Personalized Learning with Mindspark

COVID-19-induced school closures have hampered learning, especially for students from socio-economically marginalized and geographically remote backgrounds. Many schools have been unable to implement effective remote-learning models. Consequently, the students will likely suffer learning losses, especially in key areas like literacy and numeracy.

EI has Mindspark available to over 130,000 students remotely via smartphones and laptops in makeshift community labs, to mitigate these losses Using Mindspark’s PAL software, students learn at their level of competency, and teachers can monitor their progress and common misconceptions through detailed usage reports.

The coming academic year will likely be similarly unpredictable. Given this, EI believes that scaling remote Mindspark access in India and globally can help students from marginalized backgrounds and their teachers engage in meaningful learning despite school closures.

Long-term, it can help create a robust hybrid learning model, where in-class learning is supplemented and improved by Mindspark usage at home.

Learning losses due to COVID-19 will be unequal, and closely related to students’ socio-economic background. In India, 92% of primary school students are estimated to have lost one language ability, and 82% one mathematical ability during the first wave of the pandemic. Globally, before the pandemic, 53% of students from low- and middle-income countries were unable to read and understand a simple text. This learning poverty is projected to worsen to 63% now. Research from Ghana showed that 2/3rds of learning gains in foundational numeracy in a year were lost in 3-4 months of school closures; similar effects are likely across the Global South.

Though a lack of access to digital infrastructure is part of what makes remote learning difficult, the unavailability of quality remote-learning material is equally problematic. Current high-quality learning material often requires high internet bandwidth and expensive devices, unavailable to many. Furthermore, teachers’ efforts are hindered by decreased interaction with students; they are unable to gauge student progress or address their misconceptions. Overall, this will cause students to lag further behind and create greater disparities within-grade learning levels, and increase heterogeneity in classrooms, making effective teaching more difficult.

Mindspark is a PAL software that can help students learn their vernacular language, mathematics, and English as a second language. It draws on a vast database of questions to finely benchmark the initial learning level and rate of progress of every student and customize the material delivered accordingly, thus teaching-at-the-right-level at scale technologically. 

Based on this, it designs lessons tailored to each student’s needs and assigns remedial activity to reinforce basic concepts when necessary. It uses animated content to engage students and gamifies learning by giving points for successful answers. The software also asks questions to specifically test whether students understand concepts and knowledge application in real-life scenarios, instead of mechanistic learning.

Mindspark generates usage reports that give detailed and actionable insights into both individual and group learning levels. Teachers and school administrators can use these to plan lessons and provide remediation where necessary.

Prior to the pandemic, students would use Mindspark in labs set up in their schools. During the pandemic, Mindspark was made available remotely. Students can now access it via smartphones and similar devices at home, through an interface optimized for mobile usage. 

Mindspark (in vernacular languages) was created to improve learning among students from socio-economically marginalized backgrounds studying in government schools. In the last two decades, children in India have gained access to a school - 98% of students in rural India have a primary school within 3 km of their home, and India’s net primary school enrollment rate is comparable to those of developed countries. However, learning outcomes among these groups remain extremely poor and unequal. High teacher absenteeism, one-size-fits-all approaches, an emphasis on rote learning, weak transparency, and little or no remediation for students lagging behind curricular standards have contributed to this. This leads to a widespread heterogeneity in learning levels of the children within a single classroom and comparatively across school systems as well. This lack of basic competency is systematically lower for students from poorer families. In India, the learning outcomes are worse for the rural and remote regions, especially for the disadvantaged, tribal communities where the average literacy rate is 30-40% below the national average.The situation for students from marginalized backgrounds has only worsened in the pandemic due to limited access to digital technology and un-preparedness of the system for remote learning.

Adaptive technologies - which tailor the content that is delivered to each individual student’s need hold great potential, especially for those who are starting off at a lower level and may learn at a slower pace. An independent evaluation by the Abdul Latif Jameel Poverty Lab (J-PAL) found the relative impact of the programme to be much greater for low-achieving students, who were making no progress in school. Students from disadvantaged backgrounds often start from a lower baseline learning level, and therefore stand to gain the most from personalized learning. Teaching at the right level (TaRL) has been shown to be an effective solution to improve learning levels. However, making this accessible at scale in India remains a challenge due to issues surrounding teacher training, lack of ICT infrastructure, and reliance on rote-learning methods.

Mindspark directly addresses these students and enables teaching and learning at the right level by tailoring lessons to individual student needs and providing teachers with actionable insights that can be implemented/addressed in classrooms. Insights from student usage and assessment data are also regularly used to improve the questions and content available on Mindspark, as are developments from research in pedagogy. The use of Mindspark at home has helped thousands of such students continue learning meaningfully even during the pandemic. 

Educational Initiatives’ teams have undertaken extensive outreach, engaging with government officials, school administrators, teachers, communities, and parents to facilitate this. Numerous training sessions, frequent calls, the establishment of community-based labs have all been done. With further school closures likely due to the pandemic, increased engagement with Mindspark at home can be an effective tool to enable learning remediation and ensure that inequalities in learning outcomes for these groups remain minimal.

There are also second-order effects beyond immediate benefits to students and teachers.

1. The access to usage and learning data can build transparency in the education system, and help school administrators and education departments identify where targeted support is needed

2. The data generated from Mindspark usage is extremely useful for research purposes (randomized control trials of Mindspark have been conducted by JPAL, in addition to studies by other organizations)

3. The successful implementation of Mindspark can help guide policies that seek to implement education-technology interventions at scale, and ensure that cost-effective programmes with a proven track record are prioritized

Mindspark, EI’s evidence-based PAL software, has proven to improve learning outcomes for students from disadvantaged demographics, and for those starting at low learning levels. In areas with limited internet and electricity, EI had set up Raspberry-Pi servers for children to seamlessly access Mindspark.

During the pandemic, EI is helping students and their teachers to improve learning through targeted instruction and actionable insights to support students through access at their homes and in their communities. Teachers and parents are regularly informed of the student's progress and encouraged to allow children increased access to household devices for learning.

Mindspark has been used by half a million students over the past decade. Annually, it is used by about 200,000+ students globally, with the vast majority in India. Mindspark has been implemented in 300+ government schools and 5000+ after-school study centres in 10 Indian states, and is available in 9 languages (English, Hindi, Urdu, Telugu, Gujarati, Marathi, Punjabi, Tamil and Kannada).

Mindspark has shown promising learning gains in rigorous studies by multiple outcome evaluators – J-PAL, IDinsight and Gray Matters India- and in three different school contexts – after-school centers, private schools and government schools. Its in-school implementation led to the release of a guidebook by the Ministry of Education for implementing the PAL tool in state government schools. Moving forward, we intend to create a blueprint/roadmap for effectively implementing a remote learning programme for students who do not have access to schools.

Based on research: Using pedagogical research and rich data from 20 years of assessments conducted in government and private schools, Mindspark captures specific misconceptions and helps overcome them proactively while enabling learning. 

Assessment for learning: Diagnostic tests in Mindspark detect misconceptions and proficiency in concepts and put students in personalized remedial paths that consist of questions and activities.

Real-life examples: To pique interest in learning a new concept taught, Mindspark facilitates learning by creating the need to learn the concept either by showing real-life situations and/or posing a problem that is efficiently solved if the targeted concept is understood. 

Intelligent responses: Mindspark has sophisticated and intelligent responses like response-specific feedback & alerts, numeric and non-numeric parsers to interpret various forms of student answers (like 'two' or '2' or '4/2' or '2/1' are the same)

Intelligent feedback: Step By Step solver in Algebra, Exponents, Real numbers and Trigonometry allows Mindspark to provide very specific and contextually relevant feedback messages, at each step. This granular data is further used to develop insights.

Games and enrichment activities: Mindspark has pedagogically oriented games that ensure the engagement of children and enrichment activities allow children who have learned a concept to explore beyond the typical curriculum.

Teacher interface: It ensures teachers have complete control over what concepts the class needs to learn, and how that can be effectively timed (in sync with classroom teaching). Further, teachers are provided with student and class level insights with recommended action points based on class performance. 

Currently, Mindspark is used by about 200,000+ students globally – with the vast majority in India. Here, it serves children in 300+ public schools and 5000+ after-school centres across 10 states and 9 vernacular languages. It has shown promising results in multiple settings (after-school centers, private, and government schools). We work with government school students from grades 1-9, the majority of who hail from rural, and semi-urban areas, and a smaller fraction from urban areas. These children come from low-income to lower-middle-income households with a daily family income range of USD 2-10. Other beneficiaries of the programme include government school teachers, school leaders, administrative officers, and key stakeholders at the state/district level.

By the end of next year with support from the partners, EI expects to reach more than 0.3 million children using Mindspark despite school closures, by providing remote access. The new National Education Policy launched by the Ministry of Education puts special emphasis on strengthening foundational skills. Our intervention will serve as a model for remote learning and will inform government policy and reflect increased state budgetary allocations for software, training, and other requirements. Altogether, EI expects the spill-over benefits to reach approximately 10 million students in 5 years.

For the Mindspark at-home Program, our student learning indicators are measured as a function of reach and engagement and increase in learning outcomes. EI is measuring the success of the project using the following metrics:

• Reach is indicated by the number of unique students receiving the intervention through smartphone
• Parent engagement is indicated by the number of parents contacted by EI support teams via phone calls (to support intervention/any activity’s use)
• Student engagement is indicated by the total number of students actively engaged by the intervention which is composed by:
  • The number of activities/topics covered
  • The average duration of Mindspark usage per month (hours of usage for each student)
  • The average # of ‘Sparkies’ earned per month and the average % accuracy in questions answered per month
  • Improvement in learning outcomes in Language, English, and Maths
• Cumulative number of workshops/seminars with public practitioners for the dissemination of learnings
• Cumulative number of research papers produced through the proposed implementation
• Philanthropic/ Government funding mobilized for scaling the project
• Government funding mobilized for PAL in the states where the proposed implementation is deployed

Our team includes 250+ professionals combining depth of pedagogical and assessment expertise with experience from working with schools, government systems, and top-tier organizations in education, and leading academic institutions. Our specialized internal teams work on test development, data science, educational research, psychometrics, personalized learning, and IT development.

Our top executives are on the boards of several educational committees for governments at the state and central levels in India. Their insights have been sought for teacher education, implementation of ICT in education, drafting RTE model rules, formulating national assessment strategies, etc. They have also been involved in SSA’s Joint Review Mission, and the NCERT's Advisory Board for Educational Surveys. Our CEO, having grown in the organization from an associate to a CEO over the last decade, has also been recognized by the Schwab Foundation for Corporate Social Intrapreneurship. In the past, senior members from EI have been invited to present papers in prestigious national and international forums such as the World Bank (Washington DC, 2017), Harvard Social Enterprise Conference (Cambridge, 2017), National Council of Educational Measurement (New York, 2008), International Language Testing Association, (Hangzhou, China, 2008) and others.

EI’s teams have liaised with governments and taken their products across the country. EI’s implementation team has therefore been able to conduct assessments and introduce Mindspark in the remotest regions of the country. EI regularly invests in the learning and development of employees and provides platforms for them to engage with the larger educational ecosystem. The organization’s culture makes individuals strive for ownership, efficiency, and collaboration to achieve the vision of “every child learning with understanding”

Educational Initiatives is staffed by team members who belong to all major religions in India, and who come from diverse linguistic, cultural, and geographic backgrounds. Over 40% of EI’s staff are women, and many of them occupy key leadership positions within their teams. This diversity is crucial to develop Mindspark for different contexts, which makes it relatable to students across the country. 

To take Mindspark to students from marginalized backgrounds and remote regions, EI also employs hundreds of people across the country. This generates much-needed and relatively high-paying employment in places it might otherwise be scarce. Their hands-on experience working with stakeholders ranging from the local community to students helps take EI’s offerings to the last mile.

Employees in EI’s central team come from a variety of educational and professional backgrounds as well. While some bring management and pedagogical expertise with degrees from top institutes across the world, others come with rich teaching and on-field experience. The combination of the two is what has helped EI implement meaningful educational change.

Being part of the MIT Solve community will provide us at EI with the following:

• Building evidence on improvement in learning as a function of usage at home based on large scale evidence from EdTech implementation in remote/ blended/ at-home/ community usage models

• Building process discovery through documenting best practices with video and textual material

• Access to international networks and partnerships to potentially make Mindspark accessible to more students in countries across the globe.

• Access to funders at a global level to take Mindspark at home to more children across the world

Covid-19 has exacerbated learning gaps for students due to prolonged school closures. Governments have responded to this by provisioning online and technology-based learning through their education departments and collaborations with non-governmental partners.

However, these efforts are limited by a lack of access to and use of mobile devices for learning. We are adapting to and planning for the changes caused by the pandemic. Our focus remains on the improvement of foundational learning (Math, vernacular language, English) for children from socio-economically marginalized backgrounds in tribal districts. To ensure these children have access to adequate learning resources and opportunities, we require grant funding. This will help us reach many more students whose learning has suffered the most during the pandemic.

• Product / Service Distribution: Support via networks with international organizations who are willing to partner with us to ensure product distribution at India and global level to ensure reach to more number of children who are learning with understanding.
• Public Relations: Support us in branding and media visibility globally. This will ensure we cater not only to children in India but also in other countries where access to education is limited and such self-learning tools can be used well.
• Financial: Support via networks with impact investors, grantmakers, and international organizations who can support us to accelerate our efforts towards providing an equitable education to all children in India and other developing countries.

We intend to partner with:

• Non-profit/for-profit partner organizations who can help us access more children, both in India and other developing countries that face similar challenges of heterogeneity in learning levels, teacher quality, and other capacity-related challenges

• Organizations interested in co-developing a program that leverages complementary strengths and co-pitch the joint program to more funders

• MIT faculty who can help with impact evaluation of various operational models in and outside India