Your Degree Apprentices Don't Sit in Your Lecture Theatre — Stop Evaluating Them as If They Do

Bottom line up front: A degree apprenticeship is not a degree with a job attached. It is a tripartite arrangement — apprentice, employer, and training provider — in which learning is split between academic study and the workplace, governed by an occupational standard, and judged at the end by an independent End-Point Assessment (EPA) that neither the employer nor the provider controls. An end-of-module satisfaction survey, designed for a full-time campus student, measures a thin slice of this and misses the parts that actually determine whether the apprenticeship works: whether off-the-job training integrated with on-the-job practice, whether the workplace mentor delivered, and whether the apprentice is on track for the EPA. If your evaluation instrument was built for lectures, it is quietly blind to the model that is growing fastest.

And it is growing fast. In England, higher apprenticeships rose from 13% of starts in 2017/18 to around 40% by 2024/25, and roughly 46,800 people began a degree apprenticeship in 2022/23 alone — split fairly evenly between level 6 (bachelor's-equivalent) and level 7 (master's-equivalent). Across Europe, work-based and dual models — Germany's duales Studium prominent among them — occupy the same space. This is a mainstream mode of higher education, and it deserves an evaluation designed for it.

What the standard survey cannot see

Consider what a conventional module evaluation asks about: lecture clarity, assessment fairness, resources, workload. Now consider what determines whether a degree apprenticeship succeeds.

• Integration of off-the-job and on-the-job learning. In England, full-time apprentices must receive a minimum average of six hours per week of off-the-job training — the rule that replaced the older "20% of time" formula in August 2022. The quality question is not whether the six hours happened, but whether what was learned off-the-job connected to the work the apprentice was doing. A satisfaction score on a taught module cannot see this seam — and the seam is where apprenticeships succeed or fail.
• The employer and the workplace mentor. The apprentice's line manager and workplace mentor are, functionally, co-educators. Ofsted and other reviewers have repeatedly found variability in how well providers align off-the-job training, workplace mentoring, and academic content. If your evaluation never asks the apprentice about the mentoring relationship or the employer's release of study time, you are not evaluating half the programme.
• EPA readiness. The apprenticeship culminates in an independent End-Point Assessment, conducted by an assessment organisation deliberately separate from both employer and provider to keep it impartial. "Am I being prepared for the EPA?" is a live, high-stakes question for every apprentice — and it is entirely absent from a generic satisfaction form.
• The tripartite review itself. Progress is meant to be tracked through regular tripartite reviews involving apprentice, employer and provider. The health of that three-way relationship is a leading indicator of success. No end-of-module survey touches it.

The measurement mismatch, stated plainly

The deeper problem is a construct mismatch. A satisfaction survey measures the apprentice's experience of taught provision. But the apprenticeship's defining feature is the integration of provision with work — a property of the relationship between settings, not of either setting alone. Averaging a Likert score on lecture quality tells you nothing about integration, because integration is not a property the instrument was built to detect. You can have delighted apprentices and a failing apprenticeship — great lectures, absent mentoring, no EPA preparation — or frustrated apprentices in a programme that is, in the ways that matter, working.

This is why apprenticeship evaluation needs to ask relational and longitudinal questions, not just satisfaction questions: Is your off-the-job learning useful in your role? Is your employer giving you protected time? Is your mentor engaged? Do you understand what the EPA will require of you? Are the three parties actually meeting?

"Isn't this just adding survey burden?" — the objection

The fair objection is that apprentices are among the busiest learners in the system — holding down a job while studying — and that piling on a longer, richer evaluation is exactly the wrong response. Every extra question is time taken from work or study, and response rates for over-surveyed cohorts collapse.

This is a real constraint, and it argues against the standard approach, not for it. A long static questionnaire that tries to cover integration, mentoring and EPA-readiness with a battery of Likert items is precisely the burden apprentices resent. The alternative is not more questions; it is smarter ones — a short, adaptive, conversational check-in that probes only where a signal appears, and that can be timed to the apprenticeship's rhythm (a mid-year integration check, an EPA-readiness check) rather than bolted onto a module. Depth without length is the goal, and it is achievable.

Where Koji fits

Evaluating a tripartite, work-integrated programme is a good fit for a conversational, adaptive instrument rather than a static form.

• Conversational interviews that probe integration. Instead of a Likert item on "workplace support", Koji's AI moderator can ask an apprentice how off-the-job learning is landing in their role, and follow up where the answer signals a problem — surfacing mentoring gaps or release-time issues a fixed scale would flatten.
• The right question types for a relational programme. Koji's six structured types (open-ended, scale, single-choice, multiple-choice, ranking, yes/no) let you combine a quick EPA-readiness pulse with open probing of the mentoring relationship in one short instrument.
• Formative, cycle-appropriate timing. Because Koji supports mid-cycle collection, you can run a lightweight integration check partway through — while there is still time to fix an absent mentor — rather than discovering the problem in a post-hoc satisfaction score.
• Programme-level thematic analysis. Automatic theming of open text across a cohort tells a programme director whether "my employer won't release me for study" is one apprentice's grievance or a systemic pattern with a particular employer partner.

This complements, rather than replaces, the outcome data that arrives later — EPA pass rates, retention, progression. Evaluation of the lived tripartite experience is the leading indicator; outcomes are the lagging confirmation. The same interview engine powers the general research platform at koji.so, where employers running their own apprenticeship or graduate schemes gather the workplace side of the story.

The practical takeaway

Evaluate the apprenticeship, not the module inside it. Ask about integration, mentoring, employer release, and EPA-readiness; time your check-ins to the apprenticeship's rhythm rather than the semester's; theme the answers at cohort and employer-partner level; and keep the instrument short enough that busy apprentices actually complete it. The full-time-student survey was never designed for this model — and the model is now too large to keep mis-measuring.

Frequently asked questions

Why can't we use our standard module evaluation for degree apprentices? Because it was built to measure taught provision, and a degree apprenticeship's defining feature is the integration of taught provision with workplace learning. Lecture-quality scores are silent on mentoring, employer release time, off-the-job/on-the-job integration and EPA readiness — the factors that actually determine whether the apprenticeship succeeds.

What is the End-Point Assessment and why does it matter for evaluation? The EPA is the final, independent assessment of an apprentice against the occupational standard, conducted by an organisation separate from both employer and provider to keep it impartial. "Am I being prepared for the EPA?" is a central concern for apprentices, so evaluation should track EPA-readiness explicitly rather than assume a satisfaction score covers it.

What is the "tripartite" relationship? The three parties to an apprenticeship: the apprentice, the employer (including the workplace mentor), and the training provider. Progress is meant to be reviewed jointly in tripartite reviews. The health of this three-way relationship is a leading indicator of success and should be evaluated directly.

Won't a richer evaluation overburden already-busy apprentices? It would if you added more static questions. The better route is a short, adaptive, conversational check-in that probes only where a signal appears and is timed to the apprenticeship's rhythm — depth without length. That reduces burden relative to a long questionnaire while capturing more.

Does this apply outside England's apprenticeship system? The English standards/EPA vocabulary is specific, but the underlying model — academic study integrated with structured workplace learning, co-delivered by an employer — is shared across European work-based and dual programmes such as Germany's duales Studium. The evaluation principle (measure integration and the workplace relationship, not just the lecture) travels.

How does this relate to evaluating placements and internships? It is adjacent. Placements and work-integrated learning share the "the end-of-module survey doesn't fit" problem, but an apprenticeship goes further: the workplace is not a component of the programme, it is half of it, with a formal employer party and an independent terminal assessment.

---

Koji for Education evaluates degree apprenticeships as what they are — a tripartite, work-integrated programme — with short, adaptive interviews that surface mentoring and integration issues while there is still time to act. Talk to us.