Course Evaluation Evidence for an Accreditation Site Visit: Preparing for the Peer-Review Panel

Answer first: at an accreditation site visit, the peer-review panel is not checking whether you run course evaluations — almost everyone does. It is testing whether student feedback forms a credible, documented quality loop: collected systematically, analysed honestly, acted upon, and fed back to students, with an audit trail the panel can inspect and corroborate in interviews. The institutions that struggle are rarely the ones with low scores; they are the ones that cannot show what happened after the data arrived. This guide is framework-agnostic — it applies to NVAO, the UK QAA/TEF, ANECA, ASIIN, HCERES, ANVUR and any other agency operating under the European Standards and Guidelines (ESG 2015).

The site visit in context

Almost every ESG-aligned accreditation follows the same shape:

1. Self-evaluation report (SER). The institution documents how it meets the standards, citing evidence — including student feedback.
2. Desk review. The panel reads the SER and supporting evidence before arriving.
3. The site visit. A panel of peers (academics, a QA professional, a student member, sometimes a labour-market representative) spends one to several days on campus. They interview management, staff, students, and alumni, and request additional evidence.
4. Panel report and decision, usually followed by a follow-up / conditions stage.

Course-evaluation evidence surfaces at every step — but the site visit is where written claims meet live corroboration. If your SER says "we systematically act on student feedback," the panel will ask a room of students whether that is true.

What the panel actually probes

Under the ESG, student feedback maps to three standards in particular:

• ESG 1.3 (student-centred learning). Are students'' voices genuinely shaping teaching?
• ESG 1.7 (information management). Do you collect, analyse and use evaluation data reliably?
• ESG 1.9 (ongoing monitoring and periodic review). Do evaluations feed a real improvement cycle?

Translated into the questions a panel asks:

• Coverage. Which courses are evaluated, how often, and what is your response rate? Can you defend low response rates rather than hide them?
• Analysis. How do you turn hundreds of comments into findings? Who reads them, and how consistent is that reading?
• Action. Show us three concrete changes made because of student feedback in the last two years. Where is the record?
• Feedback to students ("closing the loop"). How do students know they were heard? "You said / we did" is the phrase panels listen for.
• Fairness and data protection. Is evaluation anonymous, GDPR-compliant, and free of small-cohort re-identification risk?

The failure modes that cost institutions

Panels see the same weaknesses repeatedly:

• The action gap. Collection is documented; action is not. There is no traceable line from a specific finding to a specific change.
• The unread comment pile. Thousands of free-text comments exist, but analysis is inconsistent, undocumented, or was never really done.
• Response-rate anxiety. The institution is defensive about low response rates instead of presenting non-response analysis and mitigation honestly.
• No longitudinal trail. The panel asks "did that change work?" and there is no before/after cohort evidence.
• Students who contradict the SER. The written narrative claims a closed loop; the student meeting reveals nobody sees results. This single contradiction damages credibility more than any low score.

Mapping site-visit requirements to concrete outputs

The table below maps what a panel expects to the specific artefacts that satisfy it — and how a modern, AI-native evaluation platform such as Koji produces each one. Traditional survey tools can produce several of these too, usually with more manual effort; the point is the artefact, whoever generates it.

Panel expectation (ESG)	Evidence artefact the panel wants	How Koji produces it
Systematic collection (1.7)	Coverage map: which modules, cadence, response rates with context	Standardized interviews across cohorts; response and participation reporting
Honest analysis (1.7, 1.9)	Documented, consistent thematic findings — not raw comment dumps	Automatic thematic analysis with bias-aware, standardized AI moderation (same probing questions every cohort)
Student-centred action (1.3)	Traceable "finding → decision → change" records	Built-in closing-the-loop action log linking themes to actions
Ongoing review (1.9)	Longitudinal, cohort-over-cohort trend evidence	Longitudinal cohort reporting to show whether a change worked
Closing the loop	"You said / we did" evidence shown back to students	Action tracking exportable into student-facing summaries
Data protection	Anonymity, GDPR compliance, small-cohort safeguards	EU-hosted processing; standardized, anonymised outputs

The decisive column for most panels is action and longitudinal review. Any tool can collect; few institutions can show the loop actually closing over time.

Building the evidence pack

Prepare a concise, navigable pack the panel can follow — not a data dump:

1. A one-page evaluation-cycle diagram. How feedback flows from collection to action to student communication, with owners and timings.
2. Coverage and response summary, with honest commentary on gaps and a non-response note where rates are low.
3. Three to five worked "you said / we did" cases, each showing the original finding, the decision, the change made, and — ideally — the later evaluation showing the effect. These are your most persuasive artefacts.
4. A sample thematic-analysis output demonstrating consistent, documented reading of qualitative feedback.
5. A data-protection note covering anonymity, retention, and small-cohort handling.

Keep it evidence-first and self-navigating: a panel that can find the loop quickly trusts it.

Preparing the student meeting

The student panel meeting is where written claims are corroborated. You cannot coach students, and you should not try. What you can do is make the loop genuinely visible during the year, so that when the panel asks "do you ever see the results of your feedback?", the honest answer is yes. If your closing-the-loop communication is real and routine, the student meeting becomes your strongest evidence rather than your biggest risk.

When a traditional tool is the better choice

Be pragmatic. If your institution already runs a mature, well-governed evaluation programme on an incumbent SET platform (EvaSys, Explorance Blue, Watermark, Anthology) and your loop is genuinely closing — documented actions, student-facing feedback, longitudinal trends — then your existing tool is sufficient for the site visit, and switching immediately before a review adds risk, not assurance. Accreditation panels judge the quality of the loop, not the logo on the software. The case for a modern platform like Koji is strongest when your weakness is exactly what panels probe hardest: thin qualitative insight, inconsistent analysis, and an action trail you cannot evidence.

Why panels triangulate — and what that means for you

A peer-review panel rarely trusts a single source. It triangulates: the self-evaluation report''s claims are checked against the raw evidence, and both are checked against what people say in interviews. Course-evaluation evidence is powerful precisely because it can be corroborated from three directions at once — the documented data, the documented actions, and the students in the room. This is also why fabrication or spin is so dangerous here: an inflated narrative that students contradict does more damage than an honest account of a genuine weakness you are addressing.

Practically, this means your job before a site visit is alignment, not polish. Make sure the story in the SER, the evidence in the pack, and the lived experience of students actually match. If the SER says mid-module feedback changes teaching within the same term, at least some students should be able to describe that happening. If they cannot, revise the claim — do not rehearse the students. Panels are experienced at detecting coached answers, and the attempt itself erodes trust in everything else you present. The most durable preparation is simply running a real, visible feedback loop for the year or two before the review, so that corroboration takes care of itself.

Where Koji fits

Koji is built around the parts of the site visit that most often go wrong. Its AI-moderated interviews produce richer qualitative evidence than a static Likert form; its automatic thematic analysis gives you consistent, documented findings instead of an unread comment pile; and its closing-the-loop action tracking and longitudinal cohort reporting produce exactly the "finding → change → effect" trail a panel asks for. It is EU-hosted for GDPR alignment. Koji runs on the same AI interview engine as the main Koji research platform (koji.so) used for customer and user research — the education product applies that engine to student voice.

Related Resources

• The Self-Evaluation Report (SER): Turning Course Evaluation Evidence into Accreditation-Ready Documentation
• Turning Student Feedback into ESG / ENQA Accreditation Evidence
• Programme Review & Revalidation: Turning Course Evaluation into Accreditation Evidence
• Turning Course Evaluations into NVAO Accreditation Evidence (Netherlands & Flanders)
• Course Evaluation Evidence for the UK TEF and QAA Quality Review

Ready to build a site-visit-proof feedback loop? Explore Koji for Education and see how AI-moderated interviews turn student voice into accreditation-ready evidence.