Can Course Evaluation Measure Sustainability Competences? GreenComp vs the Likert Scale

Bottom line up front: European higher education is under growing pressure to embed sustainability competences into the curriculum — the EU now has an official framework for them, GreenComp, and the labour market signals an urgent and widening green-skills gap. Programme directors increasingly need evidence that courses are actually developing these competences, not just covering the topic. The uncomfortable truth is that the dominant evaluation instrument — an end-of-term Likert satisfaction survey — is structurally incapable of producing that evidence. Sustainability competences are dispositional, applied, and value-laden; a five-point "I was satisfied with this course" scale cannot see them. This is not an argument against measuring them. It is an argument for measuring them properly.

Why this is suddenly a course-evaluation question

Three forces are converging on programme evaluation at once.

The policy framework now exists. In 2022 the European Commission's Joint Research Centre published GreenComp (Bianchi, Pisiotis & Cabrera Giráldez), the European sustainability competence framework, as a deliverable of the European Green Deal. It defines four interlinked competence areas — embodying sustainability values, embracing complexity in sustainability, envisioning sustainable futures, and acting for sustainability — each with three sub-competences. Crucially, GreenComp is explicitly about knowledge, skills, and attitudes: it frames sustainability as something learners do and value, not merely something they know.

The labour market is demanding it. LinkedIn's Global Green Skills Report 2024 found that global demand for green talent grew by 11.6% from 2023 to 2024 while the supply of green talent grew only 5.6% — a gap projected to widen sharply. The report estimates that by 2030 one in five jobs will lack the green talent it needs, rising toward one in two by 2050, and that workers with green skills already have a hiring rate roughly 54.6% higher than the overall workforce. Whatever one thinks of any single projection, the direction is unambiguous: graduate employability is becoming entangled with sustainability competence. That makes it a programme-outcomes question — and programme outcomes are what evaluation is supposed to evidence. We have argued the general version of this in why course evaluation cannot measure employability through proximal mediators.

Accreditors want the evidence. European quality assurance is moving toward outcomes and competences, and "education for sustainable development" appears with increasing frequency in institutional strategies and review frameworks. A review panel asking "how do you know your graduates can act for sustainability?" will not be satisfied by a course's mean satisfaction score.

Why a satisfaction survey cannot answer the question

The mismatch is structural, not a matter of writing better Likert items. Consider what GreenComp actually asks of a learner versus what a standard SET item can capture.

Competences are dispositional and applied. "Acting for sustainability" or "embodying sustainability values" describes how a graduate behaves and what they are disposed to do — observable through projects, decisions, and reasoning, not through a satisfaction rating. Asking "Rate your satisfaction with the sustainability content (1–5)" measures the student's affect toward the course, which is a different construct entirely. This is a textbook case of construct-irrelevant variance: the number you collect is contaminated by, and largely measures, something other than the competence you care about.
Self-reported competence is a weak proxy. Even a better-worded item — "I can apply systems thinking to a sustainability problem" — runs into the well-documented unreliability of self-assessment. Students with the least competence often lack the metacognitive basis to judge their own competence accurately, and self-ratings are inflated by the very course enthusiasm that satisfaction surveys generate. Measuring a skill by asking whether someone feels they have it is a known validity hazard, the same one we flag in measuring learning gain, not satisfaction.
Values resist a number. GreenComp's "embodying sustainability values" — valuing sustainability, supporting fairness, promoting nature — is exactly the kind of construct a five-point scale flattens into uselessness. Students know the socially-desirable answer, so the scale measures their awareness of the expected response as much as any genuine disposition.
Aggregated means erase the distribution that matters. A programme might be developing strong sustainability competence in a committed third of students and none in the rest — a result with very different implications than uniform moderate development, yet both produce the same mean. We make this argument in general in how the mean hides dispersion; for competences, where you often care about whether a threshold was crossed, it is especially damaging.

But isn't this just measuring "soft" things that resist measurement anyway?

The strongest sceptical position runs like this: sustainability competence is vague, values are unmeasurable, and any attempt to evaluate them is feel-good box-ticking that produces noise dressed up as evidence. This deserves a serious answer rather than a dismissal.

The objection is half right. Measuring competences badly — bolting a "sustainability satisfaction" item onto an existing survey — does produce noise dressed as evidence, and the sceptic is correct to despise it. But "hard to measure" is not "impossible to measure", and the competence-assessment field has decades of credible practice: authentic assessment of applied tasks, structured rubrics tied to observable behaviours, portfolio evidence, and qualitative interviewing that probes reasoning rather than affect. GreenComp itself was designed to be operationalised — its sub-competences are written as learning outcomes precisely so they can be assessed against evidence.

The honest position is therefore neither "measure it with a Likert item" nor "it cannot be measured". It is: the construct demands qualitative, evidence-based, behaviourally-grounded assessment, and the cheap survey is the wrong tool — not because the thing is unmeasurable, but because the tool is mismatched to it. The sceptic's real target should be lazy measurement, not measurement.

What good evidence looks like

For sustainability competences, defensible programme evaluation triangulates several sources — consistent with our general argument for triangulating across multiple evidence sources:

Authentic assessment artefacts — projects, cases, and decisions where students act for sustainability and the competence is directly observable.
Structured qualitative feedback that probes reasoning — not "were you satisfied?" but "describe a moment in this course where you had to weigh competing sustainability trade-offs, and what you decided" — answers that reveal whether embracing complexity actually developed.
Mapping to a shared framework — coding evidence against GreenComp's competence areas so it is comparable across courses and legible to accreditors, much as we describe for mapping evaluation to the ESCO skills taxonomy.
Employer and placement signal — whether graduates demonstrate these competences in work-integrated settings, closing the loop we discuss in the employer feedback loop for programme evaluation.

The recurring obstacle is the same one that blocks every qualitative-at-scale ambition: open-ended, reasoning-rich responses are expensive to collect consistently and brutally expensive to analyse across a whole programme. That economic constraint, not a conceptual one, is why institutions default to the satisfaction item they know is wrong.

Where Koji fits

This is precisely the constraint Koji for Education is built to relax. Sustainability competences cannot be read off a scale, but they can be surfaced through conversation — and Koji's AI-moderated conversational interviews are designed to probe beyond a number toward reasoning and example. Instead of "rate the sustainability content", the moderator can ask a student to walk through a trade-off they faced and why they resolved it as they did — the kind of response that actually evidences embracing complexity or acting for sustainability. Because the moderation is standardized, every student is probed consistently, which is what makes the resulting qualitative evidence comparable across a programme rather than an unstructured pile of anecdotes.

Koji's automatic thematic analysis then reads that full corpus and can map themes against a competence framework like GreenComp — turning hundreds of open-text reflections into structured, programme-level evidence that an accreditor can read, without a committee hand-coding transcripts for weeks. Its six structured question types (open_ended, scale, single_choice, multiple_choice, ranking, yes_no) let you keep scales where they are genuinely informative while carrying the competence evidence in the channel built for it. And its programme- and institution-level reporting is designed for exactly the cross-course, framework-aligned view that sustainability-competence accreditation requires. The same AI interview engine powers the main Koji platform for organisations researching skills and behaviour change in their own workforces.

To stay precise: Koji does not certify that a graduate has a sustainability competence — that is the job of assessment, not evaluation — and no tool eliminates the gap between what students say and what they can do. What Koji does is surface far richer, framework-mappable evidence of competence development than a satisfaction survey can, and make that evidence usable at programme scale.

The takeaway

The EU has defined sustainability competences (GreenComp), the labour market is pricing them in (LinkedIn's green-skills gap), and accreditors increasingly want evidence that programmes develop them. None of that evidence can come from a five-point satisfaction survey, because competences are applied, dispositional, and value-laden — constructs the Likert scale is structurally unable to capture. The answer is not to abandon measurement as "too soft" but to match the instrument to the construct: authentic evidence, reasoning-rich qualitative feedback, and framework mapping, collected and analysed at a scale that only modern AI-moderated tooling makes affordable. The satisfaction item was never going to answer the question. It is time to stop asking it to.

Need to show an accreditor that your programme develops real graduate competences, not just satisfaction? See how Koji for Education turns reasoning-rich student feedback into framework-mappable evidence.