What changes as learners gain experience—and why less support is not the only evidence of progress
A learner completes a task with a prompt, a checklist, a worked example or another person nearby.
What have we seen?
It may be tempting to answer by subtracting the support from the performance. The prompt helped, so perhaps the learner cannot really do it. The checklist was available, so perhaps the result does not yet count. The person succeeded with guidance, so the next test of progress appears obvious: remove the guidance and see what remains.
Sometimes that makes sense. A learner may no longer need a model beside them. They may begin to generate their own questions, organise a process that someone else previously structured or recognise a familiar pattern without being led towards it.
But learning support does not have only one useful trajectory.
A structure that was initially provided by someone else may later be created by the learner. A checklist may begin as a way of learning an unfamiliar sequence and later become a tool the person deliberately chooses. A colleague may first help someone understand how a difficult decision is made and later become one of several people whose expertise a capable practitioner knows when to draw upon.
The visible support can remain while the person’s relationship with it changes considerably.
And sometimes removing support does not simply reveal how much learning remains.
It changes what the person is being asked to do.
Some of these same tools or forms of assistance may also be used as access accommodations. In that context, their purpose is to enable a person to participate and engage in the activity.
Here, the focus is on
learning supports: prompts, models, worked examples, coaching, planning structures, checklists, guidance and other resources used to help someone develop understanding, experience or capability. As learning develops, the role of that support may change. Some may become unnecessary; some may be increasingly chosen or adapted by the learner; and some may become part of how the person works competently.
That leaves a more useful question than simply whether support should stay or go:
When learning support changes, what are we actually hoping has changed in the learner?
Removing support does not simply reveal the same capability
Suppose someone first performs a task with a written sequence and then performs it without one.
It can look as though we have held the task constant and removed only the help.
But we have also added remembering the sequence.
Remove a planning structure and the learner must now generate and hold the plan while carrying out the activity.
Remove the worked example and they must identify the relevant pattern without that comparison.
Remove the opportunity to consult someone and the activity may now require knowledge that competent practice would ordinarily allow to be distributed between people.
Remove a prompt and the person may need to notice for themselves that a particular decision point has arrived.
None of these changes automatically makes removal inappropriate. Sometimes remembering the sequence, generating the structure or recognising the decision point is exactly what the learner is developing.
But the unsupported task is not simply the supported task minus help.
It may now include additional demands involving memory, planning, attention, uncertainty, coordination or knowledge retrieval.
That matters when we interpret what happens next.
A fall in performance may tell us that the learner has not yet internalised something the support was providing. It may also tell us that the reconfigured task now draws on an additional capability that was not central to what we were originally observing.
Instructional-scaffolding research helps make sense of this distinction. In their review of scaffolding research, van de Pol, Volman and Beishuizen described scaffolding as contingent support that responds to the learner’s current understanding, changes as that understanding develops and involves a gradual transfer of responsibility.[1]
The important idea is not simply
less help.
It is
changing help in response to what the learner is becoming able to do.
An experimental classroom study involving 30 teachers and 768 students found that the effects of contingency depended partly on how frequently support was available and, for one achievement measure, on whether learners’ task effort was taken into account. The findings were mixed—not evidence that highly contingent support is always better—but they support the narrower point that support cannot be interpreted apart from the conditions in which it is given.[2]
Adaptive fading has also been proposed as a way to adjust external support in response to evidence of the learner’s developing understanding. In the specific context of collaborative argumentation, Noroozi and colleagues offer a design framework built around diagnosis, adaptive support and fading. It is a conceptual proposal rather than a test of whether the framework improves learning, and its specific recommendations should not be generalised across learning contexts.[3]
So reducing support can provide useful evidence.
But only if we are clear about what question the changed task is now asking.
Support can change without simply disappearing
Imagine someone learning to run a complex meeting.
At first, another person may provide the agenda, suggest the questions to ask and help them anticipate where the discussion might become difficult.
Later, the learner may create the agenda themselves but still talk through the difficult parts beforehand.
Later again, they may prepare independently but seek another person’s view when the situation involves unfamiliar risk or specialist knowledge.
Has the support disappeared?
Not entirely.
But its function has changed.
What was initially
instructional may increasingly become
collaborative or
strategic.
The same can happen with tools.
A learner might initially use a checklist because they cannot yet remember the sequence of an unfamiliar task. With experience, they may no longer need the checklist for routine versions of the work.
But they might still choose to use it when the stakes are high, interruptions are likely or the task contains many moving parts.
At that point, using the checklist need not mean that learning has stalled. Knowing
when memory is an unnecessary risk may itself be part of growing judgement.
That creates a problem with a simple progression such as:
support → less support → no support
Sometimes that is exactly what happens.
A worked example may no longer be useful. A prompt may become redundant. The learner may begin doing something for themselves that another person previously did for them.
But other trajectories are possible:
someone else provides the structure → the learner begins shaping the structure → the learner creates or adapts it themselves
or:
someone else tells the learner when to seek help → the learner begins recognising those moments → the learner chooses whose expertise they need and why
or:
a tool supports an unfamiliar task → the learner becomes fluent without it in routine situations → the learner deliberately returns to it when complexity or risk makes it useful
The support has not simply been removed.
Ownership, judgement and use have changed.
Independence can grow while support remains visible
Reduced support is attractive as a sign of progress because it is easy to see.
Yesterday the learner required four prompts. Today they required one.
Last month someone organised the process. Now the learner organises it.
Those changes can be meaningful.
But other changes can be meaningful too.
A learner may become better able to:
- recognise what kind of help would actually be useful;
- ask a more precise question rather than waiting for someone else to diagnose the problem;
- choose between tools rather than using whichever one was supplied;
- adapt a familiar structure to an unfamiliar situation;
- identify which part of something they can manage and where they need another perspective;
- combine their own knowledge with another person’s expertise;
- notice when a strategy that works in one situation is poorly suited to another;
- recognise when apparent success is costing so much effort that a different approach is needed.
These changes are easy to miss if independence is measured only by how little outside support remains.
A learner may continue to use a visual planning structure but no longer need another person to create it.
An employee may still ask a colleague to review a complex case while becoming much more precise about what needs scrutiny and where their own judgement is already secure.
A young person may still want someone alongside them when approaching an unfamiliar undertaking while increasingly deciding what role they want that person to play.
The visible presence of another person, structure or tool can remain while responsibility shifts substantially.
That is one reason progress is better understood as more than the withdrawal of help.
The question becomes not only:
How much support is there?
but also:
Who is noticing, choosing, adapting, deciding and directing?
What are we preparing the person to participate in and do?
There is another reason to be cautious about treating unsupported performance as the obvious endpoint.
The learning task itself is rarely the final destination.
A child organising a school project is not being prepared only to complete that particular project.
A student learning how to plan a piece of work is not ultimately being prepared to reproduce the teacher’s planning method indefinitely.
A new employee learning a process is not being prepared simply to pass an induction activity.
The immediate task can matter without becoming the purpose of development.
A more useful question is:
What will this person need to carry forward in a genuinely usable way?
Sometimes the answer includes knowledge or procedures they need to recall readily.
Sometimes it includes recognising patterns, adapting what they know or making decisions under less familiar conditions.
Sometimes it includes knowing where reliable information lives rather than memorising information that competent practice expects people to look up.
Sometimes it includes knowing when to seek expertise, how to frame the question and how to remain responsible for the decision that follows.
And sometimes the most important thing the current support makes possible is
experience.
This is where workplace-learning research becomes particularly useful.
In many forms of adult work, completely unsupported performance would itself be artificial. People routinely use calendars, templates, previous work, procedures, software, checklists, search tools and organisational systems. They ask colleagues, seek specialist advice, divide responsibility and request review.
These things are not necessarily substitutes for competence.
Often they are part of how competence is enacted.
Billett describes workplace learning as
co-participative: what people learn is shaped both by the activities and guidance a workplace makes available and by how individuals engage with those opportunities.[4]
That gives participation an important role.
People do not simply learn first and then earn the right to participate.
Participation is one of the ways they learn.
A review of workplace learning in vocational education likewise found learners drawing on observation, modelling, coaching, scaffolding, independent exploration and the wider work community rather than learning solely through instruction from one designated teacher.[5]
And experience does not make context irrelevant.
In a qualitative study of 35 first-line managers entering new workplaces, previous experience mattered, but so did job-specific training, social support and shared organisational arrangements. Those conditions changed in importance over time and differed between managers.[6]
An experienced manager still has to learn a new organisation: its relationships, systems, informal knowledge, boundaries and ways of coordinating work.
Similarly, a realist review of work-based learning in higher education identified interaction with others, guidance, reflection and challenging contexts among the elements associated with developing adaptive expertise.[7] This evidence concerns students learning through work rather than established employees, but it points toward an important destination: not merely repeating a familiar process unaided, but becoming more able to respond when familiar processes are no longer enough.
That is a very different picture of what learning is preparing someone for.
Participation can create the experience from which judgement can grow
This matters because support can change not only today’s performance but what the person gets to experience next.
Someone who can participate in a complex activity with guidance gains direct access to aspects of practice that can be difficult to learn from outside it.
They encounter exceptions.
They see consequences unfold.
They notice what experienced people pay attention to.
They discover which parts they understand and which parts behave differently in practice.
They develop relationships.
They learn what they need to ask.
They begin forming judgements that are difficult to acquire through explanation alone.
Over time, some of the support may no longer be needed.
But that reduction is not necessarily the most important outcome.
The more consequential change may be that the person can now enter situations that were previously inaccessible to their experience.
This also creates a relationship between learning and opportunity.
If learning through work depends partly on access to meaningful activities and guidance, then opportunities to participate help shape what evidence can emerge later.[4,5]
The person invited into a more complex project gains experience, relationships and new occasions to exercise judgement.
The person kept at the edge because they are not yet seen as ready receives fewer opportunities to produce that evidence.
This is an inference from workplace-learning research rather than a simple causal rule. Being helped to develop is not identical to having somewhere consequential to use that development.
For learning, that means the question cannot stop at:
Can they complete today’s task with less help?
It also has to include:
What are they now able to participate in, experience and attempt because of what has developed?
Support can expand agency—or quietly replace it
None of this means that more support is automatically better.
Learning support can become so complete that very little remains for the learner to notice or decide.
A parent can anticipate every difficulty.
A teacher can divide a task so thoroughly that the learner follows steps without understanding how the parts belong together.
A coach can move from helping someone examine a decision to quietly making it for them.
A mentor can become the author of the work.
That is not a problem because support remains visible.
It is a problem because
the learner’s part has become too small.
This is why agency provides a more useful question than absence of help.
A 2026 meta-synthesis of 14 qualitative studies in primary and secondary science education describes student agency as relational and structurally mediated. Students’ capacity to contribute, question and exercise responsibility was shaped by instructional design, authority, peer relationships and wider classroom conditions.[8] That evidence should not simply be transferred to workplace settings, but it sits comfortably beside workplace research in which learning similarly emerges through an interaction between what environments make available and how people participate.
Support can increase agency when it makes meaningful participation possible while leaving the learner something consequential to notice, interpret, decide or direct.
It can reduce agency when the supporter increasingly becomes the person doing those things.
The visible amount of help cannot tell us which is happening.
A learner can receive considerable guidance while still being the author of the important decisions.
Another can appear highly independent while simply following a tightly prescribed process that leaves little room for judgement at all.
As experience grows, progress may therefore appear not only in what someone can execute, but in what they can increasingly
author:
the questions they ask;
the structures they create;
the tools they choose;
the risks they notice;
the expertise they seek;
the interpretations they make;
and the direction they take next.
What changes as learning support changes?
There is no single correct trajectory for learning support.
Sometimes a prompt has done its work and disappears.
Sometimes guidance becomes lighter as the learner begins providing more of the structure themselves.
Sometimes a tool initially used for learning becomes part of how competent practice is done.
Sometimes another person’s involvement shifts from instruction to collaboration or specialist consultation.
Sometimes support remains unchanged for a while because it is still enabling the experiences through which further learning can occur.
And sometimes the form of support needs to change because it is now doing too much of the thinking that the learner needs an opportunity to do.
So the useful question is not:
How quickly can we remove the support?
Nor is it:
How much support can we provide?
It is closer to:
What is the support making possible now, and what is changing in the part the learner is able to take?
That change may include greater knowledge or fluency.
It may include organising something that was previously organised for them.
It may include recognising when an old strategy no longer fits.
It may include more deliberate use of tools.
It may include knowing when another person’s expertise is needed.
It may include greater ownership of decisions.
It may include access to more complex experiences from which further judgement can develop.
The next task still matters.
So does knowing what someone can now do that they could not do before.
But neither has to be confused with the larger purpose.
Learning is also preparation for what someone will eventually need to
participate in, navigate, contribute to and make decisions within.
Learning support is not necessarily the opposite of independence.
Sometimes growing independence means doing something without support that was once necessary for learning.
Sometimes it means taking over the role that the support previously played.
And sometimes it means becoming more skilful at deciding which tools, structures and people belong in competent practice in the first place.
The important change is not always that the support disappears.
It may be that the learner increasingly knows what to do with it.
Evidence note
This article focuses on learning and developmental support.
The research brought together here comes from several related but distinct traditions. Much of the instructional-scaffolding literature concerns school, university or vocational learners. Workplace-learning research provides evidence that learning in work settings is shaped by participation, guidance, relationships and organisational arrangements, but much of this literature is qualitative or context-specific. Research concerning work-based learning in higher education should not be assumed to describe established employees. The agency synthesis cited here is based on 14 qualitative studies of primary and secondary science education, with most studies conducted in North America; it should not be treated as a general account of agency in every learning setting.
Taken together, these bodies of research support a view of learning support as something whose role can change with experience, rather than something whose value can be determined simply by whether it remains visible. They do not establish that particular supports should always remain or always be faded.
References
- van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher–student interaction: A decade of research. Educational Psychology Review, 22, 271–296.
- van de Pol, J., Volman, M., Oort, F., & Beishuizen, J. (2015). The effects of scaffolding in the classroom: Support contingency and student independent working time in relation to student achievement, task effort and appreciation of support. Instructional Science, 43, 615–641.
- Noroozi, O., Kirschner, P. A., Biemans, H. J. A., & Mulder, M. (2018). Promoting argumentation competence: Extending from first- to second-order scaffolding through adaptive fading. Educational Psychology Review, 30, 153–176.
- Billett, S. (2002). Workplace pedagogic practices: Co-participation and learning. British Journal of Educational Studies, 50(4), 457–481.
- Mikkonen, S., Pylväs, L., Rintala, H., Nokelainen, P., & Postareff, L. (2017). Guiding workplace learning in vocational education and training: A literature review. Empirical Research in Vocational Education and Training, 9, Article 9.
- Wastesson, K., Fogelberg Eriksson, A., Nilsson, P., & Gustavsson, M. (2021). Conditions for workplace learning as a new first-line manager in elderly care. Vocations and Learning, 14, 205–222.
- Groenier, M., Khaled, A., Kamphorst, J., Tankink, T., Endedijk, M., Fluit, C., & Kuijer-Siebelink, W. (2025). Adaptive expertise development during work-based learning in higher education: A realist review. Vocations and Learning, 18, Article 11.
- Espinola, E. C., & Prudente, M. S. (2026). Promoting student agency in science education: A meta-synthesis. Disciplinary and Interdisciplinary Science Education Research, 8, Article 25.
