Data in the classroom: separating hype from reality

Data in the classroom: separating hype from reality

Making sense of data in the classroom can promise personalised education, streamlined teaching and easy student evaluation, but should schools pin their hopes on these automated systems?

Walk into a classroom today, and chances are it looks fairly similar to one from 50 years ago. There might be more technology – laptops and tablets are norms in schools now – but the basic format remains the same: chairs and desks facing a teacher.

That’s because, while automation has aided jobs and industries already, education has failed to keep up with recent rapid technological change, according to the head of UNESCO’s International Bureau of Education Mmantsetsa Marope. And it needs to.

But the tide may be about to turn. A 2018 report by market research company Technavio predicted the global market for artificial intelligence (AI) that harnesses the power of massive amounts of data collected in education will rise at an annual growth rate of 43 per cent until 2022.

So what will this shift towards assimilating technology into education and making use of data look like?

Professor Dragan Gasevic, from Monash University’s Centre for Learning Analytics, is currently working on the ‘Data Smart Schools’ project, which is a three-year research project funded by the Australian Research Council. This project aims to investigate the use of digital data in schools and identify ways to improve its capture and use.

Masses of digital data are generated within schools every day, but despite its potential, this data remains poorly used and understood. As part of the ‘Data Smart Schools’ project, Professor Gasevic and his colleagues  work with school communities to develop innovative data tools and techniques to make more effective use of their data.

The project outcomes are expected to include insights into the technical, informatic, organisational and social issues surrounding the (re)use of digital data in schools – then develop models of digital data ‘best practice’, leading to improved student outcomes, stronger teacher and parent engagement and better use of technical infrastructure investment.

It’s not only for the kids

Making use of data in education can also make life easier for teachers by taking care of time-consuming grunt work and administration tasks. Take automated marking assignments, for instance – that’s been around for nearly a decade now. Software and appropriate data collection can free up time for teachers to concentrate on, for instance, kids that need one-on-one help. Predictive models for identification of students at risk is another common way of using AI to analyse data in schools to produce early warning systems for school such as BrightBytes School Data Dashboard.

We’ve had Siri and Alexa to answer our questions for years now; and now AI can do the same in the classroom. This has already happened at the university level. Students taking the online Knowledge-Based Artificial Intelligence course through Georgia Tech in the US in 2014 unwittingly had their questions answered by a natural language processing system. The AI, named “Jill Watson” – a homage to IBM’s Jeopardy!-winning computer Watson – was trained on 40,000 questions and answers from previous semesters’ students. Built on similar principles as Jill Watson, Let’s Talk! Assistant is a AI-driven chatbot built for schools

The foundations of spatial analysis and pedagogy are also being used in educational settings. A group of researchers, led by Senior Lecturer Dr Roberto Martinez-Maldonado, from the Faculty of Information Technology at Monash University, have analysed the positioning patterns of teachers within a number of learning environments to develop an advanced AI system, called Moodoo, which determines the best teaching positions for educators.

The research analysed data from seven teachers wearing indoor positioning trackers and delivering three distinct types of classes to over 190 students in the context of physics education.

“Our results showed that by using Moodoo, wearable trackers revealed the kinds of learning tasks performed by students in the classroom and the appropriate positioning approaches used by teachers. These findings will create a foundation to help train novice teachers to use classroom spaces effectively or to assess the impact of the spatial design on teaching and learning outcomes,” said Dr Martinez-Maldonado.

Preparing students, regardless of resources

Cash-strapped schools don’t have to be left behind; there’s an ever-increasing pool of free apps and tools, such as Machine Learning for Kids. Like any new technology, the latest, most sophisticated software will be expensive, but will eventually drop.

These are just a few ways technology is already integrating into education. But while it may sound like technology will be an indispensable part of every student’s learning experience – not to mention the teacher’s professional life – it’s not a panacea for all educational ills. Hidden in the hype is the reality. Many of these technologies are still under development with kinks yet to be ironed out.

Professor Gasevic and his team of researchers have found that before rolling out new software across a school, it’s important to understand the influence of that technology on a classroom.

“Over the last several years and with the current COVID-19 pandemic, the role of technology in educational settings has increased significantly. From the widespread use of learning management systems to social media, interactive simulations, and learning games, the growth in technology has propelled the capacity for capturing data.

“Despite learning analytics dashboards – tools that report the results of analysis of data about learners and schools – being frequently developed by education systems and technology vendors with the goal to support self-regulated learning, there is little understanding of how the current generation of learning analytics dashboards is equipped to support the development of self-regulated learning.”

By analysing existing empirical studies about the use of learning analytics dashboards, Professor Gasevic and his team found that existing learning analytics dashboards are rarely grounded in recommendations established in educational research and practice.

“Despite the growing adoption of learning analytics dashboards, there are many limitations in the design of their systems which our research has identified. Particularly, learners and teachers find it hard to interpret the data presented in dashboards and to make use of the feedback presented in dashboards to inform future actions  We see much more promising results with learning analytics-empowered technologies for personalised feedback at scale such as OnTask ” Professor Gasevic said.

Properly developed learning analytics tools and educational technology has many values to both students and teachers in terms of academic success, student satisfaction, skills for self-regulated learning, and teacher efficiency. But additional insights drawn from data within a classroom can also offer appropriate and personalised student learning experiences.