The Illusion of Performance

In their fascinating book Sleights of Mind: What the Neuroscience of Magic Reveals About Our Everyday Deceptions,¹ three neuroscientists describe the deep truths that magic reveals about the everyday cognitive deceptions that pervade our lives. Among the many points of confluence between magic and the mind, Sleights of Mind explores how our brains interpret and “make up” the world based on limited informational input, how attention is a scarce resource that illusionists can use to their advantage, and how memory is so unreliable that our recollections of events are prone to manipulation. Ultimately, they show that illusions only work because humans are cognitively predisposed to being deceived.

In this sense, every walk of life has the potential to deceive us—yet few professions are implicated by the dissembling power of illusion more than education. After all, so much of what occurs when we learn is obscured to the outside observer, and often what we think we are seeing is distant from or even diametrically opposed to what’s really happening.

We decided to write our new book, Instructional Illusions, to highlight crucial ways that what seems to be going on in a classroom may deceive us, making certain approaches to learning appear more effective than they really are. Across its chapters, we lift the veil on 10 educational illusions to demonstrate how the surface phenomena of teaching don’t tell the whole story. With reference to evidence from the worlds of cognitive science and educational psychology, we show that understanding how learning occurs can give us the eyes to see what lies beneath—what works and what merely appears to work. Teaching is not a straightforward process but rather a multifaceted endeavor, the complexity of which is dizzying to comprehend, if one decides to look closely enough.

The ideas we explore are illusions, not myths. Though much has been written about the scourge of edu-myths like learning styles² or learning pyramids,³ the ideas we explore are not merely characterized as debunked. They are instead cases in which the outward appearance of effectiveness masks a deeper truth about what really works and why.

As such, our task is not to slay myths but to unveil aspects of education that appear to make sense but, upon closer inspection, prove to be much more complex. For when we bring an evidentiary lens to bear on instructional illusions, we reveal secrets that have been hiding in plain sight all along. In this article, we share one illusion—performance masquerading as learning—and suggest practical instructional strategies (with strong research backing) for teachers to ensure deep learning, not superficial performance.

The Performance Illusion

“We may feel we have learned something simply because it is fresh in mind, but that feeling is fleeting.”⁴

In early 1900s Berlin, a horse named Hans stunned audiences by seemingly displaying human-level intelligence. Owned by Wilhelm von Osten, a mathematics teacher and amateur scientist, Hans seemingly solved mathematical problems and recognized written words. A typical feat would involve von Osten asking Hans a math question such as “What is 3 + 4?,” at which point Hans would respond by tapping his hoof seven times. Scientists, educators, and even the German Board of Education were floored, as Hans’ responses were accurate an astonishing 89 percent of the time.⁵

Experts far and wide were convinced of Hans’ mathematical ability until psychologist Oskar Pfungst noticed something no one else had the eyes to see. He realized Hans was not doing mathematics at all but was instead responding to subtle, unconscious cues from his questioners. When someone posed a problem, they would involuntarily tense up as Hans approached the correct number of taps. Once Hans reached the correct answer, the questioner would subtly relax—often by a slight movement of the head, change in posture, or breath adjustment—which signaled Hans to stop tapping.

Though we are not suggesting human learners are akin to a foot-tapping horse, there is nevertheless a worthwhile comparison to draw: just as Hans’ performances created an illusion of understanding, students’ classroom performances can create similar illusions of learning that mask what’s really happening beneath the surface.

Here we will discuss performance and learning, two seemingly synonymous but very different phenomena, and reveal a pervasive illusion: certain teaching and learning strategies that are beneficial for performance aren’t good for, and can even impede, learning.

The Illusion

To address this illusion, we first need to define performance and learning in accurate terms. What we teach may temporarily go into our students’ memories, and during or directly after instruction they may be able to discuss ideas with classmates, answer questions, and even do their homework correctly. But it isn’t learning if what was taught can’t be recalled and used the next week, month, or year. Cognitive scientists Nick Soderstrom and Robert Bjork write:

The primary goal of instruction should be to facilitate long-term learning—that is, to create relatively permanent changes in comprehension, understanding, and skills of the types that will support long-term retention and transfer. During the instruction or training process, however, what we can observe and measure is performance, which is often an unreliable index of whether the relatively long-term changes that constitute learning have taken place.⁶

Performance is a short-term change in one’s knowledge or, as Soderstrom and Bjork write, “Temporary fluctuations in behavior or knowledge that can be observed and measured during or immediately after the acquisition process.”⁷ Some researchers refer to this as intermediate-term memory, lasting little more than a few hours. A student might remember the content during or shortly after a lesson, but it’s usually quickly forgotten. How many of you reading this have taught something and even administered a test that your students passed, only to find that after the weekend, they stared at you blankly when you mentioned it? In other words, the knowledge gained is fragile. It’s easily disrupted, error-prone, and quickly forgotten. With respect to performance, relearning is actually reacquisition. It takes almost as long to reacquire the lost knowledge as it took to acquire it the first time.

Learning, on the other hand, is a change in long-term memory (the mind’s seemingly limitless, close-to-permanent storage facility),⁸ and that’s the real goal of education: enduring changes in knowledge, skills, and behavior for long-term retention and transfer to new situations. Learning is stable; what you’ve learned deeply and durably isn’t easily forgotten.

Of course, you can’t always remember everything you’ve learned, but traces of it reside somewhere in your long-term memory. Therefore, relearning what you’ve forgotten goes much more quickly than initial learning. To paraphrase Alfred, Lord Tennyson: “’Tis better to have learned and lost than never to have learned at all.”⁹ Take riding a bicycle, for instance. If you haven’t ridden a bike in 20 or 30 years, you’ll be very shaky and will probably fall when you first get back on. However, within a short time—a lot shorter than it took to learn to ride the bike decades ago—you’re again able to ride fairly proficiently.

Performance requires very little cognitive processing, and this minimal processing is shallow.¹⁰ Often, it requires nothing more than simple repetition (this could include the verbatim copying of a text to your notes or project rather than paraphrasing or summarizing it in your own words) or rehearsal in your working memory. Think of what happens when someone tells you their telephone number. You might repeat it a few times and remember it directly after hearing it, but the chance is slim that you’ll remember it that evening or tomorrow. You’ve not processed the information in a meaningful way, and thus it doesn’t get integrated into the knowledge schemas* in your long-term memory. Shallow information processing involves a minimal level of cognitive engagement. Simply repeating or rehearsing information—often without trying to understand its meaning—doesn’t lead to deeper connections or integration of the information into existing knowledge structures.

Learning, in contrast, involves deep processing activities like elaboration, reorganization, and critical thinking. This meaningful and active engagement with the to-be-learned information leads to better comprehension and long-term retention. It requires a lot of mental effort to actively process, analyze, and synthesize information, including understanding the new information by relating it to prior knowledge (schema acquisition and elaboration) or by applying it in new contexts (transfer).

Performance is fragmentary. It very often focuses on memorizing discrete facts or pieces (fragments) of information without understanding how they relate to each other or to prior knowledge.^† It also involves processing surface-level details such as definitions, isolated dates, or formulas without the information being connected to broader concepts or patterns. This makes the acquisition process disjointed and compartmentalized.¹¹

Learning, however, is cumulative. It emphasizes connecting new information to prior knowledge such that each new piece of learning builds on what has already been learned. This is a cumulative process that strengthens understanding and retention because the learner continuously links new ideas to existing schemas in long-term memory, which leads to richer, more interconnected knowledge structures.

Finally—and this is likely the biggest problem with teaching for learning—performance is easy to see and measure, but learning isn’t. All you need to measure performance is a test or exam at the end of a week or unit. Students cram the night before, “learn” it, and then regurgitate it.

Learning is difficult to measure. Assessment of learning needs to be delayed, derived from other things, and done in combination with other learning. To measure learning, you need to cumulatively test and/or use assessments that involve analysis, synthesis, and evaluation. In other words, learning implies mastery, and measuring true mastery is complex, time-consuming, and often subjective.

We’re by no means saying that performance is bad or that it shouldn’t be measured. What we’re saying is that performance and learning shouldn’t be confused. The ultimate goal of teaching should be to ensure that students learn. You want what you’ve taught to be remembered, retrieved when needed, and applied in familiar and new situations (transferred).

Unmasking the Illusion

Just as skilled illusionists can spot the telltale movements of an amateur’s sleight of hand, expert teachers develop a refined ability to distinguish between the illusion of performance and genuine learning. This expertise isn’t mystical—it’s a carefully honed set of observational skills and diagnostic techniques.

Luckily, teachers have an arsenal of instructional techniques that we can use to stimulate learning. The two most well-known and effective sets of techniques are desirable difficulties¹² and generative learning strategies.¹³

Desirable Difficulties

Desirable difficulties are learning conditions that are often experienced by the learner as requiring more effort, but that have a positive effect on learning and the transfer of knowledge and skills. They were identified by cognitive scientists Robert and Elizabeth Bjork, whose research found that “Conditions of learning that make performance improve rapidly often fail to support long-term retention and transfer, whereas conditions that create challenges and slow the rate of apparent learning often optimize long-term retention and transfer.”¹⁴ The following five desirable difficulties focus on slowing the rate of apparent learning so that long-term retention and transfer are optimized.¹⁵ The Bjorks call it “making it difficult but in a good way.”¹⁶

1. Interleaving/variable practice: In interleaving, you vary the conditions of practice. You mix (i.e., interleave) practice on several related skills together. In other words, you don’t block practice, repeating the same type of task over and over again, but rather shuffle tasks around. For instance, a tennis player, after having acquired the skills of hitting a forehand, backhand, and volley, will alternate practice between them. After all, that is also what is needed when playing tennis.
2. Contextual interference: Contextual interference is doing the same thing often but in different situations or contexts. It is very similar to interleaving, but here you make the task environment—not the task itself—more variable or unpredictable in a way that creates a temporary interference for the learner. For instance, did you know that even studying the same material in two different rooms leads to increased recall of that material?
3. Spaced practice: This is also known as “distributed practice” and is about spacing learning over time. Instead of studying for an hour and a half, you split your studying into three 30-minute sessions with one or more days in between. Distributing practice (e.g., learning tasks, study attempts, training trials) supports long-term retention through consolidation (giving your brain the chance to let things gel) and retrieval practice (recalling what you’ve learned—more on that below).
4. Reduced feedback: Reducing feedback frequency and specificity makes life more difficult for learners during training but—as with all desirable difficulties—can enhance long-term performance. It stimulates their independence, knowing that the instructor won’t give them the answer in the end. Examples of reducing feedback are giving summary feedback at the end of a practice session or “fading” the frequency of feedback over sessions.^‡
5. Retrieval practice/practice testing: In a nutshell, practice testing challenges learners to recall what they’ve previously learned, usually as opposed to rereading. When they actively remember that information—retrieve it from their memory—they can remember it better and longer. Ideally, initial learning should reach the point of basic proficiency. When retrieval practice starts, students need to try to recall—to attempt to retrieve—the content. When students correctly recall an answer, they will be able to more easily recall it in the future, and it may improve their mental organization of the information. When students are not able to recall, there is a cognitive benefit to making a genuine attempt (even if it is incorrect) before looking up the correct answer. For example, if a student is reviewing a flash card but flips to the solution without attempting an answer first, this is no more effective than rereading.

Generative Learning Strategies

Generative learning strategies are instructional techniques that engage learners in actively constructing their understanding of new information. The Flemish use the term herkneden (re-knead), as in a lump of clay that can be re-kneaded or reshaped into something else. Learners use these strategies to make sense of new information, integrate it with prior knowledge, and generate new mental representations by transforming the information into something else. This encourages and requires deeper cognitive processing, leading to better comprehension and retention. Cognitive scientists Logan Fiorella and Richard E. Mayer identified the following eight generative strategies.¹⁷

• Summarizing: Condensing information into a brief, coherent summary
• Mapping: Creating visual representations of the relationships between concepts
• Drawing: Creating drawings or illustrations representing the information being learned
• Self-explanation: Explaining the material to oneself as if teaching it
• Teaching others: Teaching or explaining concepts to others, real or imaginary
• Generating questions: Having learners create their own questions about the material
• Imagining: Mentally visualizing concepts or scenarios related to the material
• Enacting: Acting out or physically demonstrating concepts

Generative learning activities supersede shallow performance demonstrations because they require deep, effortful cognitive processes such as selecting relevant information, organizing it into coherent representations, and integrating it with prior knowledge.

It’s worth noting that, in our experience, students are predisposed to seeing performance as “the main event” and learning as a necessary evil for performance. As teachers, we therefore owe it to our students to demonstrate (through approaches like desirable difficulties and generative learning strategies) that even though the road to real learning is more strenuous, it beats the fleeting nature of performance every time.

The lessons here invite us to reconsider the phrase “It’s just like riding a bike.” We now know that the meaning behind the adage is not entirely true. After all, returning to something already learned doesn’t result in instantaneous proficiency. Nevertheless, the sentiment is partially accurate in that remembering how to ride a bike comes easier than if you had never learned in the first place.

This reveals the importance of distinguishing between performance and learning: not only does one not equal the other, but focusing on the fragile, fragmentary nature of performance undermines durable, authentic learning.

Thankfully, unraveling this illusion is difficult but not impossible. Just ask Oskar Pfungst, who was able to solve the dilemma of Hans the mathematical horse in one simple step: he placed blinders over Hans’ eyes.¹⁸                                           

Paul A. Kirschner is an emeritus professor of educational psychology at the Open University of the Netherlands and a guest professor at the Thomas More University of Applied Sciences in Belgium. Carl Hendrick is a professor of evidence-informed learning and teaching at Academica University of Applied Sciences in the Netherlands. Jim Heal is a professor of evidence-informed educational leadership at Academica University of Applied Sciences and a cofounder of Learning Science Partners. This article was adapted from their most recent book, Instructional Illusions, which was published in September 2025 by Hachette Learning © Paul A. Kirschner, Carl Hendrick, and Jim Heal 2025. Reproduced with permission of the Licensor through PLSclear.

*Human brains automatically organize knowledge in networks of interconnected ideas. These networks are known as schemas, and the more we come to know about a particular domain of knowledge, the more sophisticated that schema becomes. (return to article)

^†To learn more about performance, and the difference between an information-rich and a knowledge-rich curriculum, see Developing Curriculum for Deep Thinking: The Knowledge Revival. This book by Paul A. Kirschner and several other experts in the science of learning is available for free at go.aft.org/2i8. (return to article)

^‡For more details on feedback, see “Principles of Instruction” in the Spring 2012 issue of American Educator: go.aft.org/ms7. In particular, see the sections on guiding student practice and on requiring and monitoring independent practice; as mastery progresses, systematic feedback is reduced. (return to article)

Endnotes

1. S. Macknik, S. Martinez-Conde, and S. Blakeslee, Sleights of Mind: What the Neuroscience of Magic Reveals About Our Everyday Deceptions (New York: Henry Holt & Co., 2010).

2. H. Pashler et al., “Learning Styles: Concepts and Evidence,” Psychological Science in the Public Interest 9, no. 3 (2009): 105–19; P. A. Kirschner, “Stop Propagating the Learning Styles Myth,” Computers and Education 106 (2017): 166–71;and P. A. Kirschner and J. van Merriënboer, “Do Learners Really Know Best? Urban Legends in Education,” Educational Psychologist 48, no. 3 (2013): 1–15.

3. P. De Bruyckere, P. A. Kirschner, and C. Hulshof, Urban Myths About Learning and Education (London: Academic Press, 2016); and J. Lalley and R. Miller, “The Learning Pyramid: Does It Point Teachers in the Right Direction?,” Education 128, no. 1 (2007): 64–80.

4. R. Bjork, “Memory and Metamemory Considerations in the Training of Human Beings,” in Metacognition: Knowing About Knowing, ed. J. Metcalfe and A. Shimamura (Cambridge, MA: MIT Press, 1994), 185–205.

5. Britannica, “Clever Hans,” britannica.com/topic/Clever-Hans.

6. N. Soderstrom and R. Bjork, “Learning Versus Performance: An Integrative Review,” Perspectives on Psychological Science 10, no. 2 (2015): 176–99.

7. Soderstrom and Bjork, “Learning Versus Performance.”

8. P. A. Kirschner, J. Sweller, and R. Clark, “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, no. 2 (2006): 75–86.

9. “In Memoriam A.H.H.,” an elegy written by Alfred, Lord Tennyson. A.H.H. stands for his friend Arthur Henry Hallum. The original is “’Tis better to have loved and lost than never to have loved at all.”

10. F. Craik and R. Lockhart, “Levels of Processing: A Framework for Memory Research,” Journal of Verbal Learning and Verbal Behavior 11, no. 6 (1972): 671–84.

11. J. van Merriënboer, P. A. Kirschner, and J. Frèrejean, Ten Steps to Complex Learning (New York: Routledge, 2024).

12. Bjork, “Memory and Metamemory.”

13. L. Fiorella and R. Mayer, Learning as a Generative Activity: Eight Learning Strategies That Promote Understanding (Cambridge, UK: Cambridge University Press, 2015); and L. Fiorella and R. Mayer, “Eight Ways to Promote Generative Learning,” Educational Psychology Review 28 (2016): 717–41. See also P. A. Kirschner and C. Hendrick, “Chapter 19: Problem-Solving: How to Find a Needle in a Haystack,” in How Learning Happens: Seminal Works in Educational Psychology and What They Mean in Practice, 2nd ed. (London: Routledge, 2024), 201-208 and P. A. Kirschner, C. Hendrick, and J. Heal, “Chapter 13: Make Something of What You’ve Learnt: Logan Fiorella & Richard Mayer on Ways to Generate Learning,” in How Teaching Happens: Seminal Works in Teaching and Teacher Effectiveness and What They Mean in Practice (London: Routledge, 2022), 141–50.

14. E. Bjork and R. Bjork, “Making Things Hard on Yourself, but in a Good Way: Creating Desirable Difficulties to Enhance Learning,” in Psychology and the Real World: Essays Illustrating Fundamental Contributions to Society, ed. M. Gernsbacher et al. (New York: Worth Publishers, 2011), 59–68.

15. R. Bjork and E. Bjork, “Desirable Difficulties in Theory and Practice,” Journal of Applied Research in Memory and Cognition 9, no. 4 (2020): 475–79.

16. Bjork and Bjork, “Making Things Hard on Yourself.”

17. Fiorella and Mayer, Learning as a Generative Activity; and Fiorella and Mayer, “Eight Ways.” See also Kirschner and Hendrick, “Chapter 19: Problem-Solving”; and Kirschner, Hendrick, and Heal, “Chapter 13: Make Something.”

18. Wikipedia, “Clever Hans,” en.wikipedia.org/wiki/Clever_Hans.

[Illustrations by Jessie Lin]