Course Evaluation Evidence for Architecture Programme Accreditation (RIBA, UIA, EAAE)

Architecture programmes sit inside one of the most layered accreditation systems in higher education, and student course evaluation is explicitly named as evidence in almost every layer of it. To satisfy a RIBA/ARB visiting board, a UNESCO-UIA validation panel, or an institutional review aligned to the European Standards and Guidelines (ESG 2015), you need standardized, longitudinal evidence that you collect student views on teaching, analyse them, act on them, and close the loop back to students. This guide maps those requirements to concrete outputs you can generate with a modern evaluation platform, and is honest about where a traditional survey tool is still the pragmatic choice.

Architecture is also where the ordinary student evaluation of teaching (SET) survey struggles most. The signature pedagogy of the discipline is the design studio and the crit (or jury), a formative, dialogic, workload-heavy experience that a five-point Likert grid captures poorly. Panels know this, and the strongest submissions pair quantitative response data with rich qualitative evidence about studio culture, feedback quality, and workload. That is exactly the gap an AI-moderated interview is built to fill.

The accreditation layers an architecture programme must satisfy

Unlike most disciplines, architecture is validated at several levels at once, and each level asks for student-feedback evidence:

• Professional/registration validation (RIBA + ARB in the UK; national registration boards elsewhere). RIBA validation is a peer-reviewed process run by visiting boards of practising architects, academics and architecture students. RIBA lists among the recognised methods for evaluation and enhancement of quality "listening and responding to the views of students," alongside programme team meetings, examination boards, annual monitoring reports and student module feedback forms. The 2021 Procedures for Validation and the Themes and Values framework structure knowledge, skills and understanding across RIBA Parts 1, 2 and 3.
• The UNESCO-UIA validation system. The UNESCO-UIA Validation Council/Board operates against 16 validation criteria: the 11 points of the EU Directive for architects plus 5 additional UIA criteria covering matters such as heritage, conservation and sustainable design. It is explicitly an evidence-based interrogation of academic standards and learning outcomes.
• EU Directive 2005/36/EC (as amended by 2013/55/EC). Article 46 sets the minimum training for architects for automatic recognition of qualifications across the EU/EEA: at least five years of full-time university study, or at least four years full-time plus a supervised traineeship, delivering 11 knowledge and ability points (Annex V.7 lists the qualifications). While the Directive governs curriculum content rather than survey mechanics, institutions use student evaluation to demonstrate that those abilities are actually being taught and experienced.
• Academic/institutional accreditation (ESG-aligned national agencies and the EAAE frame). The European Association for Architectural Education (EAAE) provides a disciplinary community and charter, while ESG Standard 1.9 (on-going monitoring and periodic review of programmes) and Standard 1.3 (student-centred learning) require systematic collection and use of student feedback.

The practical consequence: architecture schools are asked for the same underlying artefact repeatedly — a defensible, standardized trail of collect - analyse - act - report - close the loop — and re-purpose it for RIBA, UIA and their national agency in turn.

Requirement-to-output mapping

The table below maps common architecture-accreditation evidence requirements to the outputs a platform like Koji produces. The point is not that software satisfies a standard on its own — panels judge the academic substance — but that it removes the manual assembly work and makes the evidence consistent across cohorts and campuses.

Accreditation requirement	What panels want to see	Koji output that supplies it
RIBA: "listening and responding to the views of students"	Evidence you collected views on each module/studio and acted	Standardized per-module reports + closing-the-loop action log with dates and owners
ESG 1.9 annual monitoring and periodic review	Longitudinal, comparable data across cohorts and years	Cohort-over-cohort trend reporting on the same standardized instrument
ESG 1.3 student-centred learning	Feedback on assessment, feedback quality and workload	AI-moderated interviews that probe crit/jury feedback, workload intensity and studio culture in students own words
UNESCO-UIA / EU Directive 11 ability points	Assurance that Directive abilities are experienced by students	Thematically coded qualitative evidence mapped to each ability/learning outcome
Self-evaluation report (SER/SAR) narrative	Synthesised evidence, not raw exports	Auto-generated thematic summaries with representative anonymised quotes
Site visit / student meeting corroboration	Consistency between reported feedback and student panel	An auditable data trail that matches what students tell the visiting board
Data protection due diligence (GDPR/nFADP)	Lawful, minimised, well-governed processing	EU data residency, a data processing agreement, retention controls and sub-processor transparency

Why standard SET falls short in architecture (and what to do about it)

Three features of architectural education break the standard Likert survey:

1. Studio is formative and continuous. Learning happens through iterative design conversations, not lectures, so "the lecturer explained concepts clearly" barely applies. Panels want evidence about the quality and fairness of studio feedback and the crit.
2. Workload is a chronic, well-documented issue. Architecture consistently reports some of the highest student workloads in higher education. A single 1-to-5 workload item hides the texture accreditors actually probe. Conversational follow-up ("what specifically drove the crunch, and when?") produces evidence you can act on.
3. Qualitative comments carry the signal, and there are too many to read. Open-text boxes generate exactly the material a panel values and exactly the material a manual process cannot analyse consistently. Automatic thematic analysis standardises the reading.

Koji addresses these by replacing the static form with a short AI-moderated interview that adapts to each student, probes studio experience, feedback and workload with neutral standardized follow-ups, and then performs automatic thematic analysis so every comment is coded the same way across studios, cohorts and campuses. Because the moderation is standardized, it also mitigates (it does not eliminate) the inconsistency of who happens to be asking. The same conversational interview engine underpins the wider Koji platform at koji.so for general user and customer research, so the qualitative rigour is not education-specific window dressing.

Building an audit-ready evidence trail

A submission-ready architecture evidence pack typically contains:

• Standardized instruments used across all studios and taught modules, so RIBA/UIA panels can compare like with like.
• Response and coverage data per module and cohort, with honest commentary on response rates (architecture cohorts are small, so segment carefully to protect anonymity).
• Thematic summaries tied to Directive ability points and RIBA/UIA criteria, with representative anonymised quotes.
• A closing-the-loop log: what students said, what the programme changed, and how that change was communicated back — the single most persuasive artefact for any visiting board.
• A data-governance appendix: lawful basis, retention schedule, anonymity thresholds, and the DPA/sub-processor list.

When a traditional tool is still the right choice

Honesty matters to this audience, so be clear about the boundaries:

• If your school already runs EvaSys, Explorance Blue or a Qualtrics instance that your quality office trusts and your accreditors have accepted, and your only gap is administrative, ripping it out for a validation cycle you are mid-way through is rarely wise. Add conversational evidence where it is weakest (studio and workload), rather than replacing everything at once.
• If you need institution-wide statistical benchmarking across thousands of modules with mature LMS/SIS integration today, incumbent enterprise suites are strong at exactly that.
• If your immediate need is a simple, free anonymous form for a single studio pilot, Microsoft Forms or LimeSurvey will do the job — you will simply do the analysis and closing-the-loop yourself.

Koji's advantage is concentrated where architecture accreditation is hardest: turning messy, high-volume, studio-centred qualitative feedback into standardized, longitudinal, panel-ready evidence with the loop demonstrably closed.

Frequently asked questions

Does RIBA or the ARB require a specific course evaluation tool?

No. Neither RIBA validation nor ARB prescribes a product. They require evidence that you systematically collect student views, respond to them, and feed the results into monitoring and review. Any tool that produces a defensible collect-analyse-act-close-the-loop trail is acceptable; the value of a modern platform is consistency and reduced manual assembly.

How does course evaluation map to the EU Directive 2005/36/EC ability points?

The Directive defines the abilities an architecture graduate must have; student evaluation does not prove competence on its own, but panels use it to confirm that those abilities are actually being taught and experienced. Coding qualitative feedback against each ability point (and against RIBA/UIA criteria) gives you traceable evidence to cite in the self-evaluation report.

Our studios are small — how do we evaluate without breaking anonymity?

Small cohorts are the central risk in architecture evaluation. Aggregate to a level that protects identity (for example, at year or programme level rather than a five-person studio), suppress cells below a threshold, and rely on thematic summaries rather than verbatim quotes where a student could be identified. A platform with built-in anonymity thresholds makes this defensible.

Can conversational AI evaluation replace the design crit or jury?

No, and it should not try to. The crit is a formative assessment and teaching event. Course evaluation is a separate, systematic mechanism for gathering student views on the experience of studio, feedback and workload. Used together, the crit develops the student and the evaluation evidences the quality of the programme for accreditation.

Is AI-moderated evaluation compatible with GDPR and Swiss nFADP?

Yes, when configured correctly. Use EU data residency, a signed data processing agreement, data minimisation, a clear retention schedule and transparency over sub-processors. Because Switzerland holds a mutual EU adequacy decision, EU-hosted processing needs no additional transfer safeguards for Swiss schools of architecture.

How is this different from just reading our open-text comments?

Manual reading is inconsistent and does not scale: two readers code differently, and no one reads every comment in a large cohort. Automatic thematic analysis applies the same coding across every studio, module and year, which is precisely the standardization a validation panel looks for when it tests whether your quality system is systematic rather than ad hoc.

Related resources

• Course evaluation evidence for EUR-ACE engineering accreditation
• Course evaluation evidence for PSRB / professional-body accreditation
• Student feedback as ESG accreditation evidence
• Writing the self-evaluation report with course evaluation evidence
• Annual programme monitoring: turning feedback into evidence
• Institution-level evaluation reporting for QA audits