France's Course-Evaluation Mandate and the HCÉRES Process Model

If Italy is Europe's clearest example of standardising the instrument, France is its clearest example of standardising the process. The contrast is instructive - and it changes what a regulator ends up looking at.

The short answer (BLUF): French law requires every higher-education institution to run student evaluation of teaching (évaluation des enseignements par les étudiants, EEE), set out in the arrêté du 22 janvier 2014 that fixes the national framework for licence and master degrees. But the law prescribes no national questionnaire: institutions design their own. The national evaluation agency, HCÉRES, then checks - during periodic programme review - not the scores themselves but whether the process exists, whether results reach the programme's advisory board (conseil de perfectionnement), and whether they demonstrably feed pedagogical change. France thus audits uptake and closing-the-loop, where Italy standardises measurement. Each design fixes the other's blind spot and inherits its own. For anyone building a course-evaluation regime, France's model reframes the central question from "is the number comparable?" to "did the feedback change anything?"

What France actually requires

The legal anchor is the arrêté du 22 janvier 2014, which sets the national framework (cadre national des formations) for licence, licence professionnelle and master programmes. Among its provisions, it obliges institutions to organise regular student surveys and to put in place a mechanism for the evaluation of teaching by students, with the explicit aim of continuous improvement of programme quality. The obligation is on the establishment to run a process - not on the state to supply a form.

That distinction is the whole design. Because no instrument is centrally prescribed, French universities vary widely in how they collect feedback. What makes the system coherent is not a shared questionnaire but a shared accountability object: the HCÉRES evaluation framework (référentiel) for programmes requires that the formation "organises the evaluation of teaching and training by students and takes into account the results of this evaluation in its evolution." Evaluators are instructed to look for multi-year data, documented examples of pedagogical adjustments resulting from student feedback, and the integration of those results into governance - the conseil de perfectionnement in particular.

The conseil de perfectionnement: where the loop is supposed to close

The French model leans heavily on the conseil de perfectionnement - a programme advisory board that includes students and external members (often employers). Under the HCÉRES framework, this board must have "all useful documents to conduct its reflections," explicitly including student origin and enrolment, success rates, graduate outcomes, and the results of student evaluations of teaching. Its meetings have specified frequency and agendas, and its minutes are to be published.

This matters because it makes the closing of the feedback loop the audited artefact. HCÉRES does not primarily ask "what was the mean score?" It asks "show us the board minutes where the evaluation results were discussed, and the curriculum change that followed." That is a materially different - and arguably healthier - thing for a regulator to inspect. It targets the action gap directly, and it institutionalises the feedback literacy that turns data into teaching change. It also structurally embeds students as partners in governance rather than treating them as mere data sources.

The mirror image of Italy

Set the two systems side by side and the trade-off from our companion analysis of the Italian ANVUR/OPIS model comes into sharp relief.

Italy standardises the instrument: one national questionnaire, mandatory collection, comparable scores. It buys comparability and coverage; it risks compliance-driven data that are gathered and filed.
France standardises the process and its uptake: mandatory evaluation, locally-designed instruments, audited closing-of-the-loop. It buys genuine improvement orientation; it sacrifices cross-institution comparability and tolerates uneven instrument quality.

Neither has solved course evaluation. They have made opposite bets about what a regulator should hold institutions accountable for - the measurement or the response. The most sophisticated reading is that you want both properties, and that no single mandate delivers them.

The genuine strengths of the French bet

It audits the right thing. A mandate to act on feedback, checked through board minutes and documented curriculum changes, is closer to the actual goal of course evaluation than a mandate to produce a comparable number. Scores that never change anything are theatre; France at least tries to inspect for the change.

Local and disciplinary fit. Because institutions design their own instruments, a conservatoire, an engineering school and a law faculty can each ask questions that fit their signature pedagogies - avoiding the construct-compression that a national form imposes.

Governance integration. Routing results through a mixed board with students and employers turns evaluation from an HR-style rating exercise into a programme-development conversation.

"But critics argue..." - the strongest objections

"Without a common instrument, you cannot compare or benchmark anything." Correct, and it is a real loss. A prospective student or a national policymaker cannot line French programmes up on a shared metric. But the French counter-question is pointed: comparability of what? If the comparable number is a mood-laden Likert average, its comparability may be spurious. France effectively decides that a locally-valid, acted-upon signal beats a nationally-comparable but weakly-valid one. Reasonable people weigh that differently - the honest point is that you cannot have both maxima at once.

"A process mandate is easy to fake - tick the box, hold a meeting, write bland minutes." This is the sharpest objection, and it is true. Auditing process invites ceremonial compliance: a conseil de perfectionnement that convenes, notes the scores, and changes nothing while producing minutes that say otherwise. France's safeguard - HCÉRES asking for documented pedagogical adjustments over multiple years - raises the bar, but a determined institution can still perform compliance. No design is proof against bad faith; the question is which failure mode you prefer to police. Italy risks "collected but unused data"; France risks "documented but hollow process." Pick your poison knowingly.

"So which model is right?" Neither, exclusively. The blind spots are complementary. A system that only standardises the instrument can end up with comparable numbers nobody acts on. A system that only mandates the process can end up with earnest local activity nobody can compare or quality-assure. The mature answer is to secure a valid, lightweight comparable core and audit that feedback is acted on - measurement and response, not one or the other.

Where Koji fits

France's model exposes the two things a course-evaluation platform must do well: capture locally-relevant, valid feedback, and make closing the loop visible and auditable. That is precisely the brief Koji for Education is built for. Its AI-moderated conversational interviews and six structured question types let each programme ask questions that fit its discipline - the local-fit France prizes - while standardised, bias-aware AI moderation keeps quality consistent across a decentralised system, softening the "uneven instrument" weakness of a no-national-form regime. Crucially for the French audit object, Koji provides closing-the-loop action tracking and programme- and institution-level reporting, so the evidence a conseil de perfectionnement and an HCÉRES reviewer ask for - what students said, what themes emerged, what changed as a result, across multiple years - is generated as a by-product of running the evaluation rather than reconstructed after the fact. And because it is designed for GDPR-compliant, EU-appropriate data handling, it fits French and wider European governance expectations. The same conversational engine powers general user and customer research on the main Koji platform.

Koji does not claim to eliminate the process-versus-instrument trade-off, nor to guarantee that any board will act in good faith - no tool can legislate institutional will. What it can do is make the acting-on-feedback that France demands cheaper to do and harder to fake, by turning the loop into tracked, reportable evidence.

France decided that the important question is not "what did students score you?" but "what did you do about what they told you?" Whatever instrument you use, that is the question worth being able to answer.

France Mandates Course Evaluation - But Not the Instrument. The HCÉRES Process Model