Your Evaluation Instrument Assumes an 18-Year-Old. Most Adult Learners Are Not.

Answer first: The standard course-evaluation form carries an unstated assumption about who is filling it in: a traditional-age, full-time, campus-based undergraduate for whom the course is their main occupation. But a large and growing share of European higher-education learners are adults — part-time, returning, professional, modular. In 2024, 28.5% of EU adults aged 25–64 had participated in education or training in the previous 12 months (Eurostat), against an EU 2030 target of 60%. Malcolm Knowles's theory of andragogy argues that adult learners differ systematically from younger ones in self-direction, prior experience, motivation, and what they want from learning. When you evaluate an adult learner with an instrument written for an 18-year-old, you do not just miss nuance — you ask the wrong questions and then misread the answers.

What andragogy actually claims

In 1968 Knowles popularised andragogy — the art and science of helping adults learn — as a deliberate break from pedagogy, "teaching adults like children" (Research.com; EBSCO Research Starters). His framework rests on a set of assumptions about mature learners that are directly relevant to how they judge a course:

• Self-direction. Adults' self-concept moves from dependency toward autonomy; they want agency over their learning, not to be managed.
• Prior experience as a resource. Adults bring "a growing reservoir of experience" that is itself a learning resource — and they expect it to be recognised, not overwritten.
• Readiness tied to social role. Adults become ready to learn what their life and work roles require, when they require it.
• Problem-centred orientation. Adults are oriented to immediate application — they learn to solve a problem, not to bank knowledge for later.
• Internal motivation, and needing to know why. Adults are driven more by internal goals and want to understand the purpose of what they are asked to learn.

You do not have to accept andragogy as settled science (more on that below) to see the evaluation implication: an adult learner and a school-leaver are, on average, judging a course against different criteria. Ask them the same questions and the same 4.0 means different things — or, worse, the questions that matter most to the adult are not on the form at all.

Where the standard instrument mismatches the adult learner

Classic evaluation items — "The lecturer was engaging," "The pace was appropriate," "I was satisfied with the module" — quietly encode the campus-undergraduate experience. For the adult learner, the load-bearing questions are different:

• Relevance to practice, not lecturer charisma. The problem-centred adult judges a course by whether it changed what they can do on Monday morning. An instrument that asks about engagement and satisfaction, but not about applicability and transfer, is measuring a proxy the adult barely cares about. This is why the transfer-of-training question matters far more for this group.
• Recognition of prior experience. Did the course build on what the learner already knew, or force them through material they had mastered a decade ago? No standard SET item captures this — yet for adults it is central.
• Fit with a life that contains a job and dependants. Flexibility, modularity, and the practicality of assessment deadlines are not "nice to have" for a part-time adult learner; they are determinants of whether the course was good for them. A campus-normed form treats these as out of scope.
• Why it misreads dissatisfaction. When an adult marks a course low, the campus-normed instrument invites the wrong inference ("teaching quality problem") when the real signal may be "this was not pitched at someone with my experience" or "the fixed synchronous schedule was incompatible with my shift work." The number is the same; the meaning — and the fix — is completely different.

This connects to a point we have made about micro-credentials and short courses breaking the end-of-semester survey: the cadence is wrong too. Adult and modular provision often has no "end of semester" to survey at, and the once-a-term snapshot misses the problem-solving arc that matters to the learner.

The strongest counterargument

"Andragogy is a contested theory, not established fact — you are building an evaluation argument on soft ground." This is a fair and important objection, and intellectual honesty demands engaging it. Andragogy has been criticised for decades: the sharp adult/child binary is overdrawn (plenty of adults want structure; plenty of teenagers are self-directed), the assumptions function more as heuristics than as empirically falsifiable laws, and critics argue it describes a style of provision more than a distinct psychology of adulthood. We should not overclaim.

But the evaluation argument does not actually depend on andragogy being a complete theory of adult cognition. It depends on a much weaker, well-evidenced claim: that a 45-year-old career-switcher studying part-time while working is, on the dimensions that a course-evaluation instrument probes — motivation, relevance-to-work, prior experience, scheduling constraints — meaningfully different from a full-time 18-year-old, often enough that a single instrument normed on the latter will systematically under-serve the former. Even a critic of andragogy-as-theory will grant that. And the Eurostat participation data shows this is not a niche population: in the highest-participating countries — Sweden, Denmark, Estonia, Finland — more than half of all adults were in learning in 2024. The design principle survives even if Knowles's specific model does not. Put differently: the risk is not that you adopt a wrong theory of adulthood, but that you adopt no theory at all and let the instrument's hidden default — the school-leaver — stand in for every learner by omission. Even a modest correction, such as adding two or three items on workplace relevance and scheduling fit and reading adult cohorts against their own baseline, closes most of the gap without committing you to any contested claim about how adults differ.

A second objection: "One instrument for everyone gives us comparability. Segmenting fragments our data." True, and comparability has real value. But comparability that averages two populations answering questions that mean different things to each is false comparability — it produces a tidy number that conceals a mixture. The better path is a shared core of items plus segment-appropriate items and, above all, segment-aware reading of the results, so an adult cohort's scores are interpreted against adult-relevant expectations rather than a youth default.

Where Koji fits

The deep problem is that a fixed questionnaire cannot adapt to who is answering it — and adult learners are exactly the population a static form serves worst. Koji for Education addresses this at the level of the instrument itself. Its AI-moderated conversational interviews adapt their probing to the respondent: when a working professional says a module "wasn't relevant," the interviewer follows up on what they needed and did not get — surfacing the applicability-and-transfer signal that a Likert scale flattens to a low number with no explanation. Its six structured question types let you build items around relevance-to-practice, recognition of prior experience, and fit-with-life rather than recycling campus-experience stems. Segmented, programme-level reporting lets you read an adult cohort against adult-relevant expectations instead of a blended average. And its formative, mid-cycle collection fits modular adult provision that has no neat end-of-semester moment.

Koji does not resolve the theoretical debate about andragogy, and it does not eliminate the differences between learner populations — no tool can. What it does is reduce the mismatch between a one-size instrument and a diverse learner body, and surface the reasons behind an adult cohort's scores rather than leaving you to guess. (The same conversational engine powers the main koji.so platform, where teams run adaptive interviews with any audience — the education product applies it to the adult-learner context.)

The one-paragraph version

Nearly 30% of European adults were learning last year, yet your evaluation form still asks an 18-year-old's questions. Adult learners judge a course by relevance-to-practice, recognition of their experience, and fit with a working life — not lecturer charisma or campus experience. You do not need to swallow andragogy whole to accept the design implication: ask segment-appropriate questions, read adult cohorts against adult expectations, collect formatively, and stop treating a blended average as if it meant one thing.

Frequently asked questions

What is andragogy and why does it matter for course evaluation? Andragogy, popularised by Malcolm Knowles in 1968, is the theory of how adults learn, distinct from pedagogy. It holds adults are more self-directed, bring prior experience as a resource, are internally motivated, and learn to solve immediate problems — so they judge courses by different criteria, and an instrument written for younger learners asks the wrong questions.

How many higher-education learners are actually adults? A large and growing share. Eurostat reports 28.5% of EU adults aged 25-64 participated in education or training in 2024, against an EU 2030 target of 60%. In Sweden, Denmark, Estonia and Finland, more than half of adults were in learning.

How does a standard survey misread adult learners? It measures proxies adults barely care about (lecturer charisma, campus experience) and omits what they judge courses by (relevance to practice, recognition of prior experience, fit with a working life). A low score then invites the wrong inference — a teaching-quality problem — when the real issue may be poor pitch or incompatible scheduling.

Isn't andragogy a contested theory? Yes, and honesty requires saying so. The adult/child binary is overdrawn and the assumptions work more as heuristics than laws. But the evaluation argument only needs the weaker, well-evidenced claim that adult, working, part-time learners differ from full-time school-leavers on the dimensions an evaluation probes.

Doesn't one instrument for everyone give better comparability? Only false comparability. Averaging two populations who read the same questions differently conceals a mixture. Better to use a shared core of items plus segment-appropriate items, and interpret adult cohorts against adult-relevant expectations.

What should evaluation of adult learners actually measure? Relevance and transfer to practice, whether the course recognised prior experience, and fit with a life containing work and dependants — collected formatively across modular provision rather than in a single end-of-semester snapshot that adult and short-course formats often lack.

---

Evaluating part-time, professional, or returning learners? See how Koji for Education adapts to who is actually answering.