Robert Gagné’s Conditions of Learning is an instructional design theory stating that different types of learning (like facts vs. motor skills) require different instructional methods. It outlines five categories of learning outcomes and nine systematic instructional events, each designed to support a stage of information processing.
Core Assumptions of the Theory
Learning is Cumulative: Complex capabilities are built upon a foundation of simpler, prerequisite capabilities.
Different Outcomes Require Different Instruction: You cannot teach a physical skill (like flying a plane) the same way you teach an attitude (like safety compliance).
Systematic Design is Essential: Instruction is planned backwards from the terminal objective, the final capability the lesson should produce, through a structured sequence of prerequisite parts.
Cognitive-Behaviorist Blend & Key Distinctions
Is Gagné’s theory behaviorist or cognitive? It is both. Gagné served as a bridge between the two schools of thought.
From Behaviorism, he adopted the focus on measurable performance outcomes, task analysis, and external reinforcement.
Programmed learning is a close relative. It is a small-step teaching method that grew out of operant conditioning.
Gagné shares its task analysis, explicit behavioral objectives, small sequenced steps, active responding and immediate feedback.
He departs from the behaviorist program in two ways. He gives internal processing a causal role, and he insists that different categories of outcome need different methods.
From Cognitive Psychology, he incorporated internal information-processing models, recognizing that what happens inside a learner’s mind (attention, encoding, retrieval) is critical.
Gagné did not think reinforcement alone could explain learning. As he wrote in 1984, “reinforcement contingencies are not enough” to specify the situation in which learning occurs (Gagné, 1984).
Instructional designers frequently confuse two core elements of his work:
Conditions of Learning: The overarching umbrella framework. It asserts that different types of learning outcomes require specific internal and external states.
Nine Events of Instruction: The tactical, chronological steps an instructor takes during a lesson to trigger those optimal internal states.
Under this model, the instructor acts as an instructional environment engineer who carefully arranges external stimuli.
The learner is an active information processor. To encode the new material, the learner must already hold the right internal prerequisites, such as prior knowledge and cognitive strategies.
Conditions of Learning
Gagné’s framework hinges on the dynamic interplay between the mind of the learner and the environment designed by the instructor.
Internal Conditions
These are the internal cognitive states and capabilities that the learner already possesses or must activate for learning to happen. They include:
Existing Memory Schemata: Pre-existing knowledge structures stored in long-term memory.
Information Processing Capabilities: The capacity of working memory, the limited mental workspace that holds what you are thinking about right now, and control of attention.
Motivation: The internal drive and goal-orientation of the learner. For an adult learner, a critical internal condition is the perception of relevance, the belief that the upcoming content will directly solve a current problem.
External Conditions
These are the environmental stimuli and instructional actions manipulated by the instructional designer.
This includes the structure of the text, the sequence of graphics, the clarity of instructions, and the timing of feedback.
External conditions are the circumstances a designer controls. They must be mapped to the internal processing they support.
Consider a bold heading with a visual arrow (Event 1: Gaining Attention).
It focuses the learner’s working memory, the limited workspace for current thinking, on one specific stimulus. That changes their internal condition.
Crucially, external conditions must change based on the intended learning outcome. If you are teaching an intellectual skill like fraction division, your external condition must include step-by-step mathematical logic models. If you are teaching an attitude like safety compliance, your external conditions must include emotional storytelling or role-modeling.
Gagné called this the cumulative learning mechanism. External stimuli trigger the retrieval of simpler, previously mastered internal capabilities. Learners then combine them into a brand-new, more complex capability.
A teacher who reminds learners of two rules, for example, helps them combine those rules into a higher-order rule they have never met.
If either prerequisite cannot be retrieved, the combination fails. The designer’s job is then to find the missing prerequisite and teach it first.
Gagné’s Nine Events of Instruction
The nine events provide a framework for designing and delivering instruction in a structured and systematic manner.
Each instructional event is deliberately designed to trigger and support a specific internal cognitive process within the learner.
The order is deliberate. Moving through the nine events lines up the teacher’s external steps with the mental processing learners need to attend, encode and retrieve. It is a design rationale drawn from the information-processing model, not a sequence tested against alternatives.
Gain Attention: This event aims to capture the learners’ attention and engage them in the learning process. It can be achieved through the use of stimulating and relevant stimuli or by posing questions or problems.
Inform Learners of the Objective: Learners need to be aware of the specific learning objectives or goals they are expected to achieve. Clear communication of these objectives helps to focus their attention and motivate them to learn.
Stimulate Recall of Prior Knowledge: Activating learners’ prior knowledge helps them connect new information to existing mental frameworks. By reviewing relevant concepts or experiences, learners can build upon what they already know.
Present the Content: The instructional content is presented to the learners in a structured and organized manner. It should be logically sequenced, chunked into manageable units, and delivered using appropriate instructional strategies such as lectures, visuals, or multimedia.
Provide Guidance: Learners need guidance and support to understand and acquire new knowledge or skills. This event involves providing clear explanations, examples, demonstrations, and instructions to assist learners in grasping the content.
Elicit Performance: Learners are given opportunities to practice what they have learned. This active participation helps reinforce the newly acquired knowledge or skills and allows for feedback and correction if needed.
Provide Feedback: Learners receive feedback on their performance, indicating whether they have achieved the desired learning outcomes. Feedback helps them assess their progress, identify areas for improvement, and reinforce correct understanding or behavior.
Assess Performance: This event involves assessing learners’ performance to determine the extent to which they have achieved the learning objectives. Various assessment methods such as quizzes, tests, or practical exercises, can be used to evaluate their progress.
Enhance Retention and Transfer: The final event focuses on promoting long-term retention and transfer of the learned material to real-world contexts. Strategies such as providing opportunities for review, application in different situations, and promoting transfer of knowledge to other domains are employed to solidify learning.
The Internal Process Behind Each Event
From the third edition of his book onward, Gagné described the learner in information-processing terms (Gagné, 1977, 1985). The account follows the multi-store model of memory (Atkinson & Shiffrin, 1968).
Stimulation reaches the senses. Selected features pass into working memory, encoding stores them in long-term memory, and retrieval brings them back. A response generator then produces performance, while executive control processes and expectancies govern the whole flow.
Each event serves one stage.
#
Event of instruction
What the designer does
Internal process supported
1
Gain attention
Pose a problem, question or surprising stimulus
Reception (alerting; the stimulus enters processing)
2
Inform learners of the objective
State what they will be able to do
Expectancy (a goal to work toward)
3
Stimulate recall of prior learning
Prompt retrieval of prerequisite knowledge
Retrieval to working memory
4
Present the content
Deliver organized, chunked material
Selective perception
5
Provide learning guidance
Examples, models, cues, explanations
Semantic encoding into long-term memory
6
Elicit performance
Have learners apply what they have learned
Responding
7
Provide feedback
Tell learners how well they did and why
Reinforcement
8
Assess performance
Check achievement of the objective
Retrieval
9
Enhance retention and transfer
Review, varied contexts, job aids
Generalization
Semantic encoding means storing new information by its meaning, which examples, analogies and worked models support.
Getting Each Event Right
Each event has a typical way of going wrong. These tips apply Gagné’s reasoning to the common failure points.
Gain Attention: Tie the attention-grabber to the lesson content. A striking but irrelevant gimmick spends scarce working memory on the wrong thing.
Inform of Objective: State what learners will be able to do in observable terms, such as “calculate the mean from a frequency table” rather than “understand averages.”
Stimulate Recall: Treat this as a diagnostic checkpoint. If learners cannot retrieve a prerequisite, pause and teach it first.
Present Content: Sequence the material logically and chunk it. Dumping a fifty-page policy manual on a new hire overloads working memory.
Provide Guidance: Match the help to the outcome: a worked example for an intellectual skill, a demonstration for a motor skill, a respected model for an attitude.
Elicit Performance: Give practice on instances learners have not yet seen, so they must apply the rule rather than copy the example.
Provide Feedback: Say whether the response was right and, ideally, why not and how to fix it. In self-paced modules, design it in advance.
Assess Performance: Test the capability independently, without the earlier guidance, and match the test to the objective stated in Event 2.
Enhance Transfer: Build in review, varied contexts and job aids, but remember that transfer also depends on the workplace (Blume et al., 2010).
They are design guidance, not tested rules.
Not All Events Are Required Every Time
Gagné’s own position was that the events are functions to be supplied, not a script to follow in full. Instruction does not need all nine on every occasion. Learners can often supply some of the internal processing themselves (Gagné et al., 1992).
Over-specifying every step can even hinder learning. An experienced learner who already has the prerequisites does not need a recall prompt. A learner who read the objective in the syllabus does not need it repeated.
The expertise-reversal effect shows that supports which help novices can become redundant, or even harmful, for experienced learners (Kalyuga et al., 2003).
Martin et al. (2007) tested this by dismantling a lesson.
Aim: To find out which elements of a systematically designed lesson contribute to learning, by combining and removing objectives, information, practice, examples and review.
Method: College students used one of six versions of a web-delivered computer-literacy lesson: the full lesson, four versions each lacking one element, and an information-only version.
Results: The four versions that included practice produced significantly better posttest scores and more positive attitudes than those without it.
Conclusion: Practice was the element linked to better learning. No comparable benefit was reported for objectives, examples or review (Martin et al., 2007).
This is one of the few component-level tests of an events-style design. It cuts two ways.
It supports Events 6 and 7, practice and feedback, but gives no support to reading the nine events as nine equally necessary steps.
The lesson was one short topic. The participants were college students, and the lean version was a minimal-information control, not a lecture.
The theory offers no rule for deciding which event to cut for which learner. That gap is filled by the designer’s diagnosis of what learners already know.
Nine Events at a Glance
The nine events give teachers a straightforward recipe for lessons that help students acquire and recall new concepts.
The model is, perhaps, most linked to Ausubel’s perspective, with an emphasis on the role of the teacher in carefully managing the development of new schemata.
The link is clearest at the recall event. Ausubel (1960) tested whether introducing a relevant, inclusive concept before unfamiliar material, called an advance organizer, would improve learning and retention. Gagné’s recall of prerequisite knowledge shares that logic.
This is an interpretive comparison, not a tested one.
Even so, the approach stresses student activity through practice and feedback, so it does not simply center the teacher.
Five Categories of Learning Outcomes
Gagné argued that all human learning falls into five distinct categories. If you do not know which category your goal belongs to, you cannot design the correct instruction.
1. Intellectual Skills
This category involves knowing how to do something procedurally. It is broken down into a strict, cumulative learning hierarchy from simplest to most complex:
Discrimination: Telling two highly similar stimuli apart (e.g., distinguishing an “b” from a “d”).
Concrete Concepts: Identifying an object by its physical characteristics (e.g., pointing out a triangle).
Defined Concepts: Identifying an abstract class of objects or ideas by its definition (e.g., explaining what “democracy” or “justice” is).
Rules: Applying a single principle to guide action (e.g., multiplying by two, or capitalizing the first word of a sentence).
Higher-Order Rules (Problem Solving): Combining multiple rules to solve a novel, complex issue (e.g., engineering a structural bridge or writing software code).
Gagné (1984) linked this category to what cognitive psychologists call procedural knowledge, the knowledge of how to carry out a sequence of steps. A simple rule is typically learned abruptly: either the learner can apply it to any instance, or has not yet learned it.
Longer procedures are learned in stages. Practice then makes them faster and more automatic without changing what the learner can do.
2. Verbal Information
This is knowing that something is true: declarative facts, labels, and organized bodies of knowledge (e.g., stating the capital of France or memorizing product names).
It is learned by attaching the new fact to an expansive, existing web of verbal meaning.
Gagné (1984) distinguished two kinds of verbal information. Isolated facts and verbatim passages, such as the names of the months, are learned for exact recall. Extra practice makes that recall more accurate and sure.
Meaningful knowledge is learned differently, as an organized schema with main and subordinate ideas. Continued learning elaborates the schema by linking it to more concepts. That is a qualitative change. It is not just a more accurate recitation.
3. Cognitive Strategies
These are the internal, self-managed capabilities the learner uses to guide their own attention, learning, remembering, and thinking processes.
It is the learner’s internal “management system” (e.g., creating a mind map to study for an exam or inventing an acronym to remember a list).
Unlike intellectual skills, which look outward at manipulating data, cognitive strategies look inward at managing one’s own mental processing.
Gagné (1984) noted that strategies vary in generality. A broad strategy such as “break the problem into parts” may be clear on one task yet fail to transfer to a new one. The category corresponds closely to what later researchers study as metacognition and self-regulation.
4. Motor Skills
These are coordinated physical movements requiring muscle control and speed (e.g., swinging a golf club, welding a pipe, or typing on a keyboard).
Instruction requires clear behavioral demonstrations, physical or mental practice, and real-time physical feedback adjustments.
Motor skills are a separate case. Gagné (1984) kept them apart from intellectual skills because a motor skill improves in smoothness and timing only through repeating the movement itself, gradually and over a long period.
He drew on Fitts and Posner (1967), who described three phases:
Cognitive phase: The learner masters the sequence of movements.
Associative phase: The parts are fitted together smoothly.
Autonomous phase: The skill runs with little conscious attention.
5. Attitudes
An attitude is an internal state. It influences a learner’s choice of action toward things, people or events.
Examples include choosing to wear safety goggles. Another is deciding to follow ethical guidelines.
Modifying attitudes is incredibly difficult because you cannot simply read a fact to change a belief. It requires exposing learners to a respected role model, demonstrating the positive or negative consequences of the choice, or leveraging powerful emotional narratives.
Gagné (1984) saw it the same way. An attitude, he explained, is an inferred internal state that modulates behavior rather than determining it.
Telling learners what to think has consistently been ineffective.
A human model is essential. The learner watches a respected model make a choice and sees the consequences, an idea Gagné linked to Bandura’s work on observational learning.
Gagné’s Learning Outcome Category
Rough Equivalent in Bloom’s Taxonomy
Key External Instruction Strategy
Intellectual Skills
Cognitive Domain (Apply, Analyze, Evaluate)
Provide step-by-step procedural scaffolding and practice variations.
Verbal Information
Cognitive Domain (Remember, Understand)
Present information in highly organized chunks using dual-coding.
Cognitive Strategies
Cognitive Domain (Create / Metacognition)
Provide opportunities for open problem-solving and self-reflection.
Motor Skills
Psychomotor Domain
Provide precise physical demonstrations and repetitive physical practice.
Attitudes
Affective Domain
Utilize human modeling, storytelling, and consequence analysis.
Why Five Categories, Not a Taxonomy
Gagné (1984) said plainly that there is no particular reason to treat the five outcomes as a taxonomy. He set four tests for a good category of learning outcome:
Performance class: It can be defined as a class of human performance that learning makes possible.
Wide scope: It applies across many activities, whatever the learner’s intelligence, age or background.
Distinct demands: It differs from the others in the prior learning needed, how it is encoded, and what retrieval and transfer require.
Within-category rules: Principles about its learning generalize across tasks inside the category, but not across categories.
That is the key difference from Bloom’s taxonomy. Bloom ranks objectives by cognitive complexity, whereas Gagné groups capabilities by the conditions needed to learn them. Only within intellectual skills does he propose a strict hierarchy.
The two complement each other. Bloom’s action verbs help write the objective for Event 2, and Gagné’s categories decide how to teach it.
The argument comes from Gagné’s 1984 award address, summarized here.
Aim: To ask what people learn, broadly enough to cover masons, astronauts and schoolchildren alike.
Method: Argument and review. Gagné set four criteria for a category and tested five candidates against research on information processing, problem solving, motor learning and attitude change.
Results: Five categories that differ in how they are learned and how continued practice changes them. Skills and strategies are often acquired abruptly, verbal knowledge is elaborated, motor skills improve slowly, and attitudes resist direct persuasion.
Conclusion: Generalizations hold within the five categories but not between them (Gagné, 1984).
The paper was an invited award address, which the journal states was published without the usual editorial review. It is an authoritative statement of the theory by its author, not an independent test of it.
Matching the Events to the Outcome
Because outcomes differ, so do the events that carry the most weight. The table below applies Gagné’s reasoning about the categories (Gagné, 1984). It is not a prescription from his work.
Outcome
Events that carry most weight
Reasoning
Intellectual skill
3 (prerequisites), 5 (varied examples), 6 (new instances)
A rule is applied to new instances, so practice must vary
Verbal information
4 and 5 (organized, meaningful presentation), 9 (review)
Knowledge is retained when elaborated into a schema
Cognitive strategy
5 (modeled strategy), 6 (use on a novel task)
Strategies transfer poorly unless used on varied tasks
Motor skill
5 (demonstration), 6 and 7 (repeated practice with feedback)
Smoothness comes only from practicing the movement itself
Attitude
4 and 5 (respected human model), 7 (observed consequences)
Telling learners what to believe is ineffective
Gagné’s Learning Hierarchy and Task Analysis
Gagné turned the cumulative-learning idea into a practical tool: the learning hierarchy. It shows a designer what to teach first.
What a Learning Hierarchy Is
A learning hierarchy is a diagram of the prerequisites of a skill. The designer starts with the terminal objective and repeatedly asks, “What must the learner already be able to do in order to learn this?”
The questioning stops when the learner’s entry capabilities are reached.
The hierarchy maps necessary prerequisites. Its test is empirical: learners who have mastered a higher skill should have mastered the subordinate skills beneath it.
Learners who lack a subordinate skill should fail the skill above it. Gagné (1968b) set out the idea in his presidential address to the educational-psychology division of the American Psychological Association.
Task analysis is the method that produces the hierarchy. It tells a designer what to teach and in what order, and where to place a diagnostic check.
It works best where a task has a definable procedure and clear right answers.
A Worked Example: Solving a Simple Equation
Take the terminal objective “solve 3x + 4 = 19,” a higher-order rule that combines several simpler rules. Working downward, the designer asks what the learner must already be able to do.
Terminal skill: Solve a two-step linear equation. Subtract 4 from both sides (19 − 4 = 15), then divide by 3 (15 ÷ 3 = 5), so x = 5.
Prerequisite rules: Keep an equation balanced by doing the same operation to both sides, and use each operation’s inverse.
Prerequisite defined concepts: Equation, variable, inverse operation and “both sides.”
Prerequisite concrete concepts and discriminations: Tell the symbols +, −, × and ÷ apart, and read “=” as a statement of balance, not “now write the answer.”
Entry capabilities: Add, subtract, multiply and divide whole numbers.
A short check on the lowest levels before the lesson shows which prerequisite, if any, an individual learner lacks. Instruction then starts there.
Suppose a learner can do the arithmetic but cannot explain what “balance” means. The gap is at the defined-concept level. Teaching the two-step procedure will not work until it is filled.
Evidence for Learning Hierarchies
The earliest empirical work came from Gagné and his collaborators. Gagné et al. (1962) examined how learners acquired a mathematical task built from subordinate subtasks.
They reported that success at higher levels depended on prior mastery of the subtasks beneath them.
White (1974) later gave the idea its most careful test.
Aim: To set out a more rigorous procedure for validating a learning hierarchy, and to test whether intellectual skills are learned hierarchically while verbalized knowledge is not.
Method: White (1974) described nine stages of validation, amending earlier methods, and applied them to test whether the order of dependency among skills held up.
Results: The stages produced strong evidence that intellectual skills form hierarchies, whereas verbalized knowledge did not show the same hierarchical dependence.
Conclusion: The study supported Gagné’s claim that the prerequisite structure applies to intellectual skills specifically, not to knowledge in general.
The result also narrows the theory. Hierarchy logic covers intellectual skills, so a lesson built to teach isolated facts or an open-ended attitude cannot appeal to it. The abstract does not report how widely the result generalized across subject areas.
White and Gagné (1974) later reviewed the field. They proposed that research on whether hierarchies are valid could be considered virtually at an end.
That is a judgment by the theory’s founder and its main validator. It is considered, but not wholly independent.
Practical Applications
Gagné’s model is highly adaptable across various sectors. Below are concrete examples of how it is scaled into real-world environments.
Corporate Training & Microlearning
Corporate Onboarding Mistake: Dumping 50-page policy manuals onto a new hire on day one, assuming Event 4 (Presentation) is enough. This causes instant cognitive overload and poor retention.
Microlearning Modules: A 3-minute mobile module for a field sales representative can compress the 9 events:
Event 1: Push notification with a compelling question: “What if a client drops this objection today?”
Event 2-3: One-sentence objective and a prompt to recall a similar objection.
Event 4-5: An infographic (Presentation) accompanied by an audio trick (Guidance).
Event 6-7: A single interactive multiple-choice prompt with analytical feedback.
Event 8-9: A downloadable digital job aid card sent to their phone dashboard for field use.
Two large meta-analyses show what workplace training achieves, and what the events cannot control.
Arthur et al. (2003) pooled organizational training studies and found medium to large mean effects on four criteria.
Reactions: about 0.60
Learning: about 0.63
Behavior: about 0.62
Results: about 0.62
Training method, the type of skill trained and the evaluation criterion all related to effectiveness. The review did not test the nine events specifically. It shows that structured design is worth the effort, not that this structure is the best one.
Blume et al. (2010) addressed Event 9 directly. Their meta-analysis of 89 studies linked transfer to cognitive ability, conscientiousness, motivation and a supportive work environment.
Most predictors mattered more for open skills such as leadership than for closed skills such as computer software. A job aid cannot compensate for an unsupportive workplace.
Microlearning fares well too. A systematic review of 40 studies found that it generally improved learning outcomes (Monib et al., 2025). That review concerns the format in general, not events-based design, so it supports short modules rather than the nine-event sequence.
Traditional Education
Elementary Classrooms: A teacher uses Gagné by turning off the lights to gain attention (Event 1), connecting today’s long-division lesson to yesterday’s subtraction practice (Event 3), using colored chalk blocks to show the steps (Event 5), and sending children to the whiteboard to write out answers (Event 6).
Higher Education Lectures: Instead of lecturing continuously for 90 minutes, a university professor uses Gagné’s model to structure the lecture into 20-minute cycles. Each cycle begins with a provocative real-world problem statement (Event 1) and ends with a brief, 2-minute peer-sharing discussion prompt (Event 6) to verify comprehension.
Technical & Specialized Verticals
Healthcare Training: In an advanced life support simulation, educators apply Gagné’s events to teach high-fidelity resuscitation teams.
Event 5 (Learning Guidance) Example: The instructor provides an icon-based checklist next to the dummy, demonstrating an easy way to track cardiac drug rhythms.
Synchronous vs. Asynchronous Nuance: In synchronous simulation labs, Event 7 (Feedback) is delivered live by a monitoring instructor. In asynchronous medical software modules, Event 7 must be hardcoded as an automated branching scenario where selecting the wrong medication path shows a virtual patient’s vitals dropping.
Software Tutorials: When teaching software development, avoid static text. Use a split-screen interactive setup. The left side presents the code syntax (Event 4), while the right side provides a live terminal sandbox that evaluates the user’s input in real time (Event 6 & 7).
Worked Example: A Nine-Event Lesson Plan
Take an intellectual skill, rule using. By the end of a 50-minute lesson, students aged 14 to 16 should be able to calculate the mean from a frequency table. The figures below were computed, not estimated.
Event
Minutes
What happens
1. Gain attention
2
The class tallies how many siblings each of its 30 members has, then is asked, “What is the typical number?” The gap between a long tally and a single number creates the need.
2. Inform of the objective
1
“You will be able to calculate the mean from any frequency table and explain why we multiply.”
3. Stimulate recall
4
Students find the mean of the raw list 2, 4, 6, 8 (20 ÷ 4 = 5) and recall that 9 students with 1 sibling contribute 9 × 1 = 9 siblings in total. A learner who cannot do this is helped before continuing.
4. Present the content
10
The tally becomes a table (x = 0, 1, 2, 3, 4 with frequencies f = 4, 9, 10, 5, 2), chunked into three steps: build an f × x column (0, 9, 20, 15, 8), total both columns (Σf = 30, Σfx = 52), then divide (52 ÷ 30 = 1.73 to two decimal places).
5. Provide guidance
8
A think-aloud worked example, plus a common error shown and corrected: dividing 52 by the number of rows (5) gives 10.4, an impossible answer because the mean cannot exceed the largest value, 4. “The mean must lie between the smallest and largest value” is taught as a strategy.
6. Elicit performance
10
Students calculate the mean for a new table (x = 1 to 5, f = 3, 6, 8, 2, 1: Σf = 20, Σfx = 52, mean 2.6).
7. Provide feedback
4
A model answer is shown with the working, and each student’s error, if any, is named (forgot to multiply; divided by the rows) and corrected.
8. Assess performance
8
An independent item with no help (x = 10, 20, 30; f = 5, 3, 2: Σf = 10, Σfx = 170, mean 17), marked against the objective.
9. Enhance retention and transfer
3
Students will apply the method to a real data set from another subject, and a short retrieval quiz is scheduled for the following week.
Two features stand out. First, every event does a specific job, and Event 8 tests exactly what Event 2 promised.
Second, the plan is not rigid. A class that already knows the procedure can shrink Events 3 and 5. A class with weak multiplication needs a longer Event 3 and possibly a prerequisite lesson first.
Worked Example: Teaching an Attitude
Suppose the outcome is not a skill but an attitude: laboratory staff choosing, unprompted, to wear eye protection. Following Gagné (1984), telling people to do so is unlikely to work. The lesson leans on a human model and on consequences instead.
Event 1: A short, true-to-life account of a near miss gains attention.
Event 2: The objective is stated as a choice: “you will choose to wear eye protection when handling this solvent.”
Event 3: Recall of the hazard properties learners already know prepares the ground.
Events 4 and 5: A respected senior technician narrates her own choice and what followed.
Events 6 and 7: Learners make the choice in a short scenario and see the consequence of each option.
Event 8: Assessment is observation of what people choose in an unprompted lab session, since an attitude can only be inferred from choices.
Event 9: Later reminders at the point of use reinforce transfer.
Comparison with Other Learning Theories & Frameworks
To position Gagné within the broader field of learning design, it helps to see how his model complements or contrasts with other foundational frameworks.
Constructivism (e.g., Vygotsky)
The Contradiction: Pure Constructivism argues that learners should discover rules and concepts on their own by exploring complex, unguided environments. Gagné argues this is highly inefficient for novices. He believes instruction must be systematically directed, structured, and scaffolded from above.
The Complement: Gagné’s Event 5 (Learning Guidance) can be scaled back for advanced learners, transitioning into Vygotsky’s Zone of Proximal Development (ZPD). Here, the instructor provides just enough scaffolding to help the student stretch their capabilities without causing frustration.
Gagné took a position on discovery himself. A 1966 chapter of his, reported by Karpov (2003), was blunt. Expecting people to find a concept by trial and error, it argued, amounts to asking them to “behave like an ape.”
The remark targets unguided trial and error. It records a position, not a finding.
The meta-analytic record helps keep two targets apart. Alfieri et al. (2011) found that unassisted discovery lost to explicit instruction. Discovery with guidance, including feedback and scaffolding, outperformed other methods overall.
Vygotsky is routinely filed under constructivism. Karpov (2003) argues the reverse: Vygotsky held that children cannot be required to rediscover laws that humanity has already established.
That is one neo-Vygotskian’s reading, not a neutral summary. It does show that “constructivist” covers very different attitudes to explicit teaching.
Piaget and Stage Theories of Development
Gagné (1968a) argued that intellectual development can be understood as the cumulative result of learning. It is a growing structure of learned capabilities, each resting on simpler ones.
His theory is a deliberate rival to stage accounts.
On a Piagetian view, learners move through qualitatively different stages that limit what they can understand at a given age. On Gagné’s view, the limits are largely the learner’s current set of prerequisite capabilities, which instruction can supply.
The disagreement is the familiar one between maturation and learning as drivers of development. The two are not strictly exclusive.
ADDIE vs. Gagné
ADDIE (Analyze, Design, Develop, Implement, Evaluate) is a macro-level project management lifecycle framework used to build an entire course or curriculum.
Gagné’s Nine Events operate at the micro-level, detailing the exact psychological flow of a single lesson or learning module within that larger ADDIE architecture.
ADDIE is a label, not an authored model. Molenda (2003) searched for its origin and could not locate any single founding source. He concluded that it is a colloquial name for a generic systematic approach to instructional development.
Reiser (2001) traces that systems approach to the wartime training tradition from which Gagné’s work grew. That is why the two fit together easily.
Merrill’s First Principles of Instruction
Merrill (2002) distilled instructional-design theories into five first principles. Learning is promoted when learners solve real-world problems, activate existing knowledge, see new knowledge demonstrated, apply it, and integrate it into their own world.
The overlap with Gagné is clear. Activation resembles Event 3, demonstration resembles Event 5, and application resembles Event 6.
The difference is one of grain and emphasis. Merrill is task-centered and starts with a whole problem. Gagné breaks the lesson into cognitive processing steps and focuses on how external events trigger them.
Kolb’s Experiential Learning Cycle
Kolb’s cycle moves learners from concrete experience through reflection and abstract conceptualization to active experimentation.
Gagné’s framework can support it. An experiential challenge can serve as the attention-gaining event and the early performance event. The later events then formalize the rules and principles that the experience exposed.
Cognitive Load Theory (Sweller) & Kirkpatrick
Cognitive load theory starts from the same information-processing architecture as Gagné. Cognitive load is the demand a task places on working memory, the limited workspace for current thinking (Sweller et al., 2019).
Load comes in two kinds. Intrinsic load comes from the material itself, and extraneous load arises from poor presentation.
Managing both during Events 4 and 5 protects the learner’s working memory (Sweller et al., 1998).
Both theories rest on the same memory model. The agreement is consistency of assumptions, not independent confirmation.
Kirkpatrick’s four levels (reaction, learning, behavior, results) are a way of judging training, not a theory of learning.
Practitioners commonly pair Event 8 (Assess Performance) with the learning level. They pair Event 9 (Transfer) with the behavior level, meaning whether trainees do something differently on the job. That pairing is a convention, not a finding.
Alliger and Janak (1989) reviewed the scheme thirty years after its introduction. The levels remain a widely used framework, not a validated causal chain.
Evidence Check: Does the Theory Hold Up?
Verdict: Supported with qualifications.
Teaching organized around the nine events beats one-way lecturing in health-professions courses. In the few small randomized trials, the knowledge advantage is moderate. The practical-skill advantage is larger but varies widely between trials.
The claim that intellectual skills are built in a strict prerequisite hierarchy is weaker. It rests on older, mostly proponent-led studies that have not been rerun at scale.
Nine Events Beat Lecturing, but by Less Than the Pooled Figure Suggests
Level of evidence: Meta-analysis or systematic review; not yet replicated.
The claim is simple. Teaching designed around Gagné’s nine events produces better learning outcomes than conventional lecture-based teaching.
One meta-analysis turned up: Li et al. (2025). It pooled eleven studies of 825 health-professions learners, and every one compared the nine events with lecture-based teaching.
It reported a pooled standardized mean difference of 1.55 for knowledge examination scores and 1.83 for practice scores. Learning compliance and teaching satisfaction also favored the nine events.
Splitting the knowledge-score result by design changes the picture.
Four randomized trials (240 participants): an effect of 0.43 (95% CI 0.17 to 0.68).
Seven cohort studies: an effect of 2.20 (95% CI 1.08 to 3.32).
For practice scores, the same four trials gave a larger but far less consistent effect (SMD 1.39, 95% CI 0.37 to 2.41, I² 92%). The randomized and cohort estimates did not differ significantly.
Below the review sit small quasi-experiments such as Habib (2024), with 20 residents. A component-level experiment, Martin et al. (2007), found that the benefit lay with practice rather than with the full set of elements.
No independent replication or second meta-analysis turned up.
The review supports the claim, though for knowledge not at the size of the pooled figure. The randomized trials should outweigh the cohort studies, because random allocation protects against selection effects. Even so, they were unblinded and remain open to instructor and novelty effects.
The meta-analysis occupies the top rung. Its inputs are the weakness. The authors describe the included studies as generally low in quality and small in sample size. Heterogeneity was very high for knowledge and practice scores (I² of 95% and 90%).
The four trials are small. Each had 21 to 35 learners per arm, and two combined the nine events with another method, so they cannot isolate the events.
Every comparator was lecturing. The benefit may therefore come from structured practice and feedback in general, the element that carried the effect in Martin et al. (2007). It need not come from the nine-step sequence.
The effect looks real but smaller than claimed. Whether the nine-step sequence adds anything beyond other structured designs is untested.
Prerequisite Hierarchies Rest on Older, Proponent-Led Evidence
Level of evidence: Correlational or self-report study; partly replicated.
The claim is that intellectual skills are learned hierarchically. A learner masters a higher skill only after mastering the subordinate skills it depends on. Isolated verbal facts are not learned in a hierarchy.
White (1974) set out a nine-stage procedure for validating a learning hierarchy. It reported strong evidence that intellectual skills are learned hierarchically while verbalized knowledge is not.
Earlier work by Gagné and colleagues reported that success on the higher subtasks of a mathematical task depended on prior mastery of the subordinate subtasks (Gagné et al., 1962). The early results fit the hierarchy.
White and Gagné (1974) concluded from a review that research on the validity of hierarchies could be considered virtually at an end.
Recent assessment systems built on Gagné-style hierarchies report diagnostic use but do not test the dependency claim (Tang et al., 2020, 2021).
These are validation studies based on learners’ test performance. They show that patterns of mastery fit the hierarchy. They cannot show that teaching a subordinate skill causes later success.
The evidence also lacks independence. The principal authors are Gagné and a close collaborator, and no large, independent, preregistered rerun or meta-analysis was found.
The evidence supports a bounded version of the claim. For well-defined skills such as arithmetic and algebra procedures, mastery of prerequisites predicts success on higher skills. Teaching in that order is sensible. It does not establish a strict, general causal law.
In short, learners who lacked a prerequisite skill tended to fail the skill built on it, which makes prerequisite checks a reasonable design step. The strictness and generality of the hierarchy have been tested mainly by its proponents and are not settled.
Critical Evaluation
No learning theory is flawless. Understanding the limitations of Gagné’s model prevents its dogmatic misapplication.
Core Criticisms
Too Rigid and Linear: Critics argue that forcing every piece of learning through a strict 1-to-9 sequential pipeline creates an artificial, assembly-line experience.
Teacher-Centric: The framework was engineered in a time when classrooms were primarily instructor-led. It positions the instructor as the primary driver of environmental stimuli, which can minimize learner agency.
Poor for Highly Creative or Open Fields: Gagné’s framework struggles in highly abstract, artistic, or open-ended subjects (such as creative writing, abstract art design, or high-level philosophical exploration) where definitive “correct rules” or “clear procedural outcomes” simply do not exist.
Lesson-Level Only: The nine events organize a single lesson, not a course, and say little about spacing practice over weeks.
Idealized Categories: Real learning objectives often straddle the five categories, and the hierarchies behind them are validated only for particular skills.
Too Rigid and Linear
Gagné conceded part of this point. Three strands of evidence bear on it.
His own caveat: Instruction does not need all nine events on every occasion (Gagné et al., 1992).
Expertise reversal: Instruction that helps novices can become redundant or even harmful for experienced learners (Kalyuga et al., 2003).
Component evidence: In one lesson-dismantling study, practice was the element linked to better learning (Martin et al., 2007).
The criticism that survives is about choice. The theory gives no tested rule for deciding which events to drop for which learners.
The order itself is a design heuristic drawn from the information-processing model, not a sequence tested against alternatives. The flexibility is therefore real but left to the designer.
The designer must diagnose what learners already know.
Teacher-Centric
The designer, not the learner, selects the stimuli. The framework therefore casts the instructor as the driver of the learning environment.
Jonassen (1991) argued that the systematic-design tradition rests on objectivist assumptions, in which knowledge is a fixed reality to be transmitted. He set them against constructivist alternatives.
The reply from the guidance research is partial.
Kirschner et al. (2006): Strong guidance suits novices, because minimal guidance overloads working memory.
Alfieri et al. (2011): Discovery with guidance, such as feedback or worked examples, outperformed other forms of instruction.
That comparison pooled unassisted discovery with explicit instruction. It does not show guided discovery beating explicit teaching.
A fair reading is that the events fit well-defined content for novices and fit less well where learners should build their own interpretations. The theory also gives learners an active role in practice and feedback, so “teacher-centric” describes its emphasis more than its kind.
Poor for Highly Creative or Open Fields
Task analysis and hierarchies work best where a task has definable procedures and clear right answers. Three findings point to the limits.
Hierarchies: White (1974) found hierarchical structure for intellectual skills but not for verbalized knowledge.
Transfer: Blume et al. (2010) found that predictors of transfer behave differently for open skills such as leadership than for closed skills such as software use.
Attitudes: Gagné (1984) himself found them the category least well understood.
Together, these findings point one way. Tidy prerequisite logic suits closed, procedural content best. Creative writing, abstract art and philosophical inquiry have no definitive correct rules for a hierarchy to map.
The recommended lever for attitudes, the human model, therefore rests on a thinner base of evidence than the guidance for intellectual skills.
Lesson-Level Only
The strongest evidence for long-term retention concerns when practice happens. Two sources show it.
Dunlosky et al. (2013): A major review rated retrieval practice and spaced practice among the most useful learning techniques.
Events 8 and 9 mention assessment and review. They do not tell a designer how to space retrieval over weeks. A lesson that follows all nine events can still be forgotten if nothing in the surrounding course revisits it.
That places the theory inside a larger design task. It is one reason the events are often paired with macro-level frameworks such as ADDIE.
The nine events say little about either technique. Spacing is a decision for the whole course, not for a single lesson plan.
Idealized Categories
Gagné (1984) himself says the five categories are not a taxonomy. Real learning objectives often straddle them.
A nursing procedure shows the problem. It is an intellectual skill, a motor skill and an attitude at once. The rough mapping to Bloom’s domains is likewise approximate, so the table of equivalents is a teaching aid, not a classification rule.
Hierarchies have limits too. A hierarchy is a map of necessary prerequisites for a defined task. Learning hierarchies are validated for particular skills in particular subject areas.
White and Gagné (1974) is partly self-assessment by the theory’s own proponents. Its verdict that hierarchy research was virtually finished deserves caution.
The consequence is that the categories guide design decisions but do not sort objectives into neat boxes.
Contemporary Research
Recent work asks whether events-based teaching improves outcomes, whether the order of the events can be reversed, and whether learning hierarchies still earn a place. The evidence is generally of modest quality, and the weighting below reflects that.
Does Problem Solving First Beat the Nine-Event Order?
Gagné’s sequence puts presentation (Event 4) and guidance (Event 5) ahead of the learner’s own practice. In the vocabulary of these studies, that is instruction followed by problem solving.
Productive failure reverses it. Learners attempt problems first and receive explicit instruction afterwards.
Both arms contain explicit instruction, so only the timing differs. Sinha and Kapur (2021) pooled the evidence.
Aim: To find out whether concepts are learned better when problem solving comes before instruction or after it, and what moderates the answer.
Method: A meta-analysis of 45 articles reporting 53 studies and 166 comparisons, mostly in mathematics and basic science, with learners from second grade to undergraduate level.
Results: Problem solving first gave a moderate advantage for conceptual knowledge and transfer (Hedges’ g = 0.36, 95% CI 0.20 to 0.51). No difference was detected for procedural knowledge (g = −0.03, 95% CI −0.20 to 0.15).
Conclusion: Problem solving followed by instruction outperformed the reverse order for conceptual knowledge and transfer, more strongly when the productive-failure design was followed closely.
A meta-analysis sits high in the evidence hierarchy, but this one says less about Gagné than it seems. Both arms include explicit instruction, so it does not question the need for it.
The outcome closest to rule-using intellectual skills, procedural knowledge, showed no detected difference.
One caution is researcher allegiance. Kapur, a co-author, introduced productive failure, and the search began from articles citing the founding papers.
Ashman et al. (2020) tested the same question directly.
Aim: To test whether problem solving before explicit instruction beats the reverse order when element interactivity, the number of elements that must be processed together, is high.
Method: Two randomized experiments with Year 5 pupils learning about light energy efficiency. Experiment 1 had 64 pupils; Experiment 2 had 71 and raised element interactivity.
Results: Instruction first gave better performance on similar problems in Experiment 1, with no detected difference on transfer. In Experiment 2, it was better on both.
Conclusion: Where element interactivity is high, explicit instruction should come before problem solving.
Random allocation protects the comparison, but each experiment used one topic and one age group, and the samples were small. Sweller co-wrote the paper and is a leading author of the cognitive-load framework the result favors.
Weighed by design, the pooled evidence for problem solving first outranks two small experiments. Yet neither supports a blanket rule.
The meta-analytic advantage is confined to conceptual knowledge and transfer. For second- to fifth-graders, the pooled estimate favored instruction first (g = −0.09). That age range overlaps the Year 5 pupils.
For a designer, presenting content before practice remains a defensible default for novices and for material with many interacting elements.
A problem-first opening looks most plausible for older learners and conceptual goals. Neither study tested it inside a nine-event lesson, so it remains a design hypothesis.
Do the Nine Events Improve Outcomes?
The only meta-analysis of the nine events, Li et al. (2025), is weighed in the Evidence Check above. Smaller recent studies point the same way.
Habib (2024): In a quasi-experimental study of 20 postgraduate obstetrics and gynecology residents, post-test scores were higher after Gagné-based teaching than after conventional teaching.
Anietimfon et al. (2026): A quasi-experimental study of 382 Nigerian secondary-school students compared Gagné’s strategy with conventional teaching in algebra. The Gagné group did better, with a very large effect (partial η² = .58).
Both studies are non-randomized. Li et al. (2025) showed that non-randomized designs in this literature yield much larger effects than randomized ones.
A close cousin of Gagné’s approach has large observational data behind it. Frick et al. (2022) analyzed more than 172,000 learner journeys through online plagiarism tutorials and certification tests. The scale is large.
Students who chose instruction designed around Merrill’s first principles were nearly four times more likely to pass than to fail.
The data are observational. Learners chose their own version, so selection may explain part of the difference.
Learning Hierarchies in Adaptive Assessment
Gagné’s hierarchies have found a new use in diagnostic assessment.
Tang et al. (2020) used a Gagné-based hierarchy to generate test questions and learning-path diagrams for a mathematics skills assessment system. Tang et al. (2021) combined a hierarchy with fuzzy logic to find students’ learning barrier points.
Both report that the systems could diagnose learning barriers. Both are engineering validations in single settings, which is weaker evidence than a controlled trial.
The one meta-analysis supports a moderate benefit over lecturing. Everything else is consistent with the theory but carries less weight.
Thread
Evidence
Weight
Order of the events
A meta-analysis of problem solving first (Sinha & Kapur, 2021) and two small experiments (Ashman et al., 2020)
The pooled evidence outweighs the two experiments; it is confined to conceptual knowledge and transfer
Nine events versus lecturing
A meta-analysis with weak inputs (Li et al., 2025) and small quasi-experiments
A moderate benefit; randomized trials give smaller effects
First-principles design
Observational data on 172,000 learner journeys (Frick et al., 2022)
Lower; learners chose their version
Hierarchies in assessment
Engineering validations in single settings (Tang et al., 2020, 2021)
Lower; weaker than a controlled trial
Modern Relevance in the Age of AI
Is Gagné still relevant in a world of AI and self-directed digital learning? Probably, but the evidence for AI-delivered events is thin. The nine events work as a design checklist for digital modules.
Self-paced learners lack the teacher who supplies attention, feedback and review in a classroom. A module has to design those functions in. Has it gained attention, stated an objective, prompted recall, offered guidance, required responses and given feedback?
AI engines could tailor Event 5 (guidance) and Event 7 (feedback) to each student’s responses in real time, modernizing the model for personalized learning.
This remains a design hypothesis. A scoping review of 474 studies on generative-AI agents in education found that only 138 explicitly drew on educational theory (Wang et al., 2026).
The same review found that most empirical evaluations were small-scale and short-term. Until controlled studies test AI-delivered events, the claim stays unproven.
Who Was Robert Gagné?
Robert Mills Gagné (1916–2002) was an American experimental psychologist who took his doctorate at Brown University in 1940. He should not be confused with Françoys Gagné, whose differentiated model of giftedness and talent is a different body of work.
Wartime Training and Early Papers
Gagné spent the Second World War and the years after it in military aviation-psychology and training research.
He later moved through academic and research posts at Connecticut College, Princeton, the American Institutes for Research, the University of California at Berkeley and Florida State University.
His career sat where laboratory learning psychology met a practical problem: training very large numbers of people to do complicated jobs quickly. That problem shaped his central question.
What must a trainee already be able to do before the next skill can be learned?
Reiser (2001) places this wartime and post-war training tradition at the root of instructional design as a field. Gagné is one of its central figures.
Two papers in 1962 set the direction. In Military training and principles of learning, Gagné (1962a) addressed how learning psychology could be put to work in training.
In The acquisition of knowledge, he began to treat knowledge as something assembled from simpler component abilities (Gagné, 1962b).
How the Theory Developed
Looking back in 1984, Gagné explained what he had drawn from applying learning principles in schools, military training and adult professional development.
The general “laws of learning” were not enough. They did not specify the conditions of learning in enough detail to design instruction from (Gagné, 1984).
The first edition of The Conditions of Learning (Gagné, 1965) arranged learning into types from simple to complex, running from signal learning to problem solving.
By 1972 he had regrouped learned outcomes into five domains (Gagné, 1972). The five domains stuck.
The third and fourth editions recast the account in information-processing language (Gagné, 1977, 1985).
Gagné et al. (1992) then turned it into a design manual, Principles of Instructional Design, with the nine events laid out as a lesson template. The 1984 statement of the five categories remains the most accessible primary account.
Key Takeaways
Core Idea: Different learning outcomes need different instruction, so a designer first decides which kind of capability a lesson should build.
Five Categories: Intellectual skills, verbal information, cognitive strategies, motor skills and attitudes are learned under different conditions.
Nine Events: Each event of instruction, from gaining attention to enhancing transfer, supports one internal mental process.
Flexible Use: Not every lesson needs all nine events. Learners can often supply some of the processing themselves.
Hierarchies: Task analysis works backwards from the final objective to map the prerequisites to teach first.
Verdict: Supported with qualifications – nine-event teaching beats lecturing in health-professions courses, but the strict prerequisite hierarchy is untested at scale.
References
Alfieri, L., Brooks, P. J., Aldrich, N. J., & Tenenbaum, H. R. (2011). Does discovery-based instruction enhance learning? Journal of Educational Psychology, 103(1), 1–18. https://doi.org/10.1037/a0021017
Anietimfon, E. W., Zakariya, Y. F., Idris, I. O., & Falalu, M. K. (2026). Effects of Gagné’s instructional model on performance and retention in algebra among secondary school students in Nigeria. Frontiers in Education, 11, Article 1747651. https://doi.org/10.3389/feduc.2026.1747651
Arthur, W., Jr., Bennett, W., Jr., Edens, P. S., & Bell, S. T. (2003). Effectiveness of training in organizations: A meta-analysis of design and evaluation features. Journal of Applied Psychology, 88(2), 234–245. https://doi.org/10.1037/0021-9010.88.2.234
Ashman, G., Kalyuga, S., & Sweller, J. (2020). Problem-solving or explicit instruction: Which should go first when element interactivity is high? Educational Psychology Review, 32(1), 229–247. https://doi.org/10.1007/s10648-019-09500-5
Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. T. Spence (Eds.), The psychology of learning and motivation (Vol. 2, pp. 89–195). Academic Press. https://doi.org/10.1016/s0079-7421(08)60422-3
Ausubel, D. P. (1960). The use of advance organizers in the learning and retention of meaningful verbal material. Journal of Educational Psychology, 51(5), 267–272. https://doi.org/10.1037/h0046669
Blume, B. D., Ford, J. K., Baldwin, T. T., & Huang, J. L. (2010). Transfer of training: A meta-analytic review. Journal of Management, 36(4), 1065–1105. https://doi.org/10.1177/0149206309352880
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. https://doi.org/10.1177/1529100612453266
Fitts, P. M., & Posner, M. I. (1967). Human performance. Brooks/Cole.
Frick, T. W., Myers, R. D., & Dagli, C. (2022). Analysis of patterns in time for evaluating effectiveness of first principles of instruction. Educational Technology Research and Development, 70(1), 1–29. https://doi.org/10.1007/s11423-021-10077-6
Gagné, R. M. (1962a). Military training and principles of learning. American Psychologist, 17(2), 83–91. https://doi.org/10.1037/h0048613
Gagné, R. M. (1977). The conditions of learning (3rd ed.). Holt, Rinehart & Winston.
Gagné, R. M. (1984). Learning outcomes and their effects: Useful categories of human performance. American Psychologist, 39(4), 377–385. https://doi.org/10.1037/0003-066x.39.4.377
Gagné, R. M. (1985). The conditions of learning and theory of instruction (4th ed.). Holt, Rinehart & Winston.
Gagné, R. M. (1987). Instructional technology foundations. Lawrence Erlbaum Associates.
Gagné, R. M., Briggs, L. J., & Wager, W. W. (1992). Principles of instructional design (4th ed.). Harcourt Brace Jovanovich.
Gagné, R. M., & Driscoll, M. P. (1988). Essentials of learning for instruction (2nd ed.). Prentice-Hall.
Gagné, R. M., Mayor, J. R., Garstens, H. L., & Paradise, N. E. (1962). Factors in acquiring knowledge of a mathematical task. Psychological Monographs: General and Applied, 76(7), 1–21. https://doi.org/10.1037/h0093824
Habib, S. (2024). Effect of Gagne’s events of instruction on academic performance of postgraduate FCPS residents in obstetrics and gynaecology. Journal of Ayub Medical College Abbottabad, 36(4, Suppl. 1), 855–858. https://doi.org/10.55519/jamc-s4-14168
Jonassen, D. H. (1991). Objectivism versus constructivism: Do we need a new philosophical paradigm? Educational Technology Research and Development, 39(3), 5–14. https://doi.org/10.1007/bf02296434
Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31. https://doi.org/10.1207/s15326985ep3801_4
Karpov, Y. V. (2003). Vygotsky’s doctrine of scientific concepts: Its role for contemporary education. In A. Kozulin, B. Gindis, V. S. Ageyev, & S. M. Miller (Eds.), Vygotsky’s educational theory in cultural context (pp. 65–82). Cambridge University Press. https://doi.org/10.1017/cbo9780511840975.005
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75–86. https://doi.org/10.1207/s15326985ep4102_1
Li, Y., Liang, Z., Li, Z., Yu, Y., Yang, Q., & Li, X. (2025). Effectiveness of Gagné’s 9 events of instruction in health professions education: A systematic review and meta-analysis. Frontiers in Medicine, 12, Article 1522830. https://doi.org/10.3389/fmed.2025.1522830
Martin, F., Klein, J. D., & Sullivan, H. (2007). The impact of instructional elements in computer-based instruction. British Journal of Educational Technology, 38(4), 623–636. https://doi.org/10.1111/j.1467-8535.2006.00670.x
Merrill, M. D. (2002). First principles of instruction. Educational Technology Research and Development, 50(3), 43–59. https://doi.org/10.1007/bf02505024
Monib, W. K., Qazi, A., & Apong, R. A. (2025). Microlearning beyond boundaries: A systematic review and a novel framework for improving learning outcomes. Heliyon, 11(2), Article e41413. https://doi.org/10.1016/j.heliyon.2024.e41413
Reiser, R. A. (2001). A history of instructional design and technology: Part II: A history of instructional design. Educational Technology Research and Development, 49(2), 57–67. https://doi.org/10.1007/bf02504928
Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
Sinha, T., & Kapur, M. (2021). When problem solving followed by instruction works: Evidence for productive failure. Review of Educational Research, 91(5), 761–798. https://doi.org/10.3102/00346543211019105
Sweller, J., van Merriënboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10(3), 251–296. https://doi.org/10.1023/a:1022193728205
Sweller, J., van Merriënboer, J. J. G., & Paas, F. (2019). Cognitive architecture and instructional design: 20 years later. Educational Psychology Review, 31(2), 261–292. https://doi.org/10.1007/s10648-019-09465-5
Wang, N., Zou, D., Xie, H., & Qin, S. J. (2026). A scoping review of generative AI-powered agentic AI in education: Research landscape, agentic capabilities, and insights from the frontier agent paradigm, exemplified by OpenClaw. Computers and Education: Artificial Intelligence, 11, Article 100653. https://doi.org/10.1016/j.caeai.2026.100653
White, R. T., & Gagné, R. M. (1974). Past and future research on learning hierarchies. Educational Psychologist, 11(1), 19–28. https://doi.org/10.1080/00461527409529119
BSc (Hons) Psychology, MSc Psychology of Education
Associate Editor for Simply Psychology
Olivia Guy-Evans is a writer and associate editor for Simply Psychology, where she contributes accessible content on psychological topics. She is also an autistic PhD student at the University of Birmingham, researching autistic camouflaging in higher education.
BSc (Hons) Psychology, MRes, PhD, University of Manchester
Saul McLeod, PhD, is a qualified psychology teacher with over 18 years of experience in further and higher education. He has been published in peer-reviewed journals, including the Journal of Clinical Psychology.
