: Data, Algorithms and Harmony A sixth form college perspective

Feverishly chasing curriculum excellence can result in an outcomes driven, high pressure, high stakes environment that is not healthy for students and teachers who see every assessment, homework and interaction as a ​‘cliff edge’ that seals their fate. There is great skill in getting the balance right. Oldham Sixth Form College Vice Principal, Mark Crilly, outlines the implementation process adopted to gain a more efficient, timely and impactful way of working to provide targeted support and enhance their students’ experience.

Recommendation 3 Use a structured but flexible implementation process

To drive improvement while ensuring both powerful pastoral support and curriculum excellence requires balance. An overemphasis on support can lead to a lack of student ownership and resilience, while a relentless focus on outcomes can create a high-pressure environment detrimental to both students and teachers. Achieving this balance is a formidable challenge.

When I arrived at the college in 2016 it struck me that something unusual exists at OSFC – both staff and students appear to be happy. As I progressed through a range of leadership roles it became clear that the culture was positive, leadership was strong and empathetic, teaching was consistently effective and results and progression rates were high. As with any large organisation, there is always room for improvement, but any changes implemented would have to take place against this backdrop and should only serve to enhance rather than disrupt. That became the key challenge – how do we do better by improving efficiency and accuracy but keep what we’ve got? The focus had to be on radical but measured change.

As with schools, Sixth Form Colleges collect huge amounts of data about students’ performance around attendance, punctuality, assessment, likely outcomes and perceived effort and use this to help construct a picture of the student, supplemented by other pieces of information such as prior attainment, health issues and pastoral log entries. But getting it right and on time was proving to be a work intensive and inconsistent manual exercise to categorise students, determining who was at risk, who needed intervention and at what level – a very subjective process. The data itself was also sometimes flawed – in some cases it was based on perception rather than something more concrete, for others predictions based on cumulative assessment scores rather than point-in-time assessments. There had to be an easier way and it had to develop organically to avoid shocks and sudden lurches.

By combining data relating to effort, actual assessment and predicted outcomes we grew more adept at spotting trends and risks, seeing impressive interventions from Progress Tutors and Pastoral Leaders as a result. But there remained a missing component – how do we use all this data at once, in one go, to provide a risk profile for each student? We were doing this manually, why not automatically?

In trying to identify which students needed support and intervention we needed to develop a bespoke risk algorithm that would turn our data into something more usable and intuitive. We settled on risk scores that could be used to categorise levels of concern which would assist curriculum and pastoral teams in designing relevant interventions.The higher the score, the higher the risk.

So, a student that had very low effort grades, expected grades and attendance would have a high risk score and would become a key priority for swift intervention. From these scores we drew out 5 distinct categories associated with their full programme with each category allocated a distinct action:

We are now able to analyse the same data using a similar approach for courses, even to the extent of helping direct attention to those who most need it – our High Effort, Low Progress learners (HELPs) who are fully committed but need help with specific skills or deficits.

Categorising in this way has had a profound impact on how data is used to report to SLT and Governors as we can now map risk and category movement over time. We can see the impact of intervention on learners and how that changes their risk profile. We can also see the effect of the risk category on likely outcomes, especially whether students stay the course.

Our innovative use of data has yielded significant benefits. We have reduced staff workload, improved efficiency, enhanced monitoring and reporting, and ensured timely support for students. The high buy-in from staff, students and parents reflects the success of our approach. Moving forward, we will continue to explore how data can inform behaviours and share our findings with other institutions facing similar challenges. By aligning our efforts with the EEF Implementation Guidance Report, we have effectively addressed a common sector-wide issue, demonstrating the potential for the utilisation of data to inform improvement in complex educational settings.