Borderline policy assurance for a chartered UK university

Following the Office for Students’ publication on bachelors’ degree classification algorithms in November 2025, universities making changes to their classification approaches are now required to evidence that resulting awards remain credible under Condition B4. Where such changes would increase the proportion of higher degree outcomes, providers are expected to be able to evidence this assurance in advance of implementation, with affected providers required to notify the Office for Students by 31 July 2026 where such approaches are intended to continue from September 2026. The Office for Students has indicated that such cases may require independent academic review to support assurance. It was within this regulatory context that external assurance was sought prior to implementation.

The brief

A research-intensive UK university was preparing to introduce a refinement to its degree classification policy affecting students whose overall mark fell close to classification boundaries. The proposed change was intended to provide a more accurate reflection of student achievement, while maintaining consistency with previous cohorts and sector practice.

Given the Office for Students’ stated concerns regarding algorithm design and the credibility of degree outcomes, we were commissioned to provide external assurance in advance of implementation. The purpose of the work was to confirm that the revised policy operated as specified, that its impact on degree outcomes was understood in context, and that the change could be evidenced and explained if subject to regulatory scrutiny.

Our approach

The approved degree classification policy was independently reconstructed and tested using historical student data. The work examined how the proposed approach operated in practice across different classification scenarios and borderline outcomes, rather than focusing on a single route in isolation. This testing was used to check that the policy behaved as described in the approved specification and to identify any differences between the intended and implemented logic. Where interpretation varied, these points were discussed with the institution and resolved through clarification of the policy wording, ensuring that the final version was unambiguous and capable of consistent application. Once the policy specification had been finalised, further testing confirmed that the revised approach operated as intended across all records reviewed, with no discrepancies identified between the written rules and observed outcomes.

In parallel, recent external examiner feedback was reviewed to assess whether comments relating to standards, marking practices, or classification outcomes were isolated to individual programmes or indicative of wider issues relevant to the proposed change. Both structured responses and accompanying narrative comments were considered as part of this review.

Outcomes

• Our independent validation confirmed that the finalised policy operated in line with the approved specification when applied to historical student data. The modelling showed a limited and controlled impact on degree outcomes, consistent with the stated academic rationale and sector norms.

• The review of external examiner feedback indicated that concerns relating to grading standards were limited in scope and did not undermine confidence in the revised classification approach. No evidence was identified to suggest that academic standards or longitudinal comparability would be compromised.

• On this basis, external assurance was provided that the revised policy met the expectations set out under Condition B4 and was appropriate for implementation. This enabled the institution to proceed with confidence at its next examination boards, supported by documented evidence of assurance aligned with current regulatory requirements.

If you'd like support with an independent review to ensure your algorithm is working as intended, get in touch.

We offer three engagement models to suit your institutional context and timeline:

• Full assurance - our complete programme

• Snapshot assurance - validation, borderline review and fairness testing

• Annual assurance subscription - ongoing trends monitoring, fairness tracking, and student communications support

  
  

We can also tailor our support to your specific needs.