Big data in education
What if teachers only needed to press a button in order to find out in which areas their learners needed support, the tasks that were causing them difficulties, and those that they had already understood? Instead of working through set tasks at the same speed, teachers would be able to focus on deepening learners’ understandings of those aspects that were really necessary. Furthermore, they would be able to find out if the tasks they were setting required clearer explanations or were even unreasonable.
Thanks to the opportunities provided by big data (the storage, processing and analysis of huge data volumes) and the increasing use of digital media within learning processes, we currently have more data available to us than ever before. Every video that is viewed as part of an online course, for example, provides us with information about whether learners have watched it until the end or whether they repeatedly rewound specific sections. Each multiple choice test provides instant feedback about the number of learners who answered the questions correctly and how many attempts were needed to do so. Big data can also be used to demonstrate whether learners spend more time on texts or videos, whether they are involved in learning communities, and whether they prefer to work alone, in groups, slowly or at speed.
Is online learning the perfect learning system?
The situation described above has become reality in many schools, universities and on numerous courses. What does this mean in practice? Online learning can involve students sitting in classrooms. In these cases, students are often divided into small groups who work on a particular task using a computer, tablet or mobile phone. This system provides teachers with continuous information about the tasks their pupils have completed and how the class is going. As such, teachers can see – in real time – whether a class has grasped a particular topic or is still struggling, and can adapt their teaching to reflect the situation.
Online courses are also deployed in further settings, and some of these courses have tens of thousands of learners. In these cases, data analysis provides teachers with information about whether certain tasks are causing difficulties for a large number of students. This helps teachers to improve their classes as they then provide better explanations wherever necessary. Moreover, statistical data can be used to continually optimise online courses to ensure that they appeal to the largest number of students.
Large organisations gather data on learning to help them select training opportunities. They do so by combining information about which topics their employees would like to learn about with data on courses that have been recommended by other staff members, together with data from previous course registrations and on satisfaction. Normally, the aim is to develop an appropriate course for the largest number of learners and thus develop programmes that reflect the needs of as many people as possible. However, data collection can also take individual learning styles into account; hence, data on learning preferences increasingly form part of decisions on course materials. Once learning styles have been clearly identified (such as preferences for group-learning/working alone; preferences for practical or theoretical learning, text or video-based learning; learning alone or with a teacher, etc.) then data can also be collected on both the group as a whole, and on individual learners.
Programmes that ‘learn’ in this manner can enable tasks to be increasingly tailored to the needs of individual learners. These technical opportunities represent a revival of the age-old longing for a ‘learning machine’. Perhaps one day a data-fed computer program will exist that simply spits out the right form of material for each person and even sets it out in the right type of task.
The importance of the human factor
This quite mechanistic view of learning provides substantial advantages when it comes to developing tailor-made learning material. The purely technical aspects of learning, therefore, have much to gain in this respect. However, learning is rarely aimed at solely mediating academic content. As other trends demonstrate, social skills are playing an increasingly important role in learning due to their importance if people are to lead a successful and fulfilled life. Therefore, digital courses that are based on learning data are particularly fruitful when they also involve interaction with teachers and coaches who concentrate on improving social skills and personal development.
Clearly, the system described above also raises questions about data protection, especially when the focus moves from storing and analysing data about a group’s learning rhythm, to that of the individual. In many circumstances, the issues related to ‘transparent learners’ have yet to be resolved, and are particularly troublesome when it comes to children. Data on lifelong learning could also be used to provide a seemingly objective forecast of a learner’s chances of academic and professional success.
In short: the sheer volume of data that is currently available on learning and learner’s behaviour is enabling completely new approaches to be developed. These developments are making it far easier to continuously improve materials and lessons and to adapt them to the needs of both the group as a whole and to individual learners. Clearly then, as long as we keep data protection in mind and understand that human interaction constitutes an essential aspect of the learning process these systems could provide great opportunities in the future.