The rise of data-driven education

Friday, 10 November 2017

The rise of data-driven education

Here at LUCA, we like to use our blog to discover and explore how Big Data can be used in parts of society that have historically been closed off to data science. We passionately believe that data, if harnessed correctly, can be a force for good in society, and you can read more about our Big Data for Social Good initiatives here. In this latest blog, we shall see how such technologies are creating data-driven learning environments in our schools, colleges and universities.

Figure 1 : Schools are becoming increasingly data-driven
Figure 1 : Schools are becoming increasingly data-driven.

The Oxford English Dictionary definition of the word ‘knowledge’ is ‘facts, information, and skills acquired through experience or education; the theoretical or practical understanding of a subject’. Thus, one could define learning as ‘the acquisition of data’, and as students absorb this data, a vast amount of data is created at the same time. In fact, recent developments mean that more data is being created than ever before. Concrete statistics are hard to find, but estimates reveal that at least 50% of classes will be delivered online by 2019. Online learning (e-learning), mobile learning (m-learning), MOOCs and blended learning are current trends that create more data than traditional teaching methods.

Harnessing this data has the ultimate goal of improving the results of students, and there are a number of key applications. For example, data technologies can make the monitoring of student performance easier and more precise. If a student were to take an online exam, staff at the institution could not only see which questions they answered correctly, but also how long they took to read and answer each one. This can provide insights for those setting the exams, and they can adjust the questions to make them more understandable.

The rise of blended learning, where students undertake a mixture of classroom and online course, allows for even more insights. Staff can see how students perform in the two environments and tailor the balance accordingly. Analysis can be incredibly precise, including revealing whether a specific phrase in a textbook is difficult to understand. Thus, teaching methods and learning methods can be improved.

Success and dropout rates can also see the benefits of data science, as shown in the case of Arizona State University. At this American university, learning is digital and customized to the students. Staff can then help students with advice about their strengths and areas they need to work on. ASU has seen student success rates rise 13% and dropout rates have fallen by 54%, a benefit of the extra support that students receive.

Figure 2 : Using data can help reduce dropout rates in universities.
Figure 2 : Using data can help reduce dropout rates in universities.

However, this case study highlights the first limitation of the current technologies. In mathematical and scientific courses, Big Data techniques can easily be applied to the learning process, where exams are assessing quantitative, not quantitative, information. Whether such analysis can work as comprehensively for essay-based subjects remains to be seen. Another significant question is whether the learning process can simply be reduced to a series of numbers. Those weary of Big Data will argue that the relationship between teacher and student is never the same, and that a one-size-fits-all policy based on data could never work.

This is where ‘small data’ enters the picture. According to Martin Lindstrom’s book, small data is ‘The Tiny Clues that Uncover Huge Trends’. When applied to the education system, this would involve reducing national or international census-based assessments to the necessary minimum (which still maintaining anonymity). Thus, analysis of the data can be even more insightful for each individual school, as you can see the causation of a particular trend. We expect to see an increasing number of 'small data' developments in the coming months and years.

One of the potential worries of a data-driven education system is likely to be a lack of privacy. However, at LUCA, we always work with anonymized and aggregated data sets, and believe that the same standards should (and will) be applied to the education system. We are excited to see how Big Data can be used to improve this area of society, are you?

No comments:

Post a Comment