You are on page 1of 17

Improving Educational Outcomes

Through School Feeding in Urban

Rwanda

Monitoring and Evaluation Plan


Group 14

Kyla Cross, Nicole Flynn, Anna Jones

Monitoring and Evaluation Plan School Feeding In Rwanda 2

I. Introduction
Hunger is a worldwide health problem that is devastatingly rampant in low-middle
income countries (LMICs). The children within these countries are at an evident disadvantage,
and their health can be affected lifelong. Most immediately, hunger contributes to negative
school performance throughout childhood. According to the No Kid Hungry Initiative (2015) in
the United States, some of the impacts on school performance due to hunger include: inability to
concentrate, behavioral problems, lack of energy and motivation, and overall poor school
performance. The children as well as their instructors can feel the consequences of hunger in the
classroom every day. Broadly, hunger in the classroom can leave a country lacking in
educational diversity in its adult population, further propagating countrywide poverty and
continuing the hunger cycle (Glewwe & Jacoby, 1995). Other outcomes of childhood hunger can
be seen in Figure 1 and include poor school attendance, delayed school enrollment, and poor
health status that inevitably leaves children vulnerable to poor outcomes in the future.
One of the major determinants of childhood hunger in the classroom is poverty. With less
money, there are fewer resources to purchase food or grow food, leaving children feeling hungry
at school or simply not enrolling to begin with (Todd & Winters, 2011). Appendix A Figure 1
depicts a problem tree that shows many of the underlying determinants of childhood hunger, as
well as immediate and broad consequences. Additional causes of childhood hunger in the
classroom are related to agricultural ability in the region, lack of investment by the government,
distance of homes to food purveyors, and parental education about child feeding.
There are currently numerous programs used to address the problem of hunger in schools,
most notably in the United States. Though the demographics in the United States do not nearly
match those in LMICs, it still provides a good baseline for those struggling with hunger in the
classroom. The No Kid Hungry program focuses on providing breakfast to kids at school, with
the goal to reduce the stigma on hunger and provide adequate nourishment for school-day focus
(Hunger in our Schools, 2015). Summarized below are a few of the effects the breakfast program
had on students in the United States as of 2015.

On a global scale, the World Banks Global Food Crisis Response Program is using school
feeding programs to encourage school attendance (especially among girls) and as a social safety
net for emergency crises that may occur in the future. (The World Bank, 2012). Notably, Kenya
instituted the Home Grown School Meals program to feed school children as well as support
local farmers while keeping sustainability in mind. Most international school feeding programs
are aimed at eradicating extreme poverty and hunger as well as achievement of universal primary
education, the first two Millennium Development Goals proposed by the United Nations.
Some of the questions pertaining to how to solve this problem that remain unanswered
include: sustainability of school feeding programs, cross-cultural translation, and long-term
effects. In the 1960s the Chilean government began a school-feeding program, and it has been
continuously run since then (McEwan, 2013). Its effectiveness and sustainability have been
studied with positive results, but the question remains how to implement such a program in
different regions of the world like East Asia and Africa. Though the Chilean GDP has tripled
Monitoring and Evaluation Plan School Feeding In Rwanda 3

since the beginning of the school feeding program, it is hard to quantify alternative long-term
benefits of those types of interventions (McEwan, 2013). Ideally, a program could be developed
that is sustainable as well as replicative to best reduce hunger and improve childhood educational
outcomes.

II. Program Overview

Initial implementation of the program will require establishing a rapport with the
surrounding community of Mukamira in Rwanda, including governmental officials, school
administrators, and local farmers. During the first three months, six primary schools will be
recruited to participate in the intervention based on low socio-economic status and excellence of
English language curriculums. Those with an already established school lunch program will be
omitted from participation. In addition, local farmers will be recruited to supply each school with
locally grown produce to be incorporated into the school meals. Sourced foods will mimic what
is culturally relevant to Rwanda, including beans, sweet potatoes, and plantains. Farmers will be
given the initial seeds and supplies to grow the necessary crops. Furthermore, residents in
Mukamira will be hired and trained to act as health coordinators, implementing the various
testing, biweekly check-ins, and nutritional assessments throughout the intervention. We will
also hire community members to collect the produce and animal products from participating
farmers and to prepare the lunches daily. A central location will be purchased or built to act as a
kitchen in which all the meals will be prepped and cooked. The meals will then be delivered to
the participating schools by the health coordinators.
During the fourth month, the initial standardized English tests will be given to create a
baseline to measure any cognitive changes due to implementation of the program. Sample
questions from a first grade English test can be found in Appendix B. In Rwanda, students are
initially taught using their native language, Kinyarwanda. During their fourth year of primary
school, they begin to learn how to speak English (Global Partnership, 2016). This initial test is to
see how much English the students may know prior to their fourth year of primary school due to
exposures in their community and household.
Once this initial English test has been given and the results collected, the intervention
will begin. Students at the first three schools will be given lunch every day during the school
week. Lunches will be prepared by hired chefs and distributed to the students. Lunches will
consist of a combination of various nutrient dense foods including various vegetable and meat
stews served with ugali, a maize porridge. Meals may also consist of rice, bananas, chapatti,
potatoes, and various local fruits such as mangoes and pineapples (Development Plan, 2013).
Our control group will consist of three additional schools similar in socio-economic status and
culture as our exposure group. They will not receive meals until the halfway point of the school
year, around the tenth month of the intervention. A mid-year test will be given to Group A and B
to establish a comparison for the introduction of school lunches in Group B relative to post-tests
given at the end of the school year. Improvements should be seen in test scores for the second
half of the year compared to the first half of the year in Group B.
Attendance will be taken every day by the teachers for both our exposure and control
schools. In addition, a daily count will be taken of each meal handed out to a participating
student by our health coordinators. At the end of the school year, the same pre-test will be given
as a post-test to determine how the students English speaking skills have advanced over the
course of the intervention. Students will be given a survey of how they viewed their academic
Monitoring and Evaluation Plan School Feeding In Rwanda 4

achievements for the year and their opinion of the school lunch program. In addition, teachers
will be given a survey to determine their opinion of the school lunch program and how their
students preformed overall. If a correlation is made between the consumption of school lunch
and test scores, steps will be made to make this program wide spread throughout the district of
Nyabihu.
Nyabihu is a district in the Western province of Rwanda. Although Rwandas
government is making great strides to improve its education system, some districts are quite a bit
behind the performance of others. This intervention targets the primary schools of the city
Mukamira in the Nyabihu district. This district was selected due to it having one of the higher
dropout rates from primary school in 2015 at 9.4%. It also has one the highest grade repetition
rates at 24.5% (Education Statistical Yearbook, 2016). In addition, Nyabihu has some of the
most fertile soil in Rwanda due to the volcanic content. As a result, approximately 74% of the
population is employed in agriculture, mostly in the form of farming. Cash crops make up most
what is grown in the district including crops such as Irish potatoes, maize, and beans
(Development Plan, 2013). The introduction of the intervention could subsequently diversify the
crops grown by farmers in this area.
One of the main problems identified earlier is the question of sustainability of this type of
intervention. Community involvement is a major factor to the success of the intervention to last
long-term. The Rwandan government is set to allocate 22% of the national budget to the
education sector in the 2017/2018 fiscal year (Global Partnership, 2016). If this intervention is
successful, a portion of these funds can be used to fund the intervention annually throughout the
country. In addition, part of these funds will be redirected into the agriculture sector through the
sourcing of farmers to supply the needed produce. Farmers can then use this revenue to further
fund future crops for this program.
The main objective of this intervention is to determine if the implementation of a daily
school lunch program will improve the cognitive ability of students. An additional objective is to
improve the enrollment and attendance rate of students through the incentive of food security.
From these objectives, the short-term outcomes of this intervention will include improved
attendance rates, improved test scores, and diversification of crops cultivated by farmers. If
impoverished students are presented with a reliable source of food (through the daily lunches),
they are more likely to attend school to consume the meals. Likewise, the lunches will satisfy
any previous feelings of hunger, allowing them to focus more on the information presented in
class. This will allow them to learn more material, giving them the ability to score better on tests.
Foods that will be included in the meals will be rich in nutrients that are essential to brain
development and function, such as iron. This will improve the students ability to retain and
comprehend new educational material. Lastly, by sourcing farmers to grow specific crops and by
giving them the needed tools and seeds, diversification of their crops beyond cash crops should
result.
Although this intervention is not designed to initially be longitudinal, proposed long-term
outcomes include an increase in the percent of students that go on to achieve a higher education,
a diversification of the economy, and an increase in Rwandas GDP. The hope is that if the
intervention were successful, it would be implemented in both primary and secondary schools,
increasing the number of students completing a higher education. This would result in a
diversification of the economy, because students would have the skills needed to seek
employment in fields other than agriculture. Currently 74% of the population of Nyabihu is
employed in the agriculture sector. However, there is a limited amount of land available to be
Monitoring and Evaluation Plan School Feeding In Rwanda 5

farmed, reducing the number of hours individuals can work to approximately 5 a day
(Development Plan, 2013). Thus, this limits the earning potential of farmers. By employing
educated students in fields other than agriculture, the potential to grow Rwandas GDP increases
significantly.
Potential stakeholders for this intervention would include: the local and countrywide
government, farmers, teachers, the surrounding community, and funders of the program. The
government would have the highest stake in this intervention. It is currently allocating 22% of its
annual budget to education. This intervention could help insure the funds are well spent. In
addition, this intervention is parallel with the countrys interest in improving enrollment and
graduation rates, as well as improving test scores. If the long-term outcome of increasing the
countrys GDP were to be achieved, they would have additional funds to use elsewhere. The
government would also have the power to help fund this intervention, especially if it is to
eventually be used throughout all of the districts. Farmers would also have a high stake in the
success of this program. They are initially given the seeds and tools to cultivate the land, both of
which can be expensive. They would also have a reliable source of funds being exchanged for
their labor, as this program would pay for the food it receives from the farmers. Due to these
reasons, it would be in the best interest of the surrounding farmers to participate in the program.
However, they have the power to refuse participation in this program. The amount of cultivatable
land is limited due to surrounding mountains and thus limits what can be used to grow crops to
be used in the production of the school lunches. If they refuse to participate, this could
significantly hinder the success of this intervention.

III. Program Theory


The lunch program intervention is illustrated in the program impact pathway (Appendix
C). The program is broken up into pre (red) and post (blue) lunch program steps, which together
are necessary to achieve the ultimate end goal: increasing educational outcomes.

Pre-Lunch Program Implementation


The first component of the intervention begins with a focus on building relationships with
the community in order to gain strong support from farmers, schools and community members.
The success of the program depends on cooperation from all entities in the community. The
recruitment of schools to participate in the program and community members to aid in
foodservice are two of the main components to achieve successful implementation of the
program. Also, as a part of recruitment, sourcing farmers for the food for the lunch program is of
utmost importance because it will make food more affordable, nutrient rich, and ethnically
pleasing to students. There is a significant amount of research that has demonstrated the
importance of farm to school programs and its impact on students nutritional and educational
outcomes (Joshi, Azuma, Feestra, 2008).
Once recruitment is completed, training the teachers and food staff members on both
testing protocol and food safety measures is crucial prior to the program implementation.
Specifically training teachers will help make sure there is a standard method of test taking, so no
school or group of students has an advantage. In addition, once teachers are properly trained,
initial cognitive and English tests will be given to establish the baseline of students in each of the
respective schools. These baseline tests are the first of three given at the start of fourth grade,
which is when the English curriculum begins. Once baseline is established, the lunch program is
Monitoring and Evaluation Plan School Feeding In Rwanda 6

set to begin. The selection of schools and communities, along with baseline tests are the essence
of a successful program intervention (Belot and James, 2011).
There are a number of mitigating factors that have potential to interact with the
intervention pre-lunch program start. Amongst these are schools that already have trained food
service individuals who are willing to participate in the program. Also, if farmers have a good
relationship with the community and schools already, their willingness to participate will be
additive to the intervention. Despite the best efforts to counteract any negative factors, some out
of the researchers controls include crop season and weather, which can potentially inhibit
farmers participation in the farm to school lunch program. In addition, if there are biases in
teachers or schools in the program, problems or unforeseen circumstances, this can potentially
inhibit the progress of the pre-lunch program pathway.

Post Lunch Program Implementation

The benefit of school lunch programs on educational, growth, behavioral and cognitive
outcomes can be seen in many different countries and is discussed in the World Food
Programme (Bundy et al. 2009). After the successful implementation of the lunch program, the
next step is the consumption of the lunch by students. After the food at lunch has been consumed
for any amount time, the beneficial effects will lead to improved attendance and focus; students
want to attend school in order to receive the value of a nutritious lunch. Also, focus and attention
may also occur due to the fact that students who receive lunch are not concerned with other
factors such as hunger, and are more likely to turn their focus to school work. Students also have
the ability to find more time to study and learn the material, because they are not focused on
finding food throughout the day. After the program is completed, the immediate educational
outcomes can include higher scores on the English and cognitive tests. Long term benefits of the
increased educational outcomes can include a higher socioeconomic status, better career
outcomes, reduced health problems, reduced social stigmas including arrests, drug and alcohol
use, obesity and other diseases (Hauser 2004, Reynolds 2001, Ralston et al., 2008).
The ability of schools to participate in the program is essential. If the school agrees to the
program, then has poor communication and follow up with the research team, farmers or
volunteers, the program will suffer. The reception and adherence of the lunch program to
students is a crucial step in the program impact pathway. Students have the potential to control
the outcome of the program based on the positive or negative influence. For example, students
who enjoy the food and find it palatable will continue to reap the benefits, therefore having a
positive effect on the program. However, if students do not like the food or do not attend school
enough to achieve the same benefits, the program quality will suffer and have a negative impact.

IV. Evaluation Design

To best evaluate the effectiveness of our school-feeding program in Rwanda, we have


developed a pre-post quasi-experimental design with a non-randomized control group. Though
using a randomized controlled trial would provide the most statistical power, there are a number
of reasons to use a quasi design instead. For example, it would be unethical to randomize a
cluster of school children to receive school meals without providing the intervention to another
group of school children. Additionally, our school feeding program is aimed at system-level
changes through government sponsorship and sustainable community support as opposed to
Monitoring and Evaluation Plan School Feeding In Rwanda 7

individual level adjustments. We will, however, be able to randomize treatment groups to receive
the intervention at differing time points within the evaluation.
Our ultimate goal is to increase attendance and education outcomes among children in
LMICs, like Rwanda, by providing school meals and simultaneously fight childhood hunger. As
a sample for our evaluation, Donald Simeon (1998) explored school feeding in Jamaica through
two different study designs. The first study analyzed the effects of school feeding on
achievement, attendance, and physical growth and targeted the least scholastic students for
intervention. It was a longitudinal study that used baseline control in the first semester compared
with intervention in the second semester of the school year. A visual application of the results are
shown below, showing that school attendance declined for both school meal and control groups,
yet attendance in the intervention group declined less than the control group indicating potential
success of the program on school attendance (Simeon, 1999).

The second study analyzed the effects of hunger on cognition in children, comparing 3 groups of
students that had been either hospitalized for malnourishment, had stunted growth, or did not
have stunting. This study used a crossover design, allowing each child to be its own control
throughout the study by taking extensive baseline measurements. Both of the studies indicated
that school feeding likely results in improvements in school achievement, yet more research
would be needed to definitively establish causation.

In order for us to feel confident that the observed effects are in fact due to our
intervention, we will be using plausibility inference. Habicht (1999) describes the plausibility
assessment as having an effect above and beyond other external influences. Choosing an external
control group before our evaluation begins is an attempt to reach a more plausible conclusion.
Additionally, this evaluation uses a historical control group by comparing the schools before and
after the intervention. This is considered a staggered control approach, and can be a good
indicator for plausibility (Habicht, 1999). Some statements that we would hope to make at the
end of our evaluation to improve plausibility of our program include, but are not limited to (from
weakest to strongest):
Poor education outcomes fell rapidly in schools using our school feeding program
Poor education outcome did not fall in schools not using our school feeding program
There was an inverse association between intensity of the intervention in the schools and
education outcomes
The increase in school feeding coverage was compatible with the degree of improved
education outcomes (measure dose-response of the intervention through attendance and
meals eaten)
Monitoring and Evaluation Plan School Feeding In Rwanda 8

Confidence statements like those above would be useful in validating our program, as the
stronger statements help to rule out other reasons for educational success besides our school
feeding program.
Along the same lines, Levinson et al. (1999) explains that using a control group is
necessary in attributing educational outcome change to our intervention. As an example, in our
sample population, improvements in education outcomes are expected. We need a control group
to account for the maturation of our students through the school year in order for our feeding
program to be considered effective.

In choosing our sample for evaluation, we will use a clustered, randomized approach to
assess the primary effect of our intervention. Some considerations for sample size include:
number of groups being studied, the amount of change expected in the indicator, the level of
confidence needed for conclusions, and the probability of detecting a true difference (Levinson et
al., 1999).
The sample will be approximately 250 students in the control and 250 students in the
intervention group, using a total of 6 schools as a cluster. Our sample clusters will be determined
before the intervention based on a random drawing from each school involved in the program.
By using clustered samples, we are hoping to reduce measurement bias by avoiding
misclassification of cases and controls. Additionally, cluster sampling can reduce contamination
between study subjects, as the groups are geographically separated. One disadvantage of this
type of sampling includes the risk of within-group correlation, where the similarities of the
individuals within the cluster are interdependent of one another. The larger sampling frame for
the evaluation will be taken from a listing of elementary schools in Mukamira, where the
students have similar socioeconomic status, food insecurity, gender, and age.
To determine which schools would be eligible for the intervention and participate in our
evaluation, we set criteria to account for potential bias and confounding. For inclusion purposes,
we are looking for schools that have poor enrollment of primary school age, high levels of food
insecurity, and poor attendance of the children enrolled in the school. Those excluded from
participating in our intervention are schools that already have a feeding program for their
students, and schools in wealthier neighborhoods of Mukamira that have better food security. By
setting exclusion and inclusion criteria, our evaluation will be able to better adjust for
confounding factors and bias in our results.

In our evaluation, a very prominent source of confounding bias rests with the students
that do not attend school to receive the lunch. Those that refuse the intervention are more likely
to be different in some way, creating confounding within the evaluation (Habicht, 1999). One of
our best tools to help prevent against confounding is to begin the intervention with a wide array
of baseline measurements, so that we can control for those factors during statistical analysis.
Another key way to avoid confounding bias during our evaluation is to institute randomization,
which is why we are randomly selecting the schools that receive the lunch program first.
In addition to confounding bias, there is a potential for measurement bias in this
evaluation design. For example, reporting bias is a form of measurement bias that involves the
over reporting of an exposure, like hunger, to increase the chance of receiving the intervention.
This could be controlled for by randomizing which clusters receive the intervention first, so
over- or underreporting of the exposure would not impact intervention reception. Another form
of measurement bias is surveillance of those that may have a higher incidence of food insecurity
Monitoring and Evaluation Plan School Feeding In Rwanda 9

compared to the general population, making our data less externally valid. One way to prevent
both reporting and surveillance bias would be to construct questionnaires that would accurately
measure food insecurity and provide a generalizable study population.
Another potential source of bias in our evaluation comes from selection of our sample
from the children attending school. There is a possibility that there are children that arent
receiving our intervention due to some innate difference in those attending vs. absent from
school. For example, the poorest, most hungry children are most likely not attending school, and
they are the most in need of the intervention. When children fail to complete the program, loss to
follow up can occur which is also a source of selection bias. In order to prevent negative
outcomes from selection bias, we need to consider selection of controls carefully. Also, we
would want to encourage participation through the entire school year so that loss to follow up is
minimized and we can have the most accurate results.

V. Measurements
Construct to be Part of PIP Data Collection Frequency of Indicators
Measured Tool Measurement
Farmers Input Tally Once Number of
recruited to spreadsheet farmers that
cultivate crops (with contact commit to
info) growing specific
crops
English pretest Input Beginner level Once Average
English test previous English
known between
Groups A and B
Mid-term Output Beginner level Once Average written
assessment of English test English skills
English learned between
language skills Groups A and B
Mid-term Output Close-ended Once % of students
assessment of survey who enjoy the
meals provided lunches
Midterm Output Open-ended Once Qualitative
assessment of survey assessment of
meals lunches- what
aspects are
enjoyed or need
to be improved
Count of meals Output Tally Daily Average number
given to students spreadsheet of lunches given
(with name of to students per
enrolled student) school
Attendance Output Tally Daily % or enrolled
taken spreadsheet students that
(with name of attend each
student) school
Monitoring and Evaluation Plan School Feeding In Rwanda 10

End of term Outcome Close-ended Once % of students


assessment of survey who enjoyed the
meals provided lunches
End of term Outcome Close-ended Once % of students
assessment of survey that said they
education could focus
more or learned
more due to
reduced levels of
hunger
End of term Outcome Close-ended Once Average opinion
assessment for survey of student
teachers performance
between Group
A and B
Final English Impact Beginner level Once % of students
assessment English test who answered at
given least 75% of
questions
correctly

At the end of the intervention, students from both the control and exposure schools will
complete a close-ended survey as an end of term assessment of their academic achievements.
The students will be asked if they felt as though they could focus during class and if they learned
more this academic year compared to the previous year. The idea is that there should be a
difference in answers from the students between the exposure and controls schools. If students
are hungry, they are less likely to be able to focus on the material presented by the teachers. By
giving these students school lunch, their hunger should be sated, allowing the students to focus
on their English lessons rather than the physical pangs of hunger. If a difference is not seen, one
cannot conclude there is internal validity and is less likely to be true that our measurements will
be predictive of educational outcome.
In terms of measuring the impact of interest, a final first-level, standardized English test
will be given to students at the end of the academic year (Appendix B). The results of this test
will be compared to the baseline created at the beginning of the intervention in the form of a pre-
test. These results will also be compared to the tests given at the mid-point of the academic year
when the control schools began receiving school lunches. A difference should be seen between
the exposure and control schools. Group A (exposure) will have received school lunches from
day one of the study. This means the students in this group should have reduced hunger while at
school, allowing them to focus on the English lessons. Although Group B (control) received
lunches mid-year, their test scores are expected to be lower than that of Group A. If a difference
is not seen, one cannot claim that internal validity is present. Furthermore, schools were recruited
based on the distribution and characteristics of the source population. Therefore, generalization
of the results from the intervention to the source population should be possible.
Monitoring and Evaluation Plan School Feeding In Rwanda 11

Person to Activity Location of Information Forms Benchmark Management


Collect Activity to be Needed by
Information Collected Exception
Researchers Farmers Farms Coverage Tally At least Recruit
recruited to surrounding (number of spreadsheet/ three farms farms in
cultivate Mukamira farmers Year committed surrounding
crops (in Nyabihu committed contract to each cities (in
district) as % of total school Nyabihu);
available) ask why not
participating
Health Midterm Schools in Status (% of Close-ended At least Use open-
coordinators assessment Group A students survey 75% of ended
of meals who enjoy students surveys to
meal) enjoy meals determine
needed
adjustments
to lunch
preparation
Health Count of Schools in Coverage Tally 100% of Ask
coordinator meals Group A (% of meals spreadsheet students coordinator
provided (Group B provided (with name marked as why student
starting in compared to of each attending didnt
month 10) enrolled student) that day receive a
students) receive lunch;
lunch retrain
accordingly
Health Attendance Schools in Coverage Tally At least Ask families
coordinator of students Group A (% of spreadsheet 50% of of absent
taken and B students (with name students student why
present of of each enrolled absent via
those student) present phone call;
enrolled) address any
issues

There are four main areas of interests in which the intervention will be monitored. First,
the researchers of this study will perform the initial recruitment of the farmers to cultivate the
designated crops. The goal is to recruit at least 3 large farms per school from the area
surrounding Mukamira. If this benchmark is not met, researchers will recruit from rural areas
throughout the Nyabihu district. Second, the health coordinators in the exposure group will make
a midterm assessment of the meals. Students will complete both close and open-ended surveys
asking about their satisfaction of the lunches. If less than 75% of the students enjoy the meals,
adjustments will be made to the lunches based off answers from the open-ended surveys.
Health coordinators will also complete a daily count of the meals provided and the
number of students that attended school that day. Health coordinators will be provided with a
tally spreadsheet that includes the name of every student in each class. Next to the students
names will be two columns to record if the student was present and if they received a meal.
Every student that attends school should receive a meal. If the 100% benchmark is not met,
health coordinators will be asked a series of questions to determine if they need to be retrained
according to protocol.
Monitoring and Evaluation Plan School Feeding In Rwanda 12

VI. Timeline

VII. Data Usage


The information gathered from this intervention can have monumental impacts on school
lunch programs and their relationship to educational outcomes. The test scores of both the
English and cognitive tests will determine the baseline measurements for students pre-lunch
program as well as the students score post-lunch program. This type of qualitative data can be
easily compared among participating schools and students. Also, this intervention has the
potential to demonstrate the importance of the community due to the fact that the success of the
lunch program will rely on community unity and cooperation among members and volunteers for
food service, schools and their teachers as well as local farmers. Through the school lunch
program in Rwanda, it is hypothesized that there will be evidence-based outcomes that are
directly related to providing lunch and not providing lunch to certain schools based on standard
childhood nutritional guidelines. In the long term, this data can potentially be linked to country
economic growth, and family success in terms socioeconomic status based on test scores.
Lunch programs are used world-wide by many countries, but in low to middle income
countries, the need for lunch programs is still disputed. This program is designed specifically for
Rwandan schools, but it can be applicable and beneficial to many LMICs all over the world.
The results can be used to develop other lunch programs as well as other educational outcomes
can be tested. However, there is a strong need for more research on the benefits of school lunch
programs and educational outcomes, specifically for those in other grades, countries, and school
systems.

VIII. Challenges
Recruitment of schools, volunteers for food service, and farmer cooperation have the
potential to hinder the start of the program. Schools may not be interested in providing lunch
Monitoring and Evaluation Plan School Feeding In Rwanda 13

programs and farmers might not be willing to participate based on current community
involvement. Also, teachers might not have time in the curriculum to give the English and
cognitive tests on three separate occasions, which are required to evaluate the success of the
school lunch program. As with many other programs, bias has the potential to interfere with this
test taking process, and the success to which students learn certain materials, especially if there is
no uniform curriculum among schools. In addition, the analysis of the data (scores of tests) may
be time-consuming and may need to be done by investigators to avoid bias in the grading by
teachers of their students. Even after the implementation of the lunch program, adherence and
participation in the lunch program by students is essential in the hypothesis that the program will
aid in increasing educational outcomes.
In order to avoid bias, loss to follow up and lack of recruitment, it is essential to have
good rapport with each entity, be available for questions and concerns, check in weekly, and
provide motivating compensation for each group participating.

IX. References

Belot, M., & James, J. (2011). Healthy school meals and educational outcomes. Journal of
Health Economics, 30, 489504.
Bundy, D., Burbano, C., Grosh, M., Gelli, A., Jukes, M., & Drake, L. (2009). Rethinking school
feeding: Social safety nets, child development and the education sector. Washington, DC:
World Food Programme and the World Bank
Chapter 3 Designing a Monitoring System in: Levinson, F.J., Lorge Rogers, B., Hicks, K.M.,
Schaetzel, T., Troy, L., Young, C. Monitoring and Evaluation: A Guidebook for Nutrition
Project Managers in Developing Countries. Washington, DC: The World Bank, 1999.
Glewwe, P., Jacoby, H.G. (1995). An economic analysis of delayed primary school enrollment in
a low income country: The role of early childhood nutrition. The Review of Economics
and Statistics, 77(1), 156-169.
http://www.jstor.org/stable/2110001?seq=1#page_scan_tab_contents
Global Partnership for Education: Rwanda. (2016, December 29). Retrieved March 28, 2017
from http://www.globalpartnership.org/country/rwanda
Habicht, J.P., Victora, C.G., Vaughan, J.P. (1999). Evaluation designs for adequacy, plausibility
and probability of public health programme performance and impact. International
Journal of Epidemiology, 28, 10-18.
Hauser, Robert M. "Measuring socioeconomic status in studies of child development." Child
development (1994): 1541-1545.
Joshi, Anupama, Andrea Misako Azuma, and Gail Feenstra. "Do farm-to-school programs make
a difference? Findings and future research needs." Journal of Hunger & Environmental
Nutrition 3.2-3 (2008): 229-246.
McEwan, P.J. (2013). The impact of Chiles school feeding program on education outcomes.
Economics of Education Review, 32, 122-139.
file:///Users/nikkiflynn/Downloads/meals.pdf
No Kid Hungry. (2015). [Graph illustration of Hunger In Our Schools Report, 2015]. Hunger in
Our Schools, No Kid Hungry. Retrieved from http://hungerinourschools.org/img/NKH-
HungerInOurSchoolsReport-2015.pdf
Ralston, Katherine, et al. "The National School Lunch Program: Background, Trends, and Issues.
Economic Research Report Number 61." US Department of Agriculture (2008).
Monitoring and Evaluation Plan School Feeding In Rwanda 14

Republic of Rwanda Western Province Nyabihu District Development Plan. (2013). Retrieved
March 28, 2017 from
http://www.nyabihu.gov.rw/fileadmin/_migrated/content_uploads/DDP.pdf.
Republic of Rwanda 2015 Education Statistical Yearbook. (2016). Retrieved March 28, 2017
from
http://www.mineduc.gov.rw/fileadmin/user_upload/Amatangazo/2015%20Education%20
Staiatical%20YearbookF.pdf.
Reynolds, Arthur J., et al. "Long-term effects of an early childhood intervention on educational
achievement and juvenile arrest: A 15-year follow-up of low-income children in public
schools." Jama 285.18 (2001): 2339-2346.
Scaling up School Feeding: Keeping Children in School While Improving Their Learning and
Health. (2012). Retrieved March 26, 2017, from
http://www.worldbank.org/en/news/feature/2012/04/25/scaling-up-school-feeding-
keeping-children-in-school-while-improving-their-learning-and-health .
Simeon, D. (1998) School feeding in Jamaica: A review of its evaluation. American Journal of
Clinical Nutrition, 67(4), 790S-794S.
Todd, J.E., Winters, P. (2011). The effect of early interventions in health and nutrition on on-
time school enrollment: Evidence from the oportunidades program in rural Mexico.
Economic Development and Cultural Change, 59(3), 549-581.
http://www.jstor.org/stable/10.1086/658347?seq=1#page_scan_tab_contents
X. Appendices
Appendix A
Appendix B
From: Georgia CRT Online, 1st grade Language Arts Practice Test
Appendix C

You might also like