Design, validation and the SkillForge partnership at Meridian University
Meridian University will launch a Postgraduate Certificate in Data Analytics in five months. The design file consists of a two-page proposal and the partner's marketing brochure. SkillForge, a private training company, will supply six of the ten modules, all marking and all learner support. The contract is fixed-price and says nothing about marker qualifications, moderation, mark schemes or turnaround times. A pilot cohort of 34 learners has just finished. Marks were released two weeks after the university received them, and nine learners have appealed, all citing inconsistent marking between modules. SkillForge's lead marker holds no teaching qualification and has not seen the university's assessment policy. The programme leader says the pilot proves the design works and wants to recruit 300 learners for the September intake. The university's quality office has asked for an assessment of what must be in place before the intake opens, and whether the pilot can be treated as validation.
1. Which design and development requirements apply to this programme, what is missing from the design file, and can the pilot be treated as validation?
Read the worked answer
What 8.3 requires. Design and development must be a controlled process with defined inputs and outputs, reviews at appropriate stages, verification, validation, and control of changes. A two-page proposal plus a supplier brochure meets none of these.
Inputs missing: learner and employer needs, entry requirements, learning outcomes, statutory and accreditation requirements, assessment strategy, delivery model, resource and competence needs, and data from comparable programmes.
Outputs missing: module specifications, assessment criteria and mark schemes, moderation arrangements, learner support arrangements, and acceptance criteria against which the programme can be judged.
Verification versus validation. Verification asks whether the design meets its specified requirements - learning outcomes mapped to assessment tasks, workload calculated, entry requirements matched to content. Validation asks whether the programme as delivered satisfies the needs of learners and other beneficiaries. The pilot is evidence for validation, not validation itself, because no acceptance criteria were set in advance and no review of the evidence has been documented.
What the pilot shows. Thirty-four learners, nine appeals, inconsistent marking across modules. On any reasonable acceptance criterion, validation fails. Releasing the programme to 300 learners now would be a release decision taken without the evidence the standard requires.
Judgement for the quality office. Require a completed design record, a documented design review, a validation report with stated criteria and evidence, and a change-control route for the fixes arising from the appeals - before the intake opens. Delaying is a business decision; the absence of a design record is a nonconformity.
2. Set out the controls the university must place over SkillForge under Clause 8.4, and explain how the nine appeals should be used as monitoring data.
Read the worked answer
8.4 requirements. The university must control externally provided processes so that they conform to requirements: define controls and criteria for selection, evaluation and monitoring; communicate what is required; monitor performance; and retain documented information. SkillForge supplies six modules, all marking and all learner support, so it is delivering core parts of the educational product, not peripheral services.
Control points to add to the contract.
- Competence (7.2 applied to external personnel): marker qualifications or equivalent experience, mandatory training on the university's assessment policy, and records of both.
- Moderation: a defined sample, for example 10% second-marked, with all fails and all distinctions reviewed before marks are released.
- Turnaround: a stated service level, such as marks returned within ten working days of the assessment, with consequences for missing it.
- Mark schemes and assessment criteria: supplied and approved by the university before delivery, not reconstructed afterwards.
- Monitoring: a monthly performance report covering turnaround, appeal rate, moderation outcomes and complaints; a right to audit; and a scheduled review meeting.
- Data protection and academic integrity: ownership of learner data and an incident-reporting route.
Using the appeals as monitoring data. The nine appeals are the first performance data point. The correct response is to investigate the cause - were criteria applied consistently across modules, was marker guidance adequate, was there any calibration? - not to ask the partner to try harder. Findings should feed the approved-provider evaluation, the risk register and management review (9.3), and repeated failure should trigger a defined escalation path that can include suspending new enrolments with that provider.
Judgement. Outsourcing the marking does not outsource the university's accountability for the result.