Graduate Capital, Not Just Human Capital: What Employability-Aware Course Evaluation Should Actually Measure

Bottom line up front: Most employability-oriented course evaluation asks a thin question — "did this course build job-relevant skills?" — that rests on a discredited model of how graduates get and keep good work. Michael Tomlinson's graduate capital model reframes employability as the accumulation and deployment of five interacting forms of capital: human, social, cultural, identity, and psychological. Skills are just one of them. If your evaluation instrument only ever asks about the first, it is blind to four of the five things that actually determine whether a graduate thrives — and it is measuring the wrong construct while claiming to measure employability.

Why "skills" is the wrong unit of analysis

The dominant folk model of employability is pure human capital: stuff more skills into the graduate, and the labour market rewards them proportionally. The European data does not cooperate. Cedefop's evidence on skill mismatch finds that around 25% of highly qualified young adults in the EU are overqualified for their jobs, and that a large share of workers report their skills are underused at work (Cedefop, Insights into skill shortages and skill mismatch). If employability were simply a function of accumulated skills, mass graduate overqualification would not coexist with employers reporting they cannot find the right people. The mismatch is not only about how much skill graduates have — it is about the social, cultural, and psychological resources that let them convert credentials into good work.

That is precisely the gap Tomlinson (2017), writing in Education + Training, set out to close. His article, "Forms of graduate capital and their relationship to graduate employability" (ERIC EJ1133688), conceptualises employability as "largely constitutive of the accumulation and deployment of a variety of interactive forms of capital". The model has since been adopted by careers services across European universities as a practical framework — and it maps onto course evaluation far better than a skills checklist.

The five capitals — and what each implies for evaluation

Human capital — the knowledge and skills a graduate can deploy. This is the only form most evaluations touch. Evaluation question that fits: not "was the content useful?" but "can you now do something specific you could not do before?"

Social capital — the networks and relationships that provide access to opportunity and information. Courses build this through group projects, peer learning, industry links, alumni contact, placements. A course can be content-rich and socially barren. Evaluation should ask whether the course connected students — to peers, to practitioners, to a field — not just whether it informed them.

Cultural capital — fluency in the norms, language, and expectations of a profession or sector: knowing how to present yourself credibly, reading the unwritten rules. Courses build this through authentic tasks, professional genres, exposure to how a discipline actually works. Evaluation should ask whether students now understand how their field operates, not just its content.

Identity capital — the graduate's sense of professional self and the coherence of their career narrative. Courses build this when students can articulate who they are becoming and why. Evaluation should ask whether the course helped students make sense of their direction — a question no satisfaction survey contains.

Psychological capital — resilience, self-efficacy, adaptability; the capacity to rebound from setbacks and sustain effort amid uncertainty. This is arguably the strongest predictor of how graduates navigate a non-linear labour market. Evaluation should ask whether students feel more able to tackle hard, ambiguous problems — a disposition, not a fact.

Laid out this way, the diagnosis is stark: the standard end-of-module Likert form, and even most "employability" add-on questions, measure a sliver of human capital and nothing else. They are answering a different, smaller question than the one they claim to.

Critics argue: "Universities cannot measure identity and resilience with a survey"

The serious objection is methodological, and it is right on its own terms. Human capital is (relatively) concrete; identity capital, cultural fluency and psychological capital are latent, developmental, and notoriously hard to assess. A five-point scale asking "rate your resilience" is close to meaningless — it invites social-desirability inflation, suffers from response-shift bias as students'' self-concept changes, and reduces a rich construct to a number that no one can act on. If the richer capitals cannot be measured well, isn't it more honest to stick to skills?

No — because that argument confuses "hard to measure with a Likert scale" with "impossible to measure". The correct conclusion is not to retreat to the one capital a checkbox can capture, but to change the instrument. These constructs reveal themselves in how students talk about their development — the narratives, examples, and reasoning behind a rating — far more than in the rating itself. The limitation is real; the response is to abandon the reductive instrument, not the important construct. A programme that measures only what a Likert scale can hold will optimise for exactly that, and quietly stop developing everything else.

Where Koji fits — evaluating capitals through conversation, not checkboxes

Koji for Education is built for exactly the constructs a survey cannot reach, because it collects reasoning and narrative, not just scores.

• AI-moderated conversational interviews let students describe how a course shaped their networks, professional understanding, sense of direction and confidence — the raw material of social, cultural, identity and psychological capital — rather than forcing latent constructs into a number that means nothing.
• Six structured question types (open_ended, scale, single_choice, multiple_choice, ranking, yes_no) let a programme pair a capital-relevant scale with a probe for evidence — "You said this course helped you feel more ready for the field; what specifically changed?" — grounding self-report in concrete examples.
• Automatic thematic analysis turns hundreds of these narratives into structured evidence, mapping which of the five capitals a programme is (and is not) developing — programme-level intelligence an accreditation panel can actually use, aligned with what employers say they need.
• Programme- and institution-level reporting shifts the unit of analysis from the module to the graduate's development across a degree, which is where capital is actually accumulated — closing the loop with the graduate-employability conversation the sector keeps having.

Koji is deliberately modest about what this proves: conversational evidence surfaces and mitigates the blind spots of a skills-only instrument; it is still an indirect, self-reported measure that should be triangulated against destination data, employer feedback, and graduate tracer studies — never treated as proof of employability on its own. The same conversational engine powers the main Koji platform, where understanding how people develop and make decisions is the core research task.

What this changes for accreditation and programme design

For a quality-assurance office, the graduate-capital model is not just a critique — it is a design brief. If a programme claims employability as a learning outcome, an accreditation panel is entitled to ask which capitals the curriculum develops and what evidence supports the claim. A satisfaction average cannot answer that; a mapped set of capital-specific findings can. This also disciplines curriculum design: teams that name the capital each element is meant to build — a group project for social capital, a professional-genre assignment for cultural capital, a reflective portfolio for identity capital — can then evaluate against that intent rather than against a generic "was it good?" scale. The model turns a vague aspiration ("our graduates are employable") into a set of falsifiable, evidence-backed claims, which is exactly the standard a sceptical external reviewer, or a sceptical academic, should demand. It also guards against a subtler failure: optimising the curriculum for the one capital that is easy to measure while the others quietly atrophy. What you can see, you can manage; what your instrument cannot see, you cannot.

The takeaway

Employability is not a pile of skills; it is five interacting capitals, and course evaluation that measures only the first is mismeasuring the whole. Tomlinson's model gives quality-assurance and institutional-research teams a rigorous, employer-recognised, European-adopted vocabulary for what to ask. The catch is that four of the five capitals live in narrative, not numbers — which means the instrument has to change before the framework can be honoured. Ask about all five capitals, ask for the evidence behind each, and stop pretending a satisfaction score measures whether a graduate will thrive.

Ready to evaluate the whole graduate, not just the skills checklist? Explore Koji for Education to see conversational, capital-aware course evaluation in action.