Many training practices have traditionally treated ease of learning as a sign of training quality: the smoother the explanation, the easier the activities, the faster the answers, and the fewer the errors, the more successful the program appears.

But learning science asks a different question:

Does ease of performance during training necessarily mean high-quality learning after the training ends?

The answer is not always yes.

Some conditions that make training easier may improve immediate performance without necessarily producing the best long-term retention, while some carefully designed difficulties that slow performance during learning may strengthen information retention and the ability to retrieve and use it later.

This is where the concept of Desirable Difficulties, associated with the work of Robert Bjork and colleagues, emerged; it invites us to rethink the relationship between ease of training and its effectiveness.

But the concept requires careful interpretation: not everything difficult is useful, and confusing the trainee is not a learning strategy.

The central issue is:

How can we design an appropriate level of challenge that makes the trainee exert productive cognitive effort without exceeding their capacity and turning challenge into a burden that obstructs learning?

Desirable difficulties refer to learning conditions that make knowledge acquisition or performance during practice relatively more challenging, yet may improve retention and long-term learning when they are appropriate to the learner and the task.

Examples include:

• Spacing learning sessions.

• Retrieving from memory instead of merely rereading.

• Interleaving different types of problems.

• Varying the conditions and contexts of practice.

• Asking the learner to try to generate an answer before it is provided.

The paradox is that some of these practices may make the learner feel that learning has become harder.

They may make more errors.

They may take longer.

They may feel that their performance is less fluent.

But this temporary reduction in ease of performance does not necessarily mean a reduction in learning quality.

When a trainee receives the answer without needing to search memory, analyze the situation, or distinguish among alternatives, the required cognitive effort is limited.

But when the trainee is asked to:

retrieve, compare, analyze, predict, generate, make a decision, and then review the outcome of that decision;

the learning process becomes more active.

Here we can distinguish between:

Exposure

and

Active Processing

A trainee who watched ten problems being solved has not necessarily learned as much as a trainee who tried to solve the problems independently and then received feedback.

This is one of the most important points to emphasize.

The term “desirable difficulties” can be misunderstood as an invitation to make training more complicated.

That is not correct.

A difficulty is desirable only when it triggers cognitive processes that support learning and the trainee can handle it to a reasonable degree.

If difficulty exceeds the trainee’s prior knowledge or capacity to process the task, it may turn into:

• Confusion.

• Excessive cognitive load.

• Unproductive errors.

• Withdrawal from the activity.

• Reduced motivation.

• Incorrect learning.

Therefore:

Difficulty is not a training value in itself; its value depends on the learning processes it evokes.

Here we reach a delicate area in training design.

Cognitive Load Theory explains that working memory is limited and that instructional design should avoid exhausting its resources with unnecessary elements.

So how can this be reconciled with adding difficulties?

There is not necessarily a contradiction.

The goal is not to increase difficulty randomly, but to remove unproductive difficulty while preserving productive cognitive challenge.

Example:

If the trainee is learning risk analysis, asking them to choose among three similar risks may represent a useful challenge.

By contrast, presenting a poorly designed table, vague instructions, undefined terms, and distracting information is not a desirable difficulty; these are design obstacles.

A fundamental distinction appears here:

Challenge that comes from thinking about the problem may be useful.

Whereas:

Challenge caused by poor training design is not.

A common practice is to compress training into a single block of time:

five hours in one day, or five consecutive days, and then the program ends.

This format may be organizationally necessary, but it is not always optimal for long-term retention.

The Spacing Effect indicates that distributing learning or practice over time can improve retention compared with massing practice into a single period in many contexts.

Instead of:

Learn Learn Learn Test

a learning journey can be designed like:

Learn Interval Retrieve Apply Interval Retrieve New Application

The difficulty here arises because some of the ease of accessing the information decreases over time.

The learner must exert more effort to retrieve it.

That effort can be part of the process that strengthens learning.

Rereading material creates a feeling of familiarity.

Retrieval, by contrast, tests the learner’s ability to access knowledge when it is not in front of them.

So instead of saying:

......Review the previous page.............

the trainer can say:

Close the notes and write from memory the four principles we discussed.

Or:

Without referring to the model, what is the next step? And why?

The distance between the question and reaching the answer represents a form of cognitive effort.

But the degree of difficulty should be appropriate.

If retrieval is nearly impossible, the expected benefit may not materialize.

This is where feedback becomes a critical element.

In Blocked Practice, a person practices one type of problem repeatedly before moving to the next type.

For example:

AAAA → BBBB → CCCC

In Interleaved Practice, the sequence might instead be:

A → C → B → A → B → C

The first method appears easier and more organized and may produce better performance during practice.

But interleaving may force the learner to ask an additional question each time:

What type of problem is in front of me, and which strategy should I use?

This is an important real-world skill because workplace problems do not usually arrive with a label telling the employee which model to use.

Sometimes a trainer can ask trainees to attempt a solution before providing the full explanation.

For example:

Before explaining a decision-making model, the trainer presents a situation and asks:

If you were responsible, what decision would you make, and what criteria would you base it on?

The answers are then discussed before the scientific model is presented.

The trainee enters the content carrying questions and knowledge gaps they discovered themselves.

But generation needs calibration: if the trainee has no prior knowledge in the field, the attempt may become guessing rather than productive learning.

If an employee practices a skill in only one context, they may learn the association between the skill and that specific context more than they learn the general principle.

Therefore, variation can be introduced in:

• Type of client.

• Size of the problem.

• Level of information available.

• Time constraints.

• Nature of the team.

• Organizational scenario.

The aim is not to change the skill, but to train the learner to recognize the deep structure of the problem despite changes in its external appearance.

This brings training closer to the concept of Transfer of Learning.

Scaffolding is essential, especially in the early stages of learning complex skills.

The problem is not the presence of support, but its continuation after the learner becomes capable of taking on more responsibility.

Training can begin, for example, with:

Full trainer performance

then:

Trainee performance with clear steps

then:

Performance with limited prompts

then:

Independent performance

then:

Performance in a new situation

This process can be called:

Gradual Fading of Scaffolding

Its purpose is to prevent two opposite problems:

withdrawing support too early, which leads to confusion,

and

maintaining too much support, which leads to dependence.

There is no single level of difficulty that suits all trainees.

A task that represents a useful challenge for an expert may exceed a novice’s capacity, while a task that challenges a novice may be routine for an expert.

This is consistent with the idea that training effectiveness is influenced by the interaction among:

Learner Characteristics × Task Difficulty × Practice Conditions

Therefore, the designer should ask:

What prior knowledge does the trainee have?

What is the level of expertise?

How complex is the task?

How much support is available?

What is the cost of error?

And is the trainee at the stage of acquiring the skill, consolidating it, or transferring it?

A practical matrix of four levels can be used:

Zone One: Comfort

Low difficulty + high ability

The trainee completes the task quickly and with very few errors.

Risk:

Boredom, limited growth, and the illusion of mastery.

Action:

Increase variation, reduce support, or introduce a new situation.

Zone Two: Productive Challenge

Appropriate difficulty + trainee capacity allows attempt, retrieval, and correction

Indicators:

• Clear cognitive effort.

• Correctable errors.

• Ability to continue.

• Improvement after feedback.

This is usually the target zone.

Zone Three: Overload

Difficulty exceeds the trainee’s available resources

Indicators:

• Repeated basic errors.

• Not knowing where to begin.

• Complete dependence on the trainer.

• Losing the path of the task.

Action:

Break the task down, restore support, provide a worked example, or reduce the number of variables.

Zone Four: Chaos

Here, the challenge does not come from the essence of the skill but from poor design:

unclear instructions, missing information, a complex interface, unexplained terminology, or activities unrelated to the objective.

Action:

Remove the difficulty rather than asking the trainee to adapt to it.

Before adding any challenge to a program, the trainer can use five questions:

1. What cognitive process do I want to stimulate?

Retrieval? Analysis? Discrimination? Decision-making? Transfer?

2. Does the trainee have enough prior knowledge to handle it?

If the answer is no, foundational instruction or scaffolding may be needed.

3. Does the difficulty come from the task or from poor design?

If it comes from poor design, it should be removed.

4. Will the trainee receive feedback?

Difficulty without an opportunity for correction may reinforce the error instead of learning.

5. Do I expect a delayed benefit?

If the difficulty reduces performance now, what evidence do I expect of its later benefit?

Retention? Transfer? Independence? Flexibility?

If we cannot identify the benefit, the difficulty may not be desirable in the first place.

Suppose we have a five-day training program.

Traditional design

Explanation Example Similar Exercise Correction Move to the Next Topic.

Parts of it can be redesigned as follows:

Day One

Foundation of knowledge + worked examples + guided practice.

Day Two

Retrieval without referring to the material + application.

Day Three

Mix skills from the previous two days with a new skill.

Day Four

Unfamiliar scenarios + reduced scaffolding.

Day Five

An integrative problem that requires choosing the appropriate tools without telling the trainee in advance which tool to use.

Then:

After 7 Days

Brief retrieval + application case.

After 30 Days

A transfer situation or a work-related task.

In this way, training is no longer merely five days; it becomes a learning structure extended over time.

The philosophy of this article can be summarized in one rule:

Do not make training difficult just to tire the trainee; make it challenging enough to force practice of the mental or professional process you want to remain with the trainee after training.

The criterion for difficulty then becomes not:

Is the activity difficult?

but:

What is this difficulty forcing the trainee’s mind to do?

If the answer is:

retrieve, compare, analyze, choose, generate, correct, or transfer;

we may be dealing with productive effort.

But if the answer is:

searching for the meaning of instructions, resisting distraction, understanding poor design, or guessing what is required;

then we are dealing with a burden that should be reduced.

One advanced role of the trainer is not merely to facilitate content but to become capable of Challenge Calibration.

That means knowing:

When to explain?

And when to ask?

When to present the model?

And when to hide it?

When to correct immediately?

And when to allow another attempt?

When to simplify?

And when to raise the level of the problem?

When to provide scaffolding?

And when to withdraw it?

These precise decisions may have a greater impact on learning quality than the sheer volume of content.

If we take the concept of desirable difficulties seriously, institutional training may need to revisit some of its assumptions.

It is not necessarily true that:

Higher satisfaction = better learning.

Nor is it necessarily true that:

Fewer errors during training = better retention.

Likewise:

Faster activity completion = greater mastery

is not always a correct rule.

Training dashboards should therefore include indicators related not only to the trainee experience, but also to:

• Delayed retrieval.

• Independent application.

• Ability to discriminate among problems.

• Transfer of knowledge to a new context.

• Reduced dependence on scaffolding.

• Sustainability of performance.

Effective training is not training that removes every difficulty from the learner’s path, nor is it training that confronts the learner with challenges beyond their capacity.

A more mature design distinguishes between two kinds of difficulty:

difficulty that consumes the learner without teaching them,

and

difficulty that compels the learner to practice the processes needed in order to learn.

At that point, the trainer’s role changes.

The trainer no longer asks only:

How can I make the content easier?

but also asks:

Where should things be easy, and where should the trainee have to exert effort?

Removing distraction, ambiguity, and poor design is essential.

But removing retrieval, attempts, correctable errors, discrimination, decision-making, and independent application may make training more comfortable and less durable.

This is why the central principle of the article can be stated as:

Good training does not increase difficulty; it engineers it.

It removes difficulty that does not serve learning and carefully designs the difficulty that makes the trainee think, retrieve, choose, and apply, so that what happens inside the training room becomes an ability that can be used outside it.

Bjork, E. L., & Bjork, R. A. (2011). Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning. In M. A. Gernsbacher et al. (Eds.), Psychology and the real world: Essays illustrating fundamental contributions to society. Worth Publishers.

Bjork, R. A., & Bjork, E. L. (2020). Desirable difficulties in theory and practice. Journal of Applied Research in Memory and Cognition, 9(4), 475–479.

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.

Karpicke, J. D., & Roediger, H. L., III. (2008). The critical importance of retrieval for learning. Science, 319(5865), 966–968.

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.

Sweller, J., van Merriënboer, J. J. G., & Paas, F. (2019). Cognitive architecture and instructional design: 20 years later. Educational Psychology Review, 31, 261–292.