WF-11 — Academic Integrity, Appeals and Complaints

This workflow has no OBEF yield and this document is not going to pretend otherwise. There is no integrity KPI, no appeals KPI and no complaints KPI in OBEF v11.5. Manufacturing a mapping would be the kind of thing that makes a workflow pack untrustworthy, and a pack that overstates one workflow invites doubt about the twenty-five others. What this workflow does instead is decide the classification of a departure that KPI 2.2 counts, protect the item integrity that KPI 2.1's rubric tests, and close the gap that would otherwise sit in the middle of the learner lifecycle. A process architecture with holes in it is not an architecture.

1. Purpose

To ensure that allegations of academic misconduct, appeals against academic decisions, and complaints about teaching, supervision, assessment or service are received, investigated and decided fairly, within published timescales, by people who did not make the decision under challenge, and that the outcome of each case is recorded in a form that other workflows can rely on: a departure correctly classified for KPI 2.2, an assessment item correctly withdrawn or retained for KPI 2.1, and a service failure correctly fed back into the programme.

2. Scope

Scope statement. This process manages academic integrity cases, academic appeals and learner complaints from first report or submission, through investigation and decision, to closure, remedy and the feeding of outcomes into the workflows that depend on them.

Applies to. All PGDipHPE and MScHPE learners; all IoL assessments including dissertations, capstone work and workplace-based assessment; complaints about teaching, supervision, placement supervision, learning resources and IoL service; and appeals against progression, assessment and award decisions.

Does not apply to. Staff grievance and disciplinary processes, which sit with MBRU HR; research misconduct by staff, which sits with the research integrity route in WF-15; whistleblowing on institutional matters; and the routine remarking or moderation of assessment, which is WF-09. Where a complaint concerns a placement supervisor rather than IoL, this workflow triages it and hands it to WF-13 and WF-24 while retaining the learner-facing record.

Applicable requirements. MBRU academic integrity policy, academic appeals policy and student complaints policy; CAA requirements for grievance, appeal and integrity processes; ISO 21001, which requires complaints and appeals processes to be established, documented and operated within the scope of an educational organisation management system; OBEF guide v11.5 KPI 2.2 counting rules and the HEDB Dismissal Reasons lookup; data protection obligations covering case files, which contain some of the most sensitive personal data IoL holds.

Why it is in the pack, stated plainly. Three reasons, none of them a KPI. First, it is the workflow that decides whether a departure enters Students - Attrition.xlsx as a dismissal, a voluntary withdrawal or a suspension, and those are counted differently by KPI 2.2. Second, assessment security and item integrity are what make the KPI 2.1 item-recycling and reliability criteria meaningful, since an item that has leaked is compromised whether or not it has been formally reused. Third, ISO 21001 requires it, and the pilot is testing whether an ISO-shaped process architecture generates OBEF data as a by-product. Leaving a required process out because it does not score would answer that question dishonestly.

3. Trigger, boundary and endpoint

4. SIPOC

5. Accountability

Process owner. IoL Programme Director for first-instance cases, escalating to the IoL Senior Director and thereafter to the MBRU-level appeal body, so that no person hears an appeal against their own decision.

Escalation.

6. Process steps

1. Receive and log. Every allegation, appeal and complaint is logged on the day it is received, with a case reference, the date, the type and the parties. [CAPTURE] case record. Serves no KPI. Serves ISO 21001, CAA and the learner's right to a traceable process.

2. Triage and classify. Integrity, appeal or complaint; academic or service; first instance or review. Decision point. If the person who would normally handle the case has any involvement in the decision under challenge, hand it to an independent decision-maker before anything else happens. In a department the size of IoL, the pool of people with no involvement is small, and independence must be arranged deliberately rather than assumed.

3. Notify and explain. Tell the learner what has been alleged or received, what the process is, what the timescales are, what support is available and what the possible outcomes are. In a professional postgraduate cohort, an integrity allegation carries potential regulatory and employment consequences well beyond the programme, and the learner should be told that at the start, not discover it at the end.

4. Investigate. Gather evidence, including the assessment, its blueprint, its rubric and, where the allegation concerns an item, the item's reuse history from WF-09. [CAPTURE] evidence log with dates and sources.

5. Hear or consider. Give the learner a genuine opportunity to respond. Record what they said, not only what was decided.

6. Decide and record reasons. [CAPTURE] decision, reasons, decision-maker, date. A decision without recorded reasons cannot be reviewed, cannot be defended and teaches the department nothing.

7. Communicate the outcome and the right of further review. [CAPTURE] date of communication.

8. Apply the sanction or deliver the remedy. For upheld complaints, the remedy is the point of the process. Record what was actually done, not what was promised.

9. Classify any resulting departure for WF-10. [CAPTURE] the classification input: Attrition_Category and Attrition_Reason drawn from the HEDB Dismissal Reasons, Voluntary Withdrawal Reasons or Suspension Reasons lookups as appropriate. Serves 2.2, indirectly. Decision point, and the only point in this workflow with a direct numeric consequence. A learner dismissed following an upheld integrity case is attrition. A learner who withdraws voluntarily while a case is open is a voluntary withdrawal, which is also attrition but under a different reason. A learner suspended pending investigation and subsequently reinstated is not attrition, and must remain in both the KPI 2.2 numerator and denominator. These three outcomes look similar in a case file and are counted differently, which is why the classification is made here, by the people who know what happened, and recorded before the case is closed.

10. Withdraw or retain the affected assessment item. Where an item, paper or task has been compromised, withdraw it, record the withdrawal in the item-reuse register maintained by WF-09, and require a replacement. Serves 2.1 criterion 2, on reliability and on items not being recycled. An item that has circulated is compromised whether or not it has been formally reused, and continuing to use it fails the reliability sub-question as surely as recycling it fails the recycling sub-question.

11. Report themes to annual programme monitoring. Aggregate, anonymised, with numbers and cohort sizes. Serves WF-04, and indirectly 2.6, since complaint patterns and satisfaction patterns usually describe the same problem from two directions.

12. Close the case. Retention clock starts; access to the file restricted to those with a continuing need.

7. OBEF data generated

This workflow generates no OBEF numerator, no OBEF denominator and no CHEDS submission field. The table below records what it hands to other workflows, which is a different thing, and it is presented as such rather than dressed up as OBEF data generation.

Capture rule. Two things must be recorded at the time and cannot be reconstructed afterwards: the reasons for the decision, because a decision remembered is not a decision evidenced and an appeal months later turns entirely on what was written at the time; and the departure classification while the facts are fresh, because a case file six months old rarely makes clear whether a learner was dismissed, withdrew during proceedings, or was suspended and never returned, and KPI 2.2 counts those three differently. Everything else here can be reconstructed from the file.

Reproducibility test. The test does not apply in its usual form, because this workflow produces no KPI value for a second analyst to reproduce. The equivalent question is: can a reviewer, from the case file alone and without asking anyone, establish what was alleged, what was decided, on what evidence, by whom, and what followed? For an integrity case that ends in a departure, that same file must also make the attrition classification obvious. If it does not, KPI 2.2 rests on somebody's recollection, which is precisely the defect this pilot exists to remove.

8. Max-score design

There is no score to maximise here, and the table below says so rather than leaving the section empty.

What this workflow must do to protect the target.

1. Classify every case-driven departure at the point the case closes, using the controlled HEDB vocabularies, and hand it to WF-10 rather than leaving WF-10 to infer it. Owner: Data Steward with the Programme Director.
2. Distinguish suspension with reinstatement from withdrawal and dismissal, since only the latter two are attrition. Owner: Programme Director.
3. Withdraw compromised items and record the withdrawal in the item-reuse register. Owner: Quality and IQA with WF-09.
4. Arrange independence deliberately in a small department, and record who was excluded and why. Owner: Programme Director.
5. Decide within the published timescale, or record an extension with a reason given to the learner. Owner: Programme Director.
6. Report themes annually with numbers and cohort sizes, so WF-04 sees the pattern. Owner: Programme Director.

[REDESIGN] actions.

The template asks for design changes that win points. Two of the four below win no points at all and are listed because they close real risks; the first two protect points that other workflows produce. That distinction is stated rather than blurred.

Sequencing note. R1 and R2 are free, immediate, and are the only two entries in this table with a point value attached to them. R3 and R4 are here because the workflow would be incomplete without them, not because they score.

9. Indirect strategy where data cannot be collected

None of the five categories applies in the scoring sense, and asserting one would be the wrong answer.

The five categories in the charter are responses to a situation where a KPI exists but its data cannot be collected. Here there is no KPI. Category 1 does not apply because the Ministry holds no instrument for integrity or complaints. Category 2 does not apply because there is no population to sample. Category 3 does not apply because the activity plainly exists; what does not exist is an indicator that measures it. Category 4 comes closest, in that the activity is genuinely invisible to OBEF, but Category 4's sanctioned response is capture so that the activity counts, and here capture will never make it count because there is nothing for it to count toward. Category 5 does not apply because there is no reported performance to be low.

So what does this workflow protect instead?

It protects the integrity of a number other workflows produce. KPI 2.2 is the clearest case. The difference between a dismissal, a voluntary withdrawal and a suspension with reinstatement is a classification decision made inside a case file, and it moves programme-level retention by several points in a cohort of twelve. WF-10 owns the number; this workflow owns the fact.

It protects the evidence base for a KPI IoL cannot submit. KPI 2.1's reviewers assess whether a course uses methods to check accuracy and reliability, and whether items have been recycled. An assessment estate with an unmanaged leak has neither reliability nor a clean reuse history, whatever the register says.

It protects the pilot's own credibility. The charter draws a line between designing activity so that it qualifies and counting activity that does not meet the stated criteria. A pack that invented an OBEF mapping for its integrity process would be standing on the wrong side of that line in its own documentation. The honest statement, that this workflow yields nothing and is included anyway, is worth more to the pilot than a fabricated contribution would be.

And it satisfies a requirement that is not negotiable. ISO 21001 requires complaints and appeals processes within the scope of the management system. CAA requires them. A learner's right to challenge a decision that affects their qualification and their professional standing does not depend on whether a framework counts it.

Boundary check. This workflow must never:

• claim an OBEF contribution it does not have, in this document or in any committee paper derived from it;
• classify a departure to produce a more favourable KPI 2.2 value, or delay a classification until the effect on the rate is known;
• record a suspension with reinstatement as attrition, or a dismissal as anything other than attrition;
• allow a person who took the decision under challenge to investigate or decide the challenge to it;
• close a case without recorded reasons, or communicate an outcome without them;
• continue to use an assessment item known or suspected to have circulated, on the grounds that no formal reuse has occurred;
• let a complaint about a placement supervisor disappear between IoL and the host organisation without a learner-facing record and a named owner;
• use case files for any purpose other than the case, given the sensitivity of the personal data they hold.

10. Service standards

11. Records and evidence

Appendix B readiness. The Ministry cannot demand this workflow's records under Appendix B, because no KPI depends on them. That is not a reason to hold them loosely. CAA and ISO 21001 both can and will ask, and the realistic assessment is that IoL could produce individual case files today but could not produce a consistent picture of how many cases arose, of what type, how long they took and what happened to the learners afterwards, because no case register exists. That is R3 and R4, and neither is worth a point.

12. Risks and controls

13. Performance measures

The third and fourth measures are the ones that matter to the rest of the pack. The others matter to the learners, which is a better reason.

14. Change control