Wise Application of Learning Science: Spacing and Interleaving

By Erica Kleinknecht, PhD., 2020

On Giving Psychology Away via Wise Application

In a virtual discussion about teaching the other day (a fav pass time of mine, as my regular readers will not be surprised to note) a question about spacing and interleaving came up that prompted me to take a look at the literature as it’s been a while since doing so for that topic in particular. After a quick read of the articles you see noted in the reference section of this post, I decided to write up a response here on my blog, rather than let it get lost in the weeds of a Facebook group discussion, knowing how hard it is to find those again when I want to refer back to something. So here it is.

Rather than rehash the conversation, I thought I’d make this post more of a “what is” and “how to do it” kind of essay, in the spirit of Walton’s (2014) “wise intervention” framework. Wise interventions are “wise to” (as in, “in the know”) specific underlying psychological process that, when nurtured well, can help people flourish (Walton, 2014). Wise interventions are not experimental methods, nor are they clinical interventions. Rather, they are everyday applications of principles gleaned from research. They are, in other words, an attempt at making good on George Miller’s 1969 call to Psychologists to work harder at creating a body of knowledge that can be used to promote human welfare. It’s a spirit that I take to heart in my work, and the reason d’etre of this blog.

Specific psychological process: Learning, whether new content (math, science, psychology) or a new skill (violin, guitar, dance). A change in neurological connections that equates to memory or improvement.

Wise intervention: A means of using what’s known about the psychological processes to improve upon it or enhance it’s function.

Learning science: an interdisciplinary academic field aimed at creating wise interventions for learners, by applying principles gleaned from cognitive, developmental, and educational science.

 

Learning Science: Spacing and Interleaving

Back to the topic at hand, then: the concepts of spacing and interleaving come from the Learning Science literature. They are two approaches to organizing one’s time and engagement, when it comes to creating an efficient and effective studying regime. At one level, they are simple ideas and, indeed, not new. William James warned against the opposite condition of spacing, cramming, back in the late 1800s. Spacing effects are what you get when you engage in distributed, as opposed to massed, practice. The effect is that over time, you can achieve mastery in less total time compared to longer cramming sessions. Interleaving is the practice of taking multiple to-be-learned sets and mixing them up (e.g., see here for more examples).

Spacing: When the to-be-remembered information is practiced in pieces dispersed across intervals rather than all at once, in an extended session. Contrast to a cramming approach.

Interleaving: A blend of multiple, spaced, concepts or skills, such that a learner will “mix-up” content or skills in practice sessions. Contrast to a blocking approach.

The long-term benefits of spacing and interleaving are appealing: “Learn more and better in less total time? Sign me up!” The nitty-gritty of putting these ideas into practice though can appear counterintuitive and complicated though: “Doesn’t “mixing it up” open you up for confusion, interference, or incomplete learning? Why would I do that?” In thinking about to plan out your learning time, important questions about how to do it right, or to maximum effect reasonably come to mind.

Questions about Spacing:

    • Why isn’t it beneficial to push-through and do the work all at once?
    • How do you break down information into manageable units?
    • What “are” the units or where do you draw lines?

Questions about Interleaving:

    • What, exactly, do you interleave?
    • Is there ever a time when blocking is best?
    • Do you mix up topics or disciplines to be studied?
    • Do you take different aspects of the content or skill, and mix those up?

Does the research point to answers to these questions? Yes, it does, but it’s story of conditions that point to finding a “goldilocks” balance of spacing and interleaving in a manner that is just-right and that differs across learning topics. Indeed, too granular a level and the benefits wash out (e.g., Hausman & Kornell, 2014). On the other hand, complete blocking isn’t necessarily the way you want to go, either. Rather, the way you want to think about spacing and interleaving requires some macro-level considerations (how many different topics or skills are to be studied in a homework or practice session?, Brunmair & Richter, 2019) and some micro-level considerations (within a topical domain, how to mix-up the content?).

Cognition Basics: Attention and Schemas

When thinking through how to advise a learner to organize their study and practice schedules then, several considerations need to be made. In thinking about what to aim for, it can help to have some general principles about cognition and learning in mind.

First, it’s important to remember that our ability to stay intently focused on a single topic or skill waxes and wanes. Indeed, over the span of a day (i.e., waking hours) our minds wander about 50% of the time (see Amishi Jha’s 2017 TED talk for more on “How to tame your wandering mind”). Our brain needs repeated intervals of recharge, as neurons are constantly cycling through periods of activity and rest. Bouts of activity (or focused attention) must be followed by rest, and if we don’t deliberately give ourselves that time our brain takes it anyway, leading to a diminishing return when attempting to “push through.” This principle is one key reason why interleaving is a powerful process for learning: it makes the most of cycling periods of attentional focus and attentional drift (e.g., Brunmair & Richter, 2019; Carvalho & Goldstone, 2014).

Second, neurological processing acts in the service of pattern matching and prediction (e.g., van Kesteren & Meeter, 2020). We live life through ongoing series of episodes. Across these episodes, our brain responds in increasingly routinized ways, reflecting the consistencies across experiences. Our awareness of this process comes in the way of building knowledge, creating categories of understanding, and making connections from one lesson (or situation, or activity) to another. When it’s knowledge, we say that across episodes we abstract commonalities and build semantic webs (schemas) that give us a feeling or knowing or understanding when activated. Said more simply, we build schemas of categories and concepts (for more on schemas, see here ) and once this knowledge is established, it guides how we continue to build on it. When it’s an activity, we say that we go from novice to expert, or deliberate action to automatic process. Activities are also schematized though, that’s what enables expertise. In order to create schemas that work for us, we need consistency but also contrast.

Third, while on the one hand specific areas of the brain are utilized in different ways, on the other hand, any act of learning is supported by a flow of signals across many regions in the brain. We can see different flow patterns for different activities though, noting as well that when one is learning or working on improving there are common areas of activation across different task-types. A handful of examples:

Linguistic processing is supported by a feedback loop of activity across frontal cortex, motor cortex, and temporal cortex.

Visual processing is supported by a feedback loop of activity across frontal cortex, occipital cortex and temporal cortex.

Movements are supported by a feedback loop of activity across frontal cortex, motor cortex, are parietal cortex.

 

Importantly, beneficial, long lasting learning involves making meaningful connections within and across loops. That is, following the principle of encoding variability, when you can combine visual and linguistic processing you will increase the number of cues that will effectively enable later retrieval. Same holds true for combining movement and language, or movement with visualization, and so on (see van Kesteren & Meeter, 2020, for their recent short and sweet summary of these ideas, published here in open source format).

Wise application: How-to wisely space and interleave.

On the surface the first two points – the need for attention refreshing and the need for consistency to build schemas – may seem to work against each other. However, when balanced just right, they can work together quite well, and in the end they enhance point 3, encoding variability. When it comes to spacing and interleaving, at a macro level, you want to do more than one thing or study more than one topic during a study or practice session. At a micro level, you also want to wisely group and space what you are learning.

On that note of wise spacing and interleaving, we need to consider a few more issues.

The kind of mental energy (ie., specific neural pathways or feedback loops) we need to use varies from task to task when learning and some kinds of energy benefit more from interleaving than do others. From a meta-analysis of a variety of studies examining spacing and interleaving across different domains of learning, Brunmair and Richter (2019) report an overall positive effect from interleaving, however, the strengths of the effect differs across domains (they examined expository, math, and visual materials). Interleaving is particularly helpful when learning visual materials. However, interleaving can exert negligible (if not negative) effects when studying expository materials. Mathematical tasks show moderate benefit from interleaved practice. Overall, the authors conclude that when the act of interleaving is done with complex materials it can help by highlighting contrasting facets of the information. When information is highly interconnected, interleaving can disrupt important opportunities for flow or comprehension therein.

The work Brunmair and Richter (2019) review did not include activity or motor skills. Carvalho and Goldstone (2014) though report that interleaving is particularly effective for learning complex movements. I can only assume this is because each proximal “step” towards a distal goal of the activity serves a purpose and if each purpose is practiced with a pause first, it serves to highlight how each (artificially) discrete phase adds to the big picture. As I type this I am thinking about movements in dance, gymnastics, or music performance. The end goal of such movements is a complex but flowing performance when mastered, but the road to mastery must be taken one step at a time and the act of spacing and interleaving your practice culminates in a more nuanced and exceptional performance.

My speculation reflects the collective wisdom gleaned from the work mentioned here, that the practice of breaking studying and practicing up into meaningful pieces and then wisely spacing those pieces out is in keeping with attentional dynamics and with natural schematization (or abstraction) processes. Interleaving across domains (or, more importantly perhaps, coding systems; e.g., see Young & colleagues, 2019) allows your attentional resources to refresh. As well, when grouped and dispersed wisely, interleaving across domains can serve to highlight categorical boundaries thereby enhancing the effectiveness of schema-building.

What does this look like, in more practical terms? I have some ideas. In practice, you want to plan at a macro level (topics, domains, codes) and at a micro level (creating appropriate units for practice within a domain).

At the macro level, students are usually learning and practicing more than one subject at a time, so this tends to happen naturally. The work I’ve mentioned here suggests that you can finesse this process by wisely planning which topics or activities might hang together nicely. For example, it might be beneficial to have “blocking days” for math and language arts homework, and “interleaving days” for music practice, science, and social studies. That is, when deciding how to juggle typical school topics, you want to think about topics, the code-systems each topic utilizes (e.g., primarily linguistic, visual, mathematical, or movement based) and then organize your work flow so that you don’t over-exert one coding system (e.g., linguistic codes can be emphasized in more than just language arts; much of social studies and science utilize prose, too).

    • When Math and Language Arts are the topics for homework, the research above suggests that these topics are best studied in blocks. Learners still want to space out their work and take breaks, but they should plan to complete the task before moving on to the next subject.
    • For topics like Science and Social Studies, some of the work is verbal (i.e., vocabulary), some is expository (exposition and reading comprehension), some is visual, and some is mathematical. When the work is primarily verbal or mathematical then you might want to block your practice. When the work is visual or categorical though, then interleaving natural “sets” should be helpful.
    • For kinesthetic performance practice like learning an instrument, interleaving appears to be beneficial.

Once you’ve decided which aspects of your weekly homework and practice should be interleaved or blocked, then you need to decide how to organize the interleaving. What exactly do you do, here? Yang and colleagues’ (2018) recent work provides us with some sensical guidelines. In a series of 3 studies, they show benefit to interleaving bouts of studying new material with retrieval practice of old material. So what’s interleaved in a study session is “new work” (e.g., study with an appropriate elaboration technique) and “old work” (practiced via self-testing). This sort of schedule for studying respects the issues noted earlier regarding our need for attentional refresh, the process of schematization of what we encounter and learn, and it creates natural opportunities for encoding variability.

The key to effective interleaving, then, lies in knowing where the natural breaks are, and using those to create learning sets. Hausman and Kornell (2014) show that when interleaving is taken to an extreme, for example where student-participants studied interleaved sets of flashcards that included a shuffled deck of cards for studying second-language vocabulary and anatomy concepts, they performed on later tests more poorly than when the information to be learned was studied in block-sets. Based on the research reviewed here, this outcome makes sense: the content and means of study was too similar and thus it not only overly-taxed the participants’ coding systems, it also didn’t help to delineate clear category boundaries. The takeaway then is that when deciding on learning sets, each set within an interleaved study-schedule should still retain schematic integrity as a stand-alone “snapshot” of what’s being studied or practiced.

Earlier in this post I shared some questions about how to organize study time and I think I’ve covered enough material to provide answers to these questions. Let’s review :).

Questions about Spacing

Q. Why isn’t it beneficial to push-through and do the work all at once?

A. Attention needs refreshing. When you push through your mind will wander, creating a diminishing return. You are better off giving yourself a break when you feel your mind straying from the task.

Q. What “are” the units or where do you draw lines?

A. Units can be described as “new” material and “old” material.

Q. [When it’s new material] How do you break information up into manageable units?

A. Create small, but schematic sets. Think about the bare-bones structural requirements of a paragraph, as you do so: topic, body, conclusion. That is, a complete paragraph has three parts. Use this as a guide for grouping to-be-studied sets.

Questions about Interleaving:

Q. What, exactly, do you interleave?

A. Interleave learning material based on content / codes. Some content is more appropriate to interleaving than others (see above). For material appropriate to interleave, mix up codes (verbal, visual, movement). As well, interleave “new” and “old”. This serves to refresh attention as well, because you engage differently with new (re: elaboration) than with old (re: retrieval practice) material.

Q. Is there ever a time when blocking is best?

A. Yes. Math is hierarchical and complex and appears to benefit from blocking. Complex expository work, like writing essays or engaging in detailed reading comprehension exercises also appear to benefit from blocking. (You still need to engage in spacing though when you do this kind of work).

Q. Do you mix up topics or disciplines to be studied?

A. Yes, and no. See above considerations. The lines to be drawn have more to do with coding and whether natural breaks in the process of learning can be finessed via interleaving. In short: learners have to interleave across disciplines because they are learning more than one subject or skill at a time. The granularity with which they interleave depends on coding and complexity of the material though.

Q. Do you take different aspects of the content or skill, and mix those up?

A. Yes. Using an “old / new” format, and using “engagement code” formats, you want to interleave practice engagements as described above.

In Conclusion —

 When it comes to understanding learning, memory, and skill improvement at a deep level, it’s both remarkably simple and extraordinarily complex. At one level our brains are simple: neurons fire in response to signals from the environment, and following Hebb’s Rule, neurons that fire together, wire together. There are many ways to go about the work of creating these wiring patterns in the service of our learning and performance goals, but some ways appear to be more beneficial than others. Taking the time to create wisely spaced and appropriately interleaved study and practice schedules can help learners make the most of their learning time, and, in fact save them time in the end, and to great effect.

 

References

 

Brunmair, M. & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, 145, 1029 – 1052. Doi: http://dx.doi.org/10.1037/bul0000209

 

Carvalho, P. F. & Goldstone, R. L. (2014). Effects of interleaved and blocked study on delayed test of category learning generalization. Frontiers in Psychology, 5. Doi: 10.3389/fpsyg.2014.00936

 

Hausman, H. & Kornell, N. (2014). Mixing topics while studying does not enhance learning. Journal of Applied Research in Memory and Cognition, 3, 153 – 160. Doi: https://doi.org/10.1016/j.jarmac.2014.03.003

 

Miller, G. A. (1969). Psychology as a means of promoting human welfare. American Psychologist, 24, 1063–1075. https://doi.org/10.1037/h0028988

 

van Kesteren, M.T.R. & Meeter, M. (2020). How to optimize knowledge construction in the brain. NPJ Science of Learning, 5. Doi: https://doi.org/10.1038/s41539-020-0064-y.

 

Walton, G. M. (2914). The new science of wise psychological interventions. Current Directions in Psychological Science, 23, 73 – 82. Doi: 10.1177/0963721413512856

 

Yang, C., Chew, S-J., Sun, B., & Shanks, D.R. (2018). The forward effects of testing transfer to different domains of learning. Journal of Educational Psychology, 111, 809 – 826. Doi: http://dx.doi.org/10.1037/edu0000320.

 

Young, A. P., Healy, A. F., Jones, M. & Bourne Jr., L.E. (2019). Verbal and spatial acquisition as a function of distributed practice and code-specific interference. Memory & Cognition, 4, 779 – 791. Doi: https://doi.org/10.3758/s13421-019-00892-x

 

 

 

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s