Skip to main content
How Students (Think They) Learn: The Plusses and Minuses of “Interleaving”
Andrew Watson
Andrew Watson

As the school year begins, teachers want to know: can mind/brain research give us strategies to foster learning?

We might also wonder: what will our students think of those strategies?

College Students Sitting in Hallway

It seems plausible — even likely — that students will prefer the strategies that help them learn. If those strategies help, why wouldn’t students like them?

Strategies to Foster Learning

Some classroom truths seem almost to basic to say out loud. For instance:

#1: We want our students to learn several different sub-topics within any particular topic.

And

#2: Students need to practice to learn.

When teachers think about those basic truths at the same time, we often adopt a specific strategy.

We ask students to practice (that’s #2) each individual subtopic (that’s#1) on its own. So:

Students practice identifying nouns, and then they practice identifying verbs, and then the practice identifying adjectives.

Or, angles, then circumferences, then areas.

Or, backhand, then forehand, then serve.

We could represent this strategy this way: AAA, BBB, CCC. Each sub-topic gets its own discrete practice session.

But, would a different strategy be better? How about: ABC, CBA, BCA?

In other words: should students jumble different topics together when they practice?

Interleaving: Old Research, and New

The answer to that question is YES: students SHOULD jumble different sub-topics together when they practice.

For research confirmation, you can check out this study by Rohrer and Pashler.

Or, for a broader synthesis, explore Agarwal and Bain’s great book, Powerful Teaching.

Or, you might ask a pointed question: “has this strategy been tested in actual classrooms, not just in psychology research labs?”

The answer to that question is also YES.

recently published study by Samani and Pan tried this strategy in a college physics class.

Sure enough, students learned more when their homework problems were interleaved than when sub-topics were practiced one at a time.

That is: students whose practice problems covered Coulomb’s Law by itself learned less than those whose practice problems also included capacitors and composite wires.

So, we arrive at this tentative teaching advice:

No doubt, you have your students practice — either in class, or with homework, or both.

When students practice, they should work on a few sub-topics at a time, not just one.

So far, so good.

Paradox: Teaching Solutions Create Studying Problems

Let’s return to the question that opened this blog post: do students prefer the study strategy that fosters learning. (They should; after all, it helped them learn!)

Reader, they do not.

Why?

In Samani and Pan’s study (and many others), students found that effective learning strategies are more difficult.

That is: they require more thought, and frequently lead to more short-term mistakes. (Students did relatively badly on the homework before they did relatively well on the tests.)

From one perspective, this finding makes perfect sense.

If we do difficult mental work, we will struggle and fail more often. And yet, all that extra hard thinking will ultimately lead to more learning. (Soderstrom and Bjork have written a GREAT review article on this topic.)

That encouraging perspective, however, runs into a perfectly understandable alternative: most people don’t like struggle and failure.

We shouldn’t blame students for disliking the interleaving. It hurt their heads. They did badly on the homework. YUCK.

As teachers, we have the long-term perspective. We know that short-term struggle leads ultimately to greater learning.

But, most students lack that perspective. They feel the struggle and the pain, but don’t recognize the long-term benefits.

Teaching Advice 2.0

Given all these findings, how should we structure students’ practice?

I think all these findings add up to this guidance:

First: interleave practice.

Second: tell students that you are doing so, and explain why.

The language you use and the level of explanation will, of course, vary by the age of the student. But, let them know.

Third: structure grading systems to value ultimate learning more than immediate understanding.

After all, if we both require interleaved practice (which is quite difficult) and grade students on the success of their practice, we will — in effect — force them to have lower grades. They will rightly feel the injustice of this instructional paradigm.

In other words: this practice strategy — in my view — does imply a grading policy as well.

TL;DR

Students, of course, must practice to learn.

Teachers should structure their practice to cover a few sub-topics simultaneously.

We should explain why we’re doing so; “interleaving” ultimately results in more learning.

We should create grading structures that account for the initial difficulty of interleaved practice.

If we get this balance right, students will willingly face early learning challenges, and ultimately learn more.


Rohrer, D., & Pashler, H. (2010). Recent research on human learning challenges conventional instructional strategies. Educational Researcher39(5), 406-412.

Agarwal, P. K., & Bain, P. M. (2019). Powerful teaching: Unleash the science of learning. John Wiley & Sons.

Samani, J., & Pan, S. C. (2021). Interleaved practice enhances memory and problem-solving ability in undergraduate physics. NPJ science of learning6(1), 1-11.

Soderstrom, N. C., & Bjork, R. A. (2015). Learning versus performance: An integrative review. Perspectives on Psychological Science10(2), 176-199.

A “Noisy” Problem: What If Research Contradicts Students’ Beliefs?
Andrew Watson
Andrew Watson

The invaluable Peps Mccrea recently wrote about a vexing problem in education: the “noisy relationship between teaching and learning.”

In other words: I can’t really discern EXACTLY what parts of my teaching helped my students learn.

Was it my content knowledge?

The quality of my rapport with them?

The retrieval practice I require?

The fact that they slept and ate well in the days before class?

Some combination of all these variables?

Because I don’t know EXACTLY which teaching variable helped (or hurt) learning, I struggle to focus on the good stuff and eliminate the bad stuff.

I thought about Mccrea’s wisdom when I read a recent study about interleaving.

Here’s the story…

Interleaving 101

Frequent blog readers know all about interleaving, a way of organizing students’ practice.

Let’s say I teach my students about parts of speech.

Once they have a basic understanding of each one, I could have them practice each part of speech on its own.

That is: they identify nouns on Monday, adverbs on Tuesday, prepositions on Wednesday, and so forth.

Researchers call that structure “blocking” — as in “blocks of homework focusing on individual topics.”

Or, I could have my students jumble several topics together every night.

That is: Monday night they practice nouns, adverbs, and prepositions. Tuesday they practice verbs, prepositions, and conjunctions. Wednesday: nouns, verbs, and adjectives.

The total number of practice problems would remain the same, but they’d practice several parts of speech all together.

Researchers call this system “interleaving” — as in “weaving together several different topics.”

Measuring Success

Of course, teachers want to know: does interleaving work? Do students who interleave their practice learn more than students who block?

Let’s imagine two ways of answering that question

Strategy #1: ask the students.

Obviously.

Who knows more about the students’ learning than the students themselves?

Strategy #2measure their learning.

Obviously.

If students who block consistently remember more than students who interleave (or vice versa), then we have a winner.

So, what’s the answer?

Answers, and Vexing Questions

According to Samini and Pan’s 2021 study, strategy #1 yields a clear answer: students say that interleaving is harder and results in LESS learning.

Of course, that means they think that blocking is easier and results in MORE learning.

Alas, strategy #2 arrives at a contradictory result.

When we measure students’ actual learning, they remember more after interleaving than blocking.

Samini and Pan’s study gets this result. And, LOTS AND LOTS of research gets to the same result. (See Agarwal and Bain’s book for a great review of the research.)

In other words, this study points to an especially “noisy” part of the relationship between teaching and learning.

Students genuinely think and believe that interleaving interferes with learning.

However, interleaving in fact promotes learning.

How do we handle this quandary?

Tentative Solutions

In my high-school classroom, we do A LOT of retrieval practice.

Almost every day, I fire off questions and ask students to attempt an answer.

Sometimes I call on raised hands; or cold call; or have students write answers in their notebooks (I circle the room to check their responses). They might write on the board; they might consult in pairs.

I’m entirely comfortable using retrieval practice — and so are my students — because on the second day of class I showed them research about retrieval practice.

I told them:

This might feel hard at first.

But, trust me. It feels hard because your brain is working harder. And that means you’re learning more.

It’s like going to the gym. You don’t gain muscle by picking up paper clips. You gain muscle by picking up heavy things. Hard work leads to better fitness.

The same rule applies here. Retrieval practice is harder, so you’ll learn more.

Since that day, I stop every now and then at the end of an RP session and say: ” do you feel how much you’ve learned? Do you see how much retrieval practice is helping?”

In fact (I swear I am not making this up), one of my Sophomores once said: “Thank you Mr. Watson for making us do retrieval practice every day.”

I tell this story because it applies to interleaving as well.

I’ve been interleaving all year, but I haven’t (yet) explained it to my students. I plan to do so this upcoming week (or next).

My hope is: they’ll see why we’ve been bouncing back and forth from topic to topic in ways that might seem random or disorganized.

We’ve been interleaving all along.

I offer this solution as “tentative” because my context might not match yours.

For instance, if you teach younger or older students, they might not respond as mine do.

If you teach students with diagnosed learning differences, interleaving might not benefit them as much.

And so forth.

As always: consider the research findings, consider my experience, and then use your own best judgment to fit them into your classroom practice.

TL;DR

If students’ beliefs contradict research, I myself tell them about the research — graphs and all. And then I ask them to trust me.

Retrieval practice and interleaving really do work. My students know about this research pool. So far, they’re on board.

If you try this strategy, or another one, I hope you’ll let me know about your own experience.


Samani, J., & Pan, S. C. (2021). Interleaved practice enhances memory and problem-solving ability in undergraduate physics. NPJ science of learning6(1), 1-11.

The Best Kind of Practice for Students Depends on the Learning Goal
Andrew Watson
Andrew Watson

In some ways, teaching ought to be straightforward. Teachers introduce new material (by some method or another), and we have our students practice (by some method or another).

Result: THEY (should) LEARN.

Alas, both classroom experience and psychology/neuroscience research suggest that the process is MUCH more complicated.

For instance:

When we “introduce new material,” should we use direct instruction or more of an inquiry/problem-based pedagogy? *

When we “have our students practice,” what’s the very BEST kind of practice?

Around here, we typically offer two answers to that 2nd question: retrieval practice and interleaving.

Retrieval practice has gotten lots of love on this blog — for instance, here. I have written less about interleaving, mostly because we have less research on the topic.

But I’ve found some ripping good — and very practical — research to share here at the end of 2021.

“What?,” “Why?,” and Other Important Questions

Let’s start with definitions.

Let’s say I teach a particular topic today: “adjectives.” And tomorrow I teach “adverbs.” Next day, “prepositions.” Next: “coordinating conjunctions.”

How should I structure students’ homework?

They could do 20 adjective practice problems tonight. Then 20 adverb problems the next night. Then 20 prepositions. And so forth.

Let’s call that homework schedule blocking.

Or, they could do 5 adjective problems a night for the next 4 nights. And 5 adverb problems a night starting tomorrow night. And so forth.

If I go with this system, students will practice multiple different topics (adjectives, adverbs, prepositions…) at the same time. So, let’s call that homework schedule interleaving.

For the most part, when we compare these two approaches, we find that interleaving results in more learning than blocking. (Lots of info here. Also in this book.)

That’s an interesting conclusion, but why is it true?

In the first place, probably, interleaving is a desirable difficulty. Students must THINK HARDER when they interleave practice, so they learn more.

In the second place, well, we don’t exactly know. Our confusion, in fact, stems in part from an arresting truth: interleaving usually helps students learn, but not always.

Of course, NOTHING ALWAYS WORKS, so we’re not fully surprised. But if the exceptions helped explain the rule, that could be mightily helpful…

An Intriguing Possibility…

Two scholars — Paulo F. Carvalho and Robert Goldstone — have been studying a potential explanation.

Perhaps blocking and interleaving enhance different kinds of memories. And so, research produces contradictory results because researchers use different kinds of memory tests.

Specifically, they propose that:

During blocked study, attention and encoding are progressively directed toward the similarities among successive items belonging to the same category,

whereas during interleaved study attention and encoding are progressively directed toward the differences between successive items belonging to different categories.

In other words: blocking focuses students on the properties of a particular category (“adjectives”). Interleaving focuses students on the distinctions among different categories (“adjectives, adverbs, prepositions”).

And so: if I want students to DEFINE ONE topic or idea or category (“adjectives”), blocking will help them do that well.

If I want students to COMPARE/CONTRAST MANY topics or ideas or categories, interleaving will help them do that well.

To repeat the title of this blog post: “the best kind of practice for students depends on the learning goal.”

In their most recent study, Carvalho and Goldstone test this possibility.

Sure enough, they find that students who block practice do better at defining terms, whereas those who interleave practice do better at multiple-choice questions.

The study gets splendidly intricate — they work hard to disprove their own hypothesis. But once they can’t do so, they admit they they just might be right.

Caveats and Classroom Implications

Caveat #1: “one study is just one study, folks.” (Dan Willingham.)

Although, to be fair, Carvalho and Goldstone have been building a series of studies looking at this question.

Caveat #2: The researchers worked with adults (average age in the 30s) studying psychology topics.

Does their conclusion hold true for K-12 students learning K-12 topics? Maybe…

Caveat #3: Practically speaking, this research might focus on a distinction that evaporates over time.

In truth, I always want my students to know specific definitions — like “tragedy” — well. And, I want them to compare those well-known definitions flexibly to other definitions — like, say, “comedy.”

An an English teacher, I — of course! — want my students to define adjective. AND I — of course!! — want them to compare that definition/concept to other related ideas (adverbs; participles; prepositional phrases acting as adjectives).

In other words, I suspect the ultimate teaching implication of this research goes like this:

We should have students BLOCK practice until they know definitions to some degree of confidence, and then have them INTERLEAVE practice to bring those definitions flexibly together.

To be clear: I’m extrapolating, based on my classroom experience and on my reading in this field.

Until my interpretation gets more research behind it, Carvahlo and Goldstone’s research suggests this general plan:

START BY DECIDING ON THE GOAL.

If you mostly want your students to know individual concepts, have them block their practice.

If you mostly want them to bring several topics together, have them interleave practice.

As your goal changes, their homework changes too.

As is so often the case, this research doesn’t tell teachers what to do. It helps us think more clearly about the work we’re doing.

In my view, that’s the most helpful research of all.


* I think that’s a false choice; both approaches make sense under different circumstances. More on that in another blog post.


Carvalho, P. F., & Goldstone, R. L. (2021). The most efficient sequence of study depends on the type of test. Applied Cognitive Psychology35(1), 82-97.

Conflicting Advice: What to Do When Cognitive Science Strategies Clash?
Andrew Watson
Andrew Watson

Teachers like research-informed guidance because it offers a measure of certainty.

“Why do you run your classes that way?”

“Because RESEARCH SAYS SO!”

Alas, we occasionally find that research encourages AND DISCOURAGES the same strategy simultaneously.

What to do when expert advice differs?

In fact, I got this question on Thursday during a Learning and the Brain Summer Institute. Here’s the setup.

“Difficult” Can be Good

Regular readers know that desirable difficulties help students learn. As explained by Bob Bjork and Elizabeth Ligon Bjork — and researched by countless scholars — some degree of cognitive challenge enhances long-term memory formation.

In brief: “easy learning doesn’t stick.”

And so: why do spacing and interleaving help students learn? Because they ramp up desirable difficulty.

Why is retrieval practice better than simple review? Because (among other reasons) review isn’t difficult enough. Retrieval practice, done correctly, adds just the right amount of challenge.

And so, if you attend Learning and the Brain conferences (like this one on “Teaching Thinking Brains”), or if you read any of the great books about long-term memory formation, you’ll hear a lot about desirable difficulty.

Memory at Work

Cognitive scientists who don’t focus on long-term memory might instead focus on a distinct mental capacity: working memory. 

Working memory allows us to gather information — facts, procedures, etc. — into a mental holding space, and then to reorganize and combine them into new patterns and ideas.

In other words: it’s absolutely vital for thinking and learning. If students are learning academic information, they are using their working memory.

Alas, all this good news comes with some bad news: we don’t have much working memory. And, our students probably have less than we do. (For evidence, try this mental exercise: try alphabetizing the workdays of the week. No problem alphabetizing 5 words? Now try alphabetizing the twelve months of the year. OUCH.)

For this reason, effective teachers pay scrupulous attention to working memory load. Every time we go beyond working memory constraints, we make learning MUCH HARDER.

In fact, I think working memory is so important that I wrote a lengthy series of blog posts on the topic. I’m kind of obsessed. (Heck: I even wrote a book on the topic, called Learning Begins.)

Trouble in CogSci Paradise

Because both topics — desirable difficulties and working memory — provide teachers with important and powerful insights, I devoted much of last week’s workshop to them. Almost every day, in fact, we talked about both.

On Thursday, one participant asked this wise and provocative question:

Wait a minute. You’ve told us that desirable difficulties help learning. And you’ve told us that working memory overload hinders learning.

But: isn’t desirable difficulty a potential working memory overload? Don’t those two pieces of advice conflict with each other? Won’t “spacing” and “interleaving” vex working memory?

Yes, reader, they certainly might.

So, what’s a research-focused teacher to do? Team Desirable Difficulty tells us to space and interleave practice. Team Working Memory tells us to beware overload. How can we make sense of this conflicting advice?

This (entirely reasonable) question has two answers: one specific, one general.

A Specific Answer

When we consider the tension between “working memory” and “desirable difficulty,” we can focus for a moment on the adjective “desirable.”

In almost every case, working memory overload is UNdesirable.

So, if our teaching strategy — spacing, interleaving, retrieval practice, metacognition — results in overload, we shouldn’t do it: it’s not desirably difficult. We should, instead, back off on the difficulty until students can manage that cognitive load.

How do we get that balance just right?

We use our teacherly experience and insight. If I create a homework assignment with lots of interleaved practice AND ALL MY STUDENTS DO TERRIBLY, then interleaving wasn’t desirably difficult. (Or, perhaps, I taught the concepts ineffectively.)

In this case, I know the next night’s assignment should be working-memory-friendlier.

No research can tell us exactly what the best balance will be. Our expertise as teachers will guide us.

The General Answer

Researchers and teachers have different goals, and follow different practices. In brief: researchers isolate variables; teachers combine variables.

We think about stress and about working memory and about alertness and about technology and about spacing and

That list goes on almost infinitely.

For that reason, I chant my mantra: when adopting cognitive science approaches to teaching, “don’t just do this thing; instead, think this way.”

That is: don’t just DO “spacing and interleaving” because research tells us they’re good ideas. Instead, we have to THINK about the ideas that guide spacing and interleaving, and be sure they make sense at this particular moment.

Should we have students meditate at the beginning of each class? It depends on our students, our school, our schedule, our culture, our curriculum, our goals, and … too many other variables to list here.

Should we ban laptops from classrooms? Ditto.

Should high schools start later? Ditto.

Should 3rd graders learn by doing projects? Ditto.

Should students read on exercycles? Ditto.

One isolated piece of research advice can’t effectively guide teaching and school-keeping decisions. We have to combine the variables, and think about them in our specific context.

Simply put: we can’t just “do what the research says.” It’s not possible; different research pools almost certainly conflict.

Instead, we’re doing something more challenge, more interesting, and more fun.

Let the adventure begin!

Pro Tips: How To Think Like A Cognitive Scientist
Andrew Watson
Andrew Watson

Here’s an enthusiastic article from down under.

cognitive science principles

The Sydney Morning Herald reports that Victoria University has introduced an “intensive” course model. Students don’t take multiple courses over many weeks. Instead, they take one course for four weeks. Students absorb a full term of learning in one zealous month.

The students interviewed by the paper were enthusiastic. 19-year-old Alice Growden says:

I am learning a lot more; I feel like the information is easier to understand this way. It’s easier to do better. You are not slammed by four different assignments at once. It is much more balanced.

The Morning Herald’s tone (and my Twitter feed) insist on the benefits of these intensive courses. Seemingly only grouchy professors — who fret that they won’t have enough time for research — object.

Cognitive Science Principles, Take 1: The Spacing Effect

Despite this article’s enthusiastic tone, cognitive scientists will quickly doubt the benefits of this “intensive” course schedule.

After all, we have lots of research showing that spreading practice out over time creates more learning than bunching that practice all together.

For instance, Doug Rohrer looks at shorter and longer lengths of time that courses cover topics. His conclusion — in the modest language of research:

Long-term learning is best achieved when the exposures to a concept are distributed over time periods that are longer rather than shorter.

He finds this conclusion to hold even for intensive language courses, where teachers most often champion the strategy.

Many other scholars have reached this same conclusion. Nicholas Cepeda (along with Doug Rohrer, Hal Pasher, and others) worked with more than 1300 students, and retested them up to a year later.

Their conclusion? Spread learning out over time.

This idea holds even for flashcard study strategies.

Pro Tip #1: If you want to think like a cognitive scientist, beware teaching strategies that promote lots of learning in a relatively short time.

Cognitive Science Principles, Take 2: The Illusion of Knowing

As quoted above, student Alice Growden emphasizes the ease with which she learns:

“I am learning a lot more; I feel like the information is easier to understand this way. It’s easier to do better.”

Yet here again, cognitive scientists will be skeptical.

Remember this principle: easy learning doesn’t stick. Instead, teachers should foster a desirable level of difficulty.

In fact, this principle helps explain the principle above. Spreading practice out over time helps students learn better because it creates additional cognitive challenges.

The extra mental work that students do, in turn, creates more enduring neural networks to encode new memories.

Another example: rereading the textbook.

Students LOVE rereading the book, because it’s relatively easy. This study strategy gives them the illusion of knowing. They say to themselves: “I recognize that passage! I must know this!”

Alas, this illusion comforts students, but isn’t helping them learn more.

I frequently cite Nick Soderstrom’s comprehensive article distinguishing between two results of study: performance vs. learning.

Students often believe that if they “perform” well — say, they recognized everything in their notes — then they have studied effectively. Alas, higher early performance often results in less learning.

Pro Tip #2: If you want to think like a cognitive scientist, beware teaching strategies that emphasize how easy new learning will be. Easy learning doesn’t stick.