Course Evaluation and Graduate Employability: Closing the Loop on the Skills Gap

Bottom line: End-of-term satisfaction scores tell you whether students enjoyed a course. They tell you almost nothing about whether the course developed the capabilities the labour market is asking for. With European evidence showing large, persistent skills mismatches and employers forecasting that nearly four in ten core skills will change within five years, programme-level evaluation that ignores employability is measuring the wrong thing. The fix is not another satisfaction survey — it is feedback designed to connect what is taught to what graduates can actually do.

The gap evaluation isn''t measuring

There is now a substantial, well-evidenced disconnect between what higher education certifies and what the labour market absorbs. Cedefop''s second European Skills and Jobs Survey (ESJS2, 2021) found that roughly four in ten EU workers are in jobs that do not match their level of education — 28% over-qualified and 12% under-qualified — and that around one in five Europeans are underskilled at the point of hiring. These are not anecdotes; they are population-level findings across the EU, Iceland and Norway.

Meanwhile the target is moving. The World Economic Forum''s Future of Jobs Report 2025, drawing on more than 1,000 employers representing over 14 million workers across 55 economies, reports that employers expect 39% of workers'' core skills to change by 2030 (down from 44% projected in 2023, but still enormous). Technological skills — AI and data foremost — are forecast to rise fastest, but analytical thinking, resilience and lifelong-learning capacity remain the most valued.

Now compare that to a typical course evaluation. It asks: Was the instructor clear? Were you satisfied? Would you recommend this course? All reasonable questions. None of them asks whether the course built a capability an employer will pay for, or whether students can transfer what they learned to an unfamiliar problem. The instrument is structurally blind to the outcome that increasingly defines a programme''s value.

Why satisfaction is a weak proxy for employability

The temptation is to assume a well-liked course is a valuable one. The evidence cautions against it. As we discuss in our analysis of why averaging Likert scores misleads, satisfaction scores are confounded by enjoyment, workload and grade expectations — and an easy, enjoyable course can score highly while developing little. A demanding course that genuinely stretches employability-relevant skills may score lower on satisfaction precisely because it was hard. Optimising programmes for satisfaction can therefore pull in the opposite direction from optimising for graduate capability.

Cedefop''s analysis is blunt on which capabilities matter: below-average interpersonal skills such as communication and teamwork "cannot be compensated for — even by the best grades or the most relevant field of study." These are exactly the skills that a five-point "overall satisfaction" item cannot detect and a thematic reading of student reflection can.

What employability-aware evaluation actually asks

Closing the loop between evaluation and employability means changing the questions, not just the analysis. Instead of (or alongside) satisfaction, employability-aware evaluation probes:

• Transfer: Can students apply what they learned to a problem they haven''t seen before — or only reproduce worked examples?
• Skill self-efficacy: Do students feel more capable in specific, named competencies (data analysis, structured writing, collaboration) after the course?
• Authenticity: Did assessment resemble the kind of task a graduate would actually face?
• Gaps: Where do students themselves sense they are underprepared for placements, projects or work?
• Employer and alumni signal: What do those one or two years downstream say the programme should have done differently?

This is closer to programme-level quality assurance than to instructor rating — and it maps directly onto accreditation expectations. European quality frameworks increasingly ask for evidence that programmes deliver intended learning outcomes, a theme we cover in our guide to AI in European higher-education quality assurance. Graduate-outcome evidence is exactly what panels want and what satisfaction averages cannot supply.

But isn''t this just turning universities into job-training factories?

This is the strongest and most legitimate objection, and it should not be waved away. Universities exist to do more than feed the labour market: to advance knowledge, to cultivate judgement and citizenship, to teach things whose value is not immediately monetisable. A narrow employability metric, badly used, could crowd all of that out — privileging short-term, vocational skills over the durable intellectual capacities that, ironically, employers themselves rank highest.

Two responses. First, employability properly understood is not narrow vocationalism. The WEF and Cedefop data both put analytical thinking, communication, adaptability and continuous learning at the top — the very capacities a good liberal or research-led education claims to build. Measuring whether a programme develops them is not a betrayal of academic values; it is asking the programme to demonstrate the value it already claims. Second, the answer to a reductive metric is not no metric but a richer one. A single "employability score" would indeed be a travesty — for the same reason a single satisfaction score is. The goal is qualitative, triangulated evidence about capability, used formatively, not another league table.

A second objection: students are poor judges of their own future employability. Partly true — which is why self-report should be one strand, combined with authentic assessment evidence, alumni follow-up and employer input, not the whole picture.

Accreditation is already asking for this

Employability-aware evaluation is not only good practice — it is increasingly a compliance expectation. Business-school accreditors such as AACSB and EFMD/EQUIS require assurance of learning: documented evidence that programmes define learning goals, measure whether students achieve them, and close the loop by acting on the results. Broader European frameworks (the ESG 2015 standards, and national bodies like NVAO) similarly expect programmes to evidence that intended learning outcomes — many of them employability-relevant — are actually met. Satisfaction averages do not satisfy these requirements; capability and outcome evidence does.

That reframes employability evaluation from a "nice to have" into part of the audit trail. A programme that can show how it surfaces students'' self-assessed capability growth, where it identifies preparation gaps, and what it changed in response is producing exactly the closing-the-loop evidence panels look for. A programme that can only show a 4.1 satisfaction average is not.

The practical move is to add a small number of capability- and transfer-oriented items to existing evaluation cycles — and, critically, to read the open-ended responses thematically rather than averaging a "career readiness" scale that means different things to different students. Connecting course-level feedback upward to programme-level outcome claims is precisely the evidence chain that accreditation reviewers want to follow, and the one most institutions currently cannot produce from satisfaction data alone.

Where Koji fits

Koji for Education is designed for exactly this richer, capability-oriented evidence. Its AI-moderated conversational interviews can ask students to describe a problem they could now solve that they couldn''t before — and probe the answer — rather than collecting a number on a "career-readiness" scale that means different things to different students. Its six structured question types (open-ended, scale, single- and multiple-choice, ranking, yes/no) let a programme combine quick comparable items with deep open exploration in one instrument. Automatic thematic analysis then turns hundreds of reflections into the patterns a programme director or teaching-and-learning centre can act on: which capabilities students feel confident in, where they sense gaps, which courses they credit for skill growth.

Because Koji supports programme- and institution-level reporting and closing-the-loop action tracking, it can connect course-level feedback to programme-level employability evidence — the kind accreditation panels increasingly expect. And the same engine can run alumni and employer interviews, so the graduate-outcome loop is closed with real downstream voice rather than assumed. Teams that also run broader market or customer research often use the shared interview engine on the main Koji platform.

Koji does not promise to close the skills gap — that is a systemic challenge no evaluation tool can solve. What it does is make the gap visible at the level where programmes can respond, replacing satisfaction proxies with evidence about capability. If your programmes are being asked to demonstrate graduate outcomes, explore Koji for Education and start measuring what the labour market is actually asking for.