What is a spaced repetition system? And how do we use it to help with our language learning?
A spaced repetition system (SRS) helps you efficiently retain memorized information for the long term by spacing out repetitions of reviews at increasing intervals of time. The result is that you review information at the optimal time for long-term retention.
For language learning, a spaced repetition system is very useful for memorizing vocabulary because it allows you to memorize words in the most efficient way. Let’s break it down how spaced repetition works. It contains two components:
Spacing effect. Researchers have found that people remember things more effectively when practice is spread out over time — spacing effect. This makes sense because it gives time for your brain to process and store the information properly in long term memory.
For example, in school you might cram for a exam the next day, but very rarely will the knowledge stick in the long term. Meanwhile, if you space out the studying over many days or weeks, you’ll find that not only are you more prepared for the exam, but also that the information sticks better in the long run.
Lag effect. The lag effect is when the spacing interval between practices increases. Research has shown that learning is more efficient when study sessions are spaced out than when sessions are successive.
For example, the first review might be in 1 day, but the second review might be 2 days following the first review. This lag effect takes the spacing effect further saying that the longer the time between spaced out study sessions, the better performance will be.
An example of a spaced repetition system would be to follow a schedule of 1, 2, 4, 8 days. The first review is 1 day after the first exposure to the information. The second review is 2 days after the first review. The third review is 4 days after the second review, and the fourth review is 8 days after the third review.
This schedule adheres to the spacing effect because the reviews are spaced out over days, and it adheres to the lag effect because each subsequent review occurs further out in time.
Pimsleur system. Pimsleur introduced graduated-interval recall, in which vocabulary is introduced and reviewed at exponentially increasing intervals, very similar to the above schedule of 1, 2, 4, and 8 days. However, this schedule is fixed for all exercises and does not adjust to the difficulty of the exercise.
Leitner system. Leitner introduced the Leitner System, in which vocabulary are put into different boxes representing different practice intervals, such as 1, 2, 4, and 8 days. The main improvement to the Pimsleur system is that a review item is only promoted to the next box (2-day box to the 4-day box, for example) if that item is reviewed correctly. An incorrected reviewed item will be demoted to the previous box (2-day box to the 1-day box). In this way, easier material will be reviewed less frequently, and harder material will be reviewed more frequently.
SuperMemo algorithm. The SuperMemo algorithm takes a programatic approach, and schedules the next review depending on the number of times the item has been successfully reviewed — reptition number, how “easy” the item is — easiness factor, and the time between reviews — inter-repetition interval.
The SuperMemo system is called an “algorithm” because the intervals between reviews are determined by a algorithm. There are many iterations of this algorithm that are available on the SuperMemo website.
Time commitment. To properly take advantage of the benefits of a spaced repetition algorithm, you must keep up with the scheduled reviews every day. If you skip a review or let the reviews build up, then you will not review them at the optimal time. After a few missed days, it is very easy to have hundreds or even thousands of reviews build up which can be very daunting.
Only suits certain types of learning. Learning with a spaced repetition system works best with knowledge that can be broken down into simple answers, such as the memorization of vocabulary for language learning. More complex types of knowledge require the careful formulation of questions to fit into the system, typically involves breaking down a complex question into many simple questions.
Adjusting schedule to incorrect answers. Some spaced repetition systems, such as the Pimsleur, don’t adjust the schedule after an incorrect answer. This is suboptimal because you should review a question sooner if you got it incorrect, where as if you got it correct, you can adhere to the normal schedule.
What is a “correct” answer? Some popular language learning apps, such as Anki, use the a variant of the SuperMemo algorithm to schedule reviews. However, a downside is that with every review, you must manually grade your own review on a numeric scale (1-Again, 2-Easy, 3-Good, 4-Hard). This can become very tedious, especially if you’re reviewing hundreds of items per day.
Wouldn’t it be nice if you could just answer questions and have the system figure out how you did instead?
Language learners can take advantage of this by scheduling a review of a particular word or concept according to a Spaced Repetion System’s schedule.
For example, an example schedule could be to review information after 1, 2, 4, and 8 days. After the last review, the information will be considered mastered.
Apps such as ListLang can programatically do this for you, so you don’t have to manage the scheduling of reviews. ListLang can also adjust the schedule in smart ways based on the difficulty of the review.
At ListLang, we have developed a proprietary SRS schedule that combines the best of all these methods. It automatically adjusts the schedule depending on the difficulty of the review, while avoiding unnecessary and tedious user input.