Cognitive scientists have discovered a great deal of information about the mind, brain, memory and learning. While some of these findings have influenced the design and development of educational materials, few have been directly translated to practitioners and students.
A group of learning scientists recently identified easy-to-use learning techniques that can help teachers and students achieve their learning goals. We spoke with Dr. Sean Kang, an Assistant Professor and Director of the Cognition & Education lab at Dartmouth College, about distributed practice – one technique proven to be effective in improving student learning.
When we want to learn something well, studying the information or practicing the task just once is almost always inadequate. Reviewing the information or practicing at the right time is critical for durable learning. Distributed practice refers to reviews that take place after some time from the original learning event, as opposed to reviews that occur immediately following the original learning event (termed massed practice).
In the research literature, the learning advantage of distributed over massed practice is known as the spacing effect. In general, the research evidence is clear that spaced or distributed practice is superior to massed practice for long-term learning and retention. Even when the total time spent on studying or practice is equated, if the review(s) is/are spaced apart rather than massed, long-term learning is enhanced. In other words, for each unit of time spent on review, you get more bang for your buck if the review is spaced/distributed instead of massed. Often, this is translated into a warning against cramming before exams.
It should, however, be noted that the spacing effect, to be precise, refers to the benefit of spaced/distributed reviews of the same information, and does not speak to the spreading out of the study of different material over several sessions/days (e.g., studying a different chapter on each day of the week vs. studying all the chapters on one day).
There are a few different theories of why distributed practice is advantageous. According to the study-phase retrieval theory, each time you encounter an item during review, there is an attempted retrieval from memory, and if that retrieval is successful (i.e., you are reminded of the earlier occurrence) the memory becomes more resistant to forgetting. In distributed practice, gaps between occurrences of an item make retrieval effortful, which benefits memory; in massed practice, you just saw the item and it is still on your mind, so there is no need to retrieve it from memory.
Another explanation has to do with contextual variability. When information is encoded in memory, the surrounding context (e.g., what you are thinking of, how you are feeling, how the information is presented, etc.) is also encoded and can later serve as a useful cue for retrieving the information. With massed practice, the context surrounding each consecutive occurrence of an item is likely highly similar. But with distributed practice, the contexts are likely more variable due to the passage of time, resulting in the encoding of different contextual information that is more effective at cueing later retrieval. A final, perhaps more intuitive, explanation is that when you encounter the same item back-to-back (i.e., massed practice), you become habituated to it and pay less attention.
Although most of the prior studies on distributed practice examined memory for lists of random words or word pairs, the advantage of distributed practice has also been shown on tasks as diverse as spelling, phonics, grammar, recall of text/prose, and solving of math problems. It is true that the bulk of the past research has focused on memory (e.g., being able to recall the studied information), but there has been some interesting recent work showing the utility of distributed practice for higher-order learning: e.g., Haley Vlach and colleagues found that distributed lessons improved elementary school children’s ability to generalize their learning about food chains to new biomes; Melody Wiseheart and colleagues showed that college students that received a spaced online review after a lecture were better able to apply the studied concepts to novel situations than students that received a massed online review. More research is needed, but I think there is growing evidence that the benefit of distributed practice goes beyond just improved memory.
Cepeda and colleagues compared the effects of different spacing lags on different retention intervals (the time between the last review and the final test). The findings suggest that after initially learning facts, the optimal gap before having a review session is about 10 to 30 percent of the retention interval. If the retention interval is seven days, then you would be best served by having the review session one day after initial learning. If the retention interval is long (say, 70 days) then the optimal gap also grows (21 days). Yet, there is no clear-cut answer as to how far apart learning sessions should be. We will never be able to empirically test all the various combinations of factors (e.g., type of learning task/material, characteristics of the learner, the particular gaps/retention intervals being considered) that affect learning outcomes.
As for how many learning sessions there should be, again there is no simple answer. Like before, it all depends: How much information are you trying to learn? What is the learning goal? When is the test? What does each learning session consist of? The stark reality is that forgetting is ubiquitous, and the only way to preserve high levels of performance/retention is to have periodic (spaced) review sessions. The nice thing about spaced review sessions is that there are savings in relearning – i.e., you may not be able to recall every item correctly at the start of a review session (because of forgetting), but you are quicker at relearning the information. For example, if you learned a foreign language in high school but can no longer speak it fluently, you would pick up the language more easily now than if you were learning it for the first time.
I think technology can absolutely be a boon for distributed practice. Classroom instruction has traditionally adhered to a linear or modular structure – one topic is covered at a time, followed by massed practice (homework assignments are usually related to the most recently introduced topic). After a unit exam, the topic is often never brought up again. Even if teachers are aware and convinced of the benefits of distributed practice, it is easy to see why it might be inconvenient or difficult for them to deviate from the usual approach. Teachers are constrained by limited class time, and it might take a seemingly excessive amount of planning to organize the topics so they recur in a distributed fashion.
But this is where technology can play a very helpful role. If computers are used to deliver assignments, the software can be programmed to sequence the assignments in a way that ensures the student receives distributed practice. Instead of the teacher planning for practice to be spread out over time, s/he could rely on a computer program to apportion the practice questions such that the student continues to get practice on a given topic long after it is covered in class.
About Sean Kang
Sean Kang, Ph.D., is an assistant professor and Director of the Cognition and Education Lab at Dartmouth College. His research addresses the application of cognitive psychology to improve learning and memory. Dr. Kang earned his Ph.D. in Cognitive Psychology at Washington University in St. Louis.
Carpenter, S. K., Cepeda, N. J., Rohrer, D., Kang, S. H. K., & Pashler, H. (2012). Using spacing to enhance diverse forms of learning: Review of recent research and implications for instruction. Educational Psychology Review, 24, 369-378. [PDF]