Tutoring to promote learning

In Mexico the strategy that the government adopted for remote teaching during the pandemic was very heterogeneous. On the one side classes were broadcasted in the open TV daily. But depending on the state and sometimes on the school, teachers were  told to continue remote learning or it was left up to them to decide whether to do so or not. The main source of communication between teachers and students according to a survey carried out by the educational authorities, was WhatsApp. Vergara-Lope, Velásquez-Durán, Calderón and Hevia estimate that the learning loss in math during the pandemic for children in Mexico lies somewhere between 0.62-0.82 standard deviations which only points out the unprecedented learning crissi that the country is facing and that is affecting the 33  million children enrolled in the education system. When we surveyed the parents of the children that participate in our tutoring program, more than half of them reported that their children met less than three hours per week with their schoolteacher during the pandemic, these are all children with connectivity (which is required to participate in the program) and with parents that are involved enough to enroll them in the tutoring program. The true average of synchronic class time during the pandemic once we include children that didn't have connectivity must be approaching cero. There is thus a need to work on regularizing all students and more evidence based and effective interventions are needed. 

Vergara-Lope, Velásquez-Durán Calderón Hevia, “Estimation of the fundamental learning loss and learn-
ing poverty related to COVID-19 pandemic in Mexico,” International Journal of Educational Development,
2022, 88, 102515.
32

In 2021 we developed a tutoring platform in Mexico Jóvenes Ayudando a Niñas y Niños (JANN). The platform works as follows, on the one side we recruit university students to work as tutors. In Mexico, in order to graduate from university students have to meet a 480 hour social service requirement people can also volunteer to become a tutor. We formed partnerships with some of the biggest public and private universities in Mexico so that students can meet this requirement by working with us. Among these universities is the national university UNAM, the two biggest private universities ITESM and UVM as well as other private universities ITAM, UNITEC and some teacher schools. Once tutors enroll in our platform a profile is created that guides them through the process of answering a series of baseline surveys, a math exam and completing the mandatory trainings. These trainings are structured in a similar way as the coursera platform, short videos followed by questions. Tutors can also manage their social service requirements, choose how many groups they want to work with these requirements vary depending on their university. 

On the other side we have built partnerships with several state governments in Mexicoto promote the platform. Among them Aguascalientes, Baja California, Durango, Sinaloa and Yucatán.  The statal government calls a meeting with all school supervisors. In these meetings we present the tutoring program, then supervisors pass on the information to directors and teachers, who share it with the parents. Interested parents need to sign their children up for the tutoring program in our platform www.jann.mx They can also be invited by other participants to join the platform or hear about the platform through social media etc. 

Once students sign up, the also receive profile in our platform, where they have to answer the baseline surveys. About 60% of the students that sign up finish the baseline surveys and are then assigned to a tutor. We ask them to answer the baseline surveys for two reasons. First, these surveys allow us to do research on the learning process of the students as well as the optimal group configuration. Second, we have seen that students that don't answer the baseline evaluations are less likely to attend classes with their tutor and actively participate in the program. 

We assign students to groups of 5 students in total. The students are all of the same grade but of different schools. Once assigned students and tutors find two times and days in the week to meet for an hour and review math topics. Tutoring is one of the most effective interventions in education because it meets students at their level. Small groups make the tutor able to focus on each of the students and their needs compared to a normal classroom. In the first implementation of our platform in the spring semester of 2021 in a randomized control trial we randomly assigned a subset of our participants to receive tutoring. We tested all participant at the end of the semester and could measure an improvement of 0.14 standard deviations for students that received tutoring sessions during an average time of 8 weeks. Given these estimates a year of tutoring sessions could help bring students back to the levels observed before the pandemic.  

An additional and key component of JANN is research. Since we have full control of the program, JANN allows us to study several details entailed in the constitution of the groups. For instance run randomized control trials to understand the better composition of the groups whether we should assign students to groups based on their performance to facilitate teaching at the right level or randomly to enhance peer effects. Additionally, we have the perfect setting to study class dynamics in a large scale.  

After each class, tutors upload a report with the attendance of students, and rate the class from 1 to 5 according to whether they think children understood the content of class.  Additionally, tutors have to record all of their sessions and upload the recordings to our platform as evidence of their classes. Tutors and parents have granted consent for this recordings to be analyzed for research purposes.

We are planning on extracting features from the recordings of the audio of tutoring sessions. The plan is to do exploratory work using Natural Language Processing techniques to study the interactions between the students and tutors and see how these interactions relate to the success of the tutoring sessions. For instance, variables that we are considering are the tone in which the tutor speaks (encouraging, scolding, etc.), the frequency of each student’s interventions among others. Currently, to study the effectiveness of teaching we rely on the annotations done by observers (Kane, McCaffrey,  Miller & Staiger 2013).  The problem is that both measurements are subjective and potentially biased. We want to automatize the task of observing and put a “neutral” or at least “constant” algorithm to analyze what is happening in the tutoring session. Preliminary analysis of the recordings in our first implementation shows that a Latent Dirichlet Allocation algorithm is able to identify topics discussed in the classroom that are related to math as well as topics rather related to personal matters.  The idea is to further understand what aspects of the tutoring sessions foster learning.

Kane, T. J., McCaffrey, D. F., Miller, T., & Staiger, D. O. (2013). Have we identified effective teachers? Validating measures of effective teaching using random assignment. In Research Paper. MET Project. Bill & Melinda Gates Foundation.

Currently we are working with public school students mainly in the states of Aguascalientes, Baja California, Durango, Sinaloa and Yucatán in Mexico. We have surveyed parents who  report that since the pandemic started their children have received on average less than 3 hours of online classes from their schoolteachers, for many of the participants JANN has increased the lecture time in more than 50%.  

The schedule of the tutoring classes is very flexible, determined by parents and students, this has allowed students that sometimes couldn’t connect to the school class, for example, because there is only one cellphone available in the family and the parents leave for work, to be able to connect to the tutoring session (for instance later in the day or in the weekend). Another example of the benefits for the poor in the program is that in the state of Baja California, 2/3 of the teachers have been on strike since November of 2021, because of that many of our students are only receiving a lecture through our tutoring sessions, teacher strikes happen often in Mexico so we can envision this will be a future benefit as well.

On the other side around 40% of the current participants don’t have any family members that have ever attended university, the program is introducing them to a university student which can potentially broaden their social sphere as well as help shape their decisions to pursue a university degree. Additionally, 93.7% participants have reported that they would like their children to continue the tutoring sessions even when in-person classes return. The return to schools can also represent an opportunity for JANN to reach the students that we haven’t been able to reach yet, the students without connection to the internet. This could happen once children have access to the school infrastructure. 

Everyone in our team was born and grew up in Mexico. Among the team some have worked before in the field of education teaching in underprivileged schools in the state of Hidalgo in Mexico with a civil association called ENLACE rural so we have been working on similar in person projects that has a smaller scale. JANN is mainly carried out by university students in Mexico who are fulfilling their social service this population is very close to the team as we all recently graduated from university. We have been working on this solution for more than a year now and we have had a steep learning curve, that puts us now in a very good position to carry out this program.

The platform has evolved and adapted more and more to the specific characteristics of the users. Most of the changes we have implemented to the platform have been based on the feedback that we have received from the participants.

We have several channels of communication with participants that allows them to express their ideas as well as allowing us to hear them out and implement the necessary changes to the platform. Tutors and parents have an open line of communication through WhatsApp where they can communicate with a team member. Tutors also have daily office hours where a team member connects to hear them additionally at the end of every instruction period we survey all participants: parents, children and tutors and ask them for suggestions to improve the program. We then carefully read these suggestions and carry out additional interviews and focus groups if needed to understand them and implement them. Also, the leading team: Bernardo and Salome are tutors in the program which helps us live the experience and the platform interface from the tutor’s side to be able to better understand the tools that they can need to successfully teach their classes.

For instance, a good example of this is our math evaluation. I (Salome) talked to my students about the math evaluation that we were using. We were applying a subset of questions of a standardized PLANEA that was used in Mexico. My students expressed that they felt that the exam was very long, implied a lot of reading even though it was a math exam and that they felt very nervous and anxious when they were answering it.

We decided to conduct a series of interviews and a focus group with students and tutors, from tutors we learned that students were sometimes receiving help from their parents when answering the evaluation. We brainstormed with them ways to develop an evaluation that wouldn’t stress students out, and that could successfully evaluate their learning levels. We use these evaluations for research purposes and thus they are crucial in the program.

We came up with several changes to address these concerns. First, we changed the name of the test to refer to it as the honesty homework with the aim to have students see this assignment as a homework rather than an exam and thereby lower their stress levels as well as incentives to use external help when answering the evaluation. At the beginning of the homework, we would explain a set of rules to students; among them, that the main objective of the homework is to be honest and make their best effort in the tasks that they know how to solve and select the option "I don't know" in the tasks they don't know. They were also asked to not ask for help or use any other resources as part of the honesty component of the homework. Students would also be shown a picture of a dog with glasses to lower the stakes of the test. And were asked to write down their name confirming that they understood the instructions.

The test was adaptive, each topic has one multiple answer question always including an "I don't know" option. If the student answers the multiple question right then they are shown a harder version of the same topic and instead of a multiple choice answer it becomes an open ended question with the instruction to leave it blank if they don't know. In total if they answer all questions right the honesty homework has 22 questions. If they miss all multiple-choice questions then they have to answer only 11 questions.

Another change that we made to the test is that instead of having text questions for example: "José goes to the supermarket to buy flowers for his father, one flower costs 5 pesos and he has 30 pesos ¿how many flowers can he buy?" we asked "30/5=" this allowed us to focus in math learning rather than reading comprehension and made the test shorter for students. Only one of the 11 questions in the honesty homework would have a text in order to see whether students can translate a real-world problem into math. Our evaluation has a higher correlation with the PLANEA test than two different versions of PLANEA among themselves and the scores of the students in the honesty homework are distributed normal when the PLANEA test had a skewed distribution. We see this as one of the many examples of how we reaped the benefits of hearing out the feedback of our participants and of integrating their feedback to our evaluation.  Other changes that we have made include the reports of the classes, the material available for tutors in the platform, some specific trainings that were requested by participants, the monitoring system that we use for tutors and to make sure they deliver classes, the means and frequency of communication with parents and in general the setup of the platform.

We are applying to solve because we have found that we could greatly benefit from  guidance in the process of managing and promoting the platform successfully. We also see opportunities where JANN could benefit from increased exposure and joining the network of MIT Solve. 

Our team is mainly composed of economists and programmers and there are several skills that we lack and are needed in order to move the program forward in the direction that we want it to grow.

Some areas that we have identified that we can improve upon are inter team communication and teamwork. We have done an amazing job as a team and we consistently overcome the challenges that present to us. Sometimes we have challenges that arise from problems in the inter team communication and we are sure there are strategies that other teams are already applying to overcome these issues so we are looking forward to have access to the mentorship and coaching that solve teams are awarded. For instance, the communication between the human resources team that has first hand knowledge and feedback from users and the programmers team who is the one that is capable of implementing these changes in the platform, is of great importance to the project and sometimes miscommunication between the two teams can lead to challenges.

Additionally, we are planning on growing the number if users and this will require to grow the team and eventually migrate to a sustainable source of funding this could be through donations (which is currently our main goal) but also through charging some stakeholders for the service which is an idea that we haven’t explored to its ends. This is another area where we could greatly benefit from the advice of experts and coaching.

Additonally, since we are trying to find a stable source of donations we need to gain more exposure and learn to present our results in the way that is used in these settings. We consider ourselves strong in the monitoring and evaluation side of the project because we have information on all the details of the project and we are using this information to produce research results and measure the impact of the program. We have are still learning to step away from our technical language and move towards learning to present our results to more broader audiences in this sense, presenting to broader audiences, is where we can use the support in evaluation and impact measurement. Also, gaining exposure through solve can help us connect with more potential donors. And other organizations with whom we can partner. Finally, joining a network of other entrepreneurs can possibly reveal other sides of the project that we haven’t even considered pursuing yet. We have come so far thanks to the a network that we have been continuously building and that MIT as our alma mater has opened to us. MIT solve can open the doors for our team to get to know other people that are finding themselves in similar stages of their project than ours we can potentially collaborate, help each other out with advice and learn from one another. We are also looking forward to share all of what we have learned so far because we think that more initiatives like ours can boost learning around the world and really make a difference in the lives of children.  The use of technology can bring us very far in implementing solutions and we are looking forward to be able to join the network to explore more potential paths of growth with our peers.

Our solution is innovative in many ways. First, we are making the use of technology to connect students to tutors that are all spread across the country. With a very small team we have managed to work with more than 11 000 participants including students and tutors. With a small team of less than 5 part-time and full time staff. 

We have succeeded at doing this by automating most of the processes. For instance when tutors enroll they can complete all of their initial training without us interacting with them until they attend the interview. We also have automated the enrollment process of students. And the assignment of students to tutors. Students and tutors receive a message when they are assigned and they can log into their profile to find the contact information of their students.  We do a weekly monitoring that is also automated and consists of 4 steps. First, tutors create a WhatsApp group with their students and add a number that belongs to us. We use this number to monitor the weekly progress of the group and whether there is communication between students and tutors, in the case that we find no activity in the WhatsApp group we notify the tutor. Second, tutors need to fill out a report after each class where they report the attendance of their students, if a tutor doesn’t fill out the reports in one week we warn the tutor that they need to resume activities. If the tutors don’t upload their reports in two consecutive weeks we send them a message and give them 48 hours to report to us or else we end their participation in the program and replace them with another tutor. If a tutor reports that a student is not attending class we message their parents asking whether they are satisfied with their tutor and with the time and date of classes and offer to relocate them if not.  Third all parents can report in their portal whether the tutor is not in touch with them, in case a  tutor is reported we message the tutor to warn them that they have been reported and need to resume activities with their group, once a month we also message participants that regularly attend classes to ask if they are having a good time in the program and to remind them that they can report their tutor in case that the tutor is not in touch with them. This system has allowed us to reduce the number of cases where a tutor fails to deliver their classes, it also helps us identify if tutors miss communication with a parent because they wrote the whatsApp number wrong or because the parent initially didn’t identify the number if the tutor and did not reply when the tutor first reached out as well as other particular circumstances that could impede classes. 

Tutors and parents are supposed to find a time to meet if a parent cannot accommodate the times of the tutor the tutor can ask for us to relocate the parent to another group, also if a specific group of parents are not committed and their children don’t attend the session the tutor can report this and we assign them to a different group. We are also working on developing a system to keep the average attendance of each group at 5 to maximize the use of tutor resources in the program.

Each of the mechanisms has been organically developed as we implement the program and find challenges that need to be addressed. We are working towards developing a system that can sustainably and automatically solve these problems without the need of the intervention of a person. 

When we finish perfecting this tool the program will be able to grow considerably and possibly work with hundreds of thousands of students just as a platform like uber or airbnb but a platform that is directed towards aiding children in low income contexts. Once the tool is developed it can be easily replicated in other countries as the only other need for it would be to increase the user base.  We think of this project as having the potential to impact a considerable share of students in Mexico and hopefully in other countries. 

Additionally, we are asking tutors to record the classes which, as discussed before, has the potential to help us contribute to the research of class dynamics in education. We can automate the analysis of thousands of hours of classes using machine learning techniques to better understand what good tutors do and possibly develop targeted and evidence based training to train  future tutors.

Impact goals are clear and brief statements of the specific external impact you wish to create through your solution, towards which all elements in your work are oriented. We look for solutions that will have a transformational impact on people’s lives. Use this space as an opportunity to describe your impact goals and how you plan to achieve them.

In the next year we are planning to keep developing the platform to automate more of the processes that we currently use. For instance, as mentioned before, being able to use the full capacity of the group and keep the attendance at 5 students per tutor. 

We also need to keep improving on the processes that are currently automated to account for all specific cases and improve the user experience. 

We also aim at increasing our base of volunteers. While university students that are fulfilling their social service are a reliable source of tutors we could also increase our capacity to attend students if we increase the number of tutors that participate in the program and one way would be to have more volunteers. Either from tutors that are returning or new volunteers that are interested in helping beyond the social service requirement. 

We want to increase the number of participants in general students and tutors. We need to do this by promoting the public image of JANN to the public and motivating more people to join. Forming more alliances with universities and state governments as well as with other organizations and building up our social media presence.  We aim at reaching around 20 to 30 thousand students in one calendar year. This means duplicating or triplicating our current reach. 

We want to start establishing the mechanisms to reach the children that need regularization the most. The children that were disconnected during the pandemic. This might imply working closely with the school system of our different partner states to make use of the school infrastructure for them to connect with their tutoring group. 

We want to start the path towards becoming an NGO and find a stable base of donors. to maintain the operations of the program. This implies developing a robust business model, growing the team, making the necessary legal processes and gaining the skills necessary to carry out this work. 

Lastly, this year we are aiming at publishing our first research paper with the results of the Randomized Control Trial (RCT) that we are currently running. And start a new RCT by september. 

In 5 years we want to establish ourselves as the go to service for children that want to improve their math skills and the go to place for university students to do a true social service for their country as well as for volunteers that one to donate a couple of hours of their week to work towards improving education. We might have current students return as tutors to create a sustainable source of tutors. We want to expand the grades that we are attending. Currently we are working with primary 3 to secondary 1st. It would be ideal to be tutoring all the way up to high school. 

We want to augment the scope of the program to tackle reading proficiency on top of math tutoring. This will require developing new tutor training programs as well as other measurements to assess reading comprehension.

We want schools to actively participate in this program and provide their infrastructure after lecture time for students of scarce resources to connect. We aim to be working with at least 100 000  students. 

In 5 years we want to have a well developed platform that can be easily exported to other countries. 

We also want to become an education lab where different theories in education can be tested to advance the frontier of education research. That possibly works with a business model that charges governments for the service or is sustained by a constant flow of donations. We want to have quarterly reports of the advancements in the programs and milestones reached for participants. And have a considerably bigger team that can address not only the administrative need of the program but is also exploiting all the data available to help us better understand the learning process and propose evidence based policies. 

We want the results of the program to be evidenced by the career paths chosen by our students. For instance, whether they attend university, whether the pursue STEM fields, whether they have higher earnings.  By the scores of our students in standardized tests, by the increased presence and  participation of our students in math olympiads or by the improvements of the results of Mexico in international assessments. As well as by enhancing the social commitment of our former tutors.

The first indicator is the number of participants. We are also testing all the children when they start and when they finish the program which allows us to keep track of their progress during their time in the program. And report their test scores as another measurable indicator.

Other indicators that we can look at are, the percentage of classes successfully completed by tutors, the percentage of tutors that successfully finish the program. The reviews of the program of parents, tutors and students in the endline surveys. 

Also the reviews of the usage of the platform to improve processes and make them more user friendly. 

The average attendance to classes of students. All of this is data that we are currently collecting and cleaning and presenting as we need to advance and show evidence of the progress of the platform. 

In the future we can start collecting follow up data to understand whether participating in our program motivated our students to pursue a career in STEM or any other university degree. 

Activities

Math tutoring

Constant monitoring and communication with all participants. 

Outputs

Service delivery

Short term outcomes

Class attendance

Introducing children to university students and children in other schools or regions of the country as well as in other socio economic backgrounds. 

Medium term outcomes

Improvement in math scores. 

Increased social behavior 

Improvement in social skills

Formation of new friendships

Tutor as the role model or mentor

Long term outcomes. 

Motivation to pursue a degree

Feeling of self fulfillment and being knowledgeable in math. 

Increased social network

More socializing skills

Increased parental involvement

Our project consists of a scaffolding structure given to children to reinforce their learning in Math and make sure none is left behind. On the first semester, we estimated three months of tutoring to close the wide gap between an average public school and an elite private school performance on standardized test by 13%. In Mexico, this gap is 1.15 standard deviations.

As documented by (Chetty, 2019) exposing young children to successful models in different areas can actually result in these children replicating the success in elder ages. The mechanism behind this can be the one posed by (Freinet, 1966) where children, in their path of self-construction, use examples they have at hand as tools to achieve their goals. If children lack the tools to overcome obstacles in their path, they might deviate from it with long-lasting consequences. We see examples of this deviation in the form of high drop-out rates through k-12 education in Mexico. 

We believe tutoring sessions will benefit the children through three main channels: abstract learning, role model learning, and a social outlet. With abstract learning we mean the Math improvements the children will see during their time in JANN, which we have already measured. With role model learning we mean the effect of exposing kids from urban and sub-urban contexts to university students on a regular basis. This exposure is also a significant one as kids are tutored in small groups (<=5 students) and there is generally high interaction with the tutor. Rapport has been shown to boost role model effects (Krämer, 2016), for which we expect to see large effects for children from families where no close relative has been to a university. Finally, we also expect the tutoring to benefit the children through its social nature. From anecdotical evidence and interviews we know new friendship bonds were created during our past iterations about 80% of students reoorted that they would miss their tutor when their tutor left the program. The creation of this bonds is valuable in itself, and also reflects higher social behavior and opportunities, as well as an improvement in social skills. 

In the long term, we expect to see changes in several aspects. Firstly, the reinforcement in Math will prevent kids from seeing it as an obstacle and therefore prevent kids from dropping out or leaving Math aside due to fears of failure. Along with the exposure to university students, we expect both features to boost motivation to pursue a higher education degree. Moreover, success in Math boosts in some cases feeling of competence of the kids (Singer, 1986), allowing them to aim further. The widened social network along with better social skills can also result in better labor market outcomes for the kids (Yakubovich, 2005). Finally, given that the tutoring is arranged with the parents we also expect it to bolster parental involvement in kids education. 

Overall, we expect the program to be able to provide positive lifelong lasting effects on participating kids lives. Moreover, the questions we seek to answer will also be able to enlighten today's educational policies in terms of gender bias and skill composition. Answering this questions will result not only in improved education within the program but also in interested school around the globe. 

Bell, A., Chetty, R., Jaravel, X., Petkova, N., & Van Reenen, J. (2019). Who becomes an inventor in America? The importance of exposure to innovation. The Quarterly Journal of Economics, 134(2), 647-713.

Freinet, C. (1966). Essai de psychologie sensible appliquée à l'éducation. Delachaux et Niestlé.

Krämer, N. C., Karacora, B., Lucas, G., Dehghani, M., Rüther, G., & Gratch, J. (2016). Closing the gender gap in STEM with friendly male instructors? On the effects of rapport behavior and gender of a virtual agent in an instructional interaction. Computers & Education, 99, 1-13.

Singer, J. M., & Stake, J. E. (1986). Mathematics and self-esteem: Implications for women's career choice. Psychology of women quarterly, 10(4), 339-352.

Yakubovich, V. (2005). Weak ties, information, and influence: How workers find jobs in a local Russian labor market. American sociological review, 70(3), 408-421. 

We combine several technologies to achieve our goals.

Videocall technologies: tutors provide their lessons through videocall technologies adapted to the needs of parents. These range from zoom to whatsapp calls. 

IM technologies: we use Whatsapp, SMS and telegram to contact our participants. These help us remind participants they need to finish requirements before receiving/giving lessons, remind parents when their kids miss school as well as tutors when they are behind on their responsibilities, and give tutors important announcements through the program. 

Web: we also developed a web platform that uses API calls to other services to train tutors through an online course, collect surveys and test responses, and give tutors resources to and tools to prepare their lessons. 

We are making use of all of these elements to create a platform that can provide the service of very low cost tutoring sessions to students. By automating most of our processes we are being able to work with a large base of users being a small team and at a low cost.   

The larger organization is called Proeducación IAP it is currently helping us have a legal identity in Mexico. They officially hire our employees and receive donations directed to us. Some of the schools that are affiliated to Proeducación are participating in JANN but this constitutes less than 3% of our students. 

The plan is to eventually be constituted as an NGO. 

Our team is mostly conformed by women in leading positions, as well as members of the LGBT community. We are happy to have created a space where members of these groups can take leadership positions.

We are looking to expand our team and are opting to prioritize recruitment among alumni of public universities or low-cost private universities possibly among our current pool of tutors. 

Moreover, we expect JANN to become an education lab where people without the resources to run RCTs can access a huge database of information to create more insights into education. This decentralized structure will be able to provide a more democratic and equitable access to the world of education research.

Our key resources are: 

A popular platform (>40k users) that connects people interested in receiving tutoring with people interested in giving it

A network of school decision-makers interested in sending invitations to the program to kids and tutors

A collection of resources used to train tutors and evaluate tutors and kids

There are three main parties that benefit from our services:

Tutors: They have an opportunity to contribute and do their social service requirements from the comfort of their houses through a highly engaging and rewarding activity. From anecdotical reports as well as surveys, we are aware of the large value tutors lay on this experience

Schools: Elementary and middle schools see improvement in their kid's learning through almost no effort. Universities have a resource where they can send an unlimited number of students to  finish their community service requirements, which is rare for them. 

Families: Families have also reported, anecdotally as well as in surveys, the large value they receive from the tutoring lessons. This is due to the learning in the experience, as well as from giving kids a social outlet to participate in while in their houses.

We see our organization as a win-win-win solution. Through our work, otherwise wasted social capital is realized and provides benefits for all parties involved. 

So far the program has been sustained through research grants. The plan to make JANN an education lab will allow us to receive multiple grants to sustain the program through overlapping periods. However, given the very positive social response we have obtained from the general public, we envision high possibilities to be able to also sustain ourselves through donations. As an example, we partnered with a private company (PYMO) that serves as an intermediary to receive and channel donations from companies to social organizations. Despite the intermediation process, which impedes us from pitching our idea or doing any kind of direct communication with donors, we have received around 15,000 USD in donations from Mexican donors. This makes us believe that if we put more effort into receiving donations we'll be able to obtain much more resources from it. Finally, our use of IT allows this program to be very scalable, which means that as it keeps growing, our cost per participant will keep decreasing (right now it's at 0.58 usd per student per month of clases). This makes the program a continuously more and more attractive receptor for donations that seek to maximize impact. 

So far we have received around 50k in research grants (40k from JPAl, 10k from the Schultz fund, 1k from MIST Mexico). Out of all research grants we have done, around a third have been successful. 

On the other hand, from intermediated donors, we have received around 15k usd. We expect this number to grow if we are able to contact the donors ourselves.