Make It Stick

Make It Stick argues that the study methods that feel most productive in the moment often produce the weakest long-term learning. Rereading, highlighting, cramming, and repeating one kind of problem in a block can create a feeling of mastery because the material becomes easy to process. But fluency is not the same as being able to recall, choose, and apply knowledge later. Brown, Roediger, and McDaniel’s central claim is that learning sticks when practice requires useful mental effort: retrieving from memory, spacing attempts over time, mixing related types, varying conditions, generating answers before seeing them, and connecting new material to what is already known.

The book’s practical reversal is uncomfortable because it changes the signal of progress. Smooth practice may be training recognition rather than recall. Harder practice, when it is relevant, surmountable, and corrected by feedback, may feel slower while building more durable mastery. The goal is not to make learning unpleasant. The goal is retention and transfer: knowledge that remains available later and works outside the exact conditions in which it was first learned.

Familiarity is a poor test of mastery

The book’s first target is miscalibration. Learners often judge progress by how familiar material feels while it is in front of them. Rereading a chapter makes the sentences easier to recognize. Highlighting can make a page look processed. Solving a run of nearly identical problems can make a procedure feel automatic. These activities can improve immediate performance, but they can also inflate confidence because they reduce the need to reconstruct knowledge from memory.

That distinction matters because real use rarely arrives with the answer visible and the problem type labeled. A student on an exam, a professional applying a procedure, or a pilot facing a failure scenario needs access, selection, and execution. The book’s pilot simulator example makes the point sharply: knowing a procedure matters only if it can be retrieved under conditions resembling use. Testing, in that setting, is not merely evaluation. It is calibration.

The easier method feels better, so learners repeat it. The harder method exposes gaps, so it can feel like failure even when it is producing stronger learning. This is why the book treats judgment about learning as part of learning itself. Objective checks, feedback, and delayed attempts reveal whether knowledge is actually retrievable. Rereading is not useless in every circumstance; initial exposure gives retrieval something to work on. The mistake is treating recognition-based fluency as evidence of durable command.

Retrieval practice turns testing into learning

Retrieval practice means trying to bring facts, concepts, procedures, or experiences back from memory. The book’s crucial mechanism is that retrieval is not a neutral measurement event. The act of recalling strengthens the routes by which knowledge can later be accessed, interrupts forgetting, and exposes gaps that passive review can hide. A quiz can therefore be a learning event, not only an assessment.

This is why the authors emphasize low- or no-stakes testing rather than constant high-pressure exams. Flashcards, practice questions, summarizing without looking, explaining an idea from memory, and clicker-style classroom questions all fit the mechanism when they require actual recall and when errors are corrected. The point is not to attach more punishment to learning. The point is to make recall ordinary and frequent.

The Columbia, Illinois classroom experiments described by Roediger show retrieval practice in ordinary school conditions. Students answered questions on actual middle- and high-school material, with immediate feedback and no grades attached to the quizzes. Material that was quizzed more often was retained better on later chapter, semester, and delayed tests. The example matters because it separates testing from grading. The learning gain came from repeated retrieval with feedback embedded in study.

Feedback is essential. If a learner retrieves an incorrect answer and never discovers the error, practice can strengthen the wrong path or leave false confidence intact. The book’s endorsement of generation and retrieval is therefore paired with correction. Struggling to recall something after a delay is not automatically bad news; it may be exactly the work that strengthens later access. But repeated failure without correction is not a desirable difficulty.

Spacing and mixing trade smooth practice for later performance

Spacing distributes study or practice across time so later attempts require renewed retrieval. It is not simply studying longer. The point is that some forgetting makes recall effortful, and successful effortful recall strengthens future access. Cramming can raise performance quickly because everything is still active and familiar, but that immediate advantage often fades. Spacing is designed for availability later, not comfort now.

The surgical-resident study cited in the book’s notes shows that spacing applies beyond textbook facts. A daylong intensive surgical lesson was compared with shorter sessions spread over weeks. The spaced schedule produced better later retention and application. The example matters because it shows spacing in skill learning: distributed practice forces learners to reconstruct and apply the procedure again instead of relying on the momentum of one intensive session. The authors do not pretend there is one perfect interval; spacing depends on the material, the learner, and the future point of use.

Interleaving is related but distinct. It means mixing related problem types, topics, or categories rather than completing a block of one type before moving to the next. Its mechanism is discrimination. In blocked practice, the learner often knows what procedure to use because the section announces it. In interleaved practice, the learner must decide what kind of problem or example is present and choose the response. That selection step is part of the skill.

The math problem research by Rohrer and Taylor, cited in the book’s notes, illustrates this difference. Shuffled practice across problem types beat clustered textbook-style practice on later tests. During practice, blocking can feel better because the learner repeats the same move and improves rapidly. Later tests and real situations are more often mixed. Interleaving prepares the learner for conditions where recognizing the type is part of the task.

The painting-style learning example makes the same point in a perceptual domain. Students learned artists’ styles better when examples were interleaved rather than massed, even though many believed massing helped more. Seeing several works by one artist in a row can feel clarifying, but mixing artists forces comparison. The learner has to notice which features distinguish one style from another.

Variation adds another layer. It changes conditions, examples, or forms within a target skill so learning does not overfit to one version of the task. The anagram study cited in the book’s notes found that learners who practiced varied anagrams for the same target word later outperformed those who practiced the same tested anagram repeatedly. Variation helps when the goal is transfer: recognizing the underlying target across changed appearances.

Generation and elaboration make knowledge usable

Generation is trying to answer, solve, explain, or predict before being shown the correct response. The book’s claim is not that wrong guesses are valuable by magic. A prior attempt changes how the learner receives the answer. It creates a question, directs attention, and makes the later correction more meaningful. As with retrieval, the benefit depends on feedback; uncorrected errors do not become learning merely because they were effortful.

Generation converts learning from reception into search. Instead of asking only what the material says, the learner asks what follows, how to solve, or what would happen next. That attempt creates a contrast between the learner’s current model and the better answer. The contrast is where revision can happen.

Elaboration complements generation by giving new material more routes into memory. To elaborate is to explain an idea in one’s own words, connect it to prior knowledge, build an analogy, create an example, or place a fact inside a larger model. The book’s official examples include heat transfer, history, angular momentum, and baseball batting to show how abstract ideas become more available when connected to concrete prior knowledge and mental models.

The baseball batter example is useful because it shows that expertise is not just more stored information. A skilled batter uses mental models that filter cues and connect them to action. That kind of learning cannot be produced by recognition alone. It requires organized knowledge so the right cues matter and the right response becomes available under changing conditions.

Reflection combines several of these processes after experience. Asking what happened, what worked, what failed, what was learned, and what should be tried next requires retrieval of the event, elaboration of its meaning, and generation of a future adjustment. Brief writing exercises belong here when they force reconstruction and connection, not when they merely record impressions.

The difficulty has to serve the future use

The phrase “desirable difficulties” is easy to misuse. The authors are not saying that any obstacle improves learning. A difficulty is desirable only when it requires cognitive work relevant to the target performance and remains surmountable with adequate support or feedback. Confusion, excessive load, irrelevant friction, or repeated failure without correction do not become good pedagogy because they are hard. The book’s own qualifications matter: in some contexts, including memory-impaired learners, errorless learning may be important even though corrected errors can help many learners in typical educational settings.

A compact decision rule follows from the book’s argument:

1. If the goal is durable recall, add retrieval attempts after initial exposure.
2. If the goal is retention over time, space those attempts so recall becomes effortful.
3. If the goal is choosing the right method or category, interleave related types.
4. If the goal is transfer, vary conditions and examples.
5. If the attempt produces errors or uncertainty, add prompt feedback and revision.

This rule also prevents turning one technique into a universal cure. Factual recall may call most directly for retrieval practice. Inference, novel problem solving, and expert performance also need mental models, elaboration, and varied application. The method should match the future use.

The deeper lesson of Make It Stick is that learners should stop using ease as the main signal of progress. Easy practice can help at the beginning, but durable learning is built by repeatedly reconstructing knowledge under conditions that resemble later use. Retrieval, spacing, interleaving, variation, generation, elaboration, and reflection all work because they replace familiarity with active access, discrimination, and connection. The practical question after any study session is not whether the material felt smooth, but whether you can bring it back, tell when it applies, and use it when the cues have changed.