This is part 5 in a series on what we know about how we learn and how this knowledge should inform how we teach. The series is intended for teachers, students, and developers of education technology who want to be more informed about how learning works. Parts 1, 2, 3, and 4.
Most people in the U.S. will spend at least 12 years in school. Many will spend far more. How much of all that will we remember?
Not much. And one of the reasons is that we don’t tend to practice in a way that produces long-term learning. “Cramming”—bundling all is one example of poor practice. Many teachers (and many students) acknowledge that cramming is a bad idea. And yet, the basic set-up of many educational environments encourage cramming. In other words, they encourage a lot of study on a single topic, bunched up in a short amount of time.
Law school is a textbook example: in nearly every law school course, students take a single final exam that determines their entire grade for the course. Are students going to cram? You bet they are. A few weeks before the exam, students trade course outlines, and furiously create outlines of their own that try to synthesize all of the material. Then they take the exam, writing as feverishly as they can to explore all of the issues on each question.
And then? Well, that’s it. No need to remember that stuff again, especially if it’s not on the bar exam. Even if it’s on the bar exam, students will re-learn the material in the same way again, spending a couple of months studying a bewildering number of subjects (and paying thousands of dollars for the privilege of doing so). Then they will forget most of the material they learned after taking it.
What it is.
Spaced practice is very simple: distribute your learning sessions over time to maximize long-term learning. Space out your learning and training sessions. Instead of spending 8 hours training all at once, spend a couple of hours one day, wait a few days, spend a couple more hours, wait a week, spend a couple of more hours, wait a few weeks, and spend a couple of more hours.
This is going to be far more effective than spending one full day training. But it’s deceptively effective. After that one day, you’re going to feel like you’ve learned a lot. When you’ve distributed your practice, you’re not going to get the same feeling. It feels more like, “Oh… I forgot… how do I do this again?”
But it’s this forgetting part that’s key. Effective remembering depends upon forgetting. Often, educational systems put the “learning” and the “testing” very close together and never revisit them again.
How it works.
Although the idea of spacing is simple, the mechanisms behind the effect are complex. There are several likely reasons for the effectiveness of spaced practice.
Part of the explanation comes from how deeply people process the material. Learners don’t tend to pay as much attention in massed practice sessions. There are some physical changes that occur at the cellular level as well. These physical changes take time, which seems to be another underlying mechanism behind the spaced practice effect.
For spaced periods longer than a day, sleep also is likely playing a role. Sleep helps consolidate memories and integrate them into our prior knowledge. When we recall something we have (partially) forgotten, we’re also more likely to remember the more important, meaningful aspects of the information, rather than the superficial aspects of the information. This makes it easier to apply the information later on.
Spaced practice is one example of how making things more difficult now can pay dividends later.
In many cases, massed practice improves performance during practice. Or even on immediate tests. The advantage of spaced practice only becomes clear on delayed tests. This raises a tangential, but important point: immediate tests are not very good measures of learning. Don’t trust them.
Of course, not every difficulty makes things better. Having students juggle torches while trying to solve a math problem in their heads is a lot more difficult that just solving the math problem, but it’s not going to help them learn math. This would be an undesirable difficulty. Spaced practice (along with several other approaches to practice which I'll discuss later), are desirable difficulties.
How we know it works.
The idea of spaced practice seems simple. But testing it, even under controlled laboratory conditions using simple word lists, is quite complex. Psychologists have to control the time on task, have to account for primacy and recency effects (the tendency to remember the first and last things you saw), and have to account for any strategy change that their participants engage in.
There has been a ton of research in this area, spanning more than 100 years. The overwhelming bulk of this research has taken place in controlled laboratory conditions, having people memorize word lists. The most comprehensive meta-analysis in this area finds that:
Spaced practice is superior to massed practice, especially for long-term learning, and
Longer “lag” between spaced study sessions leads to longer memory for the material (up to a point).
Of course, the research hasn’t stopped at memorizing words. There’s plenty of research on classroom applications, including some research that uses more reasonable classroom controls (it’s rare for classroom instruction to be completely “massed”). The same advantage of spaced practice occurs. There’s research that illustrates the same basic effect in online courses, across a wide variety of subjects and assessments.
And the effect isn’t limited to just remembering information. The same pattern shows up in explorations of skill acquisition. Surgical training benefits from retrieval practice. So does learning how to play games.
Animal brains work in the same way. Flies, bees, and rodents all illustrate spaced practice effects. Infants seem to exhibit memory benefits from spaced practice. The advantage of spaced practice over massed practice varies, but it’s always there. And it’s usually large.
One of the enduring questions in this literature is what should the spaced intervals look like? If the question is, “how long should I space out my practice?” The answer to this question is… another question: “how long do you want to remember the material for?” In the lab, “distributed practice” can mean anything longer than a few seconds: it could be minutes or months.
If we were to break a single study session into two, the interval between study sessions should be something like 10-20% of the length of time that you want to remember something for. So, if you want to remember something for ten days, then a one-day or two-day interval might be good. If you want to remember something for five months, a three-week interval might be good.
The key is that the student still has to recognize the material. If students don’t recognize the material, then the advantage of spaced practice disappears.
How to implement it.
One of the most visible applications of the research on spaced practice improves upon the basic flashcard. Flashcard apps, such as Anki, Memrise, Supermemo, and many more, incorporate spaced retrieval practice. All of these apps try to provide students with flashcards at the optimal time to improve long-term retention.
Another implication of this research is that condensed, “intensive” classes are probably not that effective. Indeed, one research study found just that — students taking a course over 6 months dramatically outperformed students taking the same course over 8 weeks on a test of conceptual knowledge at the end of the course.
Basic recall tests like flashcards are fine to support remembering basic things. But just because a lot of the research on spaced practice uses word lists doesn’t mean we should fill our classrooms with flashcards. Other research suggests that more sophisticated ways of practicing offer substantial benefits over basic, rote memorization approaches.
Spaced practice happens “naturally” in lots of different ways. In mathematics, we might be returning to (and re-using) fundamental techniques over and over again as we progress. And these same mathematical tools might be used in a physics class or an engineering class. Writing programs that emphasize “writing across the curriculum,” or more frequent writing practice also usually incorporate more spaced writing practice. Classroom and peer discussions also offer opportunities for spaced practice. Students have to recall and integrate information to make sound arguments. So it’s difficult to say exactly how often students are returning to material they learned previously without identifying the concepts (and skills) that students need to know and which activities bring those concepts (and skills) up again.
That said, the overall structure of our educational systems seems driven by an underlying assumption that we, as students, just maintain knowledge and skills indefinitely. We took the class, we took the test, now the knowledge is “ours,” even if we don’t return to it for years and years. This is just not how learning works. We spend a tremendous amount of time teaching students things that they will soon forget.
“Semesters” and “finals” are such an inherent part of the way we think about education. But you could imagine an alternate structure, where, say, three or four months after a class has ended, students get a small amount of practice again — take a short test, solve a few problems, have a discussion section — something. Then, maybe a year later, we have the opportunity to return to do this kind of exercise again.
The research on spaced practice is certainly not the only thing we have to keep in mind when designing educational systems. But the implications of spaced practice definitely go beyond creating better flashcard apps.