Do you want to learn, or pass exams? There is actually a difference, and this cool tech tries to achieve both

IBM Research-Africa is tinkering with a way to deliver ultra-personalised learning in schools, but an exam-focused education system might trip it up

THIS year marks the deadline of the Education for All goals established in 2000 at the World Education Forum in Dakar, Senegal. The six wide-reaching education goals aim to increase access and quality of schooling, improve literacy levels and narrow gender disparities in education.

In 2012, there were 122 million children enrolled in primary school in sub-Sahara Africa, a 75% jump from fifteen years ago. It’s the biggest gain in enrolment of any region in the world.

Although the region’s primary level teachers increased by 75% over the past fifteen years to reach 3.4 million, they still aren’t enough to keep up with demographic pressures. The teacher-pupil ratio has remained constant over that period, at 1:42.

In fact, sub-Sahara Africa has about the same number of teachers as North America and Western Europe, but there, the teacher pupil ratio is about 1:14. This compromises learning quality. 

One learning assessment conducted in Kenya by advocacy group Twaweza, measuring literacy and numeracy in children aged 6-16 shows that educational outcomes are dismally low.

By grade 3, all children should be able to do grade 2 level work. But just 30% children in the third grade are able to do second grade work, and by the end of the primary cycle, one in ten children leave primary school without having grade two level competence.

It suggests that children are going to school but are actually not learning, and it also means many children entering secondary school do not have an adequate foundation in mathematics that is essential for learning and analysis, particularly in science and commerce.

It is clear that while the focus on increasing access and participation in education is largely working, there is an urgent need to address quality.

It will call for a radical rethinking of how education is conceptualised and delivered, if Africa is to benefit from the increasingly connected global economy as knowledge workers.

Digital content

Like all other sectors of the economy, education is witnessing the rapid adoption of digital content on personal devices such as mobiles and tablets.

Many African government have mulled the idea of providing primary school pupils with laptops, and some have actually managed to roll them out  - Rwanda’s One Laptop Per Child project has seen over 200,000 laptops distributed to primary school pupils, making it the third largest deployment of laptops to schools globally after Peru and Uruguay.

But the results on learning outcomes has been mixed. One study from Peru showed no improvement on test scores in mathematics or languages, and no effect on attendance, time allocated to school activities or quality of instruction in class. But they did improve students’ cognitive skills; that is, the ability to learn  – skills like reasoning, perception and attention.

The setback was, because students tend to learn faster than teachers, especially when it comes to using gadgets, most teachers had turned the laptops into nothing more than digital notebooks – used only to copy what the teacher had written on the board.

But one team from IBM Research-Africa in Nairobi is tinkering with a way to deliver ultra-personalised learning to students, that bridges that gap between traditional classroom learning delivered by a human teacher, and outside classroom learning delivered by an automated tutor on a tablet.

The Cognitive Learning Centre captures fine-grained user interactions with the digital content, and, using Big Data and analytics, can develop rich learner models that are personalised to each student.

A screengrab of the CLC, that shows student performance and engagement with the lesson. Those who need attention are flagged in red and yellow at the bottom right.

In the traditional set-up the only way a teacher could know if the pupils had understood the material was by administering a test. But the CLC informs the teacher if the students are not following along on the same page as the teacher and at the end of the day, the teacher can give personalised homework to each student, depending on how well they followed along in class.

“The teacher can select any student and “drill down” to see details in terms of learning and assessment activities and performance of the student, as well as recommendations for the student,” Dr. Kommy Weldemariam, lead research scientist for IBM Research – Africa’s education projects told Mail & Guardian Africa.

“If part of the homework requires watching a video, for example, the system records how many times a student pressed/released the seek bar, stopped, closed, forwarded, muted, zoomed in/out etc. With that, a teacher can tell which part of the video was too easy or too difficult for the student,” said Weldemariam.

The video monitor shows that most students are skipping the video between 22.5 and 32.5. Recommendations for the teacher are at the bottom right.

But the really interesting application of the CLC is its ability to capture information about the environment that students are in outside class. It has sensors that measure light quality and noise levels, and even an accelerometer that detects if the student is using it in a moving vehicle.

So if a headteacher figures out that most of the students are doing their homework on the bus on the way home, or late at night, or in a noisy, dimly lit room – he or she can adjust the school timetable so that one hour at the end of the day is set aside for doing homework in school rather than making the students carry it home.

But the biggest challenge in making the system fully work is not in its technological capabilities, but squaring with an education philosophy that focuses on exams.

An exam is a finite, once-off event, and there are many tricks schools and teachers use to “game the system” and pass national exams, for example, registering weaker learners in a different centre so that they don’t drag down the school’s mean grade, or insisting on rote learning and extensive memorisation. In that case, (and sadly for students) the CLC might be “too much” of a good thing.

“The CLC is designed as a learning companion. But many teachers are more concerned about how students score in exams, as opposed to whether they are actually learning,” said Weldemariam. 

IBM Research-Africa is also working with RTI International - a nonprofit research institute - to deploy big data analytics to capture vital statistics about schools in Mombasa County, Kenya.

“Often education data is incomplete, inaccurate and sometimes even deliberately misreported.  Using analytics and cognitive technologies, we are creating a school census hub which will minimise the effort, expense and error in collecting valuable data about attendance, performance and resources at schools.

It has the potential to completely change our understanding of the situation on the ground and what needs to be done to improve it and improve the outcomes for children.”

Related Content


blog comments powered by Disqus