For generations, education has treated attention, repetition and successful practice as evidence that learning has taken place. Research in cognitive psychology suggests a more unsettling possibility: the methods that make progress feel fastest may not be the ones that make knowledge last.
By Giani Boldeanu
Imagine two students preparing for the same examination. The first spends the evening rereading carefully highlighted notes. Each pass feels easier than the last. The names are familiar, the arguments recognisable, the pages reassuringly predictable. By the end of the session, the student feels prepared.
The second student closes the textbook and tries to reconstruct the lesson from memory. The experience is slower and considerably less satisfying. Definitions disappear. Connections break apart. Entire sections that appeared clear only minutes earlier now seem inaccessible. By the end of the session, this student feels less prepared.
A week later, the second student may remember more.
That result runs against one of the most persistent intuitions in education: that the easier learning feels, the better it is working. For generations, schools, universities and professional-training systems have operated on a broadly reasonable assumption. If people pay attention, study carefully, repeat material and perform successfully during practice, learning has taken place.
The assumption is not entirely wrong. Attention matters. Explanation matters. Repetition matters. But research on memory and skill acquisition has revealed a crucial distinction between performance during practice and learning that survives after practice ends. The two can look similar in the moment and diverge dramatically over time.
A student may answer correctly because the explanation is still fresh. An athlete may execute a movement because the drill is predictable. An employee may apply a framework because the instructor has just demonstrated it. Immediate success may reflect genuine progress, but it may also reflect the temporary support built into the learning environment.
Remove the notes, change the conditions, introduce a delay or add pressure, and a different question emerges: What can the person still do when the answer is no longer in front of them?
That may be the question education has too often neglected.
The Comfort of Familiarity
Rereading is appealing because it produces a powerful psychological effect. With every repetition, the material becomes easier to process. The words move more smoothly through the mind. Concepts that initially seemed unfamiliar begin to feel obvious.
This fluency is real, but it can be misinterpreted. Ease of processing often increases confidence more quickly than it increases durable understanding. Research on metacognition—the way people evaluate their own knowledge—shows that learners frequently use the feeling of ease as evidence that learning has occurred, even when that feeling is a poor predictor of later recall.
The problem is not that rereading never helps. A first or second encounter with difficult material may be essential. The problem begins when recognition is mistaken for mastery.
Recognition asks: Does this look familiar?
Recall asks: Can I produce it without seeing it?
Application asks something harder still: Can I use it when the problem looks different from the one I practised?
These are not equivalent achievements.
In a landmark experiment, Henry Roediger and Jeffrey Karpicke asked students to study educational passages using different combinations of repeated study and testing. Students who spent more time restudying often performed better after a few minutes. But after delays of several days, those who had practised retrieving the material retained substantially more. The activity that produced the strongest feeling of immediate preparation did not produce the strongest memory later.
This is the central illusion of learning: what feels available now may not remain available later.
The illusion is not confined to examinations. It appears anywhere people rehearse under conditions that are more supportive than the environment in which performance will ultimately be judged. A footballer can complete a passing sequence perfectly when the order is fixed and no opponent interferes. A clinician can follow every step of a procedure when the instructions remain visible. A manager can explain a decision-making model in a seminar and fail to use it when time is short, the evidence is incomplete and colleagues are divided.
Practice often contains clues that real performance removes. The material has just been presented. The correct response is expected. The sequence is known. Feedback arrives immediately. The environment quietly carries part of the cognitive load.
Performance exposes what practice can conceal.
The Productive Struggle
The idea that harder learning can produce better learning is often associated with the concept of desirable difficulties. The phrase sounds contradictory because it is meant to be. Certain conditions make practice less fluent and reduce immediate performance while improving later retention or transfer.
Retrieving an answer rather than rereading it is one example. Spacing practice across several days rather than concentrating it in a single session is another. Mixing related categories or problem types, rather than completing one type repeatedly before moving to the next, can also make practice more difficult while improving the learner’s ability to identify which response a new situation requires.
These methods share an important characteristic: they force the learner to reconstruct knowledge rather than merely encounter it again.
Retrieval practice has received particularly strong support. Reviews of the evidence have found that attempting to bring information to mind can strengthen later retention more effectively than repeated study, especially when learners receive feedback that corrects omissions and errors.
Spacing produces a related effect. When practice is distributed, some forgetting occurs between sessions. The next attempt therefore requires more effort. Yet that effort can strengthen the memory precisely because the information must be rebuilt rather than simply recognised. A major synthesis of spacing research found that the most effective interval between learning sessions depends partly on how long the information must eventually be retained. Longer retention goals generally benefit from longer spacing intervals.
More recent classroom evidence has reinforced the broader pattern. A 2025 meta-analysis covering more than 3,000 learners found a moderate advantage for distributed over massed practice, although the size of the benefit varied across studies and learning conditions.
Interleaving can work in a similar way. When learners practise one problem type repeatedly, the task itself often signals which method to use. When different types are mixed, the learner must first determine what kind of problem is being faced. That additional decision can slow practice while improving later discrimination. A meta-analysis found that interleaving can support inductive learning, but also cautioned that its benefits depend heavily on the material and context; it is not automatically superior for every task.
This qualification matters. Difficulty is not beneficial merely because it is difficult.
A badly explained lesson is not a desirable difficulty. Divided attention is not a desirable difficulty. Unnecessary complexity is not a desirable difficulty. Nor is repeated failure without guidance. Research has shown that some forms of disfluency do not reliably improve memory and may simply make learning harder.
The difficulty must be productive. It must require meaningful mental work while remaining within the learner’s capacity to succeed, particularly with feedback and support.
The principle is not “make learning unpleasant.” It is “do not confuse comfort with progress.”
Why Learners Avoid What Works
There is a practical problem with these methods: they often feel ineffective.
When students reread, performance is smooth. When they test themselves, errors appear. When they cram, the material feels available. When they space their study, forgetting becomes visible. When they practise one skill repeatedly, improvement is obvious. When tasks are mixed, performance becomes uneven.
People naturally prefer the method that supplies the clearest evidence of success. Unfortunately, immediate success can be a poor judge of future learning.
Research on what has been called the misinterpreted-effort hypothesis suggests that learners often treat mental effort as a sign that a strategy is failing. Because retrieval and spacing feel more difficult, people may avoid them even after experiencing their benefits.
This creates a strange educational paradox. The methods most likely to expose gaps in knowledge are often the methods learners trust least. The methods most likely to conceal those gaps are often the most reassuring.
A learner who struggles to recall an idea may conclude, “I do not know this.” That conclusion is partly correct—and extremely useful. The failed retrieval attempt identifies exactly where further work is required.
A learner who recognises the same idea while rereading may conclude, “I know this.” That conclusion may be wrong, but the learning environment provides no immediate reason to doubt it.
Confidence rises because ignorance has been hidden.
The discomfort of retrieval is therefore not a defect in the method. It is information.
When Practice Becomes a Performance
The distinction between learning and practice performance becomes even more important in sport and other high-pressure fields.
Blocked repetition is valuable. A beginner often needs stability before variability. A new movement, technical procedure or conceptual framework may need to be simplified and repeated under controlled conditions. Early success can establish coordination, understanding and motivation.
But practice can become deceptive when the learner improves mainly at reproducing the exercise.
A passing drill can teach passing. It can also teach the choreography of that particular drill. A tactical exercise can develop perception. It can also teach the athlete to expect the same movement in the same sequence. A simulated emergency can improve preparation. It can also become a script that collapses when reality presents a different order of events.
The issue is not whether repetition works. It plainly does. The issue is what repetition is teaching.
Human performance rarely consists of executing isolated actions in fixed conditions. It requires noticing relevant cues, selecting among alternatives, adjusting to disruption and continuing after error. These capabilities do not emerge automatically from technical fluency.
A learner can become excellent at the solution without becoming excellent at identifying the problem.
This is why variable and representative practice matter. The environment must eventually require the decisions that performance itself will require. The footballer must not merely pass accurately, but decide when, where and whether to pass. The clinician must not merely remember a procedure, but recognise when the standard sequence no longer fits. The leader must not merely know a framework, but determine whether it applies under conditions the framework did not anticipate.
The purpose of practice is not to reproduce the training environment perfectly. It is to remain useful when the training environment disappears.
What Pressure Takes Away
Knowledge that can be retrieved in quiet conditions may become less accessible under pressure. Stress can narrow attention, increase reliance on familiar responses and reduce cognitive flexibility. Fatigue can weaken concentration. Fear of error can shift attention away from the task and towards its consequences.
This is where the science of learning meets the science of performance.
Confidence is often treated as a feeling to be built. But confidence based only on successful repetition in predictable conditions can be fragile. It may vanish as soon as the environment changes. More durable confidence develops when people acquire evidence that they can make an error, recognise it, adjust and continue.
Concentration is often described as the ability to remain focused. In performance, the more valuable skill may be the ability to recover focus after it has been disrupted.
Resilience is often described as toughness. In learning, it is closer to adaptive persistence: the capacity to remain engaged long enough to modify an unsuccessful approach.
These qualities cannot be developed entirely through explanation. They require carefully designed experiences in which the learner must retrieve, decide, adapt and recover.
That does not mean every training session should simulate maximum pressure. Constant stress can impair both learning and wellbeing. Difficulty must be graduated. Challenges should increase as competence increases. Feedback should clarify rather than humiliate. Recovery must remain part of the process.
But a system that protects learners from uncertainty until the day uncertainty matters has not protected them at all.
It has merely postponed the encounter.
Mistakes Need an Editor
The celebration of failure has become fashionable in education, business and sport. People are told to embrace mistakes, fail forward and become comfortable with discomfort.
The advice is incomplete.
Mistakes do not automatically create learning. Repeating an error without understanding it can strengthen the wrong response. Failure without feedback can produce confusion, frustration or avoidance. A challenge far beyond the learner’s present ability may reveal very little beyond the fact that the task was badly calibrated.
Errors become useful when they are converted into information.
What did the learner notice? What did they miss? Which assumption shaped the response? Was the failure technical, perceptual, strategic or emotional? Did the learner choose the wrong action, or execute the right action poorly? Was the task too difficult, or did it expose a genuine gap?
A teacher, coach or leader must act as an editor of difficulty. The goal is not to remove every struggle, nor to manufacture struggle for its own sake. It is to create conditions in which errors are visible, interpretable and correctable.
Too much assistance can prevent independent thinking. Too little can turn productive struggle into noise.
The art lies in knowing when to explain, when to ask, when to demonstrate, when to wait and when to change the task itself.
The Expert Learner
Some people continue improving long after others reach a plateau. Their advantage is not always superior talent or greater motivation. Often, they have developed a better model of learning.
They do not judge preparation only by how prepared they feel. They test what remains accessible without assistance. They return to material after time has passed. They vary conditions rather than repeating only what has become comfortable. They seek feedback that challenges their interpretation rather than confirming it.
They are not merely accumulating expertise. They are becoming expert learners.
An expert learner understands that different capabilities require different forms of practice. Facts that must be recalled should be retrieved. Decisions that require discrimination should be practised across varied cases. Skills that must survive pressure should eventually be rehearsed under distraction and uncertainty. Complex performance may need to be simplified at first, but it must ultimately be reintegrated into the environment in which it will be used.
This also changes the purpose of testing. A test does not need to be only a final judgement. Low-stakes testing can become part of the learning process itself by revealing gaps, strengthening retrieval and directing what should be studied next. Research comparing common study techniques has consistently rated practice testing and distributed practice among the most broadly useful strategies available to learners.
The best assessment systems recognise two different purposes. Some assessments certify performance. Others improve it. Confusing the two can make learning either needlessly threatening or insufficiently rigorous.
What Education Is Really For
The science of learning does not prove that everything previously believed about education was wrong. The strongest findings are more precise and more useful than that.
Explanation matters, but understanding something while it is being explained does not guarantee that it can later be recalled.
Repetition matters, but repeating a task under fixed conditions does not guarantee adaptable performance.
Success matters, but success during practice may partly reflect cues and support that will not be present later.
Difficulty can strengthen learning, but only when it produces productive effort rather than confusion.
No single technique is universally superior. Retrieval without adequate feedback can be less effective than richer forms of encoding in some circumstances. Interleaving can help learners distinguish between related categories, but may be less useful for certain materials. Spacing is robust, but the interval and structure still matter. The method must fit the learner, the task and the outcome being developed.
What the evidence does challenge is the assumption that good learning should always feel smooth.
A successful lesson is not merely one students understand while the teacher is speaking. A successful training session is not merely one in which the athlete performs without error. A successful leadership programme is not merely one in which participants can repeat the model at the end of the day.
The deeper test arrives later: after some forgetting, in a different environment, without the original cues and perhaps under pressure.
For generations, education has asked a deceptively simple question: Did the learner understand?
The science of learning asks a harder one: Will that understanding still be available when it is needed?
The difference between those questions may define not only the future of education, but the future of human performance.
Sources and Further Reading
Bjork, R. A., & Kroll, J. F. (2015). “Desirable Difficulties in Vocabulary Learning.” The American Journal of Psychology, 128(2), 241–252.
Brunmair, M., & Richter, T. (2019). “Similarity Matters: A Meta-Analysis of Interleaved Learning and Its Moderators.” Psychological Bulletin, 145(11), 1029–1052.
Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). “Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis.” Psychological Bulletin, 132(3), 354–380.
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). “Improving Students’ Learning With Effective Learning Techniques.” Psychological Science in the Public Interest, 14(1), 4–58.
Karpicke, J. D., & Blunt, J. R. (2011). “Retrieval Practice Produces More Learning Than Elaborative Studying With Concept Mapping.” Science, 331(6018), 772–775.
Kirk-Johnson, A., Galla, B. M., & Fraundorf, S. H. (2019). “Perceiving Effort as Poor Learning: The Misinterpreted-Effort Hypothesis of How Experienced Effort and Perceived Learning Relate to Study Strategy Choice.” Cognitive Psychology, 115, 101237.
Mawson, R. D., et al. (2025). “The Distributed Practice Effect on Classroom Learning: A Meta-Analytic Review of Applied Research.”
Roediger, H. L. III, & Butler, A. C. (2011). “The Critical Role of Retrieval Practice in Long-Term Retention.” Trends in Cognitive Sciences, 15(1), 20–27.
Roediger, H. L. III, & Karpicke, J. D. (2006). “Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention.” Psychological Science, 17(3), 249–255.
Soderstrom, N. C., & Bjork, R. A. (2015). “Learning Versus Performance: An Integrative Review.” Perspectives on Psychological Science, 10(2), 176–199.
Weinstein, Y., Sumeracki, M. A., & Caviglioli, O. (2018). “Teaching the Science of Learning.” Cognitive Research: Principles and Implications, 3, 2.
