Adult Learning Theory at Work: What Andragogy, Kolb's Experiential Cycle and Self-Directed Learning Can and Cannot Prove

The word andragogy did not originate with Knowles. The German educator Alexander Kapp used it in 1833 to describe Plato's approach to educating adults, and the term resurfaced in English through Eduard Lindeman, a colleague of John Dewey who wrote about the informal, experience-based character of adult education in the 1920s. Knowles encountered the term decades later, introduced to it by the Yugoslav adult educator Dusan Savićević, and began using it himself in the late 1960s to name something he believed American adult-education practice was missing: a coherent alternative to pedagogy, the term for the art and science of teaching children, which Knowles thought had been applied to adult classrooms, corporate training rooms and continuing-education programs without much thought for how the audience actually differed from a classroom of children.

Knowles set out that alternative at length in The Modern Practice of Adult Education: From Pedagogy to Andragogy (Knowles, 1980), building on an argument he had already published in the preceding decade. The book did not present andragogy as a predictive scientific theory supported by a specific body of controlled experiments. It presented a set of assumptions about how adult learners differ from children, offered as a practical orientation for people designing adult education and training: teachers, corporate trainers, extension workers, military instructors. That distinction, between a tested theory and a stated set of assumptions, turns out to matter more than it first appears, because the assumptions were absorbed into corporate training practice far faster than anyone checked whether they held up as a scientific account of how adults learn differently from children.

Knowles organized andragogy around six assumptions about the adult learner, refined somewhat across successive editions of his work. An adult's self-concept moves from dependency toward self-direction, so an adult learner resists being treated as a dependent personality by an instructor. An adult brings an accumulating reservoir of life experience that becomes a resource for learning, including the adult's own experience as material to learn from, not only a backdrop to it. An adult's readiness to learn becomes organized around the developmental tasks of social roles, becoming a parent, changing careers, taking on management responsibility, rather than around a curriculum sequence imposed from outside. An adult's orientation to learning shifts from subject-centered to problem-centered, wanting to apply what is learned to an immediate situation rather than store it for later use. An adult is motivated to learn more by internal pressures, self-esteem, job satisfaction, quality of life, than by external rewards such as grades. And an adult needs to know why something is worth learning before investing effort in learning it.

Almost as soon as andragogy entered wide use, adult-education researchers began asking a basic question Knowles' own writing left unresolved: what kind of claim is this? Jerold Davenport and Judith Davenport's (1985) review in Adult Education Quarterly, "A Chronology and Analysis of the Andragogy Debate," traced more than a decade of published argument over whether andragogy described a theory, a set of methods, a philosophy of adult education, or simply a label for existing good practice repackaged under a new name. Different writers, including Knowles himself at different points in his career, answered that question differently, which made andragogy difficult to test empirically in the way a scientific theory needs to be tested: a claim has to specify what observation would count against it before anyone can check whether that observation occurs.

John Rachal's (2002) article in the same journal, "Andragogy's Detectives: A Critique of the Present and a Proposal for the Future," pushed the question further by reviewing the empirical studies that had actually tried to test andragogical assumptions against pedagogical ones, largely by comparing adult classrooms taught with andragogical methods, self-pacing, negotiated objectives, against more conventional instruction, and measuring outcomes such as achievement, satisfaction or retention. Rachal found the existing empirical base thin and methodologically inconsistent: studies varied widely in how they operationalized "andragogical" instruction, rarely isolated which of Knowles' six assumptions was actually being tested, and produced mixed results that did not add up to a demonstration that treating adults according to andragogical assumptions reliably outperforms treating them otherwise. He raised a harder problem underneath the inconsistent findings: andragogy's assumptions were framed as differences between adults and children, but there was no strong evidence that these traits track age rather than accumulated experience, motivation, or the specific learning context, meaning a young but experienced employee and an older but novice one might differ more from each other than either differs from a child, which undercuts the dichotomy the whole framework assumes.

Knowles himself partly conceded this point in later writing, recasting andragogy less as a claim that adults learn in a categorically different way from children and more as a set of assumptions that apply more strongly as a learner becomes more self-directed, experienced and intrinsically motivated, regardless of age, so that a highly self-directed teenager and a highly dependent adult returning to formal study could sit at opposite ends of the same continuum. That revision solved a logical problem (a theory of adult learning should not need its central variable to turn out to be something other than adulthood), but it also quietly gave up andragogy's original claim to be a distinct theory of adult learning rather than a more general description of self-directed, experience-based, problem-centered learning that could, in principle, apply to a motivated adolescent as well. Sharan Merriam's (2001) review, "Andragogy and Self-Directed Learning: Pillars of Adult Learning Theory," reaches a similarly measured verdict: andragogy and self-directed learning remain genuinely useful for organizing how adult educators think about their own practice, but neither has held up, on the evidence available, as an empirically validated, uniquely adult psychological theory in the way the original name implied.

David Kolb's (1984) book Experiential Learning: Experience as the Source of Learning and Development offered a different account of adult learning, built less around how adults differ from children and more around the structure of learning itself, regardless of the learner's age. Drawing explicitly on John Dewey, Kurt Lewin and Jean Piaget, Kolb proposed that learning proceeds through a four-stage cycle: concrete experience, something actually happening to the learner; reflective observation, stepping back to consider what happened and why; abstract conceptualization, drawing a general principle out of that reflection; and active experimentation, trying the new understanding out in a fresh situation, which generates a new concrete experience and restarts the cycle. Kolb also proposed that individuals develop a habitual entry point and preferred path through this cycle, producing four learning styles, converging, diverging, assimilating and accommodating, measured by an instrument he developed alongside the theory, the Learning Style Inventory.

D. Christopher Kayes' (2002) article in Academy of Management Learning and Education, "Experiential Learning and Its Critics: Preserving the Role of Experience in Management Learning and Education," reviewed the specific objections that had accumulated against Kolb's model over nearly two decades of use in management education, without concluding that the underlying idea, that experience matters to learning, should be abandoned. Kayes grouped the criticisms into strands worth separating rather than treating as one undifferentiated complaint. The Learning Style Inventory itself has repeatedly shown weak psychometric properties in independent testing, inconsistent factor structure and modest test-retest reliability, which matters because an instrument used to sort employees or students into categories needs to measure the same thing consistently if that sorting is going to mean anything at all. The four-stage cycle's claim to describe a fixed, universal sequence has also drawn skepticism; real learning episodes do not obviously move step by step through all four stages in the proposed order, and treating the cycle as a required sequence can turn a flexible descriptive idea into a rigid prescription that a training designer feels obliged to build every module around.

Kayes also raised a more conceptual objection: Kolb's model treats the learner largely as an individual processing information through internal stages, somewhat apart from the social and organizational setting the learning actually happens inside, which sits awkwardly against how much workplace learning is shaped by relationships, authority and the specific demands of a job, not by an individual's private cognitive cycle. None of this amounts to a finding that experience is irrelevant to learning, a claim almost nobody seriously defends. It amounts to a more specific and more useful finding: the particular instrument, and the particular four-stage, four-style architecture Kolb built around that broader insight, carry considerably less empirical support than their continued presence in management-training materials would suggest.

Separate from Kolb's own instrument, a much broader and more widely believed claim circulates through corporate training culture: that each person has a fixed, identifiable learning style, commonly simplified to a visual, auditory, or kinesthetic preference, often abbreviated VAK, and that training designed to match a learner's style will teach that person more effectively than training that does not. Harold Pashler, Mark McDaniel, Doug Rohrer and Robert Bjork's (2008) review in Psychological Science in the Public Interest, "Learning Styles: Concepts and Evidence," examined this claim directly and set out exactly what evidence would be needed to support it: a study would need to measure each learner's style, randomly assign learners within each style group to differently formatted instruction, and then show a specific interaction effect, learners taught in their matched format outperforming those taught in a mismatched format, while the reverse pattern holds for the other style group. This is sometimes called the meshing hypothesis, because the benefit is specifically supposed to come from the match between style and format, not merely from a general preference for one format over another.

Pashler and colleagues found almost no published study that met this design standard, and the few that came close mostly failed to find the predicted interaction. People reliably report a preference for one format over another when asked, and self-reported preference is real and worth accommodating for reasons of comfort and motivation, but preference is a different claim from the meshing hypothesis, which is specifically about whether matching format to style improves learning outcomes, not whether it feels more comfortable in the moment. The review's conclusion was blunt by the standards of an academic literature review: there was no adequate evidence to justify restructuring instruction around a learner's assessed style, despite how widely that recommendation had already spread through teacher training and corporate learning and development by the time the review was published. The gap between how confidently "teach to the learning style" circulates in training materials, certification programs and vendor product claims, and how little evidence actually supports the specific meshing claim, is one of the clearer examples in this field of popular practice running well ahead of what the underlying research has shown.

Knowles' (1975) book Self-Directed Learning: A Guide for Learners and Teachers defined self-directed learning as a process in which an individual takes the initiative, with or without help from others, in diagnosing a learning need, formulating a goal, identifying resources, choosing and carrying out a learning strategy, and evaluating the result. The definition describes a process an individual can be more or less engaged in during any given situation, not a fixed personality type a person either has or lacks, which creates an obvious measurement problem for anyone who wants to use the idea practically: if self-direction is situational and variable, how does a training department decide how much structure a given employee or cohort actually needs?

Lucy Guglielmino's (1977) doctoral dissertation at the University of Georgia attempted to answer that measurement problem directly, developing the Self-Directed Learning Readiness Scale, a self-report instrument intended to produce a readiness score usable for placement, counseling, or program design decisions. The instrument was adopted widely across adult education and corporate training, which made its psychometric properties more consequential than they might have been as a purely academic exercise. Lawrence Field's (1989) study in Adult Education Quarterly, "An Investigation into the Structure, Validity, and Reliability of Guglielmino's Self-Directed Learning Readiness Scale," applied factor analysis to the scale's response data and found a factor structure that did not match the eight dimensions the scale was designed to measure, raising a direct construct-validity concern: an instrument whose internal structure does not match its intended design may not be measuring what it claims to measure, whatever its individual questions appear to ask about. Guglielmino published a response defending the scale's development and disputing aspects of Field's analysis, and the exchange between them did not end in a clear resolution either way. The scale remains in use today, and the dispute over how cleanly it measures genuine readiness for self-directed learning, as opposed to general conscientiousness or a favorable self-perception, remains open rather than settled.

None of this amounts to a case for discarding andragogy, Kolb's cycle, or self-directed learning from workplace training design. It amounts to a case for using them honestly, as organizing heuristics with intuitive appeal and decades of practitioner experience behind them, rather than as settled science that automatically justifies whatever training format a vendor or a consultant proposes. A training designer who builds a module around a clear statement of why the content matters, draws on trainees' existing job experience as material rather than treating it as irrelevant background, and gives trainees a problem to work through rather than content to receive passively, is applying something genuinely useful from Knowles' assumptions and from Kolb's broader emphasis on experience, independent of whether andragogy survives as a distinct scientific theory of adult cognition or whether the four-stage cycle describes every learning episode in the proposed order. What the evidence does not support is the much stronger, specific claim that sorting learners by an assessed style and matching format to that style improves outcomes, and a training budget spent on that specific claim is a training budget spent on something the research has had ample opportunity, and has largely failed, to confirm.

Lesson 42 picks up where this one leaves a gap open: before any training is designed at all, an organization needs a disciplined way to establish what a workforce actually needs to learn, a separate, more diagnostic question from how adults learn once a genuine need is established. Lesson 43 then returns to the instructional design frameworks, ADDIE and Bloom's taxonomy, that this lesson deliberately left aside, applying them to a need Lesson 42 has already identified, informed by the honest view of adult learning this lesson has tried to establish rather than by the inflated version that keeps circulating in training brochures.

The checklist below is a reflection tool written for this lesson, not a reproduction of Guglielmino's Self-Directed Learning Readiness Scale and not a validated psychometric instrument. It does not produce a readiness score comparable across people, and the dispute between Field (1989) and Guglielmino over what the published scale actually measures is a reason for caution about any instrument claiming to do so, including this one. Its value is narrower and more useful for an individual reader: noticing, for one specific task currently being learned, which part of the self-directed process described by Knowles (1975) is actually happening and which part is being skipped.

Select a real or realistic workplace training program you can describe in reasonable detail, drawn from your own organization (disguised if needed), a documented case, or a program you research directly. Using that program as your basis, prepare a report of 2,000 to 2,500 words that:

1. identifies which of Knowles' (1975, 1980) six andragogical assumptions are genuinely relevant to this specific learner population, with a clear statement of why, distinguishing their practical design value from the unproven claim that andragogy is a validated theory of adult-specific learning (Davenport & Davenport, 1985; Rachal, 2002);
2. designs one training activity using Kolb's (1984) experiential learning cycle, explaining which elements of the model you are relying on and which specific critique from Kayes (2002) you are deliberately working around, rather than ignoring;
3. evaluates whether a learning-styles-based personalization approach (VAK or otherwise) should be adopted for this program, applying Pashler, McDaniel, Rohrer and Bjork's (2008) meshing-hypothesis evidentiary standard explicitly to justify your decision either way;
4. designs a short self-directed learning component for the program, and explains, referencing Knowles (1975), Guglielmino (1977) and Field's (1989) critique, what "readiness" should mean for this specific learner group and why a generic readiness score would or would not be trustworthy for making a placement decision; and
5. states explicitly where the theoretical and empirical limitations examined in this lesson should change your design, versus where you are knowingly using a framework as a working heuristic despite its thin validation, and why that distinction matters for defending the program's budget.

Use at least five academic sources, including at least four from this article's reference list.

Assessment rubric (100 marks)

1. Davenport, J., & Davenport, J. A. (1985). A chronology and analysis of the andragogy debate. Adult Education Quarterly, 35(3), 152–159.
2. Field, L. (1989). An investigation into the structure, validity, and reliability of Guglielmino's Self-Directed Learning Readiness Scale. Adult Education Quarterly, 39(3), 125–139. doi.org/10.1177/0001848189039003001
3. Guglielmino, L. M. (1977). Development of the self-directed learning readiness scale [Doctoral dissertation, University of Georgia].
4. Kayes, D. C. (2002). Experiential learning and its critics: Preserving the role of experience in management learning and education. Academy of Management Learning and Education, 1(2), 137–149.
5. Knowles, M. S. (1975). Self-directed learning: A guide for learners and teachers. Association Press.
6. Knowles, M. S. (1980). The modern practice of adult education: From pedagogy to andragogy (2nd ed.). Follett.
7. Kolb, D. A. (1984). Experiential learning: Experience as the source of learning and development. Prentice-Hall.
8. Merriam, S. B. (2001). Andragogy and self-directed learning: Pillars of adult learning theory. New Directions for Adult and Continuing Education, 2001(89), 3–14. doi.org/10.1002/ace.3
9. Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning styles: Concepts and evidence. Psychological Science in the Public Interest, 9(3), 105–119.
10. Rachal, J. R. (2002). Andragogy's detectives: A critique of the present and a proposal for the future. Adult Education Quarterly, 52(3), 210–227.

Verification notes: all ten references were checked via web search against publicly indexed bibliographic records, including publisher pages, SAGE Journals, Wiley Online Library, and ERIC, confirming author names, year, title, journal or publisher, volume and page numbers. The historical detail that the term andragogy predates Knowles, coined by Alexander Kapp in 1833 and introduced into English by Eduard Lindeman, with Knowles encountering it through Dusan Savićević, is reported in prose as a historical fact rather than as a numbered academic citation, since it is a well-documented biographical detail rather than a specific research finding requiring its own reference entry. No statistic, finding, or quotation beyond what is reported in the sources above is claimed. The Projapoti Insurance Group case, the tables, the decision scenario, the myth/fact items, the self-assessment, and the synthesis connecting this lesson to Lessons 42 and 43 are original teaching material for this series.

Lesson 42 continues Module 09: once this lesson's honest view of how adults actually learn is established, the next step is building a disciplined process for identifying what a specific workforce actually needs to learn before any training is designed.