Who Evaluates a Course Taught in Four Countries? Joint Programmes and the Alliance QA Gap

When a single module is co-designed and co-taught by lecturers in Lisbon, Munich, Tallinn and Bologna, whose course-evaluation system governs it — and against whose standards is a "good" score judged? This is no longer a hypothetical. As of the end of 2024 the European Universities Initiative comprised 64 alliances involving more than 560 higher-education institutions, and the Commission's March 2024 blueprint for a European degree is pushing them toward genuinely joint programmes. The teaching is going transnational faster than the evaluation of it. Most course-evaluation tools, and most institutional QA processes, were built to answer to one national regulator, in one language, on one academic calendar. Joint programmes break every one of those assumptions.

The short answer: evaluating an alliance course is not the same problem as evaluating a course on a branch campus, and treating it that way produces feedback that is either incomparable, non-compliant, or both.

Why this is a new problem, not an old one in new clothes

Universities have run cross-border teaching for years, and this blog has examined the comparability problem in transnational and multi-campus programmes. But that earlier case is usually one institution exporting its system to another site — one owner, one instrument, one standard, replicated elsewhere.

Alliance joint programmes are structurally different. There is no single owner. Four or more autonomous institutions, each embedded in its own national quality-assurance regime, jointly deliver one course. Each partner arrives with:

• a different national QA agency and legal framework (NVAO in the Netherlands and Flanders, the Conseil in France, a Länder patchwork in Germany, and so on),
• a different home evaluation instrument, scale, and reporting calendar,
• a different language of instruction and of feedback, and
• a different institutional culture about what student feedback is even for.

Ask the same course in five national instruments and you get five results that cannot legitimately be pooled — the benchmarking trap raised to the power of the alliance. A 4.1 in one partner's system is not a 4.1 in another's, for reasons this blog has documented at length, from scale design to measurement invariance across languages.

The regulatory scaffolding already exists — and is under-used

The good news for quality officers is that Europe has already built the framework for this, even if practice lags. Two instruments matter most.

The European Approach for Quality Assurance of Joint Programmes, adopted by the ministers of the European Higher Education Area at the 2015 Yerevan conference, was designed precisely so a joint programme could be evaluated once, against a single agreed European standard, rather than accredited separately by each partner country. It is anchored in the Standards and Guidelines for Quality Assurance in the European Higher Education Area (the ESG), whose Part 1 already requires that programmes be periodically reviewed with student feedback as evidence — a requirement this blog has argued most universities only partly meet.

The pieces, in other words, are on the table: a shared standard (the ESG), a shared accreditation route (the European Approach), a register of trusted agencies (EQAR), and a coordinating body (ENQA). What is frequently missing is the operational layer — the actual instrument that collects comparable student feedback across the partners in a way the European Approach can use as evidence. That gap is where alliances quietly fall back on stapling five incompatible national surveys together and hoping.

The comparability problem is not only statistical

It is tempting to frame this as a data-harmonisation task: agree one scale, one set of items, translate carefully, and pool. That helps, but it understates the difficulty, because student feedback carries meaning that does not survive naive standardisation.

Reference-group effects mean a student judges a course partly against the others they have taken — so the same teaching earns different ratings depending on the company it keeps in each national cohort. Response styles differ systematically across cultures: acquiescence and extreme-response tendencies vary by country, so identical satisfaction can produce different numbers. And "closing the loop" — the action stage that makes evaluation worth doing — is far harder when the authority to change a course is split across four institutions and no single programme director owns the whole thing.

None of this is an argument against alliances. It is an argument that their evaluation needs to be designed for the structure, not inherited from the single-institution past.

"This is over-engineering — just let each partner run its own survey" — the counterargument

The pragmatic objection is real and widely held: alliances are hard enough to run without inventing a shared evaluation apparatus, so let each partner evaluate its own contribution with its own tools, and let the programme board read across them qualitatively.

This is defensible for loosely coupled alliances where partners mostly teach parallel modules. But it fails for the genuinely joint course — the co-taught, co-assessed module the European degree blueprint is explicitly encouraging. Three problems surface. First, you cannot demonstrate programme-level quality to a European Approach review if your evidence is five non-comparable datasets; the accreditation route the alliance wants to use assumes coherent evidence. Second, the student experience is joint even when the survey is not — a student does not experience "the Munich 40% and the Bologna 60%", they experience one course, and fragmenting the feedback misses cross-partner problems (handover, coherence, assessment load) that only appear at the seams. Third, it entrenches the dual-purpose confusion: national accountability reporting and programme improvement get tangled across five systems with different stakes.

The honest synthesis: light-touch parallel evaluation is fine for parallel teaching, but genuinely joint delivery needs genuinely joint evaluation, built to one standard, in multiple languages, comparable by design.

Where Koji fits

An alliance needs an evaluation layer that is multilingual, standard-aligned, comparable across partners, and GDPR-appropriate wherever the students sit. This is close to a description of what Koji was built to do.

Koji runs one evaluation design across many languages, so partners in different countries answer a comparable instrument rather than five divergent ones — the practical foundation for the measurement invariance that cross-partner comparison requires. Its AI-moderated conversational interviews apply the same standardized, bias-aware moderation to every student regardless of campus, removing the human-moderator and instrument inconsistency that makes national datasets incomparable. Its automatic thematic analysis surfaces the cross-partner, at-the-seams problems — handover, coherence, uneven assessment — that fragmented national surveys structurally cannot see. And its programme- and institution-level reporting is designed to produce exactly the coherent, programme-level evidence a European Approach review expects, with EU-appropriate, GDPR/AVG-compliant data handling throughout.

None of this removes the governance work — agreeing standards and who acts on findings is a human negotiation no tool performs. But it replaces "staple five surveys together" with one comparable evidence base, which is the precondition for everything else.

Alliances are not the only organisations running research across borders and languages; multinational product and customer-research teams face the identical comparability problem. The same AI interview engine powers koji.so for that general cross-border research work.

If your alliance is heading toward joint degrees, the evaluation architecture is worth designing now, not after the first review asks for evidence you cannot produce. See how Koji for Education handles multilingual, comparable course evaluation.

FAQ

What is a European Universities Alliance? A transnational partnership of higher-education institutions funded under the EU's European Universities Initiative. As of the end of 2024 there were 64 alliances involving more than 560 institutions, cooperating on joint teaching, research and, increasingly, joint degree programmes.

Why can't alliance partners just pool their existing course-evaluation scores? Because each partner typically uses a different instrument, scale, language, and reporting cycle, and student ratings carry culture- and reference-group-dependent meaning. Pooling non-comparable data produces averages that misrepresent every partner in them.

What is the European Approach for Quality Assurance of Joint Programmes? A framework adopted by European Higher Education Area ministers in 2015 (Yerevan) that lets a joint programme be quality-assured once, against a single agreed European standard aligned with the ESG, rather than separately accredited in each partner country.

How is evaluating a joint programme different from evaluating a branch campus? A branch campus usually replicates one institution's system elsewhere — one owner, one standard. A joint programme has multiple autonomous owners in different national QA regimes delivering one course, so there is no default system to inherit and comparability must be engineered.

How does Koji support cross-border course evaluation? Koji runs one comparable, multilingual evaluation design with standardized AI moderation across every partner, produces programme-level reporting suited to a European Approach review, and handles data in a GDPR/AVG-compliant, EU-appropriate way — replacing fragmented national surveys with one coherent evidence base.