Think about the last time you traveled by plane. Before takeoff, you watched the familiar safety demonstration for the umpteenth time, courtesy of the cabin crew. As we all know, part of that demonstration shows us how to put on a life jacket. Chances are you have seen that scene dozens of times. Chances are also that, if you ever actually had to put one on, you might run into a little trouble.
Cognitive psychologist and UCLA professor Robert Bjork uses this example to illustrate one of the questions that has occupied much of his career as a researcher of memory and learning. Watching someone else put on a life jacket over and over again creates a sense of familiarity. To actually learn how to do it, it would be far more effective to try putting it on ourselves. We would have to find it, unfold it, figure out where to put our head, and adjust the straps. We would probably take longer and make a few mistakes. And it is precisely that effort that would help us remember the procedure when we needed it.
The scene on the plane brings to light a paradox that Bjork has been studying for decades. Some conditions that make a task easier produce good results while we are practicing, although their effects may fade quickly. Others introduce obstacles that slow the process and increase errors, yet help what we have learned endure and remain retrievable later on.
Bjork called these obstacles “desirable difficulties.” But… why would we want to make learning more difficult? Answering this seemingly reasonable question requires us to look more closely at what happens when we learn and, above all, at the signals we use to decide that someone has learned.
The Paradox of Desirable Difficulties
The fact that a student performs a task well while practicing it does not necessarily mean they have learned how to do it. A student may correctly solve a series of exercises immediately after an explanation and struggle with them a week later. They may also stumble during practice and yet remember what they learned better when they need it again.
To explain the difference between these two situations, Robert and Elizabeth Bjork developed their theory of disuse in the early 1990s, which distinguishes between two dimensions of memory. The first is retrieval strength: it indicates how accessible a piece of knowledge is at a particular moment. It can increase rapidly with recent practice and also decline relatively quickly. The second, storage strength, reflects how deeply that knowledge is established in memory and develops more gradually.
Think of a phone number we have just looked up. We can repeat it immediately because its retrieval strength is high at that moment. A few hours later, it may have disappeared from our memory. Something very different happens with knowledge that we have repeatedly retrieved and used over time.
This distinction helps explain one of Bjork’s central ideas: learning better does not always mean performing better while we are learning. Some conditions help us get more answers right during practice and create an immediate impression of mastery. Others force us to retrieve information, reconstruct an answer, or decide which procedure to use. Performance may deteriorate during that process, while learning becomes more durable and more transferable.
That is where the precise meaning of the adjective desirable lies. Difficulty becomes valuable when it activates processes that strengthen long-term learning. Adding obstacles for their own sake serves no purpose. For these difficulties to work, they must demand an effort that the student is capable of making and help them retrieve or apply what they are learning later on.
Four Ways to Make Learning More Difficult (and More Durable)
How can a desirable difficulty be introduced into learning? Bjork’s work, along with that of other researchers, has identified several conditions that can impair performance during practice while improving later retention and transfer. Among the most widely studied are spacing, interleaving, retrieval, and varying the conditions of practice.
Spacing to Make Ourselves Remember
Think of a student who needs to learn English vocabulary. They can review twenty words five times in a single afternoon or spread those reviews over several days. In the first case, each repetition occurs while the words are still readily accessible. In the second, some of that accessibility will have faded, and retrieving them will require greater effort. That effort matters: when information has become less accessible but can still be retrieved, the act of remembering it can produce greater learning. The so-called spacing effect is one of the most robust findings in research on memory and learning.
Interleaving to Learn How to Choose
A mathematics class might devote an entire session to solving equations of one type and move on to the next type once students have mastered the procedure. Another possibility is to mix problems that require different strategies. This second option is usually more difficult: each exercise forces students to recognize what kind of problem they are facing and decide which procedure to use. Making that distinction is part of what they will eventually have to do when no one tells them in advance which strategy applies. Research has found benefits from interleaving in areas ranging from mathematics and category learning to motor skills.
Retrieving to Learn
Rereading a text several times quickly increases its familiarity. But closing the book and trying to explain its contents forces us to search for the information in our memory. The same thing happens when we answer questions, use flashcards, or take a short test. From this perspective, a test can become a learning tool: each successful act of “retrieval” modifies and strengthens memory
. So classic study habits such as closing the book and repeating aloud what we have learned, or answering questions from someone who is “quizzing” us, are also desirable difficulties.
Varying Practice to Enable Transfer
Always practicing under the same conditions makes performance easier, although it can tie what we have learned too closely to the context in which we practiced it. Varying those conditions requires continual adjustment. A classic experiment cited by the Bjorks demonstrated this with children practicing throwing beanbags at a target. Those who practiced from different distances ultimately performed better on a later test than those who had always practiced from the exact distance used in the test. In the classroom, variation might mean exploring a concept through different examples or applying it to new problems and situations.
These four strategies share the same logic: they introduce difficulty during practice in order to trigger cognitive activity that helps preserve and use what has been learned later on. The key is getting the level of difficulty right. Because, as we will see next, for that effort to produce learning, the challenge must remain within the student’s reach.
Aprender implica conservar lo aprendido cuando ha pasado el tiempo y ser capaz de recuperarlo cuando hace falta. Y para conseguirlo, el camino puede incluir olvidos, dudas, errores y respuestas que tardan en llegar.
Difficulty Has Its Limits
There is one word in Bjork’s expression that deserves to be taken very seriously: desirable. Introducing difficulties can promote learning, but it can also hinder it. The difference depends, to a large extent, on what the student is able to do with the challenge they are facing.
A difficulty is desirable when it activates encoding and retrieval processes that promote understanding, memory, and transfer. If the student lacks the knowledge or skills needed to respond to that challenge, the same difficulty becomes unproductive.
This means that the same task can work very differently for two students. Asking students to recall the causes of the French Revolution several days after studying them may require a beneficial retrieval effort. For a student who barely knows the events or lacks the vocabulary needed to understand them, the task may simply be beyond their reach. Something similar happens with interleaving: deciding which mathematical procedure to use supports learning when the student already knows the procedures they are choosing between.
That is why prior knowledge plays a fundamental role in this framework. The Bjorks themselves point out that the appropriate level of difficulty can vary considerably according to each student’s experience and knowledge. In later work, they have also emphasized that a challenge should be demanding enough to require effort while remaining surmountable.
For teachers, the implication is important. The difficulty that teaches is found where the student can still do something with it. Designing that difficulty therefore requires teachers to know what the student already knows, what they can retrieve, and what support they need. Spacing, interleaving, or asking students to recall information thus become less mechanical strategies: their effectiveness also depends on finding, at each stage of learning, a level of difficulty that allows students to make an effort while still being able to move forward.
A School That Confuses Performance with Learning
How do we know that a student has learned? The usual answer tends to rely on what we can observe in the classroom: the student answers correctly, solves the exercises, follows an explanation, improves after practice. The problem, according to Bjork, is that these signals tell us mainly about performance at that particular moment, and performance can create a misleading impression of how firmly something has been learned.
This distinction between performance and learning occupies a central place in Bjork’s research. Performance refers to what a person is able to do while practicing or immediately afterward. Learning reveals itself over time: in how much a person retains, in their ability to retrieve that knowledge when they need it, and in how successfully they can use it in different situations. That is why a class in which students get many answers right may be producing less durable learning than one in which they hesitate, search their memory for answers, and make more mistakes.
Desirable difficulties therefore invite us to broaden the horizon through which we assess what happens in the classroom. Alongside immediate results, other questions emerge: What will students remember a week or a month later? Will they be able to retrieve that knowledge without assistance? To recognize when they need to use it? Will they be able to apply it to a problem they have never encountered before?
These questions have consequences for teaching. If the goal is for knowledge to be retrievable and usable later on, a practice that requires effort may be more valuable than one that produces a rapid succession of correct answers. It also requires us to interpret hesitation and mistakes during learning carefully: some may be the result of precisely the cognitive work we want to encourage.
Bjork’s proposal thus introduces a particularly demanding measure of educational success: what matters is how much of what has been learned remains available once the class, the exercise, and the support are no longer there.
The Test of Time
Learning means retaining what we have learned after time has passed and being able to retrieve it when we need it. Achieving this may involve forgetting, uncertainty, mistakes, and answers that take time to emerge. Desirable difficulties help us understand why that effort can be valuable.
Bjork’s paradox thus brings learning back to an experience we can all recognize: what we have had to remember, reconstruct, practice at different times, or apply in different ways tends to leave a deeper imprint. Perhaps learning well has something to do with just that: finding enough difficulty to make us think while still allowing us to move forward.
