The Study Method That Feels Wrong

Hermann Ebbinghaus spent the 1880s memorising nonsense syllables, thousands of them, on himself, and produced two findings that have survived everything since. The first is the forgetting curve: retention drops steeply at first and then flattens. The second is the one this chapter is about: material reviewed at intervals is retained far better than the same total time spent in one block.

The spacing effect has been replicated for well over a century, across ages, materials, species and time-scales. It is about as solid as psychology gets. It is also almost universally ignored by the people it would help most, and the reason why is the interesting part.

Alongside it sits a second finding, sometimes called the testing effect, which is if anything more useful. Trying to retrieve something from memory strengthens it more than re-reading it does. In a standard experiment, one group reads a passage four times and another reads it once and is tested three times; the re-readers do better on a test five minutes later and substantially worse a week later. Retrieval is not measurement. It is the study.

Both findings run into the same obstacle, which Robert Bjork named desirable difficulty. Re-reading is fluent and pleasant: the text is familiar, it goes down easily, and that fluency is subjectively experienced as learning. Retrieval practice is uncomfortable (you sit there not remembering, which feels like failure) and spacing is worse still, because by the time you come back the material has faded and the session is effortful throughout.

So learners systematically prefer the method that produces the feeling of learning over the method that produces the learning. Asked to predict their own performance, students who re-read confidently expect to do better, and are wrong. The subjective signal is not merely unhelpful. It points the wrong way.

Interleaving is the same story in a third register. Given four kinds of problem to practise, most people block them: all of type A, then all of type B. Mixing them up produces worse practice performance and markedly better test performance, because blocked practice lets you apply the same method repeatedly without having to work out which method applies, and working out which method applies is the skill actually being tested later. Blocked practice trains execution and skips selection.

What does this add up to, practically? Take six hours and an exam in three weeks. The evidence is not ambiguous: six separate hours, spread across the three weeks, most of them spent testing yourself rather than reading, with the material mixed rather than blocked. A single six-hour session the night before will produce a higher score than not studying, and a much lower one than the same six hours distributed, and it will feel far more productive while it is happening.

Two cautions, because the popular versions of this overreach. The spacing effect is about durable retention, and if you genuinely only need the material for tomorrow morning, massing is not irrational. And the optimal gap scales with how long you need to remember: for a week, days; for years, months. There is no magic interval, and the spaced-repetition software that schedules this for you is automating a principle, not adding one.

The deeper point is about the relationship between effort and evidence. Nearly every domain where people train themselves has the same trap: the version that feels like progress and the version that produces it are different activities, and only one of them is comfortable. Psychology's contribution here is not the discovery that practice works. It is the demonstration that your own sense of how well it is going is a systematically biased instrument, and that the fix is to measure instead of to feel.