Day 164 — Reading: Why Forgetting Helps — The Science of Spaced Learning

Why Forgetting Helps: The Science of Spaced Learning

AIn 1885 the German psychologist Hermann Ebbinghaus published the results of an unusual experiment in which he was both researcher and sole subject. Over many months he memorised lists of nonsense syllables — combinations such as 'DAX' or 'BOK', chosen precisely because they carried no meaning — and recorded how much effort he needed to relearn each list after intervals ranging from twenty minutes to a month. The resulting 'forgetting curve', which shows memory declining steeply in the first hours and more gradually thereafter, is the finding for which he is remembered. Less often cited is a second observation in the same work: repetitions distributed over several days produced more durable learning than the same number of repetitions crowded into a single sitting.

BFor much of the following century, this 'spacing effect' remained a laboratory curiosity. It was replicated hundreds of times, with materials ranging from vocabulary to motor skills, yet it had remarkably little influence on how schools organised teaching. Textbooks continued to present a topic, follow it immediately with a block of exercises on that topic alone, and move on. Part of the explanation may be practical — a syllabus is easier to plan as a sequence of self-contained units — but part lies in the way learners and teachers judge their own progress, a point to which later research would return.

CA persistent obstacle was that early studies used short intervals: learners typically reviewed material minutes or hours apart and were tested the same day. Whether the findings held over the months that matter in education was unclear. In 2008, a team led by Nicholas Cepeda addressed this directly in a study involving more than 1,350 participants, who learned a set of obscure facts, reviewed them after a gap of up to three and a half months, and were tested after a further delay of as much as a year. The results revealed what the authors called a 'temporal ridgeline'. For any given test date, lengthening the gap between study sessions first improved and then gradually reduced final performance, and the ideal gap grew as the test receded into the future. Expressed as a proportion of the time remaining before the test, however, the optimum shrank: roughly 20 to 40 per cent of the delay for a test one week away, but only about 5 to 10 per cent for a test a year away.

DA related line of research questioned the habit of practising one type of problem at a time. In a study published in 2007, Doug Rohrer and Kelli Taylor taught students to calculate the volumes of several unfamiliar geometric solids. One group practised in blocks, completing all the problems of one kind before moving to the next; the other received the same problems shuffled together. During practice, the blocked group performed better. One week later, the pattern reversed dramatically: students who had practised with mixed problems answered 63 per cent of test items correctly, compared with 20 per cent for those who had practised in blocks. The researchers argued that shuffling forces learners to decide which method a problem requires — a choice that blocked practice makes for them, but which a real examination does not.

EWhy should a method that feels less effective produce better results? The psychologist Robert Bjork has proposed that some conditions which slow initial learning — spacing, interleaving, and testing oneself rather than rereading — are 'desirable difficulties': they make retrieval effortful, and effortful retrieval strengthens memory more than easy recognition does. The difficulty is only desirable, Bjork stresses, if the learner is able to overcome it; material that has been entirely forgotten cannot be retrieved at all. The same idea explains why a gap can be too long as well as too short.

FThe effort involved is also misleading. When Nate Kornell and Bjork asked participants to learn the styles of twelve landscape painters, those who saw each artist's works interleaved with others identified the painters of new pictures more accurately than those who studied each artist's works in a block. Yet most participants, asked afterwards, judged blocked study to have been the more helpful. Fluency during study, in other words, is easily mistaken for learning. This may account for the conventional textbook design described earlier: teachers and students alike see smoother performance during blocked practice and draw a natural but mistaken conclusion.

GThe evidence has since been synthesised for practitioners. A 2013 review led by John Dunlosky evaluated ten common study techniques and gave its highest rating, 'high utility', to only two: practice testing and distributed practice. Highlighting and rereading, among the most popular strategies with students, were rated as having low utility. Software has made spacing easier to apply: flashcard programs based on schedules devised by Sebastian Leitner in the 1970s and by Piotr Woźniak in the late 1980s now calculate when each item should reappear. Such tools, however, work best for discrete facts such as vocabulary. How to space the learning of complex, interconnected skills remains an open question, and much of the evidence comes from studies of university students rather than younger learners.