Educational Vitamin Tablets for Honduras

Even before COVID-19, Education in Honduras faced many challenges.  These include low actual classroom time, lack of textbooks, inadequately trained teachers, lack of educational technology support including low internet connectivity, and low educational completion rates beyond primary school.  School shutdowns have made the situation worse with even higher school dropouts.  

Our solution uses Kolibri on low-cost Android tablets to place digital textbooks, video lessons, and interactive quizzes plus a digital library of books directly in the hands of students at home without the need for internet.  Using innovative syncing technology, student work and quiz results completed at home are made visible to teachers during infrequent visits to schools.   

Using this information, teachers can assign tailored lessons to student needs, including remedial material not practical in a paper environment.

Based on a Creative Commons license, official Honduran textbooks are the educational framework, enhanced by other Kolibri assets, e.g. Nasa Science.

We seek to obtain high-quality engaging instruction for students despite limited time in large classrooms oriented toward rote memorization and social promotion typically staffed by poorly trained teachers.  This is especially acute in mathematics.  Job prospects are severely limited by low education.

We work in Intibucá, one of 18 departments (states) in Honduras.  School completion rates for our local communities show only 25% completing high school; 60% completing 9th grade.  Ministry of Education has a strategy to add grades 7-9 to primary schools to improve this and to require a college degrees for teachers.  Qualified teachers are being further stretched to cover multiple grades and subject areas.  Having high-quality, engaging instructional material enhances educational outcomes.  We know this through analysis of the official student achievement tests obtained under a confidentiality agreement with the Ministry of Education.  As a result, we hope students stay in school longer and learn more job-relevant skills.

Intibucá has a population of 230,000.  With 36% of the population, this translates to 83,000 students.  Before school shutdowns, we were covering about 4000 students.  We have added 5000 with tablets for distance learning.  The population of Honduras is about 6.5 million which translates to over 2 million students.

Our solution relies on Kolibri an educational ecosystem created and supported by Learning Equality, described here: Description of Kolibri.

We started with a digital rights agreement with the Minister of Education for the official Honduran textbooks so that we would be aligned with the country's educational objectives and achievement metrics.  We used custom software to parse these PDF files into chapters and lessons fit for use by teachers.  We used Learning Equality curation tools to convert these into a "channel" or coherent set of educational material.  This is further enhanced by aligning it with video presentations from other channels available on Kolibri, e.g Khan Academy.   We created automated quizzes to support the lessons.   Randomization options in Kolibri facilitate teacher assignment of homework and in-class diagnostic exercises.

We've developed a rich set of teacher training material including videos and WhatsApp support groups.  These have been embedded into the regional MOE training organization.

We are piloting a "Data Portal" which provides reporting of teacher and student usage which can be collected via "sneakernet".  We have embedded this into the departmental MOE staff so they can monitor overall progress.

These augment existing government improvement efforts like an educational vitamin tablet.

Shoulder to Shoulder has been in rural Honduras for over 30 years.  We work collaboratively to create and operate sustainable and accessible health, nutrition, and education services. We serve in the rural area of Intibucá, Honduras, part of the UN-designated Central American Dry Corridor.  It is the third poorest department in Honduras, already one of the poorest countries in Central America.  The area is mountainous with poor soil.  The people mostly support themselves by subsistence farming of corn and beans.  

The diet for many is deficient in vital micronutrients leading to a high rate of stunting in the children of the poor with the consequence that these children do not fulfill their intellectual potential.  We provide nutritional supplements to children age 2-5 so that they are ready to learn once they arrive at school.

Our healthcare operation since 2008 serves about 70,000 people and gives us a unique perspective on all aspects of the lives of the Lenca heritage people who live there.  Many of the locals have had at most a few years of primary school; many are illiterate.   This is slowly changing but still represents a limitation of economic opportunities for the area.  This area is sparsely populated and a long way from Tegucigalpa, almost to the border of El Salvador.

We've long recognized the need for improved education and since the beginning of our organization have supported it with support to send thousands of promising students to high school and a smaller number to college and professional schools.

Honduras, although second last in a World Bank evaluation of Latin American Education is trying hard to catch up.  They aare working on joining PISA, have added college educational requirement for teachers, and are adding junior high's to most primary schools.  They spend four time the percentage GDP per capita in education versus the United States.  Its just that their GDP is so low.

We've always tried to work collaboratively with the Honduran government.  In response to their strategy to require English instruction in all their schools, we worked with the Lenca office in the Ministry of Education to create the countries first public bilingual school.  We and the parents group subsidize the operation of this school in order to provide not only English but also an enriched educational opportunity for children of the region.

Accordingly, we've used our school in Camasca not only to educate children but also as a learning lab to investigate new approaches and educational technologies that hold promise to improve outcomes.

It is in our bilingual school that we first started experimenting with KAlite, a predecessor to Kolibri.  We learned that many of the primary school teachers had little prior experience with technology and developed training and support structures. From there, we did a pilot in the adjacent colegio (high school).  

After there years of experience with KAlite, we converted our installed base of 30 installations over to Kolibri in 2017.  Along the way, we found that the Khan math provided enough benefit to high school math teachers that they were willing to explore the menus for relevant lessons.  

This did not work particularly well for grade school teachers who had to teach more than one subject.  Also, the lower the grade, the more gaps teachers found in Khan, specifically in currency-based math word problems.  We knew then that we would have to have a product that totally aligned with the Ministry of Education curriculum. Kolibri offered us the opportunity to curate exactly the digital curriculum needed by the government and utilize the additional and growing set of enrichment resources in the Kolibri ecosystem.

We formed a group of math teachers and used this group to recommend what curriculum they wanted to have available.  We've since formed similar groups of teachers by subject; one for Spanish, English, Natural science, and Social Science.  We use these groups to guide our curation efforts and better understand support needs.

Our experience in the bilingual school also helped us understand the core issue with poor literacy amount the Honduran population - no books.  While most teachers try to teach reading, the opportunity for children to practice their reading is limited to time in class since most Honduran homes do not contain books.

Using the bilingual school as a base, we experimented with passing out books to support reading contests and eventually settled on finding and curating a Kolibri channel with over 1000 digital books in Spanish.  Local teachers worked with us to review and recommend reading lists of grade-appropriate books.

Before COVID-19, we were serving over 100 teachers and 4000 students in 35 schools all in Southern Intibucá.  Using Honduran achievement scores from 2013 thru 2019, we can clearly see improvement in math scores Kolibri schools versus comparable schools not using Kolibri.  The gap in performance grows with additional experience.  Additionally, our analyses show that fewer students drop out of the bottom of the class distribution - a likely indicator for dropping out.  It's no fun being in school if you are flunking

In addition, we had several years of experience using a Beta version of Android Kolibri to support at-home study by students participating in the Math Olympics.  Once the schools were shut down, we built upon this experience to deploy over 100 tablets to groups of students.  This year, we've built on this experience and are deploying almost 5000 tablets to students.  One grant with the Inter-American Development Bank is supporting a Random Controlled Trial of 3400 students in 51 junior highs paired with a similar number of control schools.  The experiment design and evaluation are being conducted by Innovation for Poverty Action (IPA).

The objective of the RCT is to create statistical support that will enable expansion of this to Meso America.

Our Kolibri implementation directly addresses the needs of a marginalized population with low internet connectivity.  Honduras does not have the resources to build out internet capacity.  The cutting-edge features of Kolibri we are deploying create capability for Honduras to transition to a digital curriculum.  The low-cost tablets which have a five-year lifetime deliver superior engaging lessons in a cost-competitive alternative to paper books with the added benefits of providing material tailored to student needs.  The Data Portal lets education administration effectively monitor student study habits and progress in a way that can support the progress of their educational system. 

While we are still very much learning and evolving our curriculum and continuing to partner with Learning Equality on new features; we now have 8 years of experience in implementing digital curriculum.  Before Covid we were in 35 schools serving 4000 students; now are supporting distance learning model for 5000 students  in 17 different municipalities in Intibuca, Honduras - population 230,000.  When schools eventually return to classrooms, we will be covering almost 10,000 students in over 100 schools.  We have developed the ancillary but necessary training and support structures and have the support of the Ministry of Education at the highest levels.

We have statistical analysis to support superior math achievement in Kolibri schools versus similar ones without.  We are currently working with Innovation for Poverty Action to complete a Random Controlled Trial of 3400 students in 51 junior highs paired with a similar number of control schools.

Learning Equality is the developer of Kolibri itself and the core mechanisms that enable offline delivery of digital curriculum.  However, we have been working closely with them for 5 years on the development of Kolibri, especially in the Android version that runs standalone on inexpensive tablets. We work closely with them in the development of features and provide a significant testbed for their refinement.

Additionally, we have curated a national curriculum for Honduras and worked with experts to align this with other rich video material available in the Kolibri ecosystem.

We are developing roles and responsibilities for staff at the department and federal level which will provide the template for adoption as a national standard.

We are working with Inter-American Development Bank on a pilot that if successful could lead to substantial additional funding for expansion across Meso America.

Latin American education needs improvement and much of LA has poor connectivity.   The size of the internet capacity in the US now estimated at $65 Billion puts an equivalent internet buildout in Latin America beyond the reach of any Latin American government.

Intibuca is so poor, so remote, and so uneducated that our pitch to Honduras is if it can work in Intibuca, it can work anywhere.  Similarly, if it can work in Honduras, it can work across Latin America.  Education is the key to lift Latin countries out of poverty and Kolibri can do this for students and strengthen teachers as well.

Kolibri an educational vitamin Tablet to support progress.

We utilize Kolibri by learning equality. See Description of Kolibri.

It is built on open-source software using Python and Durango.

While it runs on anything, we have standardized on windows laptops as servers.  This is a compromise between price ( little higher than raspberry pi) but more easily supported in our otherwise widely dispersed low-tech environment.  We prefer SSD that provide better performance and are more resilient to drops.

We utilize Android tablets for student devices.  We used 16 Gig RCA models for classroom devices but have migrated to Vankyo 32 Gig devices for distance learning since we can fit the entire 7-9 grade curriculum plus 300 digital books on these units.  We also install MS word, Excel and Powerpoint on the tablets. We are also testing the use of smartphones for use where the family has a phone but we can't get them a tablet.

Our Honduran official channel is called CREE  ('believe' in Spanish).  It is about 30 Gig.  Loading it over the internet is beyond the capability of any rural internet service.  A loading operation was completed in Tegucigalpa but it took three days.

With help from Learning Equality, we have developed a side-loading capability using sd cards that lets us set up a tablet in about 30 minutes.  We organized a "tablet factory" and trained local students to run the scripts

We obtained permission to access the government database of achievement test scores for all schools in Intibuca for the years 2013 thru 2019 (3.5 million records).  (We are working to obtain 2020 and additional teacher observational data fields).  Our analysis of the math scores for 2013-2019 shows 10 point advantage for the southern Intibuca schools with Kolibri versus other comparable schools without.  Moreover, the gap grows over time.  

This is consistent with other studies using KAlite conducted in Guatemala by https://www.funsepa.org/

Kolibri was also used in El Salvador in an RCT comparing its use to regular extra math classes. Portal Educativo Estudiantil – Consciente

We are expanding on the base of proven math results to use a math curriculum completely aligned to national standards supplemented by Khan videos.  We also provide curriculum for an additional four core subject areas: Science, Spanish, English, and Social Sciences.  We are currently running an RCT administered by Innovations for Poverty Action that will measure Spanish and Math. 

We will continue to do additional observational analyses on the other core subjects using the achievement database. 

In addition to the metrics, some previously reported, others emerging from current project work, we have the endorsement of local teachers, parents and local political officials.  Outside of the RCT, our standard practice is to require the commitment of the principal, lead teacher, head of the parent association to a contract that requires them to purchase a large screen TV, provide replacements to tablets that become inoperative or lost, and supply local IT resources to teachers.  This is often accomplished by mayor donations to the school or parent group fundraising.  While not scientific, it does reflect the conviction of groups of generally poor parents to line up and help pay for this technology and commit to make it work.

There is a risk of the security of the student data during the research process.  We have taken care to sign confidentiality agreements with researchers.

Longer-term, there is a risk of the student data in the government databases that will be created for this data.  These are similar to the grade system currently in place, SACE.

We are currently serving about 5000 students.  This is in addition to the capacity to serve 4000 that is configured as classroom servers.  If schools open next year, we will be at least 10,000.  Additionally, we have several other granting organizations that are providing tablets and laptops but without any educational software simply trusting that internet access will generate some good.  We have a working agreement with World Vision and are working with Unicef and others to converge their platform choices to Kolibri, so that our students served in 2022 could well be 20-30,000 in Honduras

Our strategy is to build capacity and capability within the Ministry of Education staff structures.  Our project that completes in 2021 will provide this expansion road may.  

The five-year number is totally dependent on hardware grant opportunities but should be over a million in Honduras.

Our work with the Inter-American Development Bank represents an opportunity to shape a Meso-American strategy.  Of note, Shoulder to Shoulder medical operation previously participated in the pilot of an IDB funded Meso-American Healthcare initiative that was ultimately over $50 million.

Our strategy is to support the RCT and this Meso opportunity and separately to continue developing the Ministry of Education capability to understand and manage digital curriculum.

Our efforts to date have mostly focused at the department and local level within Intibuca so that we have a working replication model to expand next year. 

We are working closely with the Vice Minister of Education and expect to pilot a Kolibri-based teacher training module next year.  We have anecdotal evidence of teachers improving their domain expertise and qualifying for better positions.  We believe we can build on this and create forward paths for teachers as well as students.

We are measuring Kolibri usage and the teacher adoption curve using the Kolibri Data Portal.  Kolibri captures the transactions of all types on each device that runs it.  The Data Portal and the collection mechanism that supports it can accumulate this time-and date-stamped data.  this is showing us how many teachers are using and what material they are presenting.  Using this we can identify gaps in usage and direct support resources to address technology issues.  By placing this Data Portal into the departmental staff, we are equipping the department staff to proactively manage the instructional process in their area.

We continue to analyze the official achievement test database to measure overall progress.  This assessment system coupled with Kolibri usage data, unlike one time studies, will enable the country to manage their improvement efforts ongoing. 

Besides myself, we have 4 full-time bilingual staff in Honduras + 15 students part-time to load software

Our main strategy is to utilize existing Ministry of education staff: 3 part time

 4 part-time at the department level.

 34 part-time sIT resources embedded in schools.

180 teachers part-time this year

I retired from Procter & Gamble where I managed large-scale IT implementations and technology selections.

EA: Honduran, forced to quit school after grade 6 to support siblings, taught himself English and technology, returned to school and completed college as science teacher.  Runs technology program.

MR: Honduran, walked over 3 miles to 7th grade, learned English, taught school, won a scholarship to US university, holds degree inINternational business, returned home to help people like herself. Runs robotics program and support.

NV: Honduran, Graduate of leadership English emersion school, currently completing university on weekends, wants to be a teacher, loves to read, runs literacy program

M: Mexican but grew up in US, MS in Social work runs overall team.

Shoulder to Shoulder welcomes people of all backgrounds.  Most of our hiring is local Hondurans consistent with delivering services to Hondurans.

Our team is collegial and very high-performing as evidenced by how much they are accomplishing with such a small team.

Everyone pitches in to get things done and to care for one another

We believe that the solution we have been working on for eight years is ready for primetime exposure and additional funding.  We welcome review and feedback from MIT solve.  Being selected for recognition will help us win additional funding and help the solution convergence we hope to achieve in Honduras.

Potential funders are more likely to support recognized winning solutions.

While our core strategy is to embed the work into the existing Ministry of Education organization, we still need to grow our staff capacity.  Getting the word out is one key need we have presently.  Another is to better engage the government staff functions that are planing the future direction.  Possibly a win on the IDB project will help with this, but we still need more experienced credentialled staff to engage at the federal level.