Abstract:

Working with teachers and students during the Primary research phase (staff and student interview) of RMIT University’s Classroom of the Future initiative, one of the major insights reinforced was that strong student engagement with learning material, teachers and each other is a lead indicator of student success and, obversely, weak engagement is a lead indicator of a student ‘at risk’. Student engagement (as an audience) is also valuable feedback to teachers about the effectiveness of their learning delivery and content, uncovering opportunities for improvement. There was also expressed strong links between student engagement, authentic assessment and authentic action.

The main, high level insights were, from students, “I expect you to know what I want before I do” and from teachers and administrators, “I haven’t a full picture of what my students are doing or how they are coping”.

Most teachers have a keen sense of student engagement in a class or in one subject but the challenge is to apply an objective, holistic and baselined analysis of student engagement across a whole programme. A holistic view of student engagement is also often compromised because each teacher’s view of a student’s engagement is not always captured in a CRM where case officers or student success staff can utilise it.

Leveraging these insights and RMIT’s tech partnership with Amazon Web Services (AWS), a number of innovations were ideated including a Student Engagement Analysis tool and a Collaboration Effectiveness Analysis tool. This presentation illustrates the Engagement tool.

A team of five RMIT Computer Science students, as their final, ‘capstone’ assignment, was tasked to use AWS’ biometrics products to prototype an automated Engagement Analysis tool integrated into RMIT’s technology stack via API to the scheduling system (Allocate Plus) and CRM (SalesForce HEDA/Einstein).

The core deliverables of the tool were:
a] monitor class attendance
b] analyse student engagement
c] give teachers feedback to help optimise content and delivery and
d] identify opportunities for intervention for at-risk students. An algorithm would be written to calculate student ‘risk scores’ for the CRM

The vision was that, via a teacher’s live web portal in-class, students would be identified by a screen marker showing their attendance status and 'Risk Score' (red, amber, green). After class, teachers would be given an aggregated, whole-of-class engagement trace (graph) that could be matched against a teaching session to show where improvements could be made to content and delivery. Additionally, individual students could be compared to the whole-of-class graph to highlight discrepancies i.e. above or below the ‘mean’. If a configurable threshold above or below the mean was crossed cases would be automatically raised in the CRM for review. While the prototype was designed for in-class it should also be possible to use it with remote ‘dial-in’ students

The solution offers a unique, integrated platform to reduce the amount of data mining teachers and support staff must do to gain insights, support personalisation of student learning journeys and improve teaching and learning outcomes

The team successfully delivered a small-scale prototype and the aim of my presentation is to share the background, outcomes and simple technical architecture of the tool, the potential benefits of further development and some discussion about the tool’s impact on privacy and ethics

Keywords:

Biometric, machine learning, artificial intelligence, AI, engagement, student success, student experience, personalisation, personalization, authentic assessment.