Learning Analytics
Data, as some omnipotent digital gold, has defined the thinking of the last decade. Compared to the heightened expectations, the resulting applications sometimes might seem a bit underwhelming, but data is undoubtedly acting as the engine behind the next wave of the digital revolution. This is especially true for education, where we find the following two data-driven areas singularly exciting:
Making education truly personalised
Classic education systems try to fit students into knowledge/skill level brackets because this is the only way to provide education in a cost-effective manner. Teachers now simply don’t have time to create a tailor-made plan for everyone based on their interests and progress.
Well, digital education doesn’t have such constraints. Computers can easily prepare and continuously update personalised study plans based on real-time performance data from every single student. Those who meet difficulties get specific practice opportunities to address their challenges, while those who speed ahead can broaden their skills with expanded materials. This model moves us from formal assessments towards continuous performance evaluation, where the goal is not to test knowledge at some arbitrary intervals, but to ensure that every student succeeds.
Let’s also mention that after the initial cost of implementation, computers will be able to operate the system with a marginal extra cost per student. If we can find a smart way to fit the teacher into such a system, we will witness the unfolding of a true revolution: creating solutions that don’t force students to be “average” will suddenly become financially viable. There are a number of exciting experiments going on today that utilise machine learning to offer better assistance to students. Even so, leveraging artificial intelligence is not even necessary to reach a level of personalisation that is still unprecedented in large-scale education.
At this point the challenge can be solved with product design and innovation adaptation; we don’t even have to venture into the field of technical wizardry, as the solutions are utterly feasible from an IT perspective. Today platforms like Nearpod and Kiddom are working to build blended learning systems rooted in these principles.
Automatised personalisation is heavily dependent on data collection. The system needs data to be able to tell how a student is performing, both in class and during their individual studies. So, our job is to design tools that make sure that every activity leaves a meaningful digital footprint. But here comes a notable caveat: data can be collected only if people keep using the system extensively. To assure this, we first need to design platforms that offer tangible aid to students and teachers. The product’s primary goal should be enhancing both the learning efficacy and the experience, and utilising data can only come as second – an order of importance that should never be reversed during the design process.
Better insights for better decisions
Another exciting application of data lies in analytics. It supplies teachers with a better understanding of how their students are progressing, and helps them gauge how effective their teaching activities are. If individual studying and practising also happen on a digital platform, the collected data allows for unprecedented clarity about all the ways a student might need help. Thus, it makes planning interventions much more to-the-point.
This also creates a robust feedback loop between what happens in the classroom and how well it is internalised by the students. While in this article we focus on monitoring student performance, naturally, this is just one area of application for analytics in the context of education. A better understanding of teachers’ activities and school performance will also open new possibilities for further result optimisation.
Learning analytics and automatisation actually walk hand in hand. It can often happen that the teacher doesn’t need the system to offer automatic solutions to a student’s problem, they simply need it to provide them with a clear picture to allow for well-timed interventions. Setting up a good system relies on finding a healthy balance between two distinct things: handing over certain tasks to the software for automatisation, and creating an environment where an educator can make better, data-driven decisions on how to proceed with activities that are non-automatised, i.e., they will be carried out by a human expert.
Obviously, analytics play an important role on an institutional level too. They help us better understand how well education performs in our ecosystem on a macro scale. This opens up promising opportunities for quality assurance, and gives us a way to pinpoint problems and possibly even fix them preventively.
Let’s take a moment to note that based on the above, our future could unfold in very different ways. The pessimistic vision is a cold and technocratic world, where students interact only with computers while invisible algorithms decide what they should learn based on data they’ve collected. It’s a classic “big brother” scenario, where machines monitor every step we take. Luckily, judging by the direction the western world is heading in terms of data management and experience design, we don’t think this will come true. We believe that data handling will be conducted transparently, in agreement with students (or parents), and will enable teachers to offer a more individualised focus and a wider array of options to help their students.
Finally, keep in mind that learning analytics can create value only if teachers actually use it. Teachers will only use a tool if they can see its benefits – and if they know how to leverage it. This is what we will talk about in the next section.
How to leverage learning analytics
Computers need data, but humans need insights. In multiple research conducted at various universities aiming to explore the real-world benefit of learning analytics solutions, the same theme emerged: teachers don’t actually need a plethora of data. In fact, they only need specific questions answered, which can help them plan interventions and tailor their teaching activities to their students’ needs. In short, what they need is useful insights.
Answer the right question in the right format
Let’s start this section with one of the most important product design principles: more information is not always better. On the contrary, it often leads to information overload – even more easily in a digital environment, where attention is already a scarce resource. People need information, but what they need is specific information that can help solve their problems. To them, everything else is trivia (in the best case), or rather just noise and distraction. Right, but then how is it possible for one product to give just the right information (and in the right level of detail) to different educators who are obviously facing different challenges?
Very few people need raw data for their work, and especially not the amount of data that modern digital ecosystems are producing, where every activity leaves a footprint. As discussed above, most professionals – and teachers especially – need useful insights instead. So, what are useful insights? Basically raw data that is already processed and transformed into answers to various questions; information that helps users better understand a topic that is relevant to them, and as a consequence, enables them to come to clearer conclusions and make better decisions. It’s worth paraphrasing: an insight is only useful if it promotes clearer conclusions and better decisions – otherwise, it’s just a fun fact.
The world where teachers can simply ask complex questions from their computers while leaning back in their chair (“Hey computer, I have 30 minutes left today to do some mentoring, which student needs the most help, who should I invite for a video chat?”) is still pretty far away. Instead, now most platforms are capable of showcasing various performance indicators and insights – and it’s up to the user to make the most of them. We have assembled a specific thinking framework to help you come up with good insight designs.