The Core Principles of Adult Learning

The transfer problem

Walk through the annual corporate training budget and you will find a strange accounting. Billions of dollars spent. Completion rates above 90%. Smile sheets glowing. And then, six months later, almost nobody behaves differently at work.

The best empirical picture we have of this comes from Saks and Belcourt (2006), who surveyed 150 organizations and found about 62% of content was transferred to the job immediately after training, falling to roughly 44% at six months and about 34% at twelve months. Earlier reviews such as Baldwin and Ford (1988) had already documented that only a small fraction of what people are taught in training shows up reliably back on the job. Whatever the exact number, the direction is the same. Most corporate training decays fast.

Part of the decay is simple forgetting. Murre and Dros (2015) replicated Ebbinghaus's forgetting curve and confirmed that isolated, unpracticed information drops off steeply within days. Part of the decay is structural: the training was never designed to produce behavior change in the first place. It was designed to fit into a one-day offsite, score well on a post-event survey, and produce a completion certificate for HR.

Adult learning research has spent the last six decades studying what actually produces durable behavior change in grown-ups. The field is messy, the dominant theories are contested, and a few popular ideas have been debunked outright. What survives is a set of six principles that the empirical record keeps pointing to. Training that honors most of them tends to transfer. Training that ignores them tends not to.

Principle 1: Relevance and applicability

Adults learn what they can see themselves using. That is the core insight from Malcolm Knowles's Modern Practice of Adult Education (1970), and it is the principle most often violated by enterprise training. Knowles argued that adult learners need to know why something matters before they will invest attention, and that motivation rises sharply when the content connects to an immediate life or work problem.

The memory literature gives this a cognitive mechanism. Context-dependent memory research, most famously Godden and Baddeley's 1975 underwater diver study, showed that information learned in one context is better recalled in a similar context. The effect has been replicated unevenly and a 2021 attempt to reproduce the diver result failed, so treat it as illustrative rather than decisive. The broader point holds: content encoded in the setting and situation where it will be used is more retrievable there.

The practical implication is blunt. A leadership workshop held in a hotel ballroom with PowerPoint, two weeks before anyone has a leadership situation to apply it to, is fighting both relevance and context encoding at once. Training delivered inside the workflow, tied to a decision the learner is about to make, wins by default.

Principle 2: Self-direction and agency

Knowles's second claim, developed in his 1970 and 1980 books, is that adults resent being talked to like children and learn better when they have meaningful control over pace, sequence, and focus. The framework he gave this, andragogy, is contested (more on that below), but the self-direction component has picked up durable support from a separate literature.

Self-determination theory, developed by Ryan and Deci (2000), identifies autonomy as one of three basic psychological needs that drive intrinsic motivation, alongside competence and relatedness. When learners feel pushed, surveilled, or stripped of choice, intrinsic motivation drops and so does engagement with the material. When they feel they own the direction of their learning, motivation and persistence climb.

Most mandatory e-learning modules fail this test by design. Linear click-through courses with locked navigation, fixed pacing, and no real choice signal to the learner that the company does not trust them to steer. The result is compliance-shaped attention instead of learning.

Principle 3: Prior experience as the foundation

Adult learners are not blank slates. They arrive with years of work history, mental models, heuristics, and scar tissue. David Kolb's Experiential Learning (1984) formalized this as a cycle: concrete experience, reflective observation, abstract conceptualization, active experimentation. New learning lands when it connects to and reshapes existing experience.

The developmental psychology backing for this runs deeper. Vygotsky's Mind in Society (1978) argued that learning happens most productively in the zone of proximal development, the gap between what a learner can do alone and what they can do with skilled support. The zone is a property of the learner's existing competence. You cannot design training well without knowing what the learner already brings.

Training that treats a seasoned engineering manager and a brand-new hire to the same curriculum, at the same pace, with the same examples, is building outside the zone for both of them. The manager is bored because the material is below their ceiling; the new hire is lost because it is above their floor. Both check out.

Principle 4: Problem-centered, not subject-centered

Adults engage with training when it is organized around problems they need to solve rather than subjects to be covered. The clearest applied test of this principle comes from Barrows and Tamblyn's Problem-Based Learning (1980), which grew out of McMaster University's medical school in the late 1960s.

Servant-Miklos (2019) traces the history. McMaster's founding faculty rejected the traditional lecture-heavy preclinical curriculum and organized training around clinical cases. Students were given a patient presentation, had to identify what they did not know, researched it, and returned to work the case. The format spread to medical and professional education worldwide because the graduates performed better on applied reasoning tasks and, crucially, retained what they had learned longer.

The mechanism is roughly what Bloom's taxonomy had already predicted. Bloom (1956) distinguished lower-order knowledge (recall, comprehension) from higher-order cognitive work (application, analysis, synthesis, evaluation). Problem-centered instruction forces the learner through the higher-order rungs from day one. Subject-centered instruction often never gets there.

Principle 5: Deliberate practice with feedback

Information in does not equal skill out. The single most important finding on expert skill acquisition is Ericsson, Krampe, and Tesch-Romer's 1993 Psychological Review paper on deliberate practice. They showed that elite performance in musicians, chess players, and athletes tracks with structured practice at tasks just beyond current ability, with immediate feedback, aimed at specific improvement targets. Raw time spent matters far less than the structure of the practice itself.

The feedback half of the equation has been studied on its own. Hattie and Timperley's 2007 review, The Power of Feedback, synthesized hundreds of studies and found feedback is one of the largest influences on learning when it is specific, timely, and tied to a goal the learner understands. Feedback about the self ("good job") does little. Feedback about the task, the process, and self-regulation does a lot.

The applied version of this principle lives in medical simulation. McGaghie and colleagues (2011) meta-analyzed simulation-based medical education with deliberate practice and found an effect size of d=0.71 versus traditional instruction, a large effect in education research. The combination of repeated attempts, structured feedback, and targeted difficulty produced skilled clinicians faster and more reliably than lectures and observation.

Ask how many corporate training programs include actual deliberate practice with task-level feedback on the thing the learner is supposed to be getting better at. Sustained practice, with rapid feedback, on representative tasks, beyond the annual hotel-ballroom role-play. The answer is usually zero.

Principle 6: Social learning

Adults learn in groups, from groups, and often because of groups. Bandura's Social Learning Theory (1977) showed that observational learning, modeling, and vicarious reinforcement are central mechanisms of human skill acquisition. We watch people we respect do a thing, we try it, we watch the consequences, we calibrate.

Lave and Wenger's Situated Learning (1991) extended this with the idea of communities of practice. Newcomers learn by legitimate peripheral participation, starting at the edges of a real community doing real work and moving inward as they build competence. The learning is social, contextual, and tied to identity inside the community.

Corporate training violates this principle constantly. A cohort of strangers pulled into a ballroom for a day, receiving content from an external trainer, with no follow-up community, no shared project, and no ongoing interaction, is the structural opposite of a community of practice. Whatever the content, the format denies the learner the social scaffolding that makes skills stick.

What this means for L&D design

Stack the principles side by side and you get a sharp design brief. Adult learning transfers when the content is relevant to an immediate problem, the learner has real agency, the design builds on existing experience, the core unit is a problem not a topic, there is sustained deliberate practice with feedback, and it happens inside a community of practice.

Standard corporate training typically violates four to six of these at once. The fix is to rebuild the design around the principles, rather than bolt another module onto a broken format.

Stop doing. One-off workshops with no follow-up. Linear e-learning with locked navigation. Role-plays with generic scripts. Abstract frameworks delivered without live problems to work. Feedback that rates the learner rather than the behavior.

Start doing. Embed training in real work, on real problems the learner is already trying to solve. Give meaningful choice over pace, path, and focus. Calibrate to prior experience, and let advanced learners skip what they already have. Organize content around problems rather than topics. Build deliberate practice loops with fast, specific feedback on the task. Create or respect the community of practice the learner will actually work in.

This is the design logic that makes something like learning through play work. Games already tick most of the principles by construction: you are solving a problem, with feedback every second, at a difficulty just above your current ability, alongside a group that shares the stakes.

What is contested, and what has been debunked

A responsible read of this literature requires naming where the consensus thins out. Three things are worth flagging.

Knowles's andragogy is not settled theory. Hartree (1984) argued that andragogy is a loose collection of assumptions about adult learners rather than a testable theory, and that some of its core claims (adults are naturally self-directed, for instance) do not survive contact with actual adult learners who often prefer structured guidance. Grace (1996) criticized andragogy for ignoring the social and structural contexts learners live in. Merriam (2001) summarized decades of critique by noting that children and adults often learn in more similar ways than Knowles's framework implies. The six principles above survive because each has independent support from other research traditions, not because andragogy itself is solid ground.

Learning styles are a myth. Pashler, McDaniel, Rohrer, and Bjork (2008) conducted a comprehensive review and found no credible evidence that matching instruction to a learner's purported visual, auditory, or kinesthetic style improves outcomes. The intuition survives in L&D catalogs because it feels personalized, but it is not supported. Design for the content and the task, not for a style inventory.

Pure discovery learning fails without scaffolding. There is a version of self-direction that slides into "just let them figure it out." Kirschner, Sweller, and Clark (2006) made a forceful case that minimal-guidance instruction fails novices, who lack the prior knowledge to structure their own exploration. The self-direction principle describes meaningful agency inside a designed structure, with scaffolding intact. Ericsson's deliberate practice, likewise, requires a coach or a feedback loop, not a learner wandering through the domain alone.

Also worth naming: Mezirow's (1991) transformative learning is a useful theoretical lens for how adults undergo perspective shifts, but it does not produce effect-size claims and should not be cited as if it did. Treat it as a frame for thinking about perspective change, without leaning on it for quantitative claims.

The simulator-for-teams version

If you wanted to design for all six principles at once, the shape that emerges resembles a simulator more than a workshop. Embedded in real work. Problem-centered. Deliberate practice loops with immediate feedback. Social by construction. Responsive to prior experience. Learner-directed inside a designed structure.

Medical simulation already does this for individual clinicians. Flight simulators do it for pilots. The teamwork equivalent has been missing, and that is the category QuestWorks is trying to occupy. We call it the flight simulator for team dynamics. QuestWorks runs on its own platform, cinematic and voice-controlled, and works with Slack for install, invites, onboarding, HeroGPT coaching, leaderboards, and admin commands. Participation is voluntary, quests are not tied to performance reviews, HeroGPT coaching is private, and leaders see aggregate team trends plus strengths-based XP highlights per player rather than surveillance-style telemetry. For more on how the design choices map to the learning research, see the science behind the game and our argument for closed-loop continuous team development.

QuestWorks is one design that ticks the principles. It is not the only possible one. McMaster's problem-based curriculum ticks them. A well-run apprenticeship ticks them. A tabletop RPG session aimed at a real team problem ticks them. The research exists to sharpen the design standard, so use it that way.

How to use the principles as a buyer

The simplest way to apply this reading is to grade any training you are evaluating against the six principles. Pick training that ticks at least four of them on the merits, and be skeptical of anything that ticks fewer than three.

1. Relevance. Is the content tied to a problem the learner is trying to solve right now?
2. Self-direction. Does the learner have real control over pace, sequence, and focus?
3. Prior experience. Does the design calibrate to what the learner already knows, or treat everyone as a blank slate?
4. Problem-centered. Is the core unit of the curriculum a problem to solve, not a topic to cover?
5. Deliberate practice with feedback. Are learners repeatedly attempting the target behavior with fast, specific, task-level feedback?
6. Social learning. Is there a community of practice the learner enters and stays inside, or is the cohort a one-day rental?

Most training procurement conversations do not happen at this level of specificity. They happen at the level of vendor slides and completion rate projections. The research is clear that vendor slides and completion rates fail to predict transfer. The six principles do. The harder question is whether your L&D function is set up to choose on that basis.