAID Project(s) Evaluated (Include project name(s), 
implementer name(s) and award number(s), if applicable)
USAID/Liberia Read Liberia Impact 
Evaluation
I have real or potential conflicts of interest to 
disclose.
 Yes X No
If yes answered above, I disclose the following 
facts:
Real or potential conflicts of interest may include, but are not limited to:
7. Close family member who is an employee of the USAID operating unit 
managing the project(s) being evaluated or the implementing 
organization(s) whose project(s) are being evaluated.
8. Financial interest that is direct, or is significant though indirect, in the 
implementing organization(s) whose projects are being evaluated or in the 
outcome of the evaluation.
9. Current or previous direct or significant though indirect experience with the 
project(s) being evaluated, including involvement in the project design or 
previous iterations of the project.
10. Current or previous work experience or seeking employment with 
the USAID operating unit managing the evaluation or the implementing 
organization(s) whose project(s) are being evaluated.
11. Current or previous work experience with an organization that 
may be seen as an industry competitor with the implementing 
organization(s) whose project(s) are being evaluated.
12. Preconceived ideas toward individuals, groups, organizations, or 
objectives of the particular projects and organizations being evaluated that 
could bias the evaluation. 
I certify (1) that I have completed this disclosure form fully and to the best of my ability and (2) that I will update 
this disclosure form promptly if relevant circumstances change. If I gain access to proprietary information of other 
companies, then I agree to protect their information from unauthorized use or disclosure for as long as it remains 
proprietary and refrain from using the information for any purpose other than that for which it was furnished. 
Signature
Date 8/16/2017
X
ANNEX VI: FULL REGRESSION RESULTS| 141
ANNEX VI: FULL REGRESSION RESULTS
VARIABLES
Correct 
Familiar 
Word per 
min.
Oral 
Reading 
Fluency 
Reading 
Comp.
Listening 
Comp.
Girls -2.647*** -2.891*** -0.123 -0.025
(0.666) (1.091) (0.090) (0.060)
Rural -3.929*** -6.639*** -0.426*** -0.209
(1.316) (2.287) (0.116) (0.130)
Student Age 1.440 0.849 -0.240 0.027
(1.327) (2.012) (0.212) (0.129)
Student Age Squared -0.049 -0.030 0.009 -0.002
(0.050) (0.076) (0.008) (0.005)
Do you practice reading aloud to someone at home? 3.793*** 5.992*** 0.247** 0.151**
(0.836) (1.285) (0.113) (0.074)
Does anyone read aloud to you at home? -1.400** -0.898 -0.092 -0.099
(0.692) (1.235) (0.094) (0.069)
Language Home: English 0.561 1.421 0.159 0.190***
(0.845) (1.339) (0.108) (0.071)
Parents cannot read/write -0.752 -1.695 -0.104 -0.143*
(0.884) (1.360) (0.125) (0.079)
Do you watch television at home? -0.124 -0.519 -0.177 -0.112
(1.066) (1.794) (0.123) (0.084)
Do you listen to radio at home? -0.309 -1.348 -0.166 0.039
(0.811) (1.456) (0.119) (0.079)
Do you have electricity/current at home? 0.357 1.262 -0.071 0.038
(1.029) (1.713) (0.106) (0.089)
Did you eat before coming to school today? -0.209 -0.806 -0.059 0.073
(0.663) (0.998) (0.089) (0.065)
Female Teacher 1.064 0.230 -0.038 0.005
(1.351) (2.406) (0.135) (0.119)
Volunteer Teacher 1.023 -0.037 0.165 -0.059
(1.404) (2.213) (0.138) (0.162)
Teacher's years of experience 0.346* 0.589** 0.025 0.000
(0.182) (0.294) (0.023) (0.024)
Teacher's years of experience squared -0.007 -0.015* -0.001 0.000
(0.005) (0.008) (0.001) (0.001)
Teacher received training teach reading 0.588 0.303 -0.035 -0.189
(1.260) (2.109) (0.115) (0.120)
Teacher has lesson plan available -0.094 -1.558 -0.028 -0.072
(1.060) (1.666) (0.106) (0.114)
Teacher has High School Degree 0.302 -1.766 0.233 -0.433**
(1.948) (2.358) (0.177) (0.206)
Teacher has Certificate Degree -1.056 -3.758 0.152 -0.335
ANNEX VI: FULL REGRESSION RESULTS| 142
Correct 
Familiar Oral 
Word per Reading Reading Listening 
VARIABLES min. Fluency Comp. Comp.
(1.956) (2.356) (0.164) (0.206)
Teacher has BA Degree -3.987 -8.875** -0.542** -0.330
(2.798) (3.615) (0.244) (0.233)
Sufficient materials/ textbooks for Grade 2? 2.788 4.049 0.097 0.143
(2.171) (3.425) (0.194) (0.235)
Do you have a library or reading room? 1.273 3.316 0.243* 0.104
(1.244) (2.057) (0.123) (0.134)
Constant -0.013 12.311 2.446* 2.205**
(9.167) (13.998) (1.291) (0.842)
Observations 1,418 1,418 713 1,345
R-squared 0.082 0.086 0.070 0.049
OLS regressions. Standard errors are clustered at School level, and are presented in parentheses. *** p<0.01, ** 
p<0.05, * p<0.1
VARIABLES
Correct letter 
names per 
min.
Correct 
NonWord 
per min.
Orienta
tion to 
Print
Phonemic 
Aware
ness
Girls -4.552*** -1.215*** -0.116** -0.221*
(1.370) (0.203) (0.054) (0.118)
Rural -9.995*** -0.714* -0.143 -0.349
(3.337) (0.375) (0.163) (0.256)
Student Age 9.351*** 0.123 0.295** 0.595**
(2.188) (0.471) (0.130) (0.254)
Student Age Squared -0.313*** -0.006 -0.011** -0.021**
(0.086) (0.017) (0.005) (0.010)
Do you practice reading aloud to someone at home? 7.635*** 1.083*** 0.016 0.341**
(1.575) (0.207) (0.068) (0.136)
Does anyone read aloud to you at home? 0.816 -0.188 -0.058 -0.137
(1.522) (0.251) (0.071) (0.143)
Language Home: English 0.772 0.241 0.119 0.290**
(1.529) (0.300) (0.077) (0.141)
Parents cannot read/write -2.918 0.015 0.047 0.053
(1.821) (0.332) (0.081) (0.152)
Do you watch television at home? 1.903 0.247 -0.112 -0.138
(1.644) (0.372) (0.090) (0.187)
Do you listen to radio at home? -1.749 -0.015 -0.084 0.190
(1.429) (0.303) (0.064) (0.175)
Do you have electricity/current at home? -1.073 -0.050 0.026 -0.153
ANNEX VI: FULL REGRESSION RESULTS| 143
VARIABLES
Correct letter 
names per 
min.
Correct 
NonWord 
per min.
Orienta
tion to 
Print
Phonemic 
Aware
ness
(1.883) (0.314) (0.091) (0.177)
Did you eat before coming to school today? 1.520 0.205 0.004 -0.058
(1.315) (0.219) (0.065) (0.143)
Female Teacher 2.055 -0.139 -0.083 0.103
(2.657) (0.399) (0.143) (0.249)
Volunteer Teacher 5.477* 0.350 0.252 0.026
(3.210) (0.451) (0.170) (0.315)
Teacher's years of experience 0.666 0.069 -0.031 -0.021
(0.439) (0.053) (0.025) (0.040)
Teacher's years of experience Squared -0.013 -0.002* 0.001 0.001
(0.011) (0.001) (0.001) (0.001)
Teacher received training teach reading 2.897 0.795** 0.082 0.260
(2.412) (0.369) (0.137) (0.218)
Teacher has lesson plan available -1.857 -0.392 -0.068 0.278
(2.376) (0.308) (0.116) (0.203)
Teacher has High School Degree -2.931 1.178*** 0.066 -0.302
(3.234) (0.421) (0.170) (0.433)
Teacher has Certificate Degree -0.640 0.351 0.285* -0.205
(2.686) (0.430) (0.153) (0.404)
Teacher has BA Degree -2.446 -0.854 -0.179 -0.166
(5.421) (0.658) (0.182) (0.534)
Sufficient materials/ textbooks for Grade 2 6.336 -0.055 0.257 -0.187
(4.593) (0.425) (0.208) (0.230)
Do you have a library or reading room? 2.131 0.136 -0.039 0.342
(2.667) (0.359) (0.124) (0.216)
Constant -0.505 -0.175 0.491 -0.083
(14.932) (3.183) (0.949) (1.811)
Observations 1,415 1,418 1,396 1,405
R-squared 0.116 0.074 0.059 0.031
OLS regressions. Standard errors are clustered at School level, and are presented in parentheses. *** p<0.01, ** p<0.05, * 
p<0.1
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 144
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS
0
0.1
0.2
0.3
0.4
0.5
0.6
0 1-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80
Correct Words per Minute
All
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Oral Reading Fluency
All
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 145
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 1 2 3 4 5
Reading Comprehension
All
0
0.05
0.1
0.15
0.2
Correct Letters per Minute
All
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 146
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1-10 11-20 21-30 31-40
Correct non-Words per Minute
All
0
0.1
0.2
0.3
0.4
0.5
0.6
0 1 2 3
Orientation to Print
All
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 147
0
0.05
0.1
0.15
0.2
0.25
0 1 2 3 4 5 6 7 8 9 10
Phonemic Awareness
All
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0 1 2 3
Listening Comprehension
All
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 148
0
0.1
0.2
0.3
0.4
0.5
0.6
0 1-10 11-20 21-30 31-40 41-50
Familiar Word
All
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 1-10 11-20 21-30 31-40 41-50 51-60
Oral Reading
All
ANNEX VII: EGRA SUBTASKS DISTRBUTION GRAPHS | 149
0
0.05
0.1
0.15
0.2
0 1-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100
Letter recognition
All
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1-10 11-20 21-30 31-40
Invented Word
All
150
U.S. Agency for International Development 
1300 Pennsylvania Avenue, NW 
Washington, DC 20523